help-with-robotics

[primal shell]

So you have two basic options: use an ESC that provides the encoder information or use an add-on encoder

[dull tangle]

Second using a brushless for that RPM

[hasty rivet]

Hrmmmmmmmm....

I have a few drone motors lying around but they all have the 3 leads you would connect to an ESC. I might have to look into how to attach them to a typical motor driver chip and then potentially connect an encoder.

I messed around with the DVD/CD spindle motor last night. I transplanted the encoder from the N20 motor and found that the CD/DVD motor could spin faster! Adafruit site says about 3500RPM at max voltage, but I was getting upwards of 9000RPM at 6V. Makes me want to double check the code for the encoder to make sure it is reading correctly, but it was flying, even with a 5cm diameter 3D printed flywheel attached. Made me think that they might be the go last night. The downside is the shaft diameter is only 2mm.

An outrunner style brushless from the drones would be much more securely mounted.

[young stratus]

Posted in projects, but maybe someone here knows why this servo bonnet (attached to a Pi-Zero) has a green light that will dim 30% of the time I hit the button to move 5 servos (no load). When the green light dims to nothing, the servo movement pauses and after around 2 seconds it resumes. 70% of the time when I hit a button it just moves no problem, no pause.

[primal shell]

You could be hitting a power supply current limit

[young stratus]

Thank you madbodger! I ordered a bigger power supply, 10 amp instead of the 5v 4 amp that I bought from adafruit before. Hopefully it will help out!

[dusky flare]

Good evening to you. I'm a bit stuck and need your suggestions please. I have a Victron converter that suddenly stopped working. It turns on but no output just an error (error 24 which according to Victron is an error that occurs when the system detects a current flowing back into the AC network). Except that even when disconnecting the grid everything remains intact, the inverter still does not work and the same error is present. I suspected the AC input relay to be faulty, but an individual test proves that it works great. It does not receive any commands from the control PCB. What do you suspect?

What do you suggest me next ? Which piece could be faulty ? The Control PCB pr the connexion PCB ?

[night trout]

Hey! Have any of you used Gazebo for simulating before? I'm not sure how to design my robot in it....

[dim ridge]

I want to do linear motion for a slider (pushes a tray out and back by 0-100%). It's more of a valve. It needs to be only 5cm and it has to be quick. It should be able to extend/retract a variable length (not always fully extend/retract) e.g. 40%, 70%, etc.. I've thought about using a stepper + rack and pinion or perhaps linear guide rail. It should be under $25. I have a 3D printer so I can print some parts

[native cipher]

$25 would be tough.
There are dedicated linear servos (e.g. https://www.gobilda.com/hitec-linear-servo-50mm-stroke-30-9mm-sec-3-1kg-thrust/) but they cost much more

or you can make your own rack and pinion setup

using e.g. https://www.servocity.com/actobotics-beam-gear-rack-32-pitch-acetal/

[storm kernel]

feather

[restive pike]

Perhaps a hobby servo and a scotch yoke mechanism?

[opaque osprey]

There is a guide here - https://classic.gazebosim.org/tutorials?tut=build_model - perhaps you can model in a CAD tool of choice or in blender. Fusion360 is nice, OpenSCAD is coder friendly and FreeCAD has a decent license. You export as STL or OBJ and should be able to import those into Gazebo.

[opaque osprey]

I've made a linear actuator with a hobby servo and gt2 timing belt. The "head" was clipped to the belt. I bought the belt pulleys, but managed to just about 3d print the clip. I found it gave a larger range than a scotch yoke but was a bit trickier and needed tensioning.

[proven lotus]

Hey yall, any tips on making good battery connectors? Backstory is, I’m Frankensteining a Roomba to work with a Raspberry Pi. Thing is, I’m trying to use an iRobot battery, and need a way to get proper contact with the battery leads without using the old main board.

Its one of those spring sandwich types; pic above for reference.

[pine bramble]

What I would do is try to find the Roomba part that fits that, which if available would likely be a connector with some wires. It might be a diagnostic part that a Roomba repairman might know about. If not, and it fits directly into a slot on the Roomba and there's no connector available, you could 3D print a plastic frame to hold metal blades that slide into the slots on the battery. And if you don't have a 3D printer (I don't) you could just make one out of some bit of plastic and some epoxy, something to hold the metal blades firmly in the slots. Could be held in place by rubber bands or a tight velcro strap.

[hidden junco]

Looks spring loaded / under tension between the plates, so maybe you can just jam in a pair of copper plates with a lead soldered onto each one

[proven lotus]

I think i’m gonna do basically that, although instead of soldered copper i’ll use bits of shielding (i don’t have copper, and the shielding is pretty dang conductive)

[bronze topaz]

how can i open the library from adafruit in combination with an servo. i want to know his position to give him new orders to move only a short way. my goal is to control a servo with my keyboard and when i didnt press an key anymore to stop the servo

[vocal bear]

Most servos don't have feedback to their controlling chip

It does exist, but it's not common

[night trout]

Hey guys! So I'm simulating a robot to try to use an IMU with a Kalman filter. I'm using ROS to control it and I'm using Gazebo to simulate it. Everything works, but to drive the robot (two wheels + caster wheel) I need to use Geometry/Twist messages (x y z linear velocity + xyz angular). However for my real robot I just use left_motor_throttle and rightmotorthrottle. Is there a way for me to convert this to linear + angular velocity?

I mean I think I should be able to figure out the angular by dividing the throttle values but idk about linear, maybe like the average of both throttle values?

[vocal bear]

Linear velocity is in theory, just omega*r

[primal shell]

You can either add a potentiometer or other encoder to sense position, or use a servo with built in feedback like this one: https://www.adafruit.com/product/1404

[queen yacht]

Hi!
We are from Instituto La Salle Florida, a school in Argentina. At this moment we're building a sumobot in order to compete in our national robot league. To achieve our objectives we need to contact the best enterprises.

We need to know which are the best laser sensor for sumobot.

The sumobot will be affected by:

• mechanical vibrations
• natural light indoor
• impacts
• ambient sound
  (pitch diameter: 154cm)

We were considering using vl53l0x sensor but we need hear opinions from a experimented/engineering.

Thanks in advance,
Sumo La Salle

[lone skiff]

The ST FlightSense sensor like the VL5310X are pretty solid. If it's relevant to your requirements, they also have versions which can read out an 8x8 array of distances as a rough lidar image. AMS/OSRAM has a sensor line with similar capabilities if you run into problems sourcing the ST parts.

[dusky jasper]

I know I am being unreasonable stubborn but, I am trying to use a braccio robot arm with a pca9685.
I could use the original Arduino board with the sheild and use pyserial, but I want to know why this isn't working.
The servokit library works for a 9g servo but not the servos on the arm.
Same voltage for both servos.

The Braccio arm has sr-311 and sr-431 servos.

[primal shell]

Tenacity can be valuable in engineering. When you say the servos don't work, what happens? How are you wiring the arm servos to the PCA9685?

[dusky jasper]

The wiring is directly to the board via a servo extension cable. The plugs don't fit on the PCA9685 without encroaching on the neighbouring pins although, I bent the end ones over and tried the servo directly and then a 9g servo to prove the code.
Tenacity, all 50 cables I bought had the signal cables not terminated at the plug... That threw me for a while.

It must have something to do with the frequency, but I can't seem to find the right way to use the setter in the class

[primal shell]

My best guess is you have a missing/open voltage reference. You need to have the servo power supply connected to the servos' + and - power pins. You also need to have the PWM signal connected to the servos' signal and - pins: this also means that the signal reference and the servo power supply's negative side need to be connected together.

Standard servo frequency is 50Hz. Those servos are wide range ones, so they can respond to a wider range of pulse widths (not frequencies) than standard servos, but the standard signal should operate them, just in a smaller angular range.

The library reference shows how to set the PWM frequency: https://learn.adafruit.com/16-channel-pwm-servo-driver/library-reference

However, I suspect your problem is a wiring problem, not a PWM frequency problem.

[dusky jasper]

I will probe the original shield to get the pinout,

[dusky jasper]

Yep, pinout on the servo is different thanks @primal shell should have picked that at the start.

[sly sand]

@pine bramble I think I made a mistake buying an assembled zumo with the atmel because there seems to be only 2 pins I could use for expansion by removing the lcd but I'm pretty sure that it is very well soldered

[pine bramble]

I don't follow. Why would you need to remove the LCD? There are a lot of expansion pins (well, holes for pins) on the Zumo board, as it's designed for easy expansion. You'd just solder in some header pins and have your IO.

[sly sand]

because the manual says so to connect another board for it: "If you do not need the display, you can remove it. This frees up pin 0 (PD2) and pin 1 (PD3). These pins are the transmit (TX) and receive (RX) pins of the UART, so you can use them to establish serial communication with another microcontroller. These pins are also capable of digital I/O. These pins are the recommended pins for connecting two output channels from an RC receiver"

seems they need to be removed to put another MCU and/or use an R/C receiver like an OpenTX one

[pine bramble]

Ack. So Tx and Rx aren't available unless you remove the LCD. As a UART is pretty common that's a shame.

[sly sand]

yep and from the assembly video there is a sticker in there, a 2x14(or bigger/smaller don't remember) pin set etc

seems it would be a bit difficult to remove

I didn't find anything on zumo forums surprisingly about that so I was wondering if you had some advice

I wish I had seen the version without electronics before I bought it so I would have put a pi / arduino instead on it

[pine bramble]

Well, the big question is if the Tx/Rx pins are being used by the LCD or if they're actually free.

[sly sand]

I looked for it before asking the question but it rolled away somewhere

but I see no reason to doubt the manual

Not sure I can get a better picture with the angle and the simple focus on the camera

but there seem to be a header that use all the pins under the Zumo 32u4 writings

https://www.pololu.com/file/0J864/zumo-32u4-pinout.pdf is the pinout for it and the top 2 pins series are completely used up

and soldered

Basically just for "high-level" I'm trying to figure out how to control it from a remote so I don't to write the code of the path ahead(hence the tx sent/recv) and i'd install a pi in there for a camera and make them talk so that the PI could possibly use more powerful sensors and process the data and send it as a final PWM for the motors or for a waypoint to follow

I'm a bit disappointed the avoidance sensors only do light levels and not distance as well, because that mean I can't map a room for navigation

so far im limited to basically having a colored duct tape track

[pine bramble]

For a remote connection you might also consider using something like a USB-connected Wixel to gain a Bluetooth connection to a Gamepad.

You can add analog IRs or ToF sensors to gain distance measurement. The new Pololu sensors are extremely small and light and have a range of up to 3m.
https://www.pololu.com/category/283/pololu-digital-distance-sensors

I've used both these and the older generation Pololu/Sharp sensors and they work great.

[sly sand]

the zumo have an IR receiver but apparently it's not the same frequency as the one from arduino kits. Haven't experimented with that yet, it would be "enough" for basic controls but not really natural. a gamepad would be ok though

for the usb connection, the zumo has a small usb-c, not sure I can connect any device with a wixel with that?

[pine bramble]

Not sure either, but the Wixel is a USB connection and that's a USB connection. Pololu have excellent support so you could send an email to them and ask as I'm sure somebody's thought about adding a Wixel to a Zumo before.

Downside is the need for the USB cable, which is cumbersome on a small robot, so you might want to just wire board to board and avoid the USB connectors themselves.

[sly sand]

well the 3.3V pin is free I think

doesn't supply much amps, certainly not enough for a RPI but might be enough for a tiny MCU or maybe some basic wireless

The challenge is connectors to other stuff with the UART gone...

and I have a 40 foot usb-c data cable

so the proximity sensors can also be used as IR receivers "NEC protocol with a 38kHz modulated IR signal"

Which would make it even more important to have some rangefinder as I can turn off some IR receivers from the many the robot have to one use for remote control

[queen yacht]

anyone with sumobot experience or knowledge?

[lilac wing]

Hello, I've working on a motion control device using some stepper drivers and arduino, and I need to power it using batteries, I am using sony NPF batteries that have internal BMS, and when I plug the batteries into the voltage supply of the drivers they seems to drag a current peak over 3A I guess that trigger the BMS and the batteries gets opened until I charge them to reset the BMS, somebody knows why these happens or what can I do to avoid that current peak?

[sly sand]

Turns out the LCD wasn't even soldered like the instruction says to do 🙂 So 1 problem solved 🙂

[primal shell]

Maybe add a capacitor?

[potent haven]

hello! i am looking for a pan/tilt setup for my delta, which needs to be light, and i see mixed reviews of waveshare's pan+tilt, as well as criticism of arducam's one. any suggestions?

[primal shell]

The AdaFruit one is simple, but fairly basic. I have another one based on brushless gimbal motors that's fairly nice, but it's the opposite of light. Depending on your situation, a 3D printed one might be the way to go.

[manic depot]

can you recommand a brushless gimbal which can be controlled from the robot side ?

[primal shell]

I'm not sure what you're asking. A recommendation for gimbal motors, or gimbal controllers, or both or what.

Not that I have much to offer in any case, I only got started in that technology, and bricked my controller.

[sly sand]

?? you bricked your ESC ??!

[primal shell]

Yeah. Trying to unbrick it but no luck so far.

[primal shell]

Progress, I managed to flash a bootloader (it turns out the Vcc pin in the programming connector is powered not by the CPU power as I expected, but by the motor power, so I have to hook up both CPU power and motor power and a programmer to flash a bootloader)

[manic depot]

there are some projects around brushless gimbals bot none kept alife thought...maybe you get it done...finges crossed

[primal shell]

SimpleFOC seems to be alive

[night trout]

Hi! Quick question, to use an IMU with a Kalman filter to figure out the position of a robot, what should I use? I'm guessing the gyroscope and acceleration but I'm not entirely sure

These are my options

It's the adafruit bn0055

[pine bramble]

I'd recommend using one of the fusion modes as that takes advantage of some of the "black magic" available on the chip. Even if you only use parts of the available data.

[night trout]

Its something for school though sadly, my prof really wants me to use a kalman filter

[lone skiff]

In that case you'd probably want to use mostly the raw data from the sensor: accelometer, gyroscope, and magnetometer. The others are processed with the chip's own sensor fusion.

[pine bramble]

Yes, agreed, running a Kalman filter over an already-processed output would lead to some rather skewed results.

[manic depot]

there is a IMU Mode which does fusion of sensors for you

[night trout]

which one is it?

[pine bramble]

Yes, if you read back four messages you'd find I suggested that. @night trout has stated that this is part of a learning exercise and their professor wants them to learn to use a Kalman filter. A fusion mode on the BNO055 wouldn't be appropriate then as the Kalman filter will require the raw sensor inputs.

[night trout]

Yup, I'm starting to think about maybe first using the fusion sensor to get the tracking part right so I can then implement the Kalman filter though. I'm having a hard time implementing a Kalman filter in multiple dimensions haha

[pine bramble]

That sounds like a good plan. Getting a Kalman filter to function and provide the correct output is hardly trivial.

[night trout]

Yeah haha

[sly sand]

Finally found some structural pieces systems for making a robot

[scarlet sapphire]

I think it's broken, when I tell the servo to move it doesn't, but it does move when powered

[primal shell]

Could be the servo doesn't like the control waveform?

[hollow sun]

Hi all! I know this will likely be a dumb question and may be incredibly obvious, but regarding calibration of magnetometers, once offsets are measured through an app like MotionCal or SensorLab, do the offsets actually get written back to the magnetometer itself? Or do I need to put those offsets in my code? Thanks in advance! (For context, im using an LIS3MDL on a CircuitPlayground Bluefruit and barring any other way to calibrate this on the CP board, im having to calibrate it on a separate arduino using the MotionCal app)

[lone skiff]

It'd be part of your code or stored in some data file, yeah. The sensor itself will generally just report its raw readings, since it doesn't know what kind of calibration model you might want to use with it.

[hollow sun]

Thanks Ed! That seemed to be the missing piece of info in any tutorials i've found so far. I'm assuming something like x = x - 59.8; y = y + 4.9; to offset the x,y,z heading.

[lone skiff]

There are different models out there that can be more sophisticated than just an offset, instead doing things like fitting a 3D ellipsoid to the field, etc. "Hard and soft iron calibration" is a good keyword if you want to research things.

[still arrow]

so I'm looking at the KY-033 TCRT5000 but I'm getting conflictin information, some sites say it works up to 8mm but others say it works up to 25?

[sharp solar]

Weirdly enough, I didn't see an official datasheet anywhere, but most of the unofficial ones I got said 25mm

[still arrow]

well, hopefully it's that

there's also something about peak operating distance being 2.5mm, is this about the differences it can detect or something else?

[sharp solar]

That sounds like it's saying that it's most effective at 2.5mm

[lilac wing]

Someone with experience using planetary gearbox motors? I need to know if there is a way to reduce backlash in this kind of gearbox.

[vagrant thunder]

As far as a prebuilt gearbox goes, not easily no. There are ways to make gearboxes with lower lash, but it comes at some cost or another. If you need a low-lash gearbox, consider a cycloidal or harmonic drive?
https://www.machinedesign.com/mechanical-motion-systems/article/21831618/methods-to-minimize-gear-backlash

[lilac wing]

thanks for your answer, yes I was looking for a nema17 harmonic drive motor, but it seems there is only two options online.

[hollow wedge]

Does anybody have any experience with this Raspberry Pi Pico robot kit? https://kitronik.co.uk/collections/coding/products/5335-autonomous-robotics-platform-for-pico?variant=39517608247359

[whole rune]

is the raspberry pi motor hat ( https://www.adafruit.com/product/2348 ) compatible with these motors: https://www.adafruit.com/product/3802 ?

[primal shell]

Yes, those motors will work fine with that motor HAT

[whole rune]

i'm looking to build a solar powered rover for my nature cameras. i already have 2x 100w panels, 2x 100Ah SLA batteries in 24volt, a solar charge controller, wifi, cameras, sitting out in my field and it all works fine. i'd like to put this existing setup on a mobile platform so that i can drive it around and look at different things 🙂 currently, the gear is around 50kg (mostly the 2 SLA batteries). i'm looking for some motors + motor controller that can move that much mass. it doesn't need to move fast. i'd like to power it via the 24 volt batteries, and if possible, i have an extra raspberry pi that i'd like to use to drive it - i'd like to code up a small website where i (or others) can send the commands to drive the rover, move the cameras, etc.

[primal shell]

There's this 24V 600W motor with a built-in speed controller for $49: https://www.allelectronics.com/item/dcm-2460/24-vdc-brushless-motor-600-watt/1.html

However, you might want to gear it down for lower speed and higher torque (or use a gear motor)

[whole rune]

this is where i get confused... what do i plug this motor in to? can i connect the motor to a https://www.adafruit.com/product/2348 and drive it with a raspberry pi + python?

[slim gale]

Stepper Motors are for accurate control, for example, a 3d printer. Brushed and brushless can be used for accurate control but are usually used to spin something really fast

[primal shell]

If you wanted to use that big motor I linked to, you don't need a motor controller as it has one built-in. However, you would need something to tell the motor controller what to do. The controller accepts an input of 0-5 volts to set the motor speed. You can get this from either an analog signal (like from a DAC such as product ID 935), or a digital potentiometer like product ID 4286.

[whole rune]

happen to know if there's a raspberry pi hat i can control the motors with? i do have this ras pi hat: https://docs.sunfounder.com/projects/pisloth/en/latest/about_robot_hat.html

[primal shell]

It rather depends on which motors you're using. There are a variety of motors and controllers available.

[whole rune]

i'm wrapping my head around everything and hopefully getting closer.

i'd like to try using the https://www.allelectronics.com/item/dcm-2460/24-vdc-brushless-motor-600-watt/1.html motors and https://www.adafruit.com/product/4286 potentiometer

based on the motor specs: https://www.allelectronics.com/mas_assets/media/allelectronics2018/spec/DCM-2460.pdf i need to cut the control wire to expose the 4 inner wires and connect them to a JST plug, which i plug in to the potentiometer.

i need another jst plug to connect the potentiometer to my raspberry pi, and i have a sunfounder robot pi hat with two i2c ports: https://docs.sunfounder.com/projects/robot-hat/en/latest/hardware_introduction.html

the robot pi hat has two i2c, so i should be able to control two of those motors? is all of this accurate? thanks!

[lone skiff]

So, on the plus side, I2C is a shared bus, and the digipot boards have some selectable address pins, so you should be able to run multiple ones on the same Pi pins.

On the minus side, it's not clear from the motor documentation exactly what voltage it is presenting to the potentiometer, so it's possible that it would be out of spec for the digipot. Unlike a physical potentiometer that's just made of wires and resistors, a digital potentiometer uses transistors, so they don't like working with voltages outside their own power-supply range (which would be 3.3V for working with the Pi).

Actually, I take that back, it looks like the analog voltage can be up to 15V with this chip, which is good. The motor runs at 24V though... I'd generally expect that it would have a lower voltage for its speed controller, but the datasheet doesn't exactly specify that.

[whole rune]

i'm open to alternative motors and so forth, at the end of the day i'd like to control them via a raspberry pi + hats. i don't know enough about motors to know what i need or what to look for

a different brushless 24 volt motor: https://www.robotshop.com/en/42d-brushless-dc-planetary-gear-motor-24v-12rpm.html but i don't understand how to wire it or what controller i need, in order to control all the motors via python on a raspberry pi

[primal shell]

I'm guessing it's 5V, but I can measure it if desired

[primal shell]

That one also includes its own controller, but is controlled differently. It has two inputs, a speed input that is high frequency (15-25kHz) PWM, and a direction input. Both are compatible with 3.3V or 5V logic. It also offers a pulse output giving the motor speed at 12 pulses per revolution (open collector).

[whole rune]

which type of motor is easier to work with? or "better"?

[primal shell]

That, of course, depends on your control system and desired motor parameters.

[whole rune]

how do i control https://www.allelectronics.com/item/dcm-2460/24-vdc-brushless-motor-600-watt/1.html or https://www.robotshop.com/en/42d-brushless-dc-planetary-gear-motor-24v-12rpm.html with a raspberry pi? either one or both would be awesome. i think i'm not understanding what goes between the pi and the motor. i'm picturing a raspberry pi hat which the motors would connect to, but maybe that's not correct?

[primal shell]

For the Allelectronics one, you could use either a DAC board or a digipot. For the robotshop one, you'd need a source of a 15-25kHz PWM signal.

[whole rune]

i mean how do i connect the DAC or digpot or PWM signal source to the raspberry pi (or some compatible pi hat)?

[primal shell]

That depends on which interface you choose. Naturally, a HAT will be simplest to integrate, but there are ways to hook other peripherals to a Pi if needed.

[rancid zodiac]

Hey @primal shell! It's been a while since you last rescued one of my projects... Would you be able to help me with a bleuart issue?

[primal shell]

I'm not sure, I don't know a lot about Bluetooth, and even less abut bleuart

[rancid zodiac]

@primal shell, I did just figure out what I was stuck on. I didn’t realize that you can use bleuart.println(). I was struggling trying to use the bleuart.write, but that seems to be limited to using a buffer and I don’t know how to populate the buffer with my own data as opposed to sensor data.

[primal shell]

Oh yeah, those stream subclasses are handy

[whole rune]

i watched a couple arduino + brushless motor videos and i think i understand better now, or i understand even less 🙂 if i understand better, if i get two of these motors: https://www.robotshop.com/en/42d-brushless-dc-planetary-gear-motor-24v-12rpm.html i can wire the red and black wires of each motor to the battery bank, and the blue+yellow+white of each in to the pwm on my pi motor hat ( https://docs.sunfounder.com/projects/robot-hat/en/latest/hardware_introduction.html ). do i need a bec or esc for that motor? how do i determine which wire from motor goes in to which port on robot hat?

[slim gale]

for a vehicle such as the one you are making, I would get some ESCs for the drive motors. The red and black are motor power and the colored ones should be encoder data which is how much the motor has turned. Not %100 sure about the wires though

[whole rune]

how would i connect the motor to an esc, for example this one: https://www.amazon.com/RC-Brushless-Electric-Controller-bullet/dp/B071GRSFBD/ref=sr_1_2?keywords=30+amp+esc+brushless&qid=1657068884&sr=8-2 - the esc has less wires than the motor?

[slim gale]

I’m only one third qualified for this

[primal shell]

Both of the motors you linked to have built-in ESCs, so you don't need an external one.

[whole rune]

any idea about the wiring coming from the motor vs going to the esc? i don't want to make a costly mistake

[primal shell]

I'm not sure I understand what you're asking. You don't need an ESC with those motors.

[whole rune]

how do i wire the robotshop motor to a raspberry pi?

[primal shell]

I'm not sure. The Robotshop motor is controlled with high frequency PWM, so you'd need a way to generate high frequency PWM. I don't know if the Pi can do that on its own, or if it would need something external to do so.

[whole rune]

the robot pi hat i linked can handle the frequencies, but i'm confused about the wiring from the motor as it's different colors than every pwm wiring i've seen, like for servos and other brushless motors

[primal shell]

The PWM signal would go to the blue wire on the motor, with the ground/zero volt reference to the black wire

You'd have another signal (from the Pi or the Hat, either could work) to the white wire to control the direction.

[whole rune]

the black would go to the battery?

[primal shell]

The black would go to both the battery negative and the HAT ground/zero volt reference.

[whole rune]

the battery and motor are both 24 volt and the hat + pi are both 5v ? does that matter?

[primal shell]

The Pi is actually 3.3V. You will need to find some way to power those. However, the 3.3V signal from the Pi is sufficient to control that motor, which will recognize any voltage over 2V as a logic "high" and any voltage under 0.6V as a logic "low".

[whole rune]

yeah i was looking at a bec or something to convert the battery 24v to 5v or whatever for the pi. the robot hat i have says it takes up to 12v actually, if i end up using that.

[primal shell]

I'm not sure that motor HAT will do the job. I'm looking at the data sheet for the PCA9685 chip that generates the PWM, and it states that it can produce PWM frequencies from 24 to 1526Hz. That's pretty far from the 15000-25000Hz the motor controller wants.

[whole rune]

dang 😦

[primal shell]

I'll admit I'm a little surprised at that PWM frequency spec. That's really high, and they don't give hints as to why or how to generate it.

Is that 67W motor going to be sufficient? Your build sounded pretty heavy, but maybe you didn't want to move fast.

[whole rune]

i'm not sure to be honest, it's still unclear to me what exactly i need. it did say like 60kg stall torque for that motor. i'm totally open to any motor suggestions, i just want to be able to like ssh in to my pi to move to it around. my pi also has a camera so i can see where i'm going. the robot hat i linked has some servos so i can pan+tilt the camera which is nice. i want to swap out the camera for a separate pi camera with night vision

whatever i need to send some signals to move some big motors is fine

[primal shell]

Poking around on the web, it seems like many people have some issues controlling those Robotshop motors, but I did find one forum message where they said it would basically work with lower frequency PWM, so it might be worth a try, depending on your tolerance for dealing with weird issues.

[whole rune]

that's kind of disheartening 🙂 i didn't think finding motors would be this difficult to be honest, but it seems like most folks have been using smaller motors

[primal shell]

You might look at the Pololu 37D gearmotors. They're simple brushed motors, so easy to control, and they offer one with a 150:1 gear ratio that delivers 56kg-cm of torque and 68RPM at no load. You mentioned 60kg torque, but didn't give a radius, so I'm not sure if that's a hard figure, wishful thinking, or just the amount of weight you want to shift.

However the gear trains on these small motors aren't really designed for high torque, so you may be better off with the sort of motors designed for things like powered wheelchairs, which are bigger high torque units.

[whole rune]

i was going to print some 200mm diameter wheels

my batteries and everything are around 60kg 😅

i'd like to upgrade the batteries to lifepo3 in the future, that will reduce the weight by around half

however - i'm interested in making a semi-generic heavy duty platform, that can haul around a trailer or something in the future

[primal shell]

This kind of thing might suit your needs: https://www.ebay.com/itm/284883457632

[whole rune]

i've seen some similar projects using wheelchair motors, but ebay kind of scares me to be honest

[primal shell]

You're in an awkward realm where you need fairly beefy motors, which most hobbyist suppliers don't carry. So the remaining outlets are industrial supply (sometimes tricky to just buy a small quantity), medical supply (super expensive), or the secondary/surplus market (like that All Electronics motor I linked or eBay). They're all valid choices, just with different tradeoffs. For one-off builds, I'll often go with eBay, but if I'm designing something intended for other builders to replicate, I'll try to find off-the-shelf parts somewhere.

[whole rune]

the pololu data sheet is extensive 🤌

[primal shell]

I really like the Pololu people and products.

[whole rune]

so with the pololu, item # 4687, 4697 has 73kg-mm torque, which i'm assuming is enough to move a ~60kg mass, if i have 2x or 4x of them? also i see it comes with encoder or without encoder, i'm assuming i want with encoder, but then ultimately how do i connect this in to some kind of device that connects to a pi?

it has a 6 pin plug i haven't seen before 😅

[primal shell]

The 6-pin plug is for the version with the encoder. I think it's just an ordinary header with 0.1" spacing. The motor drive itself is just two wires (red and black). You'd need something a little more powerful than the motor HAT to drive it. The HAT is good for up to 13.5V, and you'll be at 24V. It's also good for 1.2A and you'll probably want closer to 4A. Pololu offers some suitable motor drivers and controllers for that regime.

As for encoders, that lets you monitor how fast the motor is turning. Since you'll have a camera for feedback, that may not be important.

[whole rune]

so what do i need that's more powerful?

[primal shell]

A motor driver or controller.

[whole rune]

like this: https://www.pololu.com/category/96/pololu-trex-dual-motor-controllers ?

[primal shell]

How powerful your motors need to be depends on the weight of your platform (which you know), the size of the wheels (which you also know), how fast you want to go, how much ground resistance (grass, for example takes much more effort than rolling over a smooth floor), and how much slope you need to deal with (flat ground is easier than steep hills)

[whole rune]

that one says end of life :sad:

fast: cm/s and slope, i think max might be 20-30 degrees

[primal shell]

Those are an older design, and only good to 16V. Their Simple Motor Controller line would be suitable, they're good to 24V, you can control them with I2C (the Pi would work well that way), and they offer plenty of current. They're also cheaper.

[whole rune]

i think this might be a better option? https://www.pololu.com/category/94/pololu-simple-motor-controllers

[primal shell]

Yeah, their G2 range of those would suit your use case

[whole rune]

or the jrk g2?

[primal shell]

I'm unsure what the JRK refers to here. In any case, it's bedtime in my time zone.

[queen yacht]

Hi people, does anyone know where i could buy a mapple mini (stm32f103c8t6)?.
I ve been looking everywhere and I couldn’t find it.

[little needle]

how likely is it that a lattepanda v1 would survive 5.8-6v on the 5v input

[slim gale]

Unlikely enough to put a regulator

[lilac wing]

Hi, someone that knows about a Marlin alternative for robotic arms?

[sharp solar]

Like a steadicam marlin?

Oh lol nvm I do not

[lilac wing]

Like a Pan Tilt Track Head

[molten jay]

Hello Web, any1 here?
Am a new member and new to discord.
I have the adafruit BNO055. in the meantime am reading all sensor data 9DOF and quaternions + EUL, am looking for a software to model and view my object

I have my own driver using UART to communicate with the BNO055

how to move forward from here? THANKS

[primal shell]

I'm not quite sure what you're looking for here, but you can use Processing to show rotated views of objects: https://www.e-education.psu.edu/earth801/node/489

[molten jay]

Hey, Thanks for the respond

[molten jay]

I have Euler reading, ACC, GYRO,MAG and Quaternion, i want to use the Quaternion to visualize an object lets say an airplane and move my BNO055 and see the airplane move and act the sme as the way i move my sensor

[primal shell]

I think you can do that in Processing, if you have a 3D model of your airplane to start with. There are presumably options for other languages as well, but the basic process would be similar, draw your object and apply the necessary rotations to it based on input from your sensor. Under the hood, it's generally accomplished by creating a "rotation matrix" containing numbers computed with sines and cosines of your angles, and multiplying your model coördinates by that matrix.

[molten jay]

I need to create an AHRS system using python/ROS/TensorFlow, dosnt really matter. Am pretty sure am not that one genius explorer to create my own, soo am on the search for a system that its input is Euler angels or Quaterneons and a model that i create dont know how for now and just visualize it. Yesterday i managed to run Adafruit bunny visualizer but i need something more customizable and not webSerial

https://adafruit.github.io/Adafruit_WebSerial_3DModelViewer/

[peak plume]

ROS and rviz should work for that, but seem very overkill

[molten jay]

I am looking into it, there is also something called gazebo

[peak plume]

That is mainly for simulation, afaik

[molten jay]

I need a tool to move my airplane model live

and the coordinates are given via uart

when you say simulation what do you mean?

[peak plume]

Simulation as in simulating how a robot would behave in different environments. AFAIK the sensor data would be faked to match a simulated virtual environment

Whereas rviz is afaik mainly a visualization tool for your (sensor) data. So you would intake your live UART data and write an adapter to convert that to the ROS format. Maybe that adapter already exists as a ROS package you can simply install via apt. If you link your input data correctly to your robot (or plane) model, the visualization should follow the orientation of your sensor

[molten jay]

THANK YOU VERY MUCH !! really explained it very nice, I will be looking about it today.

[molten jay]

Can I run ROS / RVIZ on windows OS? all I see is linux based os

[peak plume]

I used it on Linux, but it may be possible to run it in a VM or via WSL. I wouldn't recommend that though, the learning curve of ROS is already fairly steep, so I wouldn't stray off the beaten path

Again, ROS seems like cracking a nut with a sledgehammer for your use case. If you don't need a pretty plane model for your visualization, but can make do with just a few vectors, I would advise to go with a 3D plot from matplotlib or the likes

[molten jay]

I think for learning and developing ROS is better maybe in the future it will help me

https://www.youtube.com/watch?v=a-mfCeykmYw&ab_channel=RoboticsandROSLearning

that's exactly what I need

[peak plume]

You have been warned 😅

[molten jay]

HHHH I have you boys with me 🙂

[peak plume]

I would suggest starting directly with ROS2 foxy. It seems like there are packages for your IMU already available.
Much of the ROS1 documentation is outdated and quite confusing if you are new.

Though if you want to follow along with your linked video, that's ROS1

[molten jay]

YEs yes of course, first task is creating my own publisher and subscriber, the publisher will read uart imu raw data. Subscriber will print the data. GOOD LUCK

[peak plume]

http://wiki.ros.org/ros_imu_bno055 (ROS1)
https://github.com/flynneva/bno055 (ROS2)
You can probably use one of those to publish the sensor data

rostopic echo /imu/data (ROS1)
ros2 topic echo /bno055/imu (ROS2)
to get terminal output of your data

[molten jay]

I have my own driver, oki soo the story is like this, I have a MCU BGM220 from SiLabs and it is sampling ,ultiple sensors and data and firing all data on uart, i cant change to something else cuz it is sooooo low power and my project is portable, the hole idea from ros2 and rvis is just to connect it when needed via usb c which is also a ftdi uart and display it in ros

like a monitor once a while

[peak plume]

Ah okay, thought you were running stand-alone