#help-with-robotics

1 messages ¡ Page 2 of 1

oblique pagoda
#

hard decisions to make...

#

coming down to dfrobot devastator with metal gear dc motors (the specs I quoted) vs 1:10 r/c

cursive terrace
#

yeah, you'd probably need 25g servos or better

oblique pagoda
#

I need to find a use for these 50 kilowatts servos on digikey that are in the same series as 250 microwatts one

cursive terrace
#

has anyone used one of the Pi "hats" for robotics (in this case a CRICKIT) and have it randomly "do stuff" when a program isn't running?

outer spoke
#

Hello everyone, i have a question regarding i2c multiplexers as i have not used them before. I'm planning to use multiple AS5600 sensors for a project but the issue posed by this sensor is that its i2c address is fixed. After surfing the internet for a bit, the suggestion was to use a multiplexer. Could someone let me know if that is the best way to sort this issue or are there any better alternatives (I'm open to trying out other sensors as well).

cursive terrace
#

unless the sensor has a means to change the address individually, that's going to be your best bet

outer spoke
primal shell
# outer spoke Hello everyone, i have a question regarding i2c multiplexers as i have not used ...

The main other option is to use multiple I2C buses. If you just need a few sensors and you're using a CPU with multiple hardware I2C buses available (or software I2C is sufficient for your needs), that can save an additional chip. Otherwise the multiplexer (which basically acts as a controllable bus switch) seems to be the usual approach. As always in engineering, what is the "best way" depends on what factors you're optimizing for.

cursive terrace
#

has anyone had a microcontroller board (CRICKIT Hat, in this instnace) suddenly cause a servo to go crazy? my home-built robot arm tried to pound it's way through the desktop for no reason (no code running at all)

primal shell
#

Sounds like either the code is going off into the weeds, a signalling problem, or a power problem.

oblique pagoda
#

Is it made by us robotics ?! But more seriously did you touch it just before it did that ?

#

or is it battery operated and they suddendly dont have any charge anymore ?

primal shell
oblique pagoda
#

didnt realize the irobot movie was 2004 🤣

cursive terrace
#

no, the servo is a generic MG90 and no code running - the board is on a Raspberry Pi 3 and i got it to stop by turning off the power to the board

#

there are 3 other servos connected that were not affected

primal shell
#

It could be the servo itself is faulty too

cursive terrace
#

true - but it seems to be "fine" now (as does the board) - it kind of startled me that a project that i wasn't even dorking with decided to take on a life of it's own

oblique pagoda
primal shell
#

That reminds me of a comedy skit from way back "What to do when your childrens' paper dolls come to life and do a mindless dance of death"

oblique pagoda
#

Id be kinda freaked out honestly so I need to know why this happens 😦

primal shell
oblique pagoda
#

especially if you disconnect the pi from the socket and it has no batteries and it still does that

topaz grail
primal shell
#

Back in the days of phase control dimmers, there was a lot of broadband interference to deal with. At least the mics were usually FM.

topaz grail
primal shell
#

A bunch of college and community theater, and working with independent filmmakers.

cursive terrace
#

since i've been an "application programmer" since The Dark Ages, i find it a tad frustrating that i can't just turn on logging on the controller board 😃

cursive terrace
rocky crypt
#

i was tihnking of making a hexapod project and im split between the decision of using a raspberry pi as the mcu and something like an esp32

#

i've got a raspberry pi zero W at hand but it seems pretty slow so development on it is a big paaaaaaaaaaaaaaaaaain

#

but im not sure if something like an esp32 or an arduino would be more suitable for the task

#

i do have servo driver boards though, so the pin count wouldnt be a problem

fast tiger
#

First, the Pi Zero is an SBC, not a microcontroller . It's a completely different class of device. Second, for tasks such as motor control neither of those options are ideal.

rocky crypt
#

Right.. So what would be a more suitable option?

fast tiger
#

Probably just a more general purpose ARM board with lots of PWM outputs. Communication can be handled by an ESP32.

rocky crypt
#

Ah i dont think those would come in cheap

#

But I have a servo controller board though, I don't really need the pwm outputs

fast tiger
#

ARM dev boards? Adafruit sells dozens of variants.

#

But yeah a servo driver is fine as well.

rocky crypt
#

But I assume the code between the arm and some other microcontroller would be pretty close right?

clear mantle
#

@olive light has done a number of hexapods and will likely have suggestions

#

looks like quads. not hex

fallen pecan
# rocky crypt But I assume the code between the arm and some other microcontroller would be pr...

For example the RP2040 (available on various Adafruit Feather boards) is arm. If you're programming it with Arduino, and you're lucky, your code can be exactly the same between for example Arduino uno and RP2040.
"and you're lucky?" The parts that are very close to the hardware are probably different between different microcontrollers. For example how PWM works and which pins it's available on could be completely different

rocky crypt
#

Ah i see, got it

olive light
#

I made one hex, but never really finished it

#

in any case, you don't really need a lot of computing power, my first robots used a 16MHz 8bit mcu, and that was more than enough

#

the raspberry pi pico is a nice choice, because it has a lot of pins, and is capable of pwm output on all of them

#

but pretty much anything will do, in the worst case you can use a pca9685 for the pwm

rocky crypt
#

I've got exactly two pca9685 boards for the pwm

#

Curious if an esp8266 would be up for the task as i have that one laying aroumd somewhere

olive light
#

if you use the pca9685 with it, then sure

#

otherwise it wouldn't have enough pins to control 18 servos

#

with the pca, you only need 2 pins for i2c, and 2 pins for the remaining 2 servos (the pca only does 16)

#

power is the worst problem with a robot like that

fast tiger
#

Especially since hobby servos in particular tend to oscillate, meaning they are constantly in a half-stalled state

vagrant thunder
#

Running a pi zero with an RTOS isn’t unheard of, but it’s pretty overkill for a basic hexapod.

fast tiger
#

Indeed

vagrant thunder
#

I wouldn’t develop the software for the robot on the same pi zero you plan to run it on unless you’re VERY comfortable with CLI interfaces like an SSH terminal.

#

A 1st gen pi zero is not going to manage robot tasks well if you load it up with a desktop environment.

fast tiger
#

Rephrasing that: do not run a desktop environment on a system that is also trying to directly control motors.

vagrant thunder
#

And no, having a driver IC in between your board and motors doesn’t count as indirect control.

cursive terrace
#

although, if you're not worried about speed, a hexapod with OS + GUI would be stable at least 😏

fast tiger
#

Well, if standing in one place counts as stable then yeah...

olive light
primal shell
#

An "alternating triangles" gait is a valid way to maintain stability whilst moving

oblique pagoda
#

Im a bit confused by the tarmo4 3d printing project and why they put gyro on the BOM that doesnt have any electronics. Are they so universal they all have the same connectors and work the same no matter the MCU that drive it / can interface with motors directly ?

fast tiger
#

They aren't that universal, no. Also there's no real use for a gyro on a 3d printer. Accelerometers can be useful though.

oblique pagoda
#

no the gyro is in the instruction for 3d printing an r/c car

#

but they dont say anything about which one to get/what connectors (I suppose it's some kind of electronics) or why except it drive better at high speeds

#

so Im very confused by it. In a related video I saw it seems the gyro adjust the ride height on the wheels of both side/downforce or "something" so that it rools better

#

But still no data on which gyro to get/how to connect it etc

fast tiger
#

That could be referring to an actual mechanical gyroscope then.

oblique pagoda
#

ah I suppose they mean an r/c type gyro that connects to the ESC

mossy crest
#

oh interesting, it seems like in the R/C car world, controlled mechanical gyros are used for stabilization, not sensing. it looks like they're single axis with no gimbals, in contrast to the moment control gyros that are used for spacecraft maneuvering

fast tiger
#

Yeah that's what I was wondering.

mossy crest
#

i guess it makes them more akin to reaction wheels? though those i think those typically involve reaction torque on the rotation axis, not precession torque (which is at right angles to the rotation axis)

sly sand
#

Does anyone now a module where I can connect a RF antenna and send or receive RF signals in many frequencies, I would want to use it with a raspberrypi pico w. If you now any module please tell me(I am desperate).

fast tiger
#

That's a very generic statement, the answer to which would likely be illegal.

cursive terrace
#

or a CB maybe

lone skiff
# sly sand Does anyone now a module where I can connect a RF antenna and send or receive RF...

SDR, or "software defined radio" is probably a good keyword to research. However, they usually are USB devices rather than something that can easily interface to a Pico. If you narrow down your use case, people may be able to have better advice. "I want to broadcast TV signals" versus "I want to send 2400 baud serial on a choice of 3 nearby frequencies" would be very different scenarios.

fast tiger
#

A pico also likely won't have the computation power necessary to do anything useful with an SDR.

hushed finch
#

not necessarily help but it sanity check on a project idea I'm staring down. For a fpv style drone replace the radio receiver and controller with a pi pico receiver and pi4 ground station. idea is pico can handle simulating serial input into the flight controller. Pico will run a websocket that connects to ground control. GC can just be gimbals. or extend it out etc. video can be handled seperately so you only need to send controls and telemetry. Ground control could even be an RTK.

now on paper it makes sense within wifi range, and could be doable I think with cellular since its only controlls.

sly sand
#

Can I use a nfr52840? Cause I saw that I can only receive and transmit RF between others nfr52849, is it true? I hope not cause I have order the nfr52840 hoping I could use, if can't please tell me things that I can do with it...

lone skiff
#

It just depends what you're trying to do. If you're making you own devices that just have to talk to each other, then you have a lot of freedom, versus needing to talk to many other devices that might have different radios.

primal shell
hushed finch
sly sand
#

And what do you think about the nfr52840 for the thing I want to do, can it handle them?

cursive terrace
#

er - "sending" in the broadcast range (to a radio) can be illegal depending upon your jurisdiction -- there are certain signal-strength limits that you would need to take into account, for example; things like remote-control systems (RC cars, home autoation, etc.) have very specific bandwidths that you are allowed to use, but have the downsides that a lot of other things are also using that (which is why they have security and identification protocols)

#

aside from the "playing on the radio" bit, it sounds like you want an automation system and there are a couple of OSS projects that might be more in line with what you want to do (although doing it all yourself is a worthy project)

primal shell
sly sand
#

Can someone help I can't display nothing on my display 128×64 , this is where I bought it, https://m.pt.aliexpress.com/item/1005005368669648.html?pdp_npi=2%40dis!EUR!3%2C28€!2%2C66€!!!!!%402103253916879902318186586e2fd2!12000032766486155!btf&_t=pvid%3Aa734e7cb-26f7-4234-8d04-9b53284603ef&afTraceInfo=1005005368669648__msite__c_ppc_item_bridge__xxxxxx__1687990232&spm=a2g0n.ppclist.product.0&gatewayAdapt=gloPc2braMsite, in the back it says, GM12864-01A, and I'm using a raspberry pico W

grizzled bobcat
#

software/ code for animatronics?

primal shell
sly sand
#

Raspberrypi pico w

#

Do you mean like the idea?

#

IDE*

primal shell
#

How are you programming your Pico?

sly sand
#

With micropython in thonny

primal shell
sly sand
#

None

#

Cause I didn't find any

sly sand
#

I'm going to try it out tomorrow, I'll give you feedback

sly sand
#

and i did i like search and discoverd that because of the pins name the display is oled, it uses scl sda bl rst, dl, dc...

primal shell
#

Ah, the listing says "SPI" but the pinout has SCL and SDA which are I2C signals. The ST7567A chip itself supports both (plus a parallel interface), it has some interface select pins (SI2, PSB, and C86) which tell it which interface to implement.

sly sand
#

forget about it im using an other display...

sly sand
#

does anyone now a library for a pn532 with i2c using a raspberrypi pico w, coudn´t find none

cursive helm
sly sand
#

im using mycropython...

cursive helm
#

ah -- sorry -- I don't know about that.

sly sand
#

don't worry

cursive helm
sly sand
#

ok thanks

turbid trail
sly sand
#

Thanks, sorry for being late...

cursive terrace
#

y'all might appreciate this: i have a self-built "robot arm" that i'm programming some sequences of movements for and the program is controlled by a NeoKey1x4 off to the side -- the latest sequence, which i completely eye-balled, ended up with the end of the arm coming down on the "terminate program switch"

technically, it committed suicide i guess 😏

frank portal
#

Does anyone know how to use Adafruit with p5play on replit? I tried a few things but replit doesnt allow importing external packages and chatGPT doesn't know how to do it either. The Express is on, and the replit is working, it just isn't connecting?

cursive terrace
#

can you save files to your local computer? (I've never heard of replit before)

#

@frank portal i don't think you can use the "run" from replit - it's running in the cloud and not on your local hardware, so it's not going to be able to connect to any devices

#

i would recommend looking at the suggested editors and alternates on the Adafruit learning center

frank portal
#

Ohhh

#

Would visual studio code work?

cursive terrace
#

i've not used the JavaScript stuff (for either adafruit or vs code), so i can't provide guidance

#

it would at least have your files on the local system(s) instead of in the cloud 😃

manic depot
#

@frank portal replit is kind of a cool plattform but maybe this is the wrong place to ask. They should have a forum. I would guess you need a payed account for anything special

frank portal
#

. Okayy

#

☠️

near grove
#

Anyone using CAN bus for robotics

#

hopefully using python…

cursive terrace
naive galleon
lone skiff
naive galleon
lone skiff
#

The scan_data variable in that example code is an array of distances as a function of angle, so you can apply your own analysis logic to it. For instance, a function that would tell you that there is an obstacle in that range would be like:python def is_obstacle(scan_data): for angle in range(0, 45): # degrees if scan_data[angle] <= 500: # millimeters return True return False

naive galleon
quick silo
#

Question for any of you who have purchased Adafruit's TT Motors:

I'm looking into this (https://www.adafruit.com/product/3802) motor to be back-driven as a generator.

I'd like to know how much resistance the motor provides when being back-driven, both open circuit and when shorted.

If anyone could give me some info, it would be greatly appreciated! Thanks

TT Motor All-Metal Gearbox - 1:90 Gear Ratio

These durable (but affordable!) gearbox motors (also known as 'TT' motors) are an easy, low-cost way to get your projects moving. This is a TT DC All-Metal Gearbox ...

#

To be more specific... I need a motor with this sort of geometry and high gear ratio because I want to mount it into an elbow or knee bracket to try using it to charge wearable electronics.

primal shell
#

I had the notion to grab one, fab up an adapter and then try turning it with a torque wrench but realized the resistance is going to vary with the speed (especially in the shorted condition) so it's going to be more like a table or equation than just a couple of numbers

sacred gate
#

i think the torque required, is also based on the current flowing

#

so the more load you add, the more it will resist the spinning

#

but the higher the speed, the more voltage, and now that depends on the voltage/current profile of the load

primal shell
#

Amusingly, the given conditions were "open circuit" (zero current) and "short circuit" (zero voltage). But you're right, it would be cool to plot torque vs RPM and load.

quick silo
#

Surprised to see you, bodger

#

And yes, the resistance of the generator is the sum of losses and BEMF which is directly proportional to the current.

#

But I was just looking for ballpark numbers. There won't be an exact science to how much resistance/power generation is "optimal" because it entirely depends on what the wearer is comfortable with.

primal shell
#

I was not aware I was surprising!

kinder-german.jpg
naive galleon
#

Hello, good day. I'm currently utilizing the adafruit_rplidar library and executing this example code again from this GitHub repository: https://github.com/adafruit/Adafruit_Learning_System_Guides/blob/main/pi_rplidar/display_lidar_pi.py. I've noticed that it's running smoothly, however, when I attempted to add a print statement within the "if distance > 0" loop, it caused an errors about new scan flag mismatch or check bit not equal to 1.

I am suspecting that print() is executing too slowly which leads to the errors. Is there a way to make print() execute faster or any other alternative way to print statement in this loop just for debugging purposes?

Thanks ^^v

primal shell
sacred gate
#

ive noticed on the broadcom port, that the framebuffer text console was a major part of the slowdown

primal shell
#

It probably paints the characters bitwise

inland bear
#

Okay, lay it on me, I want to get some help with the maths of calculating the angle between two time of flight grid sensors based on a user holding their hand flat above both (pretend each is -/+15degrees off vertical, actually one is -30degrees and one is vertical ish, but they are potentially moved between uses). This hand position would be maintained for a few seconds to establish a flat baseline, and then (after chime/instruction-WAV) the user moves hand left and repeat wait, then right.
In my head I'm ignorantly like surely that's common, calculating angular offset between two planes in 3d space...but the thought of the maths scares me... I also assume ulab.numpy would be sufficient to process this on device (esp32s2)

lone skiff
vagrant thunder
#

So you have two ToF sensors and you want to measure the angle between the two? Assuming the two sensors are aimed outwards from point A in the law of cosines diagram, what other information do you have besides lengths b and c, and which angle do you need to solve for?

cursive terrace
#

from the description of the project, you could just reduce it to a 2D problem

#

that should take some of the scary out of it

inland bear
#

I want to be a bit more 3d in my head, and then probably reduce the 3d to a 2d problem. For example I want to remove the arm and leave a line / plane that represents the hand by assuming the arm extends to the end of the grid (the ToF sensors are 4x4 or 8x8 grids VL53L5CX's)

#

Thanks you three as per usual 🙂

#

It's also only "fairly" safe to say the two sensors will be parallel horizonatally, currently they aren't as manually held by soft-reusable glue dots (at least 3degrees off), but I could make a pcb a bit like laser cut panels to ensure a 30degree offset between them and horizontal parallelism. I'd rather not as it would be nice to have this buildable without special / custom pcbs, using whatever is to hand (glue dots or cardboard etc), and have the maths adapt.

#

(this is the theremin in case your half paying attention between my long drawn out postings)

night grove
#

@inland bear Are you trying to derive volume and pitch controls from hand position? On a theremin, a horizontal hand motion controls pitch and a vertical one volume (though two different hands interacting with two different antennas) -- are you aiming for the same meaning of vertical and horizontal in a single hand?

#

If so, do you want to e.g. keep a constant volume as you vary the pitch by keeping the hand at the same height above the sensors?

#

If so, it seems like what you're aiming for is mapping hand position to (x, y) coordinates. You can derive that from the (rho, theta) coordinates of (distance from a sensor, angle at that sensor.) You can get the angle via the law of cosines, as pointed out above.

random sphinx
#

I have a leg on a robot swinging forward and back several times per second and I need to know when it reaches the end of the swing (in other words, when do I need to stop the motor and reverse the swing). What would be a good way of measuring that it has reached the end of the swing?

I was thinking some kind of magnet, a pressure switch or a laser, but not sure what would work best. I am assuming a physical switch would be a bad idea as the motor would be told to stop once it reaches the switch, so it would probably drive into the switch. I thought of a physical stop (plate to stop the swing), but I'm sure that would be unfriendly to the motor.

I appreciate any suggestions on how to get this resolved.

cursive terrace
#

you can set up physical switches to be triggered before "end of swing" to account for any delays in processing the signal

#

or an optical beam trigger - use an adjustable "flange" on the leg that interrupts a sensor and you can move said flange/sensor to calibrate the end of movement that you want

vocal bear
#

I have modeled a linkage in some software. The software allows me to set the crank angle as a function of time including adding polynomial fit curves to it and then also gives me the angular velocity and angular acceleration as a function of time. Does it make the most sense to write a control algorithm that attempts to use a PID to set the position time?

flat quarry
#

Can a DMX breakout board like https://www.sparkfun.com/products/15110 be used for RS485 communication? I'm thinking of using the DMX output and cutting off an end of the cable and hooking that up to my flipdot controller. I read DMX is rs485 soooo would it work?

SparkFun ESP32 Thing Plus DMX to LED Shield

The SparkFun ESP32 DMX to LED Shield is the perfect way to send and receive DMX data whether it's coming in or out over the onboard XLR-3 jack or ArtNet.

primal shell
fast tiger
#

Ethernet is another option if you want isolation.

#

But yeah I second RS-485. And I still need to do a project with it.

primal shell
#

Heh, so do I (my electric tractor uses BAC bus, which in turn uses RS-485 signalling, to talk to its ESCs). I bought a USB to RS-485 adapter, now I need to splice in some wires to try listening in and see if I can figure out what's going on (the root problem is the ESCs slow way down when the battery voltage is higher than they expect and I've raised the battery voltage by replacing the lead acid batteries with lithium iron phosphate).

cobalt lagoon
#

hello yall. I am making a line following robot that detects a coloured line using two colour sensors. Thsese sensors must be able to detect black,white,red,green,reflective tape and blue. Can any one recomend me a colour sensor?

fresh crown
#

I've heard of using multicolored lights to bias a sensor towards sensing a particular color

primal shell
zenith crow
#

Since this shield is PWM, would this be able to run DC motors as well? (This one below specifically)

primal shell
#

No, it provides a logic level signal, you'd need a more powerful driver to power a motor

zenith crow
#

Hmm gotcha gotcha. Is there a feather option that would only require 1 wing instead of 2?

#

Or would the arduino form factor motor shield be an option?

zenith crow
primal shell
#

You don't need a wing to run servos, you can run them with ordinary GPIO pins with PWM. One single-board solution is the Crickit (product ID 3343), which supports several servos, as well as motors and a variety of other things. The drawbacks are a) it's larger that most wings, and b) it's currently out of stock. Another possibility is the motor wing (product ID 2927), and just run the servos with spare GPIOs.

zenith crow
#

Got it, thanks!

naive galleon
#

Hi guys, i'm having a trouble for obstacle avoidance (btw, I'm using using adafruit circuitpython library).
Would like to ask for any help or any suggestion for this project to be done.

So the concept is I would like to have 0 to 20deg and 340 to 360deg as my front and if there is any distance becoming less than or equal to 650 in those range (which is I categorized as obstacle), I want to stop first then rotate until no obstacle then move forward.

As you can see in the video, if the obstacle is too big it somehow register in the code, however when I make it thin, then it is not somehow registering.

CONCEPT.png photo1695282146_1.jpeg
slim kernel
#

What would be a good power supply to use at the 2-terminal block on the shield that is labeled DC Motor Power for this motor?

vagrant thunder
slim kernel
#

I have the option to use a EasyDriver A3967 instead - is this a better driver for this low voltage stepper motor?

#

I've got a CanaKit 5V 2.5A power supply, will this work with the L293D darlington driver that is in the V2 motor shield?

vagrant thunder
slim kernel
vagrant thunder
slim kernel
slim kernel
#

Using the CanaKit 5V 2.5A power supply now, does this look like the correct phase wiring?

primal shell
#

Looks right to me

slim kernel
#

We have motion! Programmed 2560 steps in single stepping mode and a travel indicator is showing 0.4" of motion. Which jives with the linear actuator spec that says 0.000156" motion per step.
[ 0.400/0.00015625 = 2560 ]

#

Now the goal is find the maximum speed possible.

#

Mcmaster says it should be able to go 0.4" per second!

#

That's 768 rpm I think.

cobalt lagoon
#

hello sirs, i have come up with an idea of making my own colour sensor

#

the blue thinng isan LDR

#

the transperent dome is a lens (focal lenght 3 cm) to focuse the light onto the LDR

#

i was thinking that the light will turn red and the LDR will take a value, then the light will turn blue then the LDR will take a value

primal shell
#

Seems workable

topaz elbow
#

anyone has idea how to make good grip wheels to this sub micro motor? shaft is 1.95mm D type

topaz elbow
primal shell
#

I'd press fit a rubber grommet onto the shaft as a first try

topaz elbow
#

it will be a nano sumo robot, so tires must be tight connected to the shaft

vocal bear
#

https://www.robotshop.com/products/solo-mini-40v-80a-dc-bldc-pmsm-acim-motor-controller
Has anybody used this controller before? It looks pretty good but I don't have the best track record with identifying these things

RobotShop USA
SOLO MINI 40V 80A DC BLDC PMSM & ACIM Motor Controller

Description Controls DC, BLDC, PMSM, EC coreless and ACIM motors Wide input voltage supply range from 6V to 40V Continuous output current of 20A, maximum current of 80A Dual core with parallel processing architecture USB, UART, and CANopen communication The SOLO MINI 40V 80A DC BLDC PMSM & ACIM Motor Controller is

grand oyster
#

not sure if this is the right place to post this. is there a way to use the pijuice with a cooler?

proven lotus
#

Anyone work with Hiwonder HTD-45 motors and their LSC-16 driver? I'm getting read timeout errors when i try to get them running with the pc software

steady tapir
#

Hi guys, I am making a 6DOF robotic arm as a project with my friends, and we are currently considering the motors. At the moment, we are leaning much more toward using 4 bipolar stepper motors and 2 servos.

As the electronics guy, I am in charge of choosing a power supply and driving the stepper motors, but I am having some problems with this. The stepper motors we currently want are rated at 2 amps per phase. At the same time, I am trying to use an A4988 driver module to drive each motor, which has a maximum output of 2.1A per phase. Could this work? I know the DM860T driver is a much more popular choice for these projects, but I think it might be overkill for what we want to do (especially in terms of price). If not, what is another good, affordable stepper motor driver?

Note: I have never worked with stepper motors before so excuse the 'simple' question 😅

native cipher
#

the real question is not what the steppers are rated for but how much power they really need for your project. E.g. ig you put the maximum planned load on your arm, how much current each stepper would need to hold/move that load?

if they only need 1 A to deal with your loads, then your motors and your drivers are perfectly fine.

naive galleon
#

Hi guys, I'm currently working on a project involving a two-wheeled mobile robot. I'm utilizing UWB for positioning, Rplidar for collision avoidance, and motors with a motor driver for the robot's movement.

At this point, I've successfully implemented the functionality to move from my current position (start) to the target position (end) using the Euclidean formula with only the integration of motors and UWB. Although I haven't reached the phase of path planning yet, I've just managed to do a simple computation of linear distance during linear movement.

So, my motor control is intended to be controlled by UWB data and Rplidar. However, when attempting to incorporate Rplidar, an error related to "Incorrect descriptor starting bytes or Wrong body size" is displayed.

I suspect that the processing might be too slow when placing Rplidar data in the main_visualization function, or there might be an issue with the code structure, such as an incorrect format. I've already attempted placing main_visualization() inside the get_data from_LIDAR(), but the problem persists.

Below is a snippet of my code, and I would greatly appreciate any suggestions or ideas on how to resolve this issue. Thank you.

📎 snippet.py
forest cipher
vocal bear
cursive terrace
#

the datasheet should provide that information

vocal bear
#

I did check the data sheet, I could not find it

fast tiger
#

"Digitally set zero position alignment" suggests that it defines the index pulse relative to that.

vocal bear
#

So that would mean that you can set it on startup?

fast tiger
#

I didn't look at the datasheet in detail, but it's either that or you can store it in persistent memory.

cursive terrace
#

the top of the sheet says "encoder can be zero-set via SPI or AMT Viewpoint™ PC "

vocal bear
#

What's a reason to have a lower resolution when a higher is available?

cursive terrace
#

memory/speed - less bytes, less overhead

steady tapir
#

Quick question: In a robotic arm I am making, the servo motors with the highest torque will be 40 kg-cm (these: https://www.amazon.nl/-/en/dp/B0C64NLF1Q) with a current draw of ≈ 4-5 amps. Would standard 22 AWG servo wires (like these https://www.amazon.nl/-/en/dp/B08728B6HB) be suitable for connecting the servo to power and to the microcontroller (Arduino Mega), or is there another type of wiring/connection method I can use that is better? I know I cannot loosely attach wires to the servo due to EMI, but I'm not really sure about how else I could optimize wiring and what type of wires to use to connect the servos to power and PWM.
Note that I will be using a power distribution board (https://www.amazon.nl/-/en/dp/B0814B5P5M) to deliver power to each individual servo, of which there will be 7

primal shell
#

I would use thicker wires than 22 AWG for that kind of current. Make sure you have fuses to protect your wiring and electronics too.

steady tapir
#

Thanks, we seem to be going down the route of using 14 AWG and simply soldering everything together - maybe excessive, but it should be reliable enough right?

native cipher
#

14 awg is an overkill.
I think 18 is quite adequate

vocal bear
#

That is, is this gearbox increasing torque? It says it's speed reducing which makes me think obviously that it is increasing torque, and I am not familiar with a Worm Drive gearbox that doesn't increase torque, but the way the ratio is described, 30 to 1, is a bit confusing to me

native cipher
#

30:1 is pretty typical for wormgear.
holding torque will be very high - for wormgear, usually you can not back drive it at all until you apply torque high enough to break the gears

vocal bear
#

That's what I thought. For some reason I started thinking the ratio was the ratio of input gear to output gear, so 30 to 1 would be 30x speed and 1/30 torque

native cipher
#

in this case, they actually have the max torque on the website: 20 Nm = 177 lb-in

vocal bear
native cipher
#

I guess it means that if the load is more than 20 Nm, you can break the gears of the gearbox.

vocal bear
#

I see that makes sense, because the pull out torque at the lowest speed is 25 Newton meters, which doesn't make any sense either

#

Excuse me 2.5 newton meters

frank spoke
#

I'm looking at power supplies from Adafruit for projects involving strings of lights and servos. Any one in particular you guys recommend?

sweet plank
#

my advice is work out your approximate current draw and then go from there

quasi plaza
#

excuse the terrible drawing

theplan.png
#

I was wondering if anyone knows the best way to go about this, I want to use an arduino to control a motor (I was thinking stepper) to pull a weight of about 2.5 kg from and too a specific height. the motor would roll up and lower a chain or rope that is attached to the object, so I dont think the chain or rope will move too far away from drive shaft. I did just try it with a nema 17 and a a2988 board but from my tests its seems really weak.

cursive terrace
#

well, any kind of gearing will help improve torque (if you've already got a motor, you could try finding a gearbox or build-your-own) -- a pulley system would also work

rose temple
#

metal gear ... ~15$

drifting sapphire
ionic python
#

Hey folks, I'm hoping that someone can help me save my sanity. I have three FS90R continuous servos connected to a Pico W (powered by USB). I cannot seem to trim these things to sit still (and if I do, they then stop going forward or backwards but I can't seem to find the perfect balance). I think I either have the PWMOut frequency wrong (currently set to 50 based on the examples in the docs) or I need to manually set the ContinuousServo min_pulse and max_pulse but I have no idea where they should be. I've been banging my head against the wall with this one for a few days. If anyone can point me in the right direction I would really appreciate it! 🙂

sweet plank
#

Have you looked to see if there is library that can help? that would be my first idea

turbid trail
#

I just wired up an FS90R and tested with this:

p = pwmio.PWMOut(board.EXTERNAL_SERVO, frequency=50)
s = adafruit_motor.servo.ContinuousServo(p)

while True:
    s.throttle = 1
    time.sleep(1)
    s.throttle = 0
    time.sleep(1)
#

mine does come to a complete halt. I don't remember adjusting the trim but maybe I did in the past

ionic python
ionic python
turbid trail
ionic python
#

yes, and I can't get it behaving as I expect (with throttle 0 being stopped, 1 being forward and -1 being backwards).

turbid trail
#

is the servo being powered at 5v and the control signal is 3.3v? that is what I am doing

#

there are number of comments in the adfruit forums that getting continuous servos to stop is not easy. However, for some reason I have not had that problem

#

I was testing this with a Feather RP2040 Prop-Maker

ionic python
#

I'm using a standard Pico W but being powered via a Powerboost 1000c

turbid trail
#

I only have a couple of the FS90R's. Interestingly sometimes when they start up they don't stop completely, but putting a little strain on them fixed that. I did adjust the trim pot on one. It's a pretty touchy adjustment but I got it to stop

ionic python
#

the weird part is that its clearly responding to the code, but I just cant seem to get it tuned right... if I trim it to stop when it should stop, it then won't go forward or backwards

turbid trail
#

is it under load?

#

mechanically?

ionic python
#

right now just wearing the horn that came with it, nothing heavy

turbid trail
#

same here

#

I modified my test program to go foward, stop, reverse, repeat. Changed throttle too, away from 1.0 and -1.0, no difference

#

did you try adjusting the trim with a program similar to mine?

ionic python
#

yup thats pretty much what I'm doing right now:

# 
while True:

    print(".")

    print("full stop")
    head.throttle = 0
    selection = input()
    print("full ahead")
    head.throttle = 0.5
    selection = input()
    print("full stop")
    head.throttle = 0
    selection = input()
    print("full reverse")
    head.throttle = -0.5
    selection = input()
turbid trail
#

I just tried adjusting the trim again, moved it 1/4 turn, and then back. Last adjustment is very tiny. I am using a jeweler's screwdriver

#

are the servos from us?

#

Fitec brand FS90R

#

apparently the old name for "FEETECH"

ionic python
#

yup this is the same name I have on mine

turbid trail
#

yes two positions, one it goes slow, but doesn't stop, other position it does stop, getting really slow when I'm getting close. maybe they are 180 degress off, not sure

ionic python
#

I think I maybe need to get my multimeter out and make sure they're getting enough power on the 5v rail as I have other devices (such as an i2c amplifier) also connected

turbid trail
#

yes, 180 degrees opposite. one position it really stops, the other it just gets slow. I put a dot on the trimpot with a marker to make sure

#

mine says 4.78 volts

ionic python
#

I'll have to come back to this... my office is also our spare room and my mother in law is asleep in there at the moment (and all my equipment is in there)

turbid trail
formal kraken
ionic python
formal kraken
#

Gotcha

noble sundial
#

its having some troubles with the click library, do i need a specific version?

valid totem
#

Anyone here ever try using the braincraft hat? I'm having an issue where it powers off the pi when I try to use the 3pin Jst GPIO ports 😞

cursive terrace
#

that sounds like a short 😱

valid totem
#

I think its cause the servo is too much for the jst port maybe?

#

cause my connector works fine when I connect it to the pi without the hat on the breadboard

#

Trying to figure out how I can safely connect this servo. I really wanted to put the camera + hat in a skull that I already have the jaw moving using a servo.

sweet plank
#

what are you using to power the whole thing?

valid totem
#

Pi plugged into the wall, servo plugged into the 3 pin jst port #13

#

thats def the problem

#

It only did it once so far, the pi still seems to be running fine, Going to test the hat itself more now

oblique pagoda
#

but what power your socket ? coal ? gas ? nuclear ?

valid totem
#

power cable that came with the pi, plugged into a big power strip / surge protector.

#

Its pretty much on the strip by itself

#

that + a light

sweet plank
#

Id be a little surprised if a single servo browns it out but its not impossible, depends what the servo is drawing

valid totem
#

thats what I thought but I posted a thread on adafruit about it and they said I shouldnt plug a servo into the jst port

sweet plank
#

doesnt the example literally say a servo?

valid totem
#

I'm using an MG90S servo

sweet plank
valid totem
#

thats what I thought

#

the servo connector looks fine, wires brand new

#

I am using a monitor with the pi

#

would that draw too much power or something?

sweet plank
#

i guess it depends how much current you can pull on those connectors but its weird they specifically suggest servos

#

they plug a servo in in one of the examples

#

what does your servo move?

valid totem
#

just noticed it does say +5v

#

is the jst connector only 3v maybe?

sweet plank
#

possibly... not sure, hopefully someone with a bit more knowledge about that hat pops up

valid totem
#

its the monitor, without the monitor connected, it can run the servo

#

must be overdrawing when all 3 are connected

cursive terrace
#

servos can pull a lot of current, so if you don't have a beefy power plug, you could experience the brown-outs from the servo

valid totem
#

I think its browning out from the monitor + servo + hat

#

I was being dumb I guess

olive light
#

I find it best to power servos separately from the electronics

sweet plank
#

oh for sure

olive light
#

you just need to have the gnd common

sweet plank
#

theres a super easy board i use all the time..

#

PCA 9685?

olive light
#

yeah, that's a pwm chip (originally a LED driver)

valid totem
#

yeah I have an extra 9V connect here, I'll have to solder up a custom connector for this

sweet plank
#

yeah i grabbed a few boards from...idk some supplier a while back and they have done a good job

valid totem
#

I have one of those too somewhere I think

olive light
#

9V batteries are anemic in terms of current

#

if you mean alkaline, anyways

#

but you can just use the pwm from your mcu, and still power the servo separately, it's the red wire

cursive terrace
olive light
#

but if it's powered from usb, that might not be able to give 2.5A

faint canopy
#

With the Servo Hat on a pi, is it possible to combine the use of the PWM library with the servo library?

#

or do they each grab exclusive use of access on that port

olive light
#

the servo library uses the pwm library under the hood

faint canopy
#

that's what I would have expected, but it doesn't seem to expose the pwm object. When dimming an LED, using servo[1].angle isn't exactly clean

#

for instances when you might want to control both servos and LEDs

olive light
#

why would you have servos and leds on the same pin?

faint canopy
#

the same hat

#

not the same pin

olive light
#

As far as I know, that thing uses the pca9685 chip for the pwm, and that chip can't set the pwm frequency per pin. For servos you need 50Hz pwm frequency, traditionally, which is way too little for good LED dimming, as it will be visibly flickering.

faint canopy
#

ahhh, there it is. Ok that makes sense

#

thank you

#

looks like it's mosfet and a native PWM pin for me!

olive light
#

you don't need a mosfet for a servo

faint canopy
#

for a led

olive light
#

right, but if you put the servo on the native pin, and the led on the hat, you won't need that mosfet

faint canopy
#

got like 12 servos and a few leds

olive light
#

makes sense then

faint canopy
#

servos on the hat, pins 12 and 14 for leds

olive light
#

if the leds are not too bright, you might get away without the mosfet anyways

#

modern leds take very little current

faint canopy
#

true, but 3.3 isn't quite enough to drive a sequin. they're pretty dim on it

olive light
#

sounds like you have a solution then

valid totem
#

So does the braincraft hat just not work anymore / is out of date?

#

I followed the guide to the letter, and I'm stuck trying to install the kernel module, everytime I try it says Unable to compile driver because kernel space is 64-bit, but user space is 32

#

And then I noticed the flag about how bulls eye may not work, but now we're on Bookworm, so am I just SOL? How do I get an old buster install ?

sweet plank
#

what bit does it fail at

valid totem
#

It doesn't work at all on bookworm, had to restart using bullseye but now I'm have to figure out libcamera cause the adafruit doc is outdated and uses rasbpistill even though that doesn't work on bullseye

#

They mention it doesn't work on bullseye, but don't point to libcamera or any other alternative

#

Its just outdated AF

solid sonnet
#

Hi guys, quick question about servo. I got a gripper set earlier that use a SG90 servo to open and close the robotic gripper. I can have it open all the way with a angle command of 100 deg and close all the way with 0 deg. So I am wondering if I have it grip something, like a water bottle. I command it to close with 0 deg. The gripper will stop at around 30 deg, then I can move the object and it wouldn't fall out, so that's good. But then the servo make an increasingly high pitch nose (that's not good I assume). Would that destroy the servo?

If so, how should I go about solving this issue? If I command it at 30, deg, there wouldn't be enough gripper force to hold the bottle. Thanks!

olive light
oblique pagoda
#

Do you need access to the raw video stream to do computer vision/autonomous navigation or you can use wifi camera that uses their own power like says a gopro and fetch the frames by http?

olive light
#

any way you can get the pixels might work, depending on what you are doing exactly and what are your requirements with speed and latency and so on

#

in case of wifi interference may also be a problem

cursive terrace
# solid sonnet Hi guys, quick question about servo. I got a gripper set earlier that use a SG90...
  1. yes, it will destroy the servo, especially if left that way for a period of time (the driver is trying to make the motor move and can't so everything is super-stressed)
  2. if you note that you can get a grip at 30 degrees, you can slightly stress the servo by closing to 30 degrees minus a "fudge factor" of, say, 5 degrees which may or may not over-drive the servo depending on the squishy-ness of the target
#
  1. as djhipu noted, you can also use some kind of feedback sensor but you're going to have to still go "past" 30 degrees for a firm grip (and you might want to think about non-slip coating on the gripper)
solid sonnet
#

I will give those a try, thanks!

ionic python
#

Hey folks, I have this MG996R 360 servo and it's working great, except I've noticed something weird... it will always return to the exact same position. Does anyone know if this is normal? I was hoping to have partial control

olive light
#

interesting, I wouldn't expect it to have any position feedback at all

ionic python
#

It's a little annoying because it makes my droids head a all jerky. I guess I could possibly improve things by tweaking the size of the gears?

cursive terrace
#

that seems like there's something in your control code that's maintaining a state and the servo is being returned to that previously stored position

primal shell
#

I wonder if you have a ordinary (non continuous rotation) servo

olive light
#

then it would just keep the position

primal shell
#

Under some conditions. I'm wondering if the drive signal is going away or something.

ionic python
#

I’m wondering if it’s something to do with the pulse size. I need to dig into the specs and check if the library matches (I’m using a picon zero)

native cipher
#

all the way around is not the same as "continuous rotation". There are servos capable of rotating more than 360 without being continuous rotation.

idle grove
#

how would I connect a A4988 stepper driver to a raspberry pi 4

cursive terrace
#

i found quite a few on-line references for that particular driver and a pi, so i would suggest trying out the ones that you're most comfortable with

oblique pagoda
#

dont forget the couplin cap and to reset the driving and direction pin when not driving the motor in your program

ionic locust
#

does anyone know if the Adafruit TB6612 can be used with servo motors? My hope is that the A1/B1 lines can be pulled high and A2/B2 grounded so that its always in "forward" direction, and then the A/B1out can drive the servo, ignoring the A/B2out lines

olive light
#

just use those pins to drive the servos directly, there is no reason you would need a driver chip for this

#

servos already have a driver chip for their motor inside them

ionic locust
#

the inputs are PWM

#

oh nm I see your point

cursive terrace
ionic locust
#

sorry EdGJr I hadn't noticed the robotics board first - I deleted the other message

oblique pagoda
#

I thought only digital servos had a driver chip in them

#

and driving them directly encourages peoples to power them from MCU too

#

most peoples seems to buy the counterfeit 90 analog series anyway

olive light
#

they still have a chip, the chip is not a microcontroller, but an analog circuit, but it's still a chip

#

usually it's an h-bridge, a generator controlled by the potentiometer, and frequency comparator that compares the generated signal with the signal received

ionic locust
#

to drive servos higher than 6V with the PCA9685 chip (on a homebrewed PCB), is it just a question of putting the higher voltage on V+ (the middle pins of the pwm outputs)?

olive light
#

you also need to have the grounds of both circuits connected

ionic locust
#

is it necessary to put a mosfet on each pwm output?

olive light
#

no

#

the servo already has a motor driver chip inside

#

the signal it takes is very low current

ionic locust
#

ok thanks for your help!

idle grove
#

Dose anyone have any tips to get more holding torque out of a nema 17

fast tiger
#

You'll have to be more specific. NEMA 17 only describes the form factor, not the characteristics of the motor itself.

#

That being said, you might be able to get your driver to shove more current through it, but you'll have to check the datasheet to see what the maximum holding current is.

fresh flicker
#

Hello experts
This question is to those of you have experience with servos:

  • I want to make a self resetting airsoft target popper. I use a steel plate of roughly 1 kilo which will have a servo attaced at the bottom to raise from horizontal to vertical and vice versa.
    How strong would such a servo need to be?
olive light
#

depends on how the force is transmitted

#

the general rule for a simple lever is the length of the lever times the weight of the thing at its end

native cipher
# fresh flicker Hello experts This question is to those of you have experience with servos: - I ...

1 kilo is a lot for servos.
Common "high strength" servos are 20kg-cm, meaning they can provide torque of 20 kg with arm of 1 cm (or, more realistically, 1 kg at 20 cm lever). I do not know how large your plate is, but it looks like it is at the limit of what such a servo can do - and operating close to the limit means that you risk breaking the servo geartrain. You can look for an even more powerful one, like 35 kg, or use a motor with some gearing.

In any case, I strongly urge you NOT to mount the plate directly on the servo. Instead, make a construction where the plate is attached by some hinge/pivot point to something solid, and the servo uses a linkage to raise/lower it. This way, when you airgun pellet hits the plate, the impact would not be transmitted directly to servo axle.

fresh flicker
#

@native cipher Ok, good advice. The plate I have in mind is about 350mm x 135mm, 2mm steel. The weight was calculated to just over 1 kg.
From reading on various webpages, it became clear that even such a relatively small weight is difficult for hobby servos to rotate vertically from one end.
Maybe the solution is to rotate around the center instead, so the plate can be its own counterweight.

native cipher
#

I'd do something like this

#

calculating the length of arm so that when the plate is horizontal, arm goes straight down

split tulip
#

Is there a way to know if a solenoid’s arm has been blocked, other than putting a detector on the thing the solenoid arm is moving?

fresh flicker
#

Thanks @native cipher

cursive terrace
olive light
#

I know it's possible to tell if the core is inside the coil or not by measuring how the current changes when you switch it on or off, but I have no idea how to do this in practice

cursive terrace
#

i don't know if you can actually measure any real difference in the electrical values across the solenoid based on where the actuator is within the housing -- any EE's have insights?

#

hmm, the "best" answer i can locate is to use a linear potentiometer or similar external measurement system

olive light
#

that's certainly going to be the most accurate and reliable

#

an optical sensor could work too

split tulip
olive light
#

as long as your motor controller supports it

primal shell
cursive terrace
split tulip
#

The third state for a servo is “jammed halfway to full extension”. :). Perhaps there’s a circuit that measures the current or voltage ramp to determine if the coil’s mag field fully interacted with the rod, IOW, some circuit that measures the conversion of potential energy to kinetic energy. [Kickback is experienced in the opposite direction, but things have to be symmetric, right?]

I think that a servo might be too slow, but I got one to test its speed. I can use gears or levers to speed it up, I guess.

fading void
#

Are servos like this easy to break via code? https://www.adafruit.com/product/1449 Sometime between opening the servo's package, calibrating via circuitpython, and testing my two way telegraph code my servo seized up.

Analog Feedback Micro Servo - Plastic Gear

It looks like a micro servo, it acts like a micro servo, but it's more than just a micro servo! We got a factory to custom-make these classic 'micro' sized hobby servos with a ...

split tulip
# fading void Are servos like this easy to break via code? https://www.adafruit.com/product/14...

I’ve read that if there is a valid PWM range per servo type; if exceeded, the motor can be deformed. If you can’t find the datasheet (which is hard for Adafruit parts since they don’t even identify the model), try to find something that’s about the same shape and size which DOES have a datasheet,

I’ve also read that you can experiment with the range to find its extents. Start at the middle of the PWM range and move slowly until the arm reaches the published extent, then back off a tiny bit. Do the same for the other direction.

native cipher
#

range of 1000 us to 2000 us should be safe for all servos.
But most plastic gear servos can be destroyed if stalled for prolonged time - I. e. if the servo is prevented from reaching the requested position by some mechanical obstacle

modest canyon
#

ah, yay! theres an adafruit discord
im eyeing the playground express + servo controller since it doesn't need soldering, but my bot has 5 servos whereas the circuit can only control 4. any suggestions?
i cannot solder, i live in a small dorm room 🥲 . i have checked local makerspaces and they either don't have one or its a ridiculous membership price

he's currently on a pi and lynxmotion ssc-32 but i have another project that would benefit from his boards more, and i'm looking to shrink + simplify a bit

cursive terrace
#

i'm using Raspberry Pi's and various `hats that have pre-soldered headers -- there are various types of the same basic variety that can be had from a lot of sources

turbid trail
#

why can't you solder in your dorm room? is there a rule?

modest canyon
#

I'll check those out, thank you!

Yeah >-> a rule against it and I don't want to risk paying a fine

modest canyon
cursive terrace
#

i have a similar restriction (apartment), so i've got a Sparkfun servo hat, which is effectively identical to the servo bonnet but pre-soldered with a "bonus" I2C STEMMA port

modest canyon
#

oh, I came across that one and thought about it!!

#

I'll take a closer look

cursive terrace
#

i would recommend using two power supplies, nevertheless -- one for the pi and one for the 'hat

split tulip
#

What circuits are used to isolate the power coming from the computer (via the serial lines: USB or FTDI programmer) from the power from the wall, such that we don’t send 1+ A back to the computer’s USB socket from the MCU (which is getting power from the wall because it needs more than 0.5A)?

I’ve been reading that USB lines are quite dirty — does this refer to the power or the signal lines?

olive light
#

see chapter 4.5 Powering Pico

#

most of this will apply to any other microcontroller as well

primal shell
#

Or ... a simple diode

fast tiger
#

If that 1A is on the ground line you definitely need an isolator.

split tulip
# primal shell This, for one: <https://www.adafruit.com/product/2107>

So the order would be PC USB (0.5A) -> FTDI -> /2107-> ESP32 ?

And the main power would be: Wall Wort (4A) -> motor driver power/GND.

And I’d tie the ground wires together after the /2107 and the wall wort?

When I’m not debugging, remove the FTDI and all else remains the same.

The only issue is that the /2107 only passes 100mA, which is less than the ESP32 (needs 310mA)… is there some way to split off the main power to feed it to the ESP32 while still accessing it’s TxD/RxD pins?

split tulip
#

Another option is to drop the USB connection and use OTA updates and read/write the serial monitor via the web (WebSerial).

primal shell
merry zephyr
split tulip
turbid trail
primal shell
split tulip
#

Say I have a couple of t-slot interruptors and am out of GPIO pins… how else might I hook that up?

I have a I2C on the GPIOs already… I thought maybe I could use an I2C circuit… are there “generic” boards, that is, that work for any digital signal and so I wire board to a t-slot? Or do I have to find a I2C version of the t-slot somewhere?

fast tiger
#

I'm a little unclear on what you are asking, but I2C GPIO expanders do exist, and some of them even have an additional interrupt pin that can be triggered when one of the pin states on the expander changes.

split tulip
fast tiger
split tulip
fast tiger
#

I'm assuming by t-slot you just mean the aluminum profile that is presumably part of the frame on your robot. Regardless, my interpretation is that you are trying to wire up a limit switch (or several), and that you want to know if there's a board you can use to poll all of those switches using I2C.

#

If that's the case, an I2C expander will be fine.

split tulip
fast tiger
#

Yeah

forest fjord
#

Hi there, i'm working on a scara arm robot project, using cncshield v3 and grbl-scara https://github.com/jared-hughes/grbl-scara.
I've hooked it up with universal gcode sender and the x and y axis seems to work fine, but the z axis doesn't seem to work. I've tried looking into the source code and found that the forward kinematics implemented at the system.c seems to account for z axis. Any ideas on what am i supposed to do to make the z axis work?

📎 config.h
olive light
#

what do you mean by "doesn't work" exactly?

forest fjord
#

the stepper for z axis doesn't move even when i change the Z coordinate in the g code

olive light
#

but is it powered?

forest fjord
#

yes

#

i almost forgot, the z axis works if i use the regular grbl 1.1

olive light
#

I don't really have experience with either personally, but the docs seem to imply that the z axis is optional, perhaps it's disabled in the scara defaults somehow?

little hedge
#

Hey, I have a question about stepper motors since I'm not too experienced with them and I'm unsure if a stepper motor is the right choice for my project, but would this NEMA-17 motor be able to handle speeds of roughly 150-210 RPM?
https://www.adafruit.com/product/324

Stepper motor - NEMA-17 size - 200 steps/rev, 12V 350mA

A stepper motor to satisfy all your robotics needs! This 4-wire bipolar stepper has 1.8° per step for smooth motion and a nice holding torque. The motor was specified to have a max ...

forest fjord
cursive terrace
vocal bear
#

How practical would it be to create a curved rail or set of rails for something to roll on? Sort of like a roller coaster but without any loops

#

Buy something I mean something small, maybe the size of a fist

cursive terrace
#

lego technic has things like turntables that come in various sizes (you can probably get "knock off" copies cheaper but you didn't hear that from me)

#

or even just a small set of bearings like you'd use to put a small turntable ("lazy susan") toegether with

native cipher
# vocal bear How practical would it be to create a curved rail or set of rails for something ...

would this work?
https://www.amazon.com/Typhoon-Frenzy-Roller-Coaster-Building/dp/B00E6HEQ3K/ref=sr_1_4?dib=eyJ2IjoiMSJ9._97A5-2xoOZEG_6WyqoXJBbHx5n0iEmIIcxmAm_uMfmhpuXDHtl2cZmrKjNVntG_nH9wJMkbiAoQomeFTR1dp79MrKRm7eAaF4kNdZnikHbJxG0GLcPyNTgZtAD6I6u51MoZh3RpqUaKTLHj8ParayxDK_T4LEvbZv-8WUM1eDceO9q9KZRWb_lQN_fgWEt7gVg63PPEsyj-1q6yp3dmokgrx6wHMd6jbr0O6wqSrt5JUHWhafAhP5G9VfsEFlJ7I27LVnp0jEpJE7sTExcXzHoYSub_iHJYgikf-2uHmMg.WkbKBTYn6K8V2KEaUWpwNi8Hbs_mEczf1auzOmfPeUw&dib_tag=se&keywords=knex+roller+coasters&qid=1709001747&sr=8-4

K'Nex 51438 TYPHOON FRENZY ROLLER COASTER BUILDING SET Building Kit...

Bring a wave of fun from the ocean to your living room with the new Typhoon Frenzy Roller Coaster Building Set! Build with over 640 K'NEX parts & pieces like rods, connectors and track, and add motorized power to send the silver coaster car twisting and turning through over 19 feet of swirling, n...

#

actually here is a better link
https://knex.parts/subject/roller-coasters/

Knex.Parts

K’NEX roller coasters are one of the most popular types of K’NEX building sets. These sets allow builders to construct their own miniature roller coasters complete with loops, twists, turns, and other thrilling features. K’NEX roller coasters come with a range of specialized parts, including coaster cars, tracks, supports, and other accessories,...

light badger
#

hey, is anyone able to help with the coding of my robotics project. Instead of stopping the motors when the red target is detected the 2nd time, I want it to pause for 5 seconds and then go forward again. Can anyone help? would be appreciated

📎 message.txt
cursive terrace
split tulip
#

In the seesaw_servo code of the Adafruit_seesaw lib, why are the min & max microsecond pulse widths multiplied by 3.27? Is this factor related to the PWM frequency somehow?

native cipher
#

This seems to be conversion from pusle duration to values used for analog write.
E.g. pulse duration of 1000 us = 1ms corresponds to duty cycle of 1/20 =0.05 (because servos use 50 hz frequency, hence 20 ms cycle). Assuming that analogWrite uses 16-bit numbers, 0.05*2^16=3276.8 Thus, 1000 us corresponds to value of 3276.8 for analog write

split tulip
split tulip
#

At a 50 hz frequency, I should see a rising edge every 20 ms, and then it should fall 1-2 ms later. But the oscilloscope shows a rising edge every 1 ms. It’s almost as if the frequency call was ignored, and it used 100000 hz instead.

This is using the Adafruit_seesaw and seesaw_servo libs. Default servo.attach(11) — that is, no min/max set, so it defaults to 1000..2000. And a servo.write(0) — so it should be the 1000 us duty cycle

[Could I be using the scope wrong? Maybe the ATtiny chip is bad?]


// test ATtiny over I2C to servo

// For I2C Bus
#include <Wire.h>

#include "Adafruit_seesaw.h"
#include "seesaw_servo.h"

// define a I2C wire that will be assigned to pins 14,15
#define I2C_SDA 14  // runs on pins 14,15
#define I2C_SCL 15
TwoWire I2C_BUS0(0);  // use I2C bus #0

/// ATtiny
Adafruit_seesaw ATtiny(&I2C_BUS0);
#define SERVO_PIN 11  // ATtiny pin for servo
seesaw_Servo servo(&ATtiny);

int errorCount;
void setupDevices() {
  // start I2C
  I2C_BUS0.setPins(I2C_SDA, I2C_SCL);  // side effect: ATtiny uses this bus
  Serial.println("I2C: pins set");

  // start ATtiny
  if (!ATtiny.begin()) {
    Serial.println("Error.  ATtiny is refusing to start.");
    errorCount++;
  } else {
    Serial.println("ATtiny began");
  }

  //servo.begin();  // not needed if ATtiny already started
  servo.attach(SERVO_PIN);  // 1000,2000 are the default widths

  if (errorCount) {  // FATAL ERROR! HANG the ESP
    Serial.printf("FATAL: %d error(s)\n", errorCount);
    while (1) delay(100);
  }
}

int angle, delta = 10;
void loop() {
  Serial.printf("angle: %d\n", angle);
  servo.write(angle);
  angle += delta;
  if (!angle || angle == 180) {delta = -delta;}
  delay(300);
}

///////////////////////////

void setup() {
  Serial.begin(921600);
  delay(1000);
  while (!Serial) {};
  Serial.println("TESTING: ATtiny(servo)...");

  setupDevices();

  Serial.println("Looping...");
}
IMG_1027.jpg
native cipher
#

which ATtiny chip is it? could it be that the Arduino core for it it doesnt support changing pwm frequency?

split tulip
jade lichen
#

https://www.youtube.com/watch?v=4X9ORR-z_tY isn't this neat? can someone tag me in some literature and videos on how to make my TTS in OpenDAN translate over to animatronics

YouTube
DisneyResearchHub
Bend-It: Design and Fabrication of Kinetic Wire Characters

Elastically deforming wire structures are lightweight, durable, and can be bent within minutes using CNC bending machines. We present a computational technique for the design of kinetic wire characters, tailored for fabrication on consumer-grade hardware. Our technique takes as input a network of curves or a skeletal animation, then estimates a ...

▶ Play video
honest badger
#

hey, so, to clock a certain CMOS chip i'm using in a project, I need a stable square wave whose frequency I can set on-the-fly in the 300 kHz to 3MHz range. Ideally I want something with a digital serial input that determines the output frequency. Any suggestions for parts that can do this?

olive light
honest badger
#

So, I've been using an arduino nano for my project, and unfortunately there are some frequency limitations w/ the timer. There's some trickery you can do to get higher frequencies out of the timer divisions, but that ends up messing with other functions like delay() and millis() which are important for my application. An external DCO feels like the simplest answer in my context, I just need to source the right part.

olive light
#

I was thinking about using a dedicated microcontroller for this

#

then you don't care about delay and millis

#

iirc you should be able to do 8MHz pwm on the atmega328

honest badger
#

A single DCO would be cheaper and smaller and require less time programming imo

#

I've got some pretty tight space and cost limits here

olive light
#

if you say so

#

I would just go with an attiny or something like samd9

#

but if you can find a cheaper part, all the better

mossy crest
honest badger
#

also, the big green breakout thing is empty and just represents how things would be routed

mossy crest
honest badger
#

for context, one kink is that I need the clock to be controllable with a 1v/octave control voltage source. I would use the arduino to read an input voltage and translate that to a frequency at which I would set the clock output.

olive light
#

why do you need 3MHz? that's way beyond hearing range

mossy crest
honest badger
#

I've been thinking about this timer division thing-- here's what I have

const uint32_t kHz = 1000;
const uint32_t MHz = 1000000;
class rfOsc(){
  public: 
    // Hardware info
    uint32_t systemClockFreq = 8*MHz;
    int counterBits = 16;
    int prescalerDivisor = 1;
    // Timer calculations
    float systemTickPeriod = 1/CPUclockRate;
    float timerClockFreq = systemClockFreq/prescalerDivisor;
    float timerTickPeriod = 1/timerClockFreq;
    float timerRolloverDur = tickPeriod*10^counterBits;

    uint32_t defaultFreq = 720*kHz;
  
  // Calculate the number of ticks needed to produce a given
  // frequency, then set the timer period to that number.
  void setFreq(float frequency){
    uint32_t freq = frequency;
    float period = 1/freq;
    uint32_t ticks = round(timerTickPeriod/maxFreqPeriod,0);
    tim.setPeriod(ticks)
  }
}
olive light
#

float

honest badger
#

gonna make some changes to the types here for space obvs, this is a first draft

#

also i need to set up a timer handler

mossy crest
honest badger
#

another relevant factor here is that I just started learning C++ in January and have never worked with hardware on this level before. I understand how a DCO works, and I sort of understand how the Arduino timer registers work. But after a couple days of trying to figure out how to get an RF rate square wave out of the arduino, I feel like I'm maybe halfway there. Which is why my original question was "does anyone know of a DCO that works at these rates"-- from my perspective, that solution would have been trivial to implement

#

i don't know what an ISR is or what it means to double buffer a compare register

olive light
#

I would start by searching at lcsc, mouser or digikey

mossy crest
honest badger
olive light
#

microcontrollers are cheaper

mossy crest
#

with a 8MHz CPU clock (Arduino Nano), frequency accuracy will be limited toward the upper end of your desired range, so i guess there's that drawback

fast tiger
#

What are we actually trying to accomplish here?

mossy crest
#

looks like 1V/octave control of a speech synthesizer's pitch?

olive light
#

then use a 48MHz microcontroller

honest badger
#

I've looked a little into using an STM32 but that presents all the same challenges as the Arduino but Bigger

mossy crest
honest badger
#

since this is for musical applications, i'd express it in cents https://en.wikipedia.org/wiki/Cent_(music)

Cent (music)

The cent is a logarithmic unit of measure used for musical intervals. Twelve-tone equal temperament divides the octave into 12 semitones of 100 cents each. Typically, cents are used to express small intervals, to check intonation, or to compare the sizes of comparable intervals in different tuning systems. For humans, a single cent is too small ...

fast tiger
primal shell
#

Amusingly, I'm building a circuit for musical applications, I'm currently looking at a wide word width non-integral frequency divider implemented on an FPGA.

honest badger
#

so, the chip i'm using also has an internal clock whose freq is set by an RC pair on one of the pins. so i could just do some current injection at that point to set the internal oscillator. however, the arduino's dac output is only 8 bits, which poses some potential tuning problems

mossy crest
#

yeah on a 8MHz AVR with low-level timer programming, you get 4MHz, 2MHz, 1.33MHz, 1MHz… at least when using the CPU clock. the ATmega32U4 has a high-speed Timer4 that can support up to 64MHz operation (from an internal PLL, i think)

honest badger
#

i've also considered using a 4046 vco chip and somewhat complicated 1v/octave circuit to tune it. but that requires a lot of calibration to get right, and i'm not really sure how well it will behave at RF rates

mossy crest
#

4046 is a classic broadcast FM demodulator, i think. no idea how well it does running its VCO open-loop

honest badger
#

here's the 1v/oct tuning circuit.

Screenshot_2024-03-23_at_3.34.52_PM.png
#

Q1 and Q2 here are doing the linear to logarithmic response shaping

mossy crest
#

ATmega328 doesn't have a true DAC, only PWM, so you'd still want to filter and buffer its "analog" output if you want to feed it to a VCO

honest badger
#

at this point using Timer1 on the nano is still on the table, but it's pretty far beyond my abilities right now. i don't know how to configure the timer to do what i want, or how to turn that into a square wave output to one of the pins

mossy crest
#

at 8MHz, your Timer1 tuning increment is 250ns. that might get you acceptable frequency accuracy near 300kHz, but not near 3MHz

honest badger
#

in this case they're using timer 2, which is 8 bits. but I could just change it to use timer 1, right?

mossy crest
#

if you’re only going as low as 300kHz, you might only need 8 bits

#

you’d have to change the registers and constants for Timer1, but the idea is similar

honest badger
#

right now this code doesn't use a prescaler and limits you to 256 increments ```c
void Geekble_Oscillator::Divide_8MHz_by(uint16_t Divide_Variable)
{
// Output Frequency = 8MHz / Divide_Variable
// Divide_Variable 1 ~ 256
if (Divide_Variable < 1)
{
Divide_Variable = 1;
}
if (256 < Divide_Variable)
{
Divide_Variable = 256;
}

OCR2A = (Divide_Variable - 1);
TCNT2 = 0;
}```

#

prescaling to 4MHz and using timer1's 16 bit counter would provide more definition, right?

mossy crest
#

if OCR2A is zero, the output toggles on every clock cycle, giving you 4MHz

#

you could also use Timer1 at higher resolution but lower frequencies, then multiply it up using the 4046 as a multiplying PLL (sorry, i don’t know how to do that off hand)

mossy crest
#

you can get a bit more resolution by giving up on having a 50% duty cycle, i think

honest badger
#

So, if I'm trying to get the timer to reset when its value matches OCR1A, I want these two bits to be 10, right?

Screenshot_2024-03-23_at_8.23.37_PM.png
#

I'm unclear on what it means to toggle OC1A

#

err, wait, since this table describes what these bits do in Phase Correct and Phase and Frequency Correct PWM modes, but i'm using CTC mode, then I can just leave these as zero?

mossy crest
#

the WGM bits determine the counting sequence, while the COM bits determine how the output pins change

#

for your application, you probably want WGM 15, and set OC1A to toggle

honest badger
#

If it's toggling, does that mean I'm actually getting half the CPU clock rate, even with the prescaler set to 1?

#

I don't really need PWM for my case

#

the main thing i need is tight freq control

Screenshot_2024-03-23_at_9.03.13_PM.png
honest badger
#

oh boy. yeah i don't think this is gonna work

Screenshot_2024-03-23_at_9.19.08_PM.png
#

I need a much smoother curve than that for musical purposes

#

i.e. probly gonna have to switch to an SMT32 if this is how i end up clocking the speech chip

primal shell
#

Yeah, integral division can only go so far. Which is why I eventually shifted to non-integral division, but I'm doing that with additional hardware.

merry zephyr
#

Hi, i have weird problem that i cant solve this. I have RPi 4B, PCA9685. Im controlling servos with Mediapipe library and picamera2. While my code must move servo its happening, but my camera freezing for time that my servo is moving. Does anyone had similar problem or can help me with solving this?

cursive terrace
mortal tide
#

I'm looking for an off-the-shelf but very hacker friendly robot chassis with SLAM. Any recommendations? Something like this looks interesting but the lack of documentation and a community turns me off https://ca.robotshop.com/products/athena-professional-robot-platform

RobotShop Canada
Athena Professional Robot Platform

Description Athena Professional Robot Platform Offers rich configuration and standard hardware interfaces Comes with brand new VSLAM and traditional SLAM based on lidar Equipped with a depth camera, ultrasonic sensor, cliff sensor, and bumper Convenient and flexible for a more narrow environment Supports physical magne

olive light
#

didn't irobot had a version of roomba dedicated for hacking?

mortal tide
#

I thought that was decades ago but I will research - thanks

olive light
#

I'm kinda frozen in time here

#

exactly a decade

#

doesn't seem to have a lidar

mortal tide
#

I love some assembly! VIAM seems to be a cloud based product. I'd prefer that all code reside in the robot chassis.

#

Related, do I recall correctly that Intel ended the Jetson series?

olive light
#

well, I meant the hardware, of course you would use your own code

mortal tide
#

ah

olive light
#

it seems to be compatible with ROS, which has its own SLAM libs

mortal tide
#

For hardware I'm all set. I have a couple of robot chassis lying around waiting for SLAM. I had thought it would save time to buy a chassis with SLAM, but perhaps I'm wrong. The bottom line is that I have some budget and I want to get SLAM as quickly as possible

olive light
#

if you can run ROS on those robots, you should be able to get started easily

#

I don't have experience with ROS myself, so I can't help with it much

mortal tide
#

easy enough to drop in a RPI but I thought I'd need something more powerful. Thanks for the advice!

olive light
#

it doesn't have to be fast...

#

you will probably see the limitations for yourself once you start with it, and then you can decide on upgrades

mortal tide
#

good point. I need to learn more before I can upgrade. I need to learn more so that I can ask more intelligent questions 🙂

mortal tide
fresh tinsel
#

Is it possible to drive a timing belt with continuous servo motor?

olive light
#

as long as you don't need precise timing on it...

#

I assume you are not asking about an actual combustion engine here?

native cipher
mortal tide
#

In general, a belt shouldn't care what kind of motor is driving it, as long as the motor has enough torque to turn the load. You will have to find appropriate pulleys that fit the motor and the belt.

prisma garnet
#

Looking for a straightforward way to build a talking skeleton. Ideally the hardware would analyze an audio source (OGG file in storage) and move a servo accordingly to control the jaw. Does Adafruit have anything "off the shelf" that can do this? I'm not seeing anything that I think would work straight away. Wondering if one of the prop featherwings can do it? I don't care if it's raspberry pi based or arduino or something else - would just like a "plug and play" if that exists.

#

Extra credit if it can also control a second servo for neck movement and flash LEDs for eyes. 😀

#

Also now I'm wondering if an audio line in would work instead of pre-recorded audio.

primal shell
#

At a minimum, you could just use the current volume of the audio

cursive terrace
#

you might try looking at the "sound reactive" projects in the learn guide as a start

olive light
#

People are not opening and closing their jaws when they speak, we do it with the lips, the tongue and the vocal chords, which the skeleton lacks. You can easily test it by putting your finger against your jaw and trying to speak - you will notice that it opens and then barely moves at all when you speak.

prisma garnet
#

Yeah I think it would have to move on noise and close on silences or pauses between words.

#

I didn't know if adafruit had something like that with processing onboard. I know they have some prop related boards but didn't see anything like this.

prisma garnet
#

Would be great if they had this in something like their Prop Maker RP2040 board.

primal shell
#

You have two basic choices: one is to digitize the audio and then do processing on it to extract the information you want. The other is to have a simple envelope detector (not much to it, a diode, capacitor, and resistor) that outputs the loudness of the input sound, then just read that with an analog port.

sly sand
#

@primal shell sir i need your help with my project

primal shell
sly sand
primal shell
#

Generally we stay on the server so multiple people can benefit from the conversation

sly sand
#

oh yeah sure

#

i want to connect an ESP8266 wifi module to a raspberry pi zero 2 W

#

i only those 4 basic ones "VCC to 3.3v" "GND" "TX-RX" and "RX-TX", what about the RST, CH-PD, GPIO2 and GPIO0?

image.png
vocal bear
#

Oops wrong one

vagrant thunder
vagrant thunder
sly sand
unborn fulcrum
#

Hello!
I have an issue with a project I am working on.
I am using an IMU with an NXP filter to get a quaternion, for which I am using for other purposes. I am using the LSM6DSOX + LIS3MDL IMU. I am using the NXP filter to calculate the quaternion.
My issue that the yaw value also seems to go crazy. I have recalibrated my IMU with MotionCal multiple times, with no luck. I have noticed that if I read the IMU values and put them through the NXP filter at a slower pace (using a delay), that the problem is a lot less severe, but still gets worse over time. In my application, I cannot use the delay to fix the issue.
Any idea of what could be happening here?

native cipher
#

I'd also try using other data fusion algorithms, such as madgwick - I had good luck using code from https://github.com/kriswiner

GitHub
kriswiner - Overview

One-man maker shop with a focus on appallingly small, value-added products. Specialties include motion sensing and motion control applications. - kriswiner

sly sand
#

is this correct?

image.png
sly sand
#

Can anyone help me please

inland bear
# sly sand is this correct?

This image is blurry and it is not possible to decypher what the pins are called on the esp8266. You have also changed your circuit layout completely without mentioning why you have additional wiring or what you are now trying to achieve compared to the previous circuit diagram. These things make it harder to help you

sly sand
#

it's just a screenshot from a website, not actually mine. what he did there is. GND to GND, VCC to 3.3V, CH_PD to a resistance of 10k and then connect it with the same pin as VCC, TX to RX, RX to TX, RST to SDA and GPIO0 to SDL

cursive helm
inland bear
#

Also agree, I still don't think you've made it clear why you are trying to use an esp8266 with the piZero2W as that pi has wifi too. If you are planning to use the esp for some custom functionality then you probably need to gain quite a bit more knowledge about the ESPs, the protocols you plan to use, wiring schemes, and generally the project requirements (break them down into smaller pieces/steps), to the point where you could answer most of the questions you have, or are able to ask more specific/detailed ones.

#

Similarly the screenshot and your last reply doesn't add any context to your question of "is this correct"

sly sand
#

will it help if i gave you the purpose of my robot?

inland bear
#

The purpose of your project would be useful, it helps a great deal to know why as well as how you plan to achieve something

cursive helm
sly sand
#

so my robot is to analyse a network. Using the IP protocol, the robot can send requests to IP addresses on the network to check if a device is active. It can also use the ARP protocol to obtain the MAC addresses of connected devices, which uniquely identifies each device on the network.

Once a device is detected, the robot can query the SSDP service to get additional information about that device. The SSDP protocol allows devices to broadcast announcements on the network, indicating their presence and providing details about their features and services.

By gathering this information, the robot can create a list of active devices on the network, providing details such as IP addresses, device names, and software versions. This information can be useful for network management, detecting unauthorized devices, or troubleshooting connectivity issues.
does that help?

inland bear
#

It does!
I don't think of that as a robot, but that might be my short-sightedness... It sounds like the piZero2W which has onboard wifi could send requests and listen for those packets on the network, and assuming they are sent by machines (not all will send as much as you'd like, identifying software versions gets complex). You could then present that on a display, or send an email alert, or host a status webpage on the pi / externally or whatever your heart desires (time and effort permitting).

sly sand
#

the only thing will make it a robot cause i have to make it move, The idea may not be the most practical or viable option but I can't change it now cause I already submitted my project idea as a robot, so i'll be using an arduino uno with a l293d shield, 4 dc motors and ultrasonic sensors to move and control it maybe by phone, I made stuff so complicated when the idea is so easy to make

#

so what do you think? should i remove the wifi and bluetooth modules and if i removed them how that will work

cursive helm
#

The Rapsberry Pi zero 2 W already has Bluetooth and Wifi so it is not clear why you need to add any hardware to support them. You also can drive your motors with the Pi using something like this https://www.adafruit.com/product/4280

#

But to be clear, I am not trying to redesign your project.. just trying to understand what you are trying to do.

sly sand
#

No, I really appreciate this, the adafruit dc and stepper motor is not available in my country for now. I will remove the wifi and bluetooth cause i didn't know about it. the raspberry pi will only have the OLED display 128*64 I2C then to show them if the connection was successful and an ethernet module (i don't know why but the prof required it so i have to add it). and the other componants in the arduino

#

now comes another 2 issues for me which are the communication between the pi and arduino, UART or I2C. And the other issue is the power supply

#

can you help me with that?

cursive helm
#

I would expect that a UART connection will be much simpler to implement. I'll defer to others to advise you on power options. There are any details to consider.

sly sand
#

is it okay for the UART communication even though there's and I2C display on the Pi?

#

i've been thinking of using a power bank 5v output between 2.0A to 3A, maybe Anker if i find it. what do you think ?

cursive helm
#

That should be fine to power the Pi and Uno, but you may want a separate battery pack to power the motors.

#

something like 4 AA baterries.

sly sand
#

yes i will use 2 Li-iion 3.5V each. is that fine?

cursive helm
#

That is what I have on mine 😉

sly sand
#

i don't have to use any resistance or anything right? just a switch

cursive helm
#

For the motors? As long as they can run on the voltage, nothing is needed. The Pi needs a regulated supply.

sly sand
#

but I have a question. when i will turn on the power connected to the motor shield, isn't the pi gonna try to turn on too?

cursive helm
#

That depends on how it is connected. And this is getting beyond my "comfort zone" in providing help. Hopefully other more knowledgable folks will step in.

sly sand
#

it's ok, really appreciate your help

cursive helm
#

You're welcome. Good luck!

sly sand
cursive helm
#

You should be able to use both the UART and I2C interfaces at the same time.

sly sand
#

oh I have a question, the display will also need to use the I2C pins, will it be any problem if i connected to whatever GPIO and then in coding make it an I2C ?

cursive helm
#

I don't really understand the question. You can have multiple I2C devices using the same set of I2C pins. I have only used the default I2C pins on my Raspberry Pi's.

sly sand
#

there's more than one I2C pins on the pi?

#

i only saw the I2C on the GPIO02 and GPIO03

cursive helm
#

That is all I have ever used. If I understand correctly, some Raspberry Pi's do support more than 1 I2C bus, but I have not tried it.

#

But I have often had multiple I2C devices connected. As long as they have unique addresses, then can share the same I2C bus.

sly sand
#

I think that would be beyond my level of studying

cursive helm
#

I don't think it is necessary. The default I2C pins should work fine.

sly sand
#

yeah because we just like i will use them for the arduino

cursive helm
#

Besides, I think it best to use UART to talk to the UNO

sly sand
#

i also think so cause is the easiest

#

but isn't the I2C display will not work because there's UART communication?

cursive helm
#

I don't understand... the display uses I2C and is independent of the UART.

sly sand
#

yeah yeah that's what i'm asking about

#

like the display uses I2C right? then the communication between the pi and arduino uno doesn't have a relation with that?

cursive helm
#

That is correct. I'm sorry, I have to go offline. I hope others can help.

sly sand
#

what's the correct, which option? thank you bro for your knowledge and sorry for the disturb

cursive helm
#

Glad to help... I meant that it is correct that the display and UART have no relation to eachother.

sly sand
#

ah ok. thank you so much

#

@inland bear Sorry, do you have some knowledge about the powering the Pi? like mate just said i can do power bank to power the Pi and use another battery for the motors.

inland bear
#

not an expert here, but my experience was that as long as you join the grounds (the negative terminal of the combined battery and the ground pin from the Pi and another to arduino ground pin) you'll be able to have reliable communication (all communication levels reference a shared ground level), and have different voltage levels for different components. Also battery bank should be fine, but you'll possibly find that the standard 2.4amps isn't enough for all the things.

sly sand
#

oh really? I thought it was enough cause i calculated the sum of the current of all components and it was about 1100 mA then took 8000mAH of the power bank and divided it by 1000 mA and gave me 7.2 H and also 1100mA is under 2400mA, i think that would be fine right?

inland bear
#

Might be fine then

inland bear
sly sand
#

oh that helps. thank you so much

craggy cargo
#

Im using the adafruit MotorKit() (from the adafruit_motorkit library) and it connects successfully. But when trying to power one of the motors (kit.motor1.throttle = 1), I’m getting an OSError: Errno 121 Remote I/O Error.

Is this usually a problem because of a malfunctioning motorkit, the motors, or the connection between them?

primal shell
#

What platform is this on? Most microcontrollers aren't running an OS and wouldn't produce OS errors.

craggy cargo
primal shell
#

That's probably an I2C wiring error of some sort (a problem with pull-up resistors, missing reset line, etc.)

cursive terrace
vocal bear
#

I spoke with support for this item, and they told me that this encoder has a differential output despite the product page and the data sheet being kind of In conflict with each other. The product page suggests single ended with 0.2V as low and 5 as high. I'm trying to figure out how to pick off the Z output and use it as a 0. Thoughts? https://www.omc-stepperonline.com/nema-23-closed-loop-stepper-motor-1-2nm-166-7oz-in-with-magnetic-encoder-1000ppr-4000cpr-23hs22-2804-me1k

📎 23HS22-2804-ME1K_Full_Datasheet_1.pdf
olive light
cursive terrace
#

ow (first thing that came to mind, but my back hurts today, so...)

olive light
#

don't want to be near when the system crashes and all the motors start to spin at once

sly sand
#

Hi guys, I have a question. There’s a motor shield l293d sit on the arduino Uno. And I want to connect the arduino uno with the raspberry pi zero w and I need gnd between them. Can I use the GND pins from the motor shield ?

turbid trail
#

you could easily test this with a multimeter as well

sly sand
#

Thank you sir

grave briar
#

Hi, my friends and I are making a project of which a part is a robotic arm that can pick up and place objects. I wanted some help regarding robotic kinematics, to design the algorithm for pick and place actions. If someone has experience with this could you kindly help us a bit as we have a deadline approaching soon.

cursive terrace
marble jacinth
#

I want to drive a few motors from a ~9v rechargable battery, I'd also like to drive an esp32 controller from the same battery - what do I need to do in order to do that? I'm looking at the TB6612 breakout for driving my motors, but what would I need to do to pull power from the 9v safely for an esp32 (not chosen yet) microcontroller?

olive light
marble jacinth
#

I was looking at a NiMH, but if there are better options I'm happy to reconsider

olive light
#

what capacity?

marble jacinth
#

2000mAh

olive light
#

that might work as long as the motors are really small

#

you will need a buck converter to bring the voltage to 3.3V

#

(or down to 5V and let the dev board you use bring it down to 3.3V with the ldo it has built in)

marble jacinth
#

thanks - so is the worry that running the motors could drop the battery voltage and brown out the microcontroller? but if I have a big enough capacity battery all I need is the buck converter - both that and the TB6612 motor supply can be wired directly to the battery together?

olive light
#

as long as your motors are rated for 9V, yes

#

the esp32 chips are a bit of a pig for current, they can take even half an amp with the wifi on

#

oh, but with a buck concerter, you will actually need only 1/3 of that at 9V

native cipher
#

side remark: 9v NiMh will be bulky (they are nominally 1.2v, so 9 v is probably 6-8 cell). It might be easier to use a 2-cell LiPo (nominal 7.4v) instead - unless you really need max power for motors.

LiPo can also typically produce more amps of current than NiMh of ocmparable size.

turbid trail
topaz grail
grave briar
#

Has anyone tried implementing SLAM? I need help figuring out the steps to do the same for a project my friends and I are doing. We'd prefer doing it the easiest way since we're kind of short on time.

olive light
#

usually people just use ROS and the ready libraries it has, from what I can tell

sly sand
#

I have a problem, the black usb cable is for the powering and it’s working, the hdmi is working too but the white usb cable for my mouse, is on when I turn on the raspberry li but I can’t use it like when I move it, it doesn’t move on the screen. Am I doing something wrong?

IMG_1902.jpg
olive light
sly sand
marble jacinth
gleaming tree
#

Hellloo

#

I have two 18650 batteries so can I connect the to esp32 VIN pin via the 5v output of the l850n motor driver

#

Is this fine?? I am clueless,I am new to the world of esp32

fresh crown
#

Two batteries would add up to a peak of 8.4v I think.

native cipher
gleaming tree
native cipher
#

oh.
my first recommendation on using l298n is: don't use it.
It is very inefficient, so it easily overheats, and requires this huge heatsink.
the only good thing about it is that it is dirt cheap, but if you can afford to spare extra 5 bucks, get a better driver

#

I can suggest some options if you want

gleaming tree
#

Okay can u pls suggest some

#

Thank u

#

Also another issue I have been facing is that when I power on my ESP 32 it connects to the Wi-Fi properly and all but when I open the web server for the camera the stream is not showing

#

I can control the inbuilt LED tho

#

So does not seem like a connection issue

native cipher
# gleaming tree Okay can u pls suggest some

how many motors and what size? what is their stall current?
what voltage? are you using two 18650 connected in series (which would be 7.4v nominal, up to 8.4 v when fully charged)?

gleaming tree
#

700mA is the stall current and there will be 4 motors connected

gleaming tree
native cipher
#

(it is for two motors, so you will need two of them).
But it doesn't have 5v out, so you would need to add buck converter

native cipher
gleaming tree
placid parrot
#

Hello, I am looking for a 4 channel 12V BLDC motor driver with at least 2 -2.5A each channel for Raspberry Pi Zero or FeatherWing form factor ? Does Adafruit have such a board ?

primal shell
# placid parrot Hello, I am looking for a 4 channel 12V BLDC motor driver with at least 2 -2.5A ...

Generally, brushless motors are operated with "electronic speed controls" (ESCs). Common ones accept radio control style PWM signals for speed and direction, but there are some that can accept asynchronous serial or SPI input. Generating PWM with a Pi can be problematic as it's running an operating system and can have issues with timing critical things. Many Feathers can generate appropriate PWM easily enough, however. I haven't seen a lot of multichannel ESCs, normally you just use one for each motor. AdaFruit used to offer a huge (200A) one, these days I get them from radio control dealers, they're commodity items.

placid parrot
#

@primal shell thank you for your reply. I was planning on using a Raspberry Pi Zero for this use case. What other option would you recommend to use instead?

primal shell
#

You can use a Pi Zero, along with something like a B-G431B-ESC1 for each motor, modified for SPI control. My personal preference is Teensy boards for R/C stuff, along with PJRC's PulsePosition library to generate the PWM waveforms. In the Feather form factor, my go-to is the Feather M4 Express, it's an older board but with solid support. However, the Feather ecosystem contains a wide variety of boards and architectures, the RP2040 is powerful and cost effective, some of the others offer useful things like built-in WiFi.

olive light
#

there are also pwm shields for the raspberry pi

primal shell
#

Good point: offloading the PWM to specialized hardware is a solid approach

#

And the ability to use common, inexpensive ESCs is handy

placid parrot
#

@primal shell thanks for your advice. This is my BLDC motor Pinout. What is your recommendation?

IMG_3103.png
#

How should it be connected to the raspberry pi zero ?

#

I need to connect 4 BLDC motors to the raspberry pi Zero and control them simultaneously

#

I am looking for an easy plug and play solution if that’s possible

primal shell
#

Ah, that looks like a BLDC motor with a built-in ESC. You'd presumably hook a GPIO capable of PWM (or a PWM HAT) to the PWM input, and another GPIO to CW/CCW if you want the ability to reverse direction. I'm not sure what BRAKE does for that motor. Note that you should verify if those inputs support 3.3V logic.

terse forge
#

For a robot do you typically recommend dc motors or continuous servos? Seems like a servo could be run straight off an mcu which seems like a perk. @native cipher @primal shell

native cipher
#

I almost always use DC motors.

primal shell
#

I like the Teensy for R/C control because it supports the PJRC PulsePosition library. I like the M4 Express for smooth Arduino integration.

#

As for robot use, it depends on what I'm doing. For steering and actuator type duties, servos are a natural choice. For propulsion, I like DC motors (brushed or brushless).

native cipher
#

one can make a very small robot using continuous servos for wheel, but they are slow and are only suitable for really small robot

olive light
#

small robots are easier anyways

terse forge
terse forge
terse forge
primal shell
#

I don't know, I don't use Bluetooth. I think there are some Feather boards with Bluetooth.

native cipher
#

I have never looked at PulsePosition library, but controlling servos and motors using ESP32 is easy enough

#

(and if necessary, you can also set up quadrature encoders and closed loop mootr control, too)

terse forge
terse forge
#

Do you guys do things more in C or Python?

#

Python def makes life easy...

cursive terrace
#

the "obvious" recommendation with a lot of the adafruit stuff is to use CirciutPython, esp. on these types of boards

olive light
#

note that bluetooth is only fully supported on nrf52 with cp at the moment

inland bear
#

Could I combine a tracked tank, with a triangular body and this Crickit project https://learn.adafruit.com/crickit-flippy-robot to make something that would traverse a 6inch door rise? The existing tracked tank is only approximately 7inches by 9-wide (3 high maybe currently)

#

Last image shoes the traversal target, it's basically an I-beam that the greenhouse is sat on, plus the door ridge

#

Also has a battery comparment for 3cells in belly, so powerbank+cables are potentially temporary

primal shell
#

Maybe? Depends on grip/traction, wheel size, etc. It's a lot easier if you make a ramp for it.

inland bear
primal shell
#

There are ways to improve odds too, like adding grippy material to the vertical part of the sill.

inland bear
#

Oh that's also a great idea, relatively non damaging (rental)

#

Is there an Adafruit (or unofficial ) control app that uses usb UART like the blue fruit connect type thing? (Obvs no BT on esp yet, but maybe next week so wanted to try usb wired or WiFi but thought no app yet for WiFi)

turbid trail
#

not sure exactly what you want, but the controllee side on the mcu is just taking the UART bytes, so you could use that part and tether it over USB to some simple Python script. But I don't know of an app that would have preprogrammed buttons that send stuff over USB serial from a phone

#

<< wifi robot control app >> as a websearch comes up with some apps

#

I think

#

@inland bear 👆

inland bear
dense trench
#

How can I properly test an IR reflectance sensor that I plan to use as a line sensor for a robot? I'm using a Pololu QTRX-RC sensor, and the instructions say to drive the gpio line high for 10us, change it to input, then measure how long it takes until it's read as a 0. Their example oscilloscope traces show it taking 2.6ms over a black surface and 123us over a white surface.

When I try it though, it always takes the same amount of time no matter what's in front of it. The first time I read it after resetting the mcu it takes about 200us, and every time after that it takes about 100us. Does it need to be in a really dark environment or something to work properly?

native cipher
#

how close do you hold it to the surface?
normally it shoudl be very close to the surface measured, about 5mm is the typical distance

inland bear
#

Has anyone used the bluefruit connect app? I need victims test-pilots 😇
I've got a web based version and I need feedback / guidance on what people expect. I'm particularly keen to speak to any whose used the x/y/z or acceleration/quaternions/magnetometer and also charting (iOS only as far as i'm aware), but generally keen to ask people things they've done with bluefruit connect. Replies to this message or pings greatfully appreciated

dense trench
native cipher
#

try moving it closer and see if it helps.
Also note that what our eyes perceive as black might in fact be surprisingly reflective : a shiny black book cover can reflect more than 50% of incoming light.

inland bear
#

Actually that sensor says it has independent brightness control (independent of supply voltage) and that very reflective things can be picked up at max 30mm, but way less for less reflective

dense trench
#

I don't have the CTRL pin connected to anything since the website says "while driving it high (or allowing the board to pull it high) turns on the emitters with the board’s default (full) LED current." I'm able to see it through my phone camera but maybe it's not as bright as it could be since I only give it 3.3V? I don't know how it can give the LED a specific amount of current independent of voltage, like it'd need to somehow control a resistor to moderate the current it gives to the LED think_pepe

inland bear
native cipher
#

they most likely have a constant current led driver on the board

olive light
dense trench
#

Ahh gotcha, I didn't know that was a thing

dense trench
# native cipher try moving it closer and see if it helps. Also note that what our eyes perceive ...

It doesn't seem to matter how close it is to the object. The first picture is with a white laminated paper and the second is with a black cotton shirt. I started with the object touching the sensor then gradually moved it away while re-reading it and always got the same results in the high 90's.

The third picture is the first reading after resetting the mcu. That one is always close to 200us every time I reset it. I'd hope that could give me a clue as to what's going on but I don't see anything there. It's running the same code every time.

qtrx_over_white_laminated_paper.png qtrx_over_black_shirt.png qtrx_first_after_reset.png
native cipher
#

soemthing is obiously wrong, but I can't see what it could be

dense trench
#

Oh I noticed something. The blue line always crosses just under 1V at about 120us whether it's the first or a later reading. In any reading after the first, that's when the software reads the blue line as 0 then sets the green line to 0, which is what I've been measuring. In the first reading though it takes another ~70us for that.

The blue curve is still the exact same every time though distressed

proper island
#

Hello, I am having issues with a BNO055 sensor mounted on a robot. The problem is as follows: I have it configured in OPERATION_MODE_IMUPLUS with POWER_MODE_NORMAL, where I am using the accelerometer and gyroscope (that's why I chose this mode). I will also include the parameter sheet so you can see the other configurations I have for the sensor: bno055:
ros__parameters:
ros_topic_prefix: "bno055/"
connection_type: "uart"
uart_port: "/dev/imu"
uart_baudrate: 115200
uart_timeout: 0.1
data_query_frequency: 100
calib_status_frequency: 0.1
frame_id: "imu_link"
operation_mode: 0x08 #0x0C
placement_axis_remap: "P2"
acc_factor: 100.0
mag_factor: 16000000.0
gyr_factor: 900.0
set_offsets: false # set to true to use offsets below
offset_acc: [0xFFEC, 0x00A5, 0xFFE8]
offset_mag: [0xFFB4, 0xFE9E, 0x027D]
offset_gyr: [0x0002, 0xFFFF, 0xFFFF]
## Sensor standard deviation [x,y,z]
## Used to calculate covariance matrices
## defaults are used if parameters below are not provided
variance_acc: [0.117, 0.117, 0.117] # [m/s^2]
variance_angular_vel: [0.14, 0.14, 0.14] # [rad/s]
variance_orientation: [0.1159, 0.1159, 0.1159] # [rad]
# variance_mag: [0.0, 0.0, 0.0] # [Tesla]
The issue I am experiencing is that when the robot rotates on its own axis, meaning it only uses the gyroscope and not the accelerometer, it seems like the frame is not being updated. I can see the robot rotating, but the reference system remains static. I am not sure why this is happening and I need help because it causes the robot to lose track of its position since its odometry is a fusion of the information from the IMU and the wheel encoders.

dense trench
#

I got an N20 motor from Pololu and I'm confused about its RPM. At 100% duty cycle and 6V (which it's rated for), the output shaft rotation looks pretty slow and the math following the encoder output I observed puts the RPM at around 97.2 (which looks about right). The website says its no-load RPM should be 420, but I don't know where they get that from because my motor doesn't come close to it. Does anyone have pointers for using this kind of motor? I don't know if I'm doing something wrong or if that measurement on their website is in some super ideal situation or if my motor is just slow. https://www.pololu.com/product/5134

primal shell
mossy crest
#

do they mean RPM at the motor shaft, or at the gear reducer output?

dense trench
mossy crest
dense trench
dense trench
mossy crest
native cipher
visual shell
#

Does anyone have experience with Robot kit for Kids ? I did some basic research and found 4 possible path
Lego Spike Prime
Engino Robot https://shop.engino.com/search?q=Robotics&type=product
Vex Go https://www.vexrobotics.com/go
Makeblock mBot
Does anyone could help me understand better which one is easiest to learn and which one allow flexible project "ex: robot holding a pencil to draw on table"

This is for a smart 7 years old that already know how to program using Scratch blocks.

cursive terrace
#

my "concern" with the engino and vex go kits is that they're proprietary and you get locked into their ecosystem and they don't look like they expand beyond what that manufacturer makes

lego is "hackable" (i'm using it myself) and infinitely expandable, but I'm relying on 3d printing to make motor connections; makeblock looks more like a "traditional" entry point and if you your kid ends up wanting more, well, you're not locked into something proprietary

fast tiger
#

Lego is absolutely proprietary, but if you can do what you need with legos, go for it.

cursive terrace
#

well, yes, it's proprietary in one sense, but it's infinitely expandable (and 3d printable)

heck, if i had the money (and if my wife would even let me 😀 ), i'd get the Makeblock mBot Ultimate -- that looks pretty bad-*ss

olive light
forest cipher
#

You can make them, just don’t call them legos

dense trench
# mossy crest yeah i wouldn’t trust the driver output unless you’ve viewed it on a scope

I powered the motors directly with the power supply to give it a constant 6V, and both the power lines and the encoder output look the same on the scope as when powering it with 100% duty cycle through the motor driver. I tried this with a second one and get the same results so maybe they are behaving correctly. I think it'll be fine because they're basically the exact motors that a lot of people use for combat and sumo robots at the size I'm targeting. Actually, looking at a clip of someone's bot using similar motors, the speed looks like it matches what I'm seeing.

azure copper
#

I have a question about a dual h-bridge driver like this (https://www.amazon.com/dp/B00WSN98DC?ref=cm_sw_r_cp_ud_dp_ZQ1QGD0BRPS70SP1NTKG&ref_=cm_sw_r_cp_ud_dp_ZQ1QGD0BRPS70SP1NTKG&social_share=cm_sw_r_cp_ud_dp_ZQ1QGD0BRPS70SP1NTKG&skipTwisterOG=2)

What is the expected state when both forward and reverse are brought high at the same time?

HiLetgo BTS7960 43A Dual H-Bridge Stepper Motor Drive Current Limit...

Instruction: This driver uses Infineon chips BTS7960 composed of high-power drive full H-bridge driver module with thermal over-current protection. Double BTS7960 H-bridge driver circuit, with a strong drive and braking, effectively isolating the microcontroller and motor driver! High-current 43A...

#

I imagine you want to avoid that, but human error occurs, and I'm curious what you would expect to have happen

primal shell
#

I would expect both high (or low) side transistors to be active at the same time, which would tend to put a low resistance connection across the motor windings, causing the motor to brake (instead of coast)

azure copper
eternal falcon
#

Question
i want to use 5 solenoids (or linear actuators) to remotely press 5 buttons

I know i need 5 solenoids: https://www.adafruit.com/product/3992
and at least a control board

anyone know of any methods to multiplex one or two control boards to 5 solenoids?

(i'm a software dev, so circuits are new to me)

Medium Push-Pull Solenoid - 5V or 6V

Solenoids are basically electromagnets: they are made of a big coil of copper wire with an armature (a slug of metal) in the middle. When the coil is energized, the slug is pulled into the ...

#

i really only need to drive 1 solenoid at a time

eternal falcon
olive light
eternal falcon
primal shell
eternal falcon
primal shell
#

That's the essence of engineering, there's more than one way to do things

eternal falcon
#

as a software dev, it sometimes hurts me to acknowledge that the cheaper option to "bad code" is to buy a bigger machine

inland bear
# eternal falcon as a software dev, it sometimes hurts me to acknowledge that the cheaper option ...

It's a bit like knowing your tools... There are differences between types of hammer, and for specific tasks one may outshine the others by a country mile (a long way). It always pains me a bit when people don't know how to use excel except for recording tables of numbers, it has sooo many useful features if you look, but still it should'nt be relied upon for data science due to it's limitations (max rows and 17significant figures in calculations, all silently truncated without telling user).
It pains me too to realise spending weeks optimising C code could instead be matched by an increase in CPU for way less than ÂŁ100, which when compared to your time / wage is usually a better deal, but then again if you're selling millions of units then hardware cost scales up too quickly.

terse forge
#

Is it common to use a 2s LiPo battery meant for RC cars to power the motors in a robot?

Something like this for example
https://m.media-amazon.com/images/I/71G2uXbJY2L._AC_SL1500_.jpg

Seems like it's designed for the high current load of motors and if I just get one of these Deans connectors and solder some wires to it I can feed it into the TB6612 controller?

olive light
#

btw, you can also use the balance connector, just use the edge pins

#

(as long as the current is not too high for those wires)

terse forge
#

It's just a basic TT motor, so 3-6v. I'd like to give them 6v.

olive light
terse forge
#

oh so 8.4v for 2...which would probably be too much for 6v max rated motors..

olive light
#

that makes two in series from 6.6 to 8.4V

#

you need an ubec

terse forge
#

a who the what now

olive light
terse forge
#

hmmm

#

Do you know one that can let me do 6V?

vagrant thunder
primal shell
# terse forge Do you know one that can let me do 6V?

This one should work https://www.pololu.com/product/2859

Pololu 6V, 2.5A Step-Down Voltage Regulator D24V22F6

Output voltage
Typical max current output1
Input voltage range2

	6 V
	 2.5 A
	 6.4 V – 36 V

1 Typical continuous output current at 24 V in. Actual achievable continuous output current is a function of input voltage and is limited by thermal dissipation. See the output current graphs on the product pages for more information.
2...

terse forge
terse forge
zenith solar
#

Hi, I'm building a project involving a lot of (4) stepper motors and I'm very paranoid about if I'm buying the correct thing. I've got the DRV8833 from Adafruit (https://www.adafruit.com/product/3297), and a NEMA 17 stepper motor. The stepper has a rated voltage of 2.8V, a current of 1.68A, and a resistance of 1.5Ί. The DRV has a 1A current limiter, but Adafruit says that it can't output more than 2A per chip. Would it work, or would I have to get a more powerful driver?

vagrant thunder
zenith solar
#

I'm not really after torque - its a stepper motor music box

vagrant thunder
#

As long as your steppers don’t stall, should be fine. Can’t speak for how much torque a music box would need…

zenith solar
#

Enough to vibrate a little bit

#

Thank you!

terse forge
#

So I need better wheels. Since my proof of concept works well enough however, I want to upgrade to 4 wheel drive or 6 wheel drive. Is there any reason not to go straight to 6 wheel drive?

#

Another thing - Are there popular algorithms for collision avoidance? Currently my brilliant algo is to turn left 😂

#

The current wheel is something like this

image.png
#

Completely plastic and very cheap

primal shell
#

You can wrap a wide rubber band around it to give it more grip

native cipher
#

as for grippier wheels, what motors do you use?

#

and what wheel diameter you are looking for?

cursive terrace
#

you could see if there's different "tires" available for those hubs, or you may just have to switch wheels (but I second the rubber band notion -- definitely cheaper!)

terse forge
terse forge
#

Right now to turn one wheel goes forward while the other wheel reverses

terse forge
terse forge
terse forge
native cipher
native cipher
# terse forge Cool I didnt know what the steering types were called. I'm basically inspired b...

you can take a look at this page, listing various drivetrain types. This is intended for First Tech Challenge (FTC) robots, but may be useful for other robots too:
https://gm0.org/en/latest/docs/common-mechanisms/drivetrains/index.html
It doesn't, however, cover mars rover type of drivetrains.

Game Manual 0
Drivetrains

This section will cover the heart of any robot, the drivetrain. The purpose of the drivetrain is to facilitate the movement of the robot, and thus is crucial to the overall function of the robot. I...

terse forge
terse forge
chrome charm
#

whats the difference between fusion 360 and gazebo for creating a walking robot for example

olive light
#

one is a cad program and the other is a simulator

#

they have completely different functions

chrome charm
#

then*

olive light
#

possibly, yes

#

there is some work involved in defining the robot for gazebo

#

not sure if you can import from fusion

chrome charm
#

aw dangit

#

still ill try to leanr both of them thx

olive light
#

I would start by playing with the robots that gazebo already has defined

chrome charm
olive light
#

I think they at least have Spot and Anymal

chrome charm
#

awesome thx

olive light
chrome charm
olive light
#

gazebo is just a simulator, you have to write the code that controls your robot yourself

#

many people use ROS for that

chrome charm
olive light
#

it's a really deep and complicated subject

fast tiger
#

Yeah, "learn <thing> in <short time>" tutorials are mostly useless.

chrome charm
#

Ah i see well its still good too start somewhere

#

Ill still try to learn more after I finish it

olive light
#

all the lectures are recorded and transcribed, and all the exercises are available, so you can do it at home

terse forge
#

I'd like to make a robot chassis for a rover using a laser cutter. It can do any plastic or wood. What are some good options? I'd love to use aluminium but sadly it doesnt do metal

native cipher
#

5mm plywood is cheap and easy option.
for something stronger (but also a lot more expensive), good choice is 3/16" acetal, aka Delrin.

primal shell
#

Be careful with "any plastic", some plastics (such as PVC) can outgas products that are toxic and can damage the laser and optics.

#

I like aluminum myself, but my laser cutter doesn't cut it either, so when I want aluminum parts made, I send them out to a service. I've used both Pololu and SendCutSend with good results. They can do optional services such as bending, installing PEM nuts, threading, anodizing, powder coating, etc. as well.

remote-control-tx-initial.jpg
olive light
#

I recently used jlcpcb for aluminium parts, and they even anodized them

primal shell
#

That looks great!

olive light
#

(just the legs, the case is extruded aluminium)

primal shell
#

I was guessing the case was off the shelf. Did it come painted?

olive light
#

yes, though I had to repaint it after cutting the holes for the servos, because I mutilated it a bit with my hand tools

#

I'm now waiting for a pcb that will replace the top plate, because I couldn't be bothered to cut holes into it, and it will be lighter that way

primal shell
#

That's about par for the course for me too

native cipher
olive light
#

it's about 1/3 of what other fab houses wanted for the same thing

native cipher
#

quite reasonable

olive light
#

honestly I have no idea how jlcpcb makes any money with prices like that

sly sand
#

Hi. Is this the right place to ask about the wiring for a small robot project ?

olive light
#

no promises, but if anyone knows, they might answer

sly sand
olive light
#

sure

sly sand
#

this is all the components i will use for this project and i know where every component shiuld be placed, but the issue is that "the L293D motor shield" take all the arduino pin. so my idea was to expand all the pins from the arduino "except TX/RX" to a breadboard and then place all the components there including the motor shield "except the ones that will be placed with the motor shield". do you think this is a good idea, and if not please suggest another one, my budget ended and idk what should i do foer this. Thank you in advance

image.png
primal shell
#

Another approach is to populate the motor shield with stacking headers so you still have access to all the signals

sly sand
olive light
#

this is a stacking header

#

it has a female header on top, and male on the bottom

turbid trail
sly sand
turbid trail
#

more likely a wire will come loose and it will stop working

sly sand
#

Okay sir I’ll do that, and will it cause any problems or any changes in the coding?

turbid trail
#

I don't think so, it's the equivalent wiring just done on the breadboard

#

electrically it can be the same

#

what country?

#

Korea?

turbid trail
sly sand
sly sand
turbid trail
#

you have a lot of things that need to go on the breadboard anyway, so doing it as you suggested seems fine

sly sand
#

This one?

IMG_3039.png
turbid trail
#

The motor shield will not fit on the breadboard by itself, but you can use M-F jumpers from the pins on the bottom to the breadboard, and then M-M jumpers from the breadboard to the Arduino

sly sand
#

Yes that’s what I want to do, can it work?

turbid trail
#

sure it can work. you might want to use tape or something to prevent the jumpers from slipping off. I'd say mount everything on a piece of wood or similar so it's all immobilized and one thing moving will not pull on another. Unfortunately there are no mounting holes on the motor shield, it looks like

#

I take that back, I see one in the upper left corner, maybe there is another

sly sand
#

Yes I customised a piece of wood to fit in the shield inside.
Another question please, can I not plug the TX/RX of the motor shield to the arduino? I searched and it said the motor shield doesn’t use them. Is it correct?

turbid trail
#

could you give me a product link for the motor shield? Probably not but way to know is to look at the product documentation and/or the schematic

#

From that schematic you can see TX and RX are not being used, and A0-A5 are not being used (they don't have to be used for analog, they are fine as digital pins too).

sly sand
wary copper
#

I have a QT Py ESP32S3 (https://www.adafruit.com/product/5426) and seem to have broken a diode off the PCB. I've attatched images to where its supposed to be. I have tried repairing it, but with just a soldering iron and no PCB tools like a hotplate, it is proving very difficult. I assume the diode is an essential component as the device won't power up or show up on any computer. On one device which I luckly was able to repair the same issue, it fixed the problem

Adafruit QT Py ESP32-S3 WiFi Dev Board with STEMMA QT

The ESP32-S3 has arrived in QT Py format - and what a great way to get started with this powerful new chip from Espressif! With dual 240 MHz cores, WiFi and BLE support, and native USB, this ...

#

Does anyone know if I can just bridge the connection with solder or if that will fry the board

inland bear
# wary copper Does anyone know if I can just bridge the connection with solder or if that will...

Ideally I wouldn't just bridge it. maybe you can fit a different larger diode by bending the legs around? I just had to do this yesterday for the diode above your missing one (mine - D2 - separates battery and 5V, whereas yours - D1 I believe - separates USB VBus and the 5v rail leading to the 3v3 LDO).
I'd prefer to not suggest bridging it, but maybe it would be okay if you never have a battery attached. Best to ask in the forums where there are adafruit staff responding.

wary copper
#

i do use a battery but with device i wont now

#

ill ask in the forums tho

inland bear
terse forge
shy coral
#

Anyone familiar with Tower Pro servos?
I have a MG90D (360) and upon stopping it, it reverses before coming to a halt, how can I prevent this?
I have checked the signal with oscilloscope, and there is nothing wrong with it. it goes from 1ms duty to 1.5ms just fine, but the motor still "reverses" slightly before stopping. (stalled or not)
Same when turning off the output (eg, logic low on pwm pin) and given the motor is stalled, but much less.

olive light
#

if you mean it ovrshoots and corrects, that's normal, especially with large inertia

shy coral
shy coral
#

It seems to depend on where internally it stops, different "locations" have different backlash. Feels like the internals is trying to regulate it despite bing a Continuous 360° variant...