help-with-hw-design

[hushed smelt]

Did you put any kind of breakout for the SWD pins?

otherwise you might have to bodgewire breakout directly from the legs of the stm32 :/

[distant raven]

Easy enough compared to like.. this for instance

The uson8 package was really hard to bodge but it worked lol

The wire size is like 38awg

The diameter is about the same as 8mil wide

[hushed smelt]

good luck with that. yes you'd definitely need a microscope for that kind of work, and some really thin wire.

[woven bluff]

is it safer to desolder SOIC-16 with braids than hot air?

[supple pollen]

It's more a matter of personal preference, what equipment you have, its capabilities, your capabilities, and what your end goal is.

[woven bluff]

I have both

end goal: not tearing off a pad

[supple pollen]

If you're desoldering it for re-use elsewhere, you want to remove it intact. If you just want it off the board, the easiest approach is sometimes to just cut all the leads off and unsolder them one at a time.

[worldly schooner]

In regards to not tearing pads, I find hot air a bit more reliable, but it certainly comes with other risks.

[supple pollen]

I just used hot air to remove a component last night, but I did a poor job of it and ended up scorching the PCB.

[woven bluff]

I don't have something small enough to cut SOIC leads

[worldly schooner]

An exactoknife tends to work pretty well with some patience.

[supple pollen]

I've been known to use a Dremel type tool and cut-off disc for that task, but then there's metallic dust to clean up.

[woven bluff]

seems really easy to cut into solder mask and copper

[worldly schooner]

It certainly is great for that too.

If you’re not looking to keep the ic since you’re cutting the leads anyways, I find cutting it towards the ic more horizontally works great for preserving the pcb underneath.

[woven bluff]

wouldn't cutting horizontally apply shear force to pads?

[supple pollen]

It would, but normally the cutting is done cold when the copper adhesive is strong. They come off more easily when desoldering because heat weakens the adhesive.

[worldly schooner]

It also depends on how you’re cutting. If you simply exert straight pressure as opposed to a sawing or wiggling motion, I could see some excessive shear forces being applied.

[vast flume]

Yeah they're in the JTAG connector right?

[elder peak]

There's also the "Quickchip" magical desoldering alloy stuff.

[worldly schooner]

Chipquik is great, can confirm.

My go-to for the rare case where I need both the chip and pcb preserved.

[distant raven]

This is also why I use low temp paste on prototypes so I can use chips across multiple versions

[dry pelican]

Low temp paste is the best paste

[distant raven]

I’m using sac305 for production now because of people turning LED boards on full power and desoldering everything off the board

💀

[dry pelican]

That seems like a problem

I say that anything that reaches over 100C is a problem. Unless that part is meant to reach 100C

[distant raven]

RGB LEDs get really hot at full brightness

Which is why they’re usually on large boards with huge copper fills

50x 3535 neopixels can draw up to 15W of power

Dissipated over like a 93x32mm board is really hot lol

incoming pun

[elder peak]

I've learned a lot of useful things about power supplies, switching, heat management, batteries, et al driving RGB LEDs.

Oh yeah, no pun. SURPRISE

[distant raven]

LEDs are just fascinating because you think “oh it’s just light” but light use nearly always a byproduct of energy dissipation/release

I actually can’t think of a situation where light isn’t generated by a change in energy level

[elder peak]

True, but I guess the LED case is maximally weird compared to mere gas discharge or combustion.

I guess the fun part about working with COB strips is that there are really quite bright COB strips and they tend to be less of a heat problem on their own but supplying them with power in a small chassis then becomes the hard part.

So would you say that when you need a high temperature solder, SAC305 tin work well?

[distant raven]

Yeah, though I tend to prefer even higher temp paste than sac305

Sac305 melts ~235°C, some higher temp stuff I use melts closer to 250°C

But I think ran out of that

Actually I didn’t. I guess I had sac305 all along.. hmm 🤔

But yeah, sac305 is great

[woven bluff]

I cannot get BQ25606 to work.

does this cap has to be the value stated in datasheet?

it says 4.7uf but I don't have any so I put 10uf there instead.

[distant raven]

It’s possible that the 10uF causes the turn on time to be too slow

[woven bluff]

the logic part is working, but charging current is only 30mA

[distant raven]

What voltage divider did you use?

Or values did you use I should say

[spice turtle]

Hold on, https://www.ti.com/tool/BQ25606EVM-772#tech-docs shows that its uspposed to go to resistors

[woven bluff]

what divider? on TS pin?

[spice turtle]

one sec.

[distant raven]

On Regn

[spice turtle]

1. only shows this pin if you are using temp sense

2)4.7uF

but Its an LDO output so I think you should be OK

[woven bluff]

there is no resistors on REGN

[spice turtle]

even the eval board shows no caps 🤔

[distant raven]

Anyway, 30mA makes me think it’s failing the check after the regn LDO starts

What resister did you pick for ILim?

240?

[woven bluff]

220

and I shorted D+ and D-

so it should be recognized as DCP

[spice turtle]

QFN packages are difficult to solder, did you check every pin?

[woven bluff]

I don't have small enough probe but I did visual check seems ok

[distant raven]

I think shorting D+/D- might be your issue

Where in the data sheet did it say you could do that?

[woven bluff]

it just say BC 1.2

[woven bluff]

the next closest thing I have is 2.2uf

[distant raven]

https://www.ti.com/lit/an/slvae17a/slvae17a.pdf

DCP needs to be shorted through a resistor according to this

Is that what you’re doing?

[woven bluff]

no, just a trace

it did not say resistor value, could be 0ohm

[distant raven]

Perhaps.. I’d maybe try to put like a 10k resistor across DP/DM pins if you can. See if that changes anything

Though technically it should just default to 500mA based on what the data sheet says

[oblique solstice]

Hey is anyone here an USB expert? I've been breaking my head. I tried making an USB-c extension cord by soldering a female to an extra phone cable - and it doesn't work.

Also I tried cutting a keyboard USB and attaching a USB-c to it and it also doesn't work

Then I soldered another female to leftover keyboard USB male and if you insert the extension - it works

[distant raven]

Is it a USB-C to USB-C cable?

[oblique solstice]

Essentially yeah

[distant raven]

The cable not working for the keyboard is likely because there isn’t CC lines connected on the keyboard itself

[oblique solstice]

Like left - works. Right - doesn't work

[distant raven]

Left works because usb c is going from type C to type B connector.

Harder to go the other way around

The USB C cable plugged in is going to expect CC lines connected which they likely are not in the right side

[oblique solstice]

But like there's only 4 pads on both male and female that I got

[distant raven]

🤷‍♂️

[oblique solstice]

Also in the USB-c cable I cut open there's just 4 lines

[marsh nest]

The only observation I can make here is that the USB A to C adapter is in different "directions" - On the left, the USB C end is going into the computer, while on the right the USB C end is going out of the computer. (Yes, the cables are bidirectional, but you get the idea)

[distant raven]

I suppose it’s possible that the D+/D- lines aren’t connected together on the adapter

[oblique solstice]

Sorcery

I've also got like a hub module - like this one

My expectation was that I'd just solder it to a USB-c plug or a piece of cable ant that would work - but if it doesn't I'm in trouble

Although it's usb2 so I guess as long as I pass it through an appropriate dongle it will sort itself out...

I guess the problem is wither with the dongle or the female - cause they ain't connected yeah - but then if I turn it it still doesn't work. Pff.

[oblique solstice]

But yeah - thanks guys! I think I've figured it out - it probably sees which direction it's in by which pins are coupled and trips.

[oblique solstice]

So yeah it would make sense - but the solution is basically get my hands on a in-end or an 8-pad plug and connect em myself

[supple sky]

I was looking for a Fusion 360 electronics schematic for an RP2040, but all I see are the 3d files. Are these not available? I just wanted to dry up my project to learn schematic design.

Like, the Feather RP2040 to be more clear.

[unreal flax]

Here they are (linked from the Downloads section of the Learn guide): https://github.com/adafruit/Adafruit-Feather-RP2040-PCB

[supple sky]

Ohhhh, okay! Thank you so much.

[vast flume]

Hey guys, not too sure where to post this question since it involves both hardware and software. A few days ago I made a library for using the BMP581 with an stm32 and it worked fine. I then got a new pcb, which I assembled and tried running the same library on (and even programs that previously worked), and it no longer works correctly. Also, when I use the HAL_I2C_IsDeviceReady function I get HAL_BUSY.

[distant raven]

Probably something blocking running

[vast flume]

I ended up running an i2c scan and I can't seem to find the device. Could bad pullup resistors cause this?

[distant raven]

It’s very possible

[marsh nest]

That's what I would check first

[distant raven]

Or there’s no electrical connections at all to the MCU pins

[marsh nest]

Yep. You can probe your traces and/or do a visual inspection to make sure everything is connected and there's no bridges.

[vast flume]

I tried swapping out the sensor for another unit & resoldering the sensor & pullup resistors to no avail 😦

Heres how the pcb looks and also a blank pcb

[unreal flax]

I'd probably be most suspicious of U2, since the pads aren't available for visual inspection. But checking continuity with a multimeter would generally be the way to go.

[marsh nest]

Almost looks like you may have some bridging on U1, just below the 1. But yeah, I would worry about u2 too

[vast flume]

Moved everything to a new pcb bc I couldn't figure it out and now the voltage regulator (U5) is shorting/overheating 😦

[unreal flax]

Yup, continuity-check time...

[vast flume]

What points should I test at? 😅

[unreal flax]

Likely a power-ground short somewhere, so those rails would be step one.

[vast flume]

Ok

It would be everything downstream of that regulator though right?

[unreal flax]

Yep. If the input power were shorted, the regulator wouldn't be running at all.

[vast flume]

Yeah

[vast flume]

I did a continuity check on the u5 linear regulator, and the gnd and 3v3 pins seem to be connected

I see no way for this to have happened though. Could it just mean the regulator is bad?

[distant raven]

Could be shorted somewhere else

[vast flume]

Ended up replacing the regulator, and reseated almost all of the components. Still getting the same issue 😦

[distant raven]

Did you check for shorts on a bare board?

[vast flume]

I will but I doubt it since the previous board worked in this sense. Just acted weird with the sensor

Just checked, no shorts on a bare board

[dry pelican]

There is a possibility that a component died

or that there is a solder bridge underneath a component

[woven bluff]

I swapped to another BQ25606, still did not work

maybe the PCB is shorted somewhere?

[vast flume]

Is there a way that I can check my caps to make sure they aren't broken? In theory if one broke wouldn't it short to ground?

[marsh nest]

Check the resistance of the capacitor. If it's zero, well, there's your short.

Actually. Nevermind

You already have a short somewhere, so that won't help

[vast flume]

How would that not help?

[marsh nest]

Ok. So. Checking the resistance on one side to the other isn't going to check the resistance on the shortest path, it will check the resistance on the easiest path.

So if it's a cap between 3.3v and gnd, and there's a short anywhere between 3.3v and ground, reading the resistance between the points on either side of the cap will end up as 0.

[vast flume]

Ahh I see

How should I track down this short then? 😅

[worldly schooner]

A faulty cap can be hard to trace. If visual inspection fails to single out any significantly discolored or cracked capacitors, you’re in for a tough time.

The other method for locating a short would be to use a thermal camera to see if any traces are heating up significantly more than the others, but you’ll need a pretty expensive one to have enough resolution to be able to single it out.

[vast flume]

The linear regulator was what was heating up the most

[marsh nest]

My first step is always eyeballs. See if I can spot any bridging.

Second step is to identify components where the pins could be bridged (Your U1 and J1 would be my first checks) and make sure that each pin is not shorted to its neighbor. Reference for these is nice and useful because sometimes there's a few GND pins or whatever next to each other that you don't have to worry about - And sometimes there's a power pin next to a ground pin that's probably the best thing to check first

[vast flume]

I tried removing u1 and j1 and testing for shorts, got one both times

Same with u2

[marsh nest]

My next step would be to double check and make sure the pads are cleaned (That's sent me in circles a few times in the past), and then, one by one, remove any component that could possibly be shorting it until you find your answer.

Alternately - and I have done this on more than one occasion - set aside the problem board and start a new one. Put one component on. Test for shorts. Repeat until you find a short or it's done.

I'd probably start with the tricky items first, but that's just me

[vast flume]

Jeez this is going to take a while 😅

Seems easier to just reorder all of the components lol

[supple pollen]

There are also testers that transform very low resistances into audible tones, you can use to trace back and forth to find where the lowest resistance (and therefore the short circuit) is.

[distant raven]

The diode/continuity mode on my multimeter yells at me on shorts or resistance under like 10ohm

[supple pollen]

That's handy, but for short tracing, I find a milliohmmeter more useful

[distant raven]

Any particular reason?

Finding more precisely where the short is since resistance is a function of conductor size and length?

[supple pollen]

Moving around on a PCB trace, being able to tell the difference between 150mΩ and 160mΩ is very useful to know if you're getting closer to the actual short.

The audio ones make it easy, as you can keep your eyes on the probes and hear the tone rising and falling as you move around.

[distant raven]

Mine isn’t really that precise, just a fairly steady beep

I’m usually not working on anything that I can’t see the short on

But I’m sure one day it’ll be really useful to see where shorts are more accurately

[supple pollen]

Accuracy is useful, but "more quickly" is what I value

[distant raven]

Makes sense

[unreal flax]

Need a tiny little magnetometer probe to follow the current path to the short... 😁

[supple pollen]

That doesn't work as well as you might imagine, as the return current path is often nearby, and the fields tend to cancel out.

[unreal flax]

Good point.

[vast flume]

Thats what I've been using

[fervent lance]

There are also those things that vaporize flux on the pcb, it works like a cheaper thermal camera

[scarlet jay]

https://tenor.com/view/dog-gif-24988609

Hello Everybody!

You might wonder what a man like me does o keep you up at night.

I take care of your brain while you sleep but I need some HELP

HELP ME PLEASE

• How can I make a 3.3 to 40V circuit without losing quality of DAC (freq/wave/amplitude)?

• Where should I properly put a BI MINUS on my BIOIMPEDANCE circuit?

[supple pollen]

The short answer is "an amplifier". A detailed answer would need to know what the range is (for example, is the 40 volts from 0 to +40V, or from -20V to +20V, or something else. Also, the minimum frequency, the maximum frequency, the load impedance (or load current), and what the limits of fidelity are.

[idle python]

whawhat isiss the red highlighted thing

what is*

[supple pollen]

That is an NPN bipolar transistor

[idle python]

where am i supposed to get those

[supple pollen]

It is being used as a switch, to allow a digital signal to control the "enable" input of the voltage regulator. They're widely available from many sources, including AdaFruit https://www.adafruit.com/product/756

It is, of course, not necessary if you don't want to switch the regulator on and off that way.

[idle python]

good

why is there a polarised capacitor of 470uf tho

and on the sim7600 datasheet i sawit uses a 330uF polarised capacitor instead

these 2 modules run on the same voltage

so why do they have different capacitors

[unreal flax]

The size of the capacitor is generally related to how much short-term variation in current draw there is (like needing more power suddenly to transmit), not the voltage.

[dry pelican]

You can usually slap as much capacitance as you want on a circuit. Too little is usually a problem. However you can have too much capacitance and it can lead to inrush/overload issues

[worldly schooner]

As for the reasoning behind using a polarized capacitor, they tend to be cheaper than ceramic caps by capacitance and voltage rating. You certainly could use a large ceramic cap in its place, but a 470uf ceramic cap is not cheap…

That’s a number that, while physically possible and commercially available, is pretty rare for anything above 4v rating, so you really do need a polarized cap for those high energy storage applications.

[idle python]

what is equivalent series resistence in tantalum capacitors?

i wanna get 470uf polarised capacitor for my sim module

[spice turtle]

basically a caps internal resistance. Think of it as a resistor in series with a capacitor. This is why paralleling caps lowers the overall ESR of a system

https://www.digikey.be/fr/articles/simple-explanation-of-capacitor-esr

Generally speaking, ceramics have low ESR and are used a lot in SMPS's. Aluminum Electrolytics are used as well, but the higher the ESR, the bigger the cap. You can get low ESR Aluminum electrolytics that are smaller but they may cost more. Tantalums are kind of a mix between the two. They used to be recommended a lot in SMSPs circuits, now not so much

[formal sand]

My recommendation would be to look for tantalum capacitors with the same capacitance rating and the necessary voltage rating on digikey and mouser, make a list of the esr of like 3 of them and average them together

[idle python]

@formal sand i found two that have the same voltage rating and capacitance but they have different esr

one is 0.4 ohm at 100khz and other is 25 m Ohm at 100khz

what does this mean

[supple pollen]

That means the regulator tries to keep the voltage at its feedback pin equal to 0.92 volts. The voltage divider composed of R2 and R4 scales the output voltage to the feedback voltage, so by regulating the feedback voltage, the regulator also sets its output voltage (which is determined by the ratio of R4 and (R2 + R4))

[idle python]

ohhhh ok

[formal sand]

Huh interesting, they are more than an order of magnitude off

[idle python]

i will use a circuit simulator so i can make the circuits

cause i dont know nothing about making circuits

[supple pollen]

A simulator is a good learning tool, but I found I learned the best by actually building circuits. Some of the educational offerings (Snap Circuits, "150 In One", Lectron, Gakken, Brick R Knowledge, etc.) are a great way to get started, or just a solderless breadboard and a handful of common components will take you far. Knowing some of the basic concepts like how a voltage divider works is super useful for understanding a wide variety of circuits.

[woven bluff]

what temperature and air setting do you usually use when soldering QFN with hot air gun?

I'm learning EE mostly from datasheets and application notes

[idle python]

set it under the maximum operating temperature of the qfn chip

[woven bluff]

operating temperatures are much smaller than even the lowest solder melting point

sometimes datasheet supplies a reflow profile but not always

too bad OSH park does not offer PCBA

[idle python]

jlcpcb and pcbway offer PCBa

jlcpcb is very expensive for assembly on both sides of the pcb tho

and they dontlet you assemble only one board you have to buy 2

[woven bluff]

I know, but they charge a lot of castellated holes

OSH does not

[formal sand]

That's a good way to learn

[woven bluff]

I suspect OSH does not do castellated holes properly. I think they just strike a milling blade through because there are leftover copper flakes still attached to the castellated holes.

[formal sand]

With a hot air gun is easier to regulate heat than an oven, you have to be careful that you don't go much higher than the maximum reflow temps specified in the datasheet, but in general i go by melting point of my solder and the size of the part. The bigger the part the more air I put in.

But in general it is a sliding rule that you have to learn based on your solder and size of part

[woven bluff]

4mmx4mm VQFN with thermal pad, the datasheet does not provide a reflow profile

BQ25606

I used 400C 10% air and I think I burnt the chip.

[idle python]

problem is how would i work with ICs

they dont exist in circuit simulators

[woven bluff]

they do in manufacturer provided sims

[formal sand]

I did some research but im not on my pc, I got some numbers for you

[woven bluff]

like LTspice

[formal sand]

I found this information when I was looking at the info in page 38 related to the mls peak temp

[formal sand]

Yeah most manufacturers have ltspice simulations on their website if you look for the part number, if it is a jellybean part you just use whatever is already in your simulator

Especially with digital ics it can become a real challenge to simulate so you might need to breadboard and do your measurements manually

[woven bluff]

also you cannot use breadboard for SMPS

[formal sand]

You can

Not a regular breadboard but you can

[woven bluff]

<100KHz maybe

[idle python]

how could i calculate the voltage output from voltage feedback

voltage feedback min: 0.900V
0.900V will output 4.75V
voltage feedback max: 0.946V
0.946V will output 18V
how many volts do i get from 0.92

nvm i found out

[spice turtle]

365F seems to work well and I set my dial on "3 to 4"

[spice turtle]

what IC are you using? most of the time they tell how

[idle python]

mp1482

i didnt realise it tells you how to in the datasheet

well this makes more sense now

vbat has a voltage of 3.7v

which is the typical voltage for sim modules

[spice turtle]

https://www.monolithicpower.com/en/design-tools/design-tools/dc-dc-designer-online.html

theres also this tool

what an interesting tool 🤓

[idle python]

do they have mp1482

[spice turtle]

yes. On the list to the left, it pops up

[idle python]

what kinda diode is this

nvm its a schottky

wait what

its a zener

but the symbol is schottky

nvm i thought the symbol was schottky\

ok i need help

i addeda zener diode in this simulator

but i dont know how to set these options

i want them to match these

https://octopart.com/mmsz5231bt-onsemi-58698887

[torn saddle]

Hello , Good Morning. I am looking for someone to layout a PCB for a Arduino Zero hat. Is this the place to ask if someone want to layout PCB board? The schmatic is in KCad. Is someone interested in making PCB payout?

[formal sand]

How much are you willing to pay?

[torn saddle]

$50- $100 is that fair?

[woven bluff]

this kind of job is paid by hour

[torn saddle]

I am ok, with pay by hour, If interested ,,,,,

[woven bluff]

you can post your schematic here so people can see how complex is it.

[torn saddle]

ok I will post today

[scarlet jay]

can it be like this?

like an open pad module where I will put WIRAH! 🙂

[formal sand]

What is this project for? College?

[supple pollen]

Yes, just set it up as test points, pads, or terminal blocks. You can then either install headers or terminal blocks and hook the wires to those, or just solder the wires directly to the pads.

[latent grotto]

Im looking at the circuit diagram for the esp32 s2 SAOLA dev board. A lot of resistors are marked as 0 ohms

https://dl.espressif.com/dl/schematics/ESP32-S2-SAOLA-1_V1.1_schematics.pdf

?

I was wondering if anyone knows if they are actually 0 ohms

or left out on the board

Because I am very confused on whats happening

Another instance can be found near the USB port on the schematic

Its too hard to see whats going on in the image

[dry pelican]

I think they usually use 0 ohm resistors as jumpers but there isn't really any need for jumpers when the components do not allow traces underneath them

maybe for testing or prototyping. Or some sort of way to allow the end user to disable features

[formal sand]

Zero ohm resistors are an interesting topic. They have a lot of uses as crazy guy described. I have seen them be used generally during the development of beta units. In the case where you might want to add an actual resistor later on for some sort of application. You will see them sometimes connected to io lines of mcus. In some cases you might want to replace the ones near mcus with switching signals for a certain value of resistance later on based on emc testing. Limiting current of switching signals reduces interference to other equipment.

Other times they might be useful to remove a certain function or probe a signal

[scarlet jay]

thank you!!! 🙂

[hidden pulsar]

hi im looking to find a suitable contact pad for these pogo pin connectors

would these brass sheets work? i did a bit of research and found this to be pretty similar to typical implementations

[rustic flume]

These pogo pin connectors are likely gold plated. the brass is not and will oxidize depending on moisture in the air, which will eventually impede good contact.
Have you considered desigining a circuit board and ordering it with gold plated contacts?

[hidden pulsar]

hmm that would be a first for me so i hadnt really considered

i also need to embed either of these into a custom 3d printed design

[rustic flume]

check this out: https://youtu.be/qhr6h7wuqo0?t=707

it seems gold plating is a common service offered by these service companies and if you only need a small PCB they are apparently also quite inexpensive.

[hidden pulsar]

right thanks

i was thinking i dont want to overcomplicate things though since theres not much circuit involved

[rustic flume]

they have a lot of shapes

[hidden pulsar]

o yeah okay ill look at those again

problem is im not finding any big pads here

i need it to be because the pin touching down wont be very accurate

[rustic flume]

maybe a piece of stainless steel

brass will be OK for a while but only if the environment is nice

and it probably needs a clean after a while with some abrasive

[hidden pulsar]

gotcha

it is a one off application project

but stainless steel might be smart

[hidden pulsar]

another q, what might be a good sheet material for attracting to an electromagnet. needs to be relatively lightweight and not too expensive preferably. also needs to support some wear

[rustic flume]

i think your only real option is steel

which isn't exactly lightweight

be aware some stainless steels are not magnetic at all

the cheap mild steel commonly sold as construction material has excellent magnetic properties

If you want a very abrasion resistant surface, i would consider a spring steel piece, it's hardened thus very abrasion resistant.

[supple pollen]

There are parts made for the purpose. Here's one example https://www.mill-max.com/products/new/low-profile-spring-loaded-target-connectors

[spice turtle]

How would you interpret C1 and C2? In the datasheet its called out for cap "CC" to be 1uF. Would it be safe to say that caps C1 and C2 are probably 1uF?

I suspect they might be decoupling caps, but I feel they would be called out as "0.1uf". Odd that Cout,Cin and CFF are given the proper units 🤔

[hidden pulsar]

thanks for the suggestion. i need the contact pad to be very large though as the connection wont be very accurate

[supple pollen]

"105" would be 10*10⁵ = 1000000pF, or 1000nF, or 1µF. "104" is the common 0.1µF designation.

[spice turtle]

thanks. also I used a super react and it did something cool, lol.

[supple pollen]

That was cool, I didn't expect that

[spice turtle]

Got one left 🙂

[supple pollen]

I did not even know super reacts were a thing

[spice turtle]

me either. Must be a new feature on discord

[supple pollen]

Apparently it comes with a subscription

[spice turtle]

I got two free ones

[ivory beacon]

Hey, I'm quite clueless as to where to go with this but I sense this is not the right destination, got myself a fancy bluetooth module pcb from a car that is not working, it has hundreds of capacitors transistors and more ic's than hands on my fingers. How might I go about troubleshooting this that wouldn't take a week of unsoldering and testing half of it?
If you've got advice on where better to ask this I'm all ears, thanks.

[unreal flax]

This is one of those "if you even have to ask the question, this is probably not a task you want to tackle" situations, I think.

[ivory beacon]

I mean, I know how to check components.. however some of them need to be unsoldered to check and the quantity of them makes this a task that would take a week

was largely wondering if there's a speedier way to determine a fault, I tried powering it and seeing heat concentration, but nothing

and obviously there's no visual indications of faults either

[supple pollen]

When you say "is not working", what are the symptoms? Are there any LEDs? Does it broadcast any Bluetooth signal? What other inputs and outputs does it have? Is there a CPU? Is there a debug port or test points?

[formal sand]

The way I troubleshoot pcbs is "what is not working?" and "does anything look broken in the area that is not working?" Usually that gets me like 70% of the way there.

[ivory beacon]

No leds but it features a fiber optic line for communication, which seems to power up on boot but never again. No Bluetooth signal is the mode of failure. There is a cpu and several other semi large ics, I can send a Pic tomorrow as it's all packed up already. No debug or test points available

My best guess is the issue stems from the fiber line, as to my knowledge that is what data transfers over, so it should talk to main entertainment ecu and kick on the Bluetooth.. But it's not exactly apparent where that bit starts or ends. As it's a multi layer pcb I can't even follow the traces

Maybe I'll notice something I've missed with fresh eyes in the morning

[formal sand]

You might want to post some clear pics of it here, we could take a look and see if there is something obvious

I'd say that for vehicle based stuff my guess would be a cracked component or solder joint

[low anchor]

so this page https://learn.adafruit.com/introducing-adafruit-stemma-qt/seeed-studio-grove suggests that STEMMA and Grove are cross-compatible, but it doesn't seem clear if a Grove cable will fit into a STEMMA connector or the other way around

[rigid fiber]

OK, I am really getting sick of JLCPCB not having the parts in stock. Every day I go to submit my order, there's another part that's no longer in sufficent stock. So I spend a day finding a new one (EasyEDA Std search sucks raw amperage), redoing the routing, double checking teh routing, only to find that a different component is out of stock.

[formal sand]

Welcome to designing anything after covid

You should have seen how it was in 2021...

[hushed smelt]

For some reason I saw that going very differently than panelizing every single one and wasting all that space...

[rustic linden]

oof

[hushed smelt]

would that have been so hard?

i must have missed some option...

i must have messed up the dimensions... it's massive.

🤣 i'm dead, this is so dumb.

[hushed smelt]

that's not a work light, that's a space station.

i don't even know what to do with this thing now. can't use it as a fume extractor platform it's way to big. maybe some kind of table? ideas appreciated.

yup, i entered the diameter (250mm) as the radius. it made the radius 250mm, it's twice as big as it should be.

[spice turtle]

It doesnt seem like anything looks wrong

[pale umbra]

make an angel halo with white led

[formal sand]

Yeah this thing is really dense, I might take a look after work on my pc to see if I see anything off

[formal sand]

Hahahahha use it, double the light

[supple pollen]

Mario cloud style wall light?

[supple pollen]

No wonder the original full-circle version was so expensive! I had guessed you had calculated the number of LEDs ahead of time.

[hushed smelt]

😖

Obviously, I never caught the mistake through all the revisions. Never even second guessed it. There's no banana for scale in EasyEDA it's just a grid.

[supple pollen]

NXP CPU, Xilinx FPGA, and what looks like at least 3 switching regulators. I'd look at the power supply first. I'd also check the solder joints on stuff like the large white connector and fiber launchers.

[supple pollen]

Maybe you could angle them somewhat so they overlap (horizontally so they don't block each others' light) to make a smaller circle? I could draw it if I'm not being clear (having no luck finding a pre-existing drawing of what I have in mind)

[hexed mural]

Underneath a round table as accent lighting might work. I assume you'd have to dim them down otherwise it'd be brighter than the sun lol

Could you make a cool clock out of them?

[supple pollen]

There, drafted it myself since I couldn't find one

[worldly schooner]

Did you mistake the OD for ID 😂

Ohhh diameter vs radius. That would do it…

[hushed smelt]

OD is supposed to be 250mm. Then redesigned it to be an arc and measured the arc radius at 250mm. :/

[worldly schooner]

Yikes. I guess you have a giant ring light 3d printing project on the horizon haha

[hushed smelt]

https://tenor.com/view/bob-ross-painting-bob-ross-paint-gif-4952931

I might as well make a chandelier out of the thing it's so huge.

Put it on a winch and lower it down to the workbench whenever I need light. It's gonna end up being like 10 lbs of PLA.

[hexed mural]

Make sure to put a quick release on it to drop it on your enemies

[hushed smelt]

I honestly thought about that

[hushed smelt]

That's actually not a bad idea. If I could figure out the mechanics of it. I think OddJay just did an iris like that on show & tell this week.

It's impractically too big. I'm gonna have to redesign it and do it again. Use these for an infinity mirror or something.

[worldly schooner]

Might make for a fun floor display perhaps

[hexed mural]

Circular DDR Mat, is that anything

[hushed smelt]

for break dancing maybe, sounds like a neat thing.

[worldly schooner]

I could totally see this be cool for it’s own game thing with some sort of motion sensing controller(s).

[hushed smelt]

giant frisbee

[worldly schooner]

Spin the bottle, digital version 😂

Or if you like pseudo-gambling, a digital roulette table heh

I could see a lot of applications for this as an addition to a round gaming table tbh

[supple pollen]

Angle them, rotate them, and you have a fan that flings both air and light.

[worldly schooner]

Kinda makes me want to design something similar for a rectangular table now. Just general scorekeeping or status indicators to start…

Wait, no. I need a table first.

[supple pollen]

https://www.team-mediaportal.com/media/com_mtree/images/listings/o/2381.jpg

[hushed smelt]

That is actually one of the best implementations I've ever seen of LCARS, very nice.

[supple pollen]

I did an image search and chose one that looked good to me

[elder peak]

You know, I was just thinking about how I really ought to put the fancier LCD I've got on my ESPHome node in my geekroom and how fun it would be to make it LCARS.

[low anchor]

can someone clarify what 'number of positions' on a switch means? I'm looking for two SPDT switches in one package and supposedly this https://www.digikey.com/en/products/detail/bourns-inc/7829H-1-051LF/3759166 has 2 positions but it's just one switch, so I'm confused. (I'm assuming number of positions is like the same as it is for a connector, but evidently that's not right, so I don't know what it is now)

but if it's the number of positions the switch can be in, isn't a 1-position switch just a wire?

[formal sand]

The number of positions in this case means the number of contacts that common is connected to

[rigid fiber]

OK, I ordered just the PCBs plus a stencil from JLCPCB, and the parts mostly from LCSC, plus some form Mouser & Digikey.
Though now I'm getting like 10-100 of some of the SMT things for like pennies each. Way more than I need for 5 PCBs.

[spice turtle]

Start an excel sheet with your part inventory so you can keep track

If you got common values, you'll end up using them

[formal sand]

Agree, it will come in very handy if you do a lot of pcb design

And try to buy your ics as you are designing your schematic because they could be gone by the time you do layout

[pale umbra]

make a xena the warrior priestess disc that fly back to the owner with propellers

[supple pollen]

Visicalc?

[spice turtle]

wow thats...old lol. I had to look it up

[supple pollen]

Normally I just say "spreadsheet", instead of naming some particular flavour

[hushed smelt]

think i've settled on a large illuminated lazy susan for display pieces. seems most practical. been wanting something like that anyway.

[supple pollen]

That should look nice!

[pale umbra]

if you put magnet on the display pieces you might even be able to use leds to know with 1 look what slots are used and which arent

I almost wish I had one 😄

[hushed smelt]

There won't be slots. Will use a 24" clear plexiglass top with opaque film on top.

looked up motorized lazy susan designs and came across a really nice swiveling bezel Adam Savage recommends. that'll do nicely.

apparently people use them for R2D2 swivels. a lazy susan should be easy comparatively.

gotta turn this big L into a W somehow 😅

[pale umbra]

if it turns fast enough could have done a mower too I guess 🤣

like to cut trunks

[hushed smelt]

in other news figured out my workshop lamp isn't dying, the USB power supply is. plugged it into a different power supply and everything is good.

https://tenor.com/view/really-justin-timberlake-close-up-judging-you-intense-judgment-gif-11823466

[hexed mural]

At least you got a set of light-up boomerangs!

[formal sand]

Light boomerang is cool

[elder peak]

I think I'll fix my electronics workbench light this weekend. Because why wouldn't I, after the previous one died, assemble a workbench light out of some LEDs and a switching power supply that I'd designed? Except for the long and epic tale about zapping one of them, the parts shortage, and getting all set with the final version except for the want of a 0.2 ohm 0805 sized resistor.

[hushed smelt]

pretty much sums up what i went through too except i made the light 2x bigger than the housing 😛

i didn't forget the resistors though, they're on the other side of the pcb. :/

[supple pollen]

You can make a 200mΩ resistor out of a short length of wire and tack it to 0805 pads. I ended up fixing my worklight too, but it needed a new ballast.

[echo zephyr]

is it possible to solder on gold fingers of a pcb? like solder one gold finger to another?

thinking of doing that because i couldn't find any edge connector that will work for the gold finger i need (every connector with 40pins and 1.27mm pitch is out of stock)

[unreal flax]

Sure, those pads will take solder just fine.

Though if you're doing things side-to-side, it will probably end up somewhat mechanically fragile to bending the solder joint.

[echo zephyr]

hm, i guess i'll get the connectors from adafruit instead

[woven bluff]

are there SMPS chips that regulate by output current instead of voltage?

[unreal flax]

Yes, constant-current regulators. You'll often see that mode used in LED drivers, for example.

[woven bluff]

most of them are linear regulators, I need a buck convertor

[spice turtle]

Theres quite a few buck SMPS Drivers out there for LEDs. Substituting the first feedback resistor for LEDs also works. Your LED current is then dictated by your feedback voltage divided by the resistor

Look at NPC3065. Its an older chip but it will do what you want

NCL30160 is one Ive used in the past

MP2489 is also another

MP24894 if you need more current

[idle python]

what type of capacitor can be 470uf, small (not taller than for example regular rpi 4 gpio pins) and polarised?

[elder peak]

@woven bluff The PAM2804 and AL8861 are two switch-mode buck converters that are cheap, have been available lately, and don't demand too much out of you.

[waxen halo]

If I'm using a constant current LED driver, could I attach two LEDs in parallel to one of the outputs with the expectation that it would split the available current?

[unreal flax]

That's a somewhat risky circuit, because LEDs will have slightly different forward voltages, so the one with the lower voltage will tend to grab most of the current. At best you'll get uneven brightness, at worst you'll blow out the LEDs in sequence.

[waxen halo]

Are tolerance differences enough to worry about that? In other words if I only am bundling similar LEDs together is it still a concern?

[unreal flax]

The V-I curve for LEDs is pretty steep. There would be some range in which it would work, but forward voltages also vary with temperature, etc. so it's the sort of thing that's could work for a while then suddenly break, too. If you need parallel connections, you should have a resistor in series with each one too to absorb the variance. With a constant-current supply it's much more typical to have multiple LEDs in series, though.

[waxen halo]

Yeah, unfortunately I have the board already and the needs changed such that I want double the number of outputs and there isn't really the voltage to chain. Trying to take outputs that were meant for 1 led and put 4 on it (they don't need to be very bright). I have 5 volts so I could maybe chain 2 reds, but not white leds.

For this particular project it's relatively permanent so led burn out is likely outside the risk I want to take

[waxen halo]

How would the resistor to absorb the variance work? Am I trying to measure the fV for each and put resistors to even them out?

[unreal flax]

No, I meant just to use the same resistor value for each, just to flatten the V-I curve, since there would be a combination of a linear resistance and a nonlinear diode response. Then, for instance, if one LED had a 0.1V forward-voltage difference, you might end up with it being only 10% brighter instead of 100% brighter. But you'd lose some power efficiency since the regulator would be running at a higher voltage than it would with only the LED.

[idle python]

why does it have 6 bad connections (red dots)

[spice turtle]

Falstead id strange like that. You cant go across, you have to go dot by dot

[idle python]

then i get the errors capacitor loop with no resistance

the circuit does work if the wires go across

nvm now it doesnt

[dry pelican]

If you want parallel caps, reduce them into one capacitance that is the sum of all other caps.

Oh

That also makes sense

Software should be able to handle parallel caps

[idle python]

from https://falstad.com/circuit/directions.html:
Capacitor loop with no resistance! – it’s not allowed to have any current loops containing capacitors but no resistance. For example, capacitors connected in parallel are not allowed; you must put a resistor in series with them. Shorted capacitors are allowed.

found a demonstration of parallel capacitors

heres the description of that

i dont think this is supposed to happen but idk

[distant raven]

You’ve essentially made a voltage multiplier

Just missing another diode though..

[idle python]

i wanted to see what this really does

[distant raven]

Oh

The zener diode clamps the voltage

Or it’s supposed to anyway

[idle python]

this is what VBAT label is

i dont know why it appears twice

[distant raven]

Likely an over sight

[idle python]

vbat label has voltage of 3.8 and able to give 2a max

basically what the sim module uses

[distant raven]

Yeah, the zener diode is just to ensure that the voltage doesn’t go over the max rated for the vbat pin, it does that by “clamping” it down

[regal lodge]

how am I doing so far... 😐

full view... any tips appreciated...

[worldly schooner]

Without actually knowing what you’re connecting, can’t say I can offer anything that would really benefit you, but it looks like good progress?

[regal lodge]

low speed

arduino (mega) based

[supple pollen]

That'll generally be an electrolytic capacitor

[supple pollen]

I look at all those pads and am reminded of the DIP-64 packaged 68000 chip.

[regal lodge]

close in a way, edge connectors

[dry pelican]

That looks like a job that would benefit from an autorouter

[woven bluff]

I find it takes as much time to tune an autorouter as routing manually.

[supple pollen]

I like the look of hand routing enough that I just hand route all my boards. Like LadaAda, I find it a kind of pleasant and meditative exercise.

Although apparently the RGB LED glasses were a challenge

[regal lodge]

where it started

how it's going...

[dry pelican]

I'm just saying that since it's low speed and there are a lot of connections that don't need specific parameters, an autorouter might help. But also hand routing can be kind of relaxing or satisfying like solving a puzzle.

[regal lodge]

the 2 layer version of the board was auto routed...

[supple pollen]

That's not too bad for autorouting. 6/6 spacing?

[regal lodge]

not sure, didn't make it

[supple sky]

Is the capacitor here just to make the neopixels run more smoothly?

[woven bluff]

that cap should be the SMPS output filter cap.

you don't need it if you have a decent +5V source

[supple sky]

What do you consider a decent +5v source? I am still designing and have lots of different options so I want to get something that I can not worry about.

[woven bluff]

one of those off-line DC adapter with large current rating should do

like your phone charger, with at least 2A rating

[supple sky]

Oh nice, okay. That's easy enough.

[woven bluff]

also depends on how long is your stripe

[spice turtle]

generally its extra filtering just in case you dont have good 5V source. Also supplies more current as you light up more LEDs

[supple sky]

Well, it's actually a few different sections so a 24 LED ring, and two 50 LED strips. I really wanted to use one source but not sure if that's the move.

[woven bluff]

check the max led current, multiply by the number of LEDs

I'm not familiar with this type of LED, I think they are mostly indicator instead of high power light source.

[supple sky]

Oh I am using neopixels, this is just an example someone gave me to fix my undervoltage issue lol. I was powering all of that off from my boards power.

[woven bluff]

capacitor just help start-up load transient, if you have voltage dropping it's likely your 5V source does not have enough current rating

[compact frigate]

RP2040 has two SPI controllers. Does sdioio occupy one of them?

[woven bluff]

sdio is not SPI

there is no native support of sdio on rp2040

you could try to program the pio to emulate sdio

STM32F4 has native sdio

[supple sky]

So I tried with and without the capacitor, it seems to be working exactly the same. Does that seem right?

I also didn't use the resistor, is that bad?

[distant raven]

They’re recommendations to make the controller and LEDs last longer

[rigid fiber]

Do people have a recommended brand of solder paste and applicator?

[dry pelican]

I've used chipquik and MG chemicals pastes

but I haven't found a good tub at a good price. Those are just syringe tubes

Is this good routing for an RP2040? The 3.3v line is only 0.2mm and I'm not sure if I need to make it thcker

Here's the stuff around it

[distant raven]

That’s permissible as long as you tent via on the board

[supple pollen]

I'd change those acute angles to right angles (in the old days, sharp corners like that acted as "acid traps" during board etching and caused problems: I'm unsure if it's an issue these days, but I still avoid doing it)

[distant raven]

yeah, and it's not really a good idea to put a trace between pads of the same net like you've got for 3.3v as it can make board houses think there is a QC issue during assembly.

[dry pelican]

I'm more of asking about the current carrying capability of those traces and if the thin traces will cause any problems.

[distant raven]

nah, it's unlikely you'll use more current. but I still suggest removing that trace between the 3.3V pads, it's much better to not have it.

[supple pollen]

I normally use wider traces for power, but just the width of the pads for the branches going to the pads themselves. However, the current draw here is low enough it's not likely to be an issue (I believe the chip has an onboard regulator for its core voltage anyway)

[distant raven]

it does

the RP2040 is nice in that regard

it's nice in a lot of regards, but not having to manage ICs for both core and IO voltages is nice.

[dry pelican]

I did that.

The RP2040 does not use much power, but I've seen designs that have some thicker traces going to it

Thanks for the tips

[fervent lance]

I have a DC buck converter with an input voltage of 10V-30V and an output of 5V. Would it still provide a usable output with a 7.4V (Lipo) input?

And, one other question, can a 7.4v Lipo handle a 130 brushed motor (~8V) and a Pi Pico W (L2983D drivers)?

[spice turtle]

This is 8.4V Max (4.2V/cell) to 6V min (3V/cell). Generally yes, you can get pretty close to your input voltage and still work. Id test it with a variable powersupply to make sure

[spice turtle]

Only thing that looks concerning is VSYS max, which is 5.5VDC. You may need a pre-regulator/.

[woven bluff]

chipquick

[supple pollen]

I've tried the i-Extruder for applying paste, but it turns out it doesn't do too well with thick paste, so perhaps T5 particle size and thinned with extra flux is the way to go (the extra flux also makes it harder to put too much solder down, which is easy to do with surface mount). I have a personal fondness for leaded solder, but it's not for everybody. While there is 36/62/2 paste available, I've only seen it in a T3 mesh (too coarse for syringe application, IMHO), but SRA SSWS-T5-15G is a nice formulation with RMA flux.

[split obsidian]

Question, if you just need to run the motors with a marlin board, for like...let's say an ender 3, but not use any bed or hotend heaters...how much amperage would one need to run 4 motors? How about just one motor? I'm interested in repurposing the board and motors for another project but want to downsize the power supply to something much smaller if that's at all possible

[worldly schooner]

That would depend on how much load the motors are driving and what kind of drivers you use to drive the motors. What kind of project is it?

The removal of a hotbed usually reduces the amperage requirement significantly, but it’s still good to keep some extra headroom to be able to keep up with the spikes needed for steppers.

[spice turtle]

I think the Tmc2208 drivers (and the like) are rated for 2A max each, but dont forget, its not constant. They are only powering steppers when needed. Theres usually 5x of the drivers, so if you figure 2A each worst case, Id say 10A.

"The TMC2208 is capable of driving up to 2A of current from each output (with proper heatsinking). " dont forget the proper heatsinking part 😉 lol

thats actually not a bad Idea either. The old boards pretty much have the makings of a PID controller with a display.

[split obsidian]

Basically turning a gentle load very slowly. Think those camera slider projects? Load would be about that

[supple pollen]

There are a lot of factors involved, what I would do is measure it and see, maybe stall it to see what the stall current is and allow for that in case anything gets stuck.

[idle python]

way too big but i found a tantalum with 470uf its pretty small

now how do i make pcb antennas

[idle python]

whats the point of these resistors

i want to convert the usb signal from a sim module to SPI so i can use the spi with a raspberry pi 4 or esp32

but i dont know if i just simply wire this without adding any other components or whatever else

[acoustic flicker]

it's to prevent being Hi-Z

[twilit mango]

Greetings! I am interested in purchasing some Schottky diodes from Digi-Key, and my searches so far are not making this a reasonable task for me. So... I'd like some help. I also don't really know enough about them to be able to put together a better search than I currently am. My use case is being able to include them in wiring between, for example, a Feather and an external in-line LiPoly charger (such as the solar charger). I managed to find one in my housemate's kit, but I am certain they will come in handy again (if only to replenish the kit). (Please reply-ping; I'm also in a meeting that might delay my response.)

[rustic linden]

For a Schottky diode, LadyAda's diode guide suggests the 1N5818, with this being a direct equivalent to the linked model https://www.digikey.com.au/en/products/detail/onsemi/1N5818G/1474209

sorry if my typiing is a bit weird, going through a hupo

@twilit mango

[twilit mango]

Oooooh, thank you. I didn't even think to look into a guide. That number sounds familiar to me, so I'm betting you're right.

[twilit mango]

Really appreciate it! Thanks so much! (Searching for that version returns only 32 results, which is so much more reasonable than 5k 😄 )

[supple pollen]

There's the easy way (copy one from somewhere else), and the hard way (learn up on RF and antenna science, impedance, striplines, etc.) and roll your own.

[rustic linden]

Glad to be of help! It might be a little outdated, it seems her wiki hasn't been updated since 2016, but it seems like a decent starting place. I wonder if LadyAda would consider doing an updated version of the part finder as like, a learn guide or a blog post idk

[scarlet jay]

problems, can help? 🙂

[twilit mango]

That was the exact one I used previously, so it makes sense to get that one as a replacement. I couldn't remember what I had used. You were super helpful!

[supple pollen]

Looks pretty basic, I'd verify that the computer is connected to the right port (the one the programmer is on), and that the programmer is properly connected to the target (there are some oddball connections required in some cases, for example some programmers need a sample of the processor supply voltage to generate the correct signal levels)

[idle python]

forget it ill just add an antenna connector

[idle python]

do the resistors affect signal speed

like what would be the difference with resistors and no resistors

[worldly schooner]

A resistor pull up in this application appears to be required.
Typically, chips like these use an open-drain to switch the signal, which means that all the chip does is connect and disconnect the signal line to ground. In this configuration, if you were to not include any pull-up resistors, the signal would stay low and never go high. If you connect it directly to VCCIO, you risk a short circuit.

[worldly schooner]

The resistance does affect rise and fall times of the signal, which does significantly affect your maximum speeds. Lower resistance decreases your rise times (aka more speed) at the cost of increased power consumption.

[idle python]

wow bruh this is complicated

[worldly schooner]

When in doubt, stick to the recommended application circuit. Harder to go wrong that way.

[idle python]

@worldly schooner what about this

the tiny image above ft220 is this

[idle python]

updated version

basically this is for a esp32 cam i wanna make the esp32 send video data with lte to a rtmp server

thats why there is usb spi bridge in this schematic

i dont know hwo i will make a driver that sends ip tcp udp packets through spi

maybe i can use some parts from wifi.h?

[scarlet jay]

nothing works, tried Microchip studio and both Arduino IDE to bootload but none works

[supple pollen]

Or you can use a chip antenna.

[supple pollen]

What programmer are you using?

[scarlet jay]

Atmel ICE

it is short of voltage, just 0.1-0.3 from atmel

but that isnt the problem, I connected it

didnt work

[supple pollen]

You're using the ICE to supply the chip's power?

[scarlet jay]

yes

because its with pins but not through usb sometimes

i have 4 diff boards

[supple pollen]

I don't think that works, I think you have to supply power to the chip separately. Possibly a clock too.

[scarlet jay]

tried, doesnt work

[supple pollen]

You keep saying "doesn't work" but not giving any details. You realize that isn't much to go on.

[spice turtle]

So Im guessing you are using them as a power "OR" circuit? So chose between battery OR 5V? Or are you using it to stop 5V from feeding back into the charger?

I know someone else answered your question, but theres a special case use called an "ideal diode" where you can share the load. Just want to mention it in case you need to use such a thing in the future:
https://ww1.microchip.com/downloads/en/Appnotes/01149c.pdf

[spice turtle]

you usually get this error if you have CLK and DAT backwards.

or the whole 5 pins.

[scarlet jay]

tried doing vid but it takes a lot to upload but everything from tutorial and nothing works 😄

is there a way to save the boards if they are unbootable?

@spice turtle

its basically feather m-0

[dry pelican]

What is a mother?

In your context?

Motherboard?

Or typo?

[scarlet jay]

typo

[dry pelican]

Use the esp32-cam's spi pins

Why are you using a usb to SPI bridge?

[spice turtle]

de-solder the chip basically. Its very hard to brick an IC. Ive bricked programmers before, by letting it hit 120VAC but thats another story...

I doubt you bricked it.

[scarlet jay]

okay!

[idle python]

esp32 cam doesnt have an USB port or USB pins which is kinda dumb, so ill have to use SPI instead to transfer data in and data out

i know it seems like im trying to reinvent the wheel and i could just use some portable lte router and tape it to the box i put the esp32 in, but that would defeat the purpose of the tiny size, and theyre pretty expensive, plus it wouldnt be a learning experience and it wouldnt be that interesting to make

there was also the alternative of using UART for internet but it has rly slow transfer speeds

[idle python]

like these things?

i dont know what kind would be suitable for cellular connection

[unreal flax]

The short answer is that you shouldn't be messing with anything cellular except certified modules.

[idle python]

https://www.quectel.com/wp-content/uploads/2022/12/Quectel_EG800Q-EU_LTE_Standard_Specification_V1.1.pdfquectel ec800 has been certified in europe

[unreal flax]

I meant "don't mess with the antenna that it's passed certification with".

[idle python]

ohh

alr

yall ever search on baidu for chinese modules that are very unknown

[worldly schooner]

Yes, but I wouldn’t trust those for anything wireless unless I know exactly what it is and whether it’s certified.

[idle python]

i guess ill just use external antenna connector then

[idle python]

i can use this

it comes with a sim800

[supple pollen]

Yes, while those look simple, I just buy them, there's a lot more engineering in them than you would think, as cellular operates on a wide range of frequencies. Additionally, having the antenna a little distant from your circuitry is good, as it both improves antenna performance and reduces problems of the cellular signal interfering with your circuitry. I'm speaking from painful experience here...

[idle python]

well

still not sure about this

im about to design the pcb but i wanna know if theres anything obviously wrong with this

[supple pollen]

While I don't see anything obviously wrong, it is a little confusingly laid out

[idle python]

i can agree

i didnt have a lot of space

for the step down part i tried to copy waveshares step down module

but the symbol for the mp1482 had weirdly laid out pinouts

and i dont know how to change the positioning of those

[regal lodge]

might finally be done with this board.... anyone see any issues? it is a 4-later but inner planes are power/ground

mostly concerned about anything tombstoning or something along those lines

[hexed mural]

Out of curiosity, what's the "JLCJLCJLCJLC" text there for?

[flat vigil]

it designates where JLC puts their order info

[regal lodge]

^

I put it there so it will be under/hidden by the USB port

not a huge deal since it's just a glorified breakout board, but it only takes a few seconds to place and is free soooo

[lone basalt]

I tend to leave mine wherever they put it tbh, doesn't really effect me and it's useful to know for like customer support tickets

[regal lodge]

with JLC?

[lone basalt]

Yeah

[regal lodge]

I mean, this is a once off board, if I order it again it'll be because I need a revision

[lone basalt]

Especially for limited run boards for like prototyping stuff

[regal lodge]

it's for proof of concepting a chip

[lone basalt]

Which chip?

[regal lodge]

why would I need the order number on the board when I can look at order history to get it XD

[lone basalt]

I have like 4 different accs, (one school, 2 personal because my google acc became unlinked at some point, and one for projects related to my hackathon) so tracking down the order number from the history can be a challenge

[regal lodge]

fair enough

[lone basalt]

That reminds me, I need to design a minimal board for the rp2040

or should I say a board with 5 of them + a usb hub built in

[regal lodge]

wouldn't that just be the 2040 zero

idk if it's full open source, the schematic is at least, https://www.waveshare.com/wiki/RP2040-Zero

could use it as a base?

there's an even smaller one, forget the name of it, but it omits the USB port to get even tinier lol

[lone basalt]

Yeah, I think I know how to do it, rpi published a minimal hardware design guide

The hardest part is finding resistors and stuff that match their specifications which are in stock lol

[regal lodge]

people are using the zero to mod the Nintendo Switch 😅

[lone basalt]

My general idea is to build something like a Kubernetes cluster using picos (ie, custom rp2040 boards) and make them do some calculation

[twilit mango]

To stop it from feeding back into the charger. It's a super rare case, that, but better to be avoided.

[dry pelican]

Bitluni has done some work on clustering the $0.10 risc-v micros. Rp2040s would be more expensive. But also there's more software support

[supple pollen]

That ground parallelogram looks a little funky, I'd just do that as a pour

[brittle birch]

Hey Guys is this the right channel to post pictures of hardware designs?

[twilit mango]

This works or #show-and-tell also works. If you're looking for guidance, feedback or assistance, here is where you want to post. If you simply want to show them off, here or show-and-tell are both excellent options. Please choose one only and post there - we discourage cross-posting in multiple channels. Thanks!

[brittle birch]

Thank you very much for the quick response!

[spice turtle]

So I did a bit of digging, that uses the BQ24074, right?

Theres actually blocking diodes already internal to it

Following the path from Bat, its blocked by Q2.

[twilit mango]

Interesting. Because I was instructed to add the Schottky diode by Limor directly. 🙂

[spice turtle]

On the Out Pin?

[twilit mango]

On the JST cable between the Feather and the "load out" JST connector, or whatever it's labeled as.

I spliced it into the power line on the cable.

[spice turtle]

🤔 one sec

gonna butcher up a diagram

like this?

[dry pelican]

Yes

Sometimes there's also a diode from usb power to prevent power going back into the USB port

But usually voltage on usb is higher when plugged in so diode is reverse biased with usb plugged in

[twilit mango]

Yeah, I think so. Looks right.

It's this charger: https://www.adafruit.com/product/4755 (Which I think is what you said earlier.)

[spice turtle]

Have 5V main go through the charger instead of direct. then you use the chargers powerpath circuit, ie:

[twilit mango]

Oh right, there's no 5V going into the feather.

Only the 3.3V from the battery or the solar panel.

Through the JST connector.

[spice turtle]

Oh then you totally dont need the diode then.

[twilit mango]

OK, so here's the issue, I was told.

The battery gets charged by the solar panel when it is lower than 100%. Obviously. BUT..... The Feather has charging circuitry as well. The reason for the diode is to block the Feather charging circuitry from managing to somehow back-charge the battery as well. It's VERY unlikely to happen, but it is technically possible. Ergo, diode.

[spice turtle]

Are you using a specific feather or can I look at any one I can find?

(or are they laid out the same?)

The MCP73831 has reverse discharge protection, ie, it cant backfeed from the battery. I dont think the above problem would really happen. The solar charger has a separate connection for the battery, as does the feather. If you were using the feathers battery connection I could see it being an issue, but..why would you? The Solar charger has that connection already.

[regal lodge]

Idk what you mean.

[twilit mango]

They're basically the same in terms of the power and charging. You use the JST connector on the Feather because that's how you power the Feather with the battery/solar panel........

For what it's worth, I'm not going to be much help with the how and why here. I don't have a deep grasp on a lot of this, and was simply doing what I was told by Limor. I think I'm the one who came up with the question as to whether it could ever cause an issue with double-charging occurring, though. Which would have meant I put the idea in her mind. She did not dig into the hardware situation when I asked, and told me to use a diode off the top of her head. It's possible she simply didn't know without checking that the hardware already handles it.

[spice turtle]

But the Solar Charger can do that if you use the connectors for BATT and LOAD. Power path works like this: it selects what ever Input is greater and uses that, think of it as an "OR" circuit logically. The TI datasheet puts it nicely:

"The BQ2407x features dynamic power path
management (DPPM) that powers the system
while simultaneously and independently charging the
battery. The DPPM circuit reduces the charge current
when the input current limit causes the system output
to fall to the DPPM threshold; thus, supplying the
system load at all times while monitoring the charge
current separately. This feature reduces the number
of charge and discharge cycles on the battery, allows
for proper charge termination and enables the system
to run with a defective or absent battery pack."

Tangent: some power path chargers have an integrated boost converter so the output always stays at 5V.

[spice turtle]

Thats alright! 🙂 Ive had to do power path stuff before too and it can be daunting. Using a diode is the quick go to, but, since you have power path, you dont have to do that. Using the solar charger (or any charger with power path) eliminates the need for diodes.

I would probably still run this all by her (its a lot). But Im pretty certain you dont need a diode.

[twilit mango]

Thanks for digging into this! I didn't put a diode in the project I built that actually uses it, so this is good to know. 😄

The diode cable was for future use.

[twilit mango]

I'll run it by her if I think about it in the future, but for right now, it's all good. Thank you again!

[spice turtle]

Glad to help 🙂

[supple pollen]

This structure

[regal lodge]

That is filled though

I put traces so that I couldn't accidentally route a different trace there and slice the ground plane.

[supple pollen]

I couldn't tell from the pic, which is why I brought it up

[regal lodge]

Yeah, they are orange, same as rest of the layer

Oh, also, I usually work with zones not filled.

[supple pollen]

I do too, just leaving the dotted outline on, to reduce clutter on the screen. However I hadn't had occasion to put traces in a filled area before so I hadn't seen that effect.

[regal lodge]

oh another reason I do it is I mostly make open source hw so people like having the option

[tropic ember]

Hello, I am trying to identify the model of this TFT LCD and use it, any help will be apriciated

[spice turtle]

I have one of these. Its going to be an ILI9488 display most likely

Use 3.3V on VCC BUT the IO must be 3.3V

You can use 5V on main supply but the IO must be 3.3V

Oh, it looks like someone desoldered the on board regulator (U1)

http://www.lcdwiki.com/3.5inch_SPI_Module_ILI9488_SKU:MSP3520

[tropic ember]

Thank you so much! Can you help me with the touch which doesn't seem to work for some reason?

[spice turtle]

I havent used the touch portion myself, but I would suggest checking: connections first, that bent pin, and then the library used. If you can read the chip and compare it to one used in the Arduino library, that would also help

[tropic ember]

Graphically it works, the screen works, the touch I need to see how it works

[spice turtle]

They should also use the same IC

[tropic ember]

IC?

[spice turtle]

Integrated circuit. The little Black thing in an SSOP package.

[idle python]

why does jlcpcb charge so much for pcb assembly on both sides of pcb?? 50 dollars just for service, not even the components which i have to pay double for because jlcpcb wont let you assemble only one board

my keyboard is going insane

is pcbway cheaper?

[supple pollen]

Two-side assembly is a lot harder, as the bottom is no longer flat when it goes through the P&P a second time, and they have to glue the parts down (or solder twice) so they don't fall off when you flip the board over.

[worldly schooner]

More often people will try to assemble all of the smaller components on one side of the board, leaving the easier hand soldering work to do on their own.

Even solder paste application is harder when the board has parts on the back haha

[formal sand]

Not having to populate boards on both sides also helps with pick and placing if you go professional

[elder peak]

tl;dr: the manufacturer will flip out.

[formal sand]

They won't, they will just raise the cost while they rub their hands together

[idle python]

screw gravity

they should just install a pcb manufacturing warehouse in space

jlcpcb it costs 50 dollars for both sides and on pcbway it costs only 30

only reason im using easyeda to make this board is because i can just easily grab whichever component i want and i dont have to import the footprint and symbol like i have to with kicad

[broken zenith]

Hi! I've been working on a revision of an ATtiny841-based PCB that fixes some bugs and adds a built-in USB to serial converter. I'm not sure if I did it right, though, and I would love some input. The chip in question is the CH340E, and it's in the top right corner of the attached schematic. The RTS pin of that chip is also directly connected to RESET of the MCU, which I don't think is right, but I can't figure out the correct way. I'm also aware that it's a very messy schematic, but I'm still learning and am also open to general tips on that front 🙂

Please ping me with any questions/comments.

Thanks!

[idle python]

what capacitor size are these? 1205?

for comparison the tinier ones are 0603

[surreal tangle]

I've never done anything with an ATtiny, but that diode from RTS to RESET is pointing the wrong way if you want RTS going low to bring RESET low. Again, knowing nothing about the ATtiny, do you need to have a pullup on RESET, or does it have one internally? Also, the 1N418W datasheet shows a max 1V forward voltage, so you might want to use a schottky diode instead...

[worldly schooner]

1206 maybe?

[broken zenith]

Thanks for your advice, I'll take a closer look at those things.

[formal sand]

Prob 1206

[idle python]

alright

[idle python]

is it dumb if i have DATA_EN instead of SIM1_DATA

nvm i fixed it

i forgot a resistor

[formal sand]

That antenna connector looks far from your ic

[idle python]

its supposed to be close?

[formal sand]

Very close

[formal sand]

What ic is that one?

[idle python]

quectel ec800

i couldnt find the european version and china shares lte bands with europe anyway so it might work

[formal sand]

Generally you want to find the hardware design guidelines from the manufacturer of the ic. They will tell you how to do certain things like spacing of your antenna from your chip, matching networks, thermal recommendations, etc

Specially with antenna stuff you have to be careful because you might design something that you think it works with an antenna that you picked put the antenna doesn't match the transceiver.

You will end up with a device that can barely talk to the lte antennas

[dry pelican]

And other components need to be close to ICs

It makes routing easier and improves performance

Like put the inductor near the buck converter and place capacitors for filtering and decoupling near devices

[idle python]

would this atenna work

nvm

how do i find an antenna for this

octopart is terrible and it doesnt even specify the vswr

[formal sand]

Octopart will not specify the vswr

Vswr is a measurement of how well tuned is your antenna to your system at a certain frequency

[idle python]

oh

well where do i search for antennas then

ima look on digikey

[idle python]

also what is cable insertion loss

[supple pollen]

That's how much energy is absorbed by the cable between your components (in this case, presumably between your board and the antenna)

[idle python]

would it have anything to do with the frequency range of the antenna?

[supple pollen]

More accurately, it's generally related to the frequency of the signal. In general there will be higher losses at higher frequencies.

[idle python]

how come quectel make their own antennas but dont even talk about what quectel antenna you should use with a quectel sim module

why the need for multiple frequency bands

like the one on the left has a range from 700-2690 but the other has a range from 690 stops at 960 and skips everything from 960 to 1710?

[supple pollen]

There are a lot of cellular channels in use by different carriers.

https://www.wilsonamplifiers.com/blog/frequencies-by-provider/

[formal sand]

it is because it depends a lot on your end application, in some devices you might not even have the space for an external antenna, in others you might want to build it into the pcb.

Usually carriers have their own frequency allocations, you might want to first look into the frequencies that your module supports and then select an antenna that might work ok in those bands. In the end it is never going to be perfect because RF systems are a compromise.

[supple pollen]

Rarely are you going to want a cellular antenna on the same PCB with the rest of your electronics

[formal sand]

I've worked in some medical devices where we had to do it due to size constrains. it was not pretty...

[elder peak]

Okay, so I guess a 4 years ago, I placed my first PCB order.

[idle python]

yes but i dont understand the frequency bands

i dont get why theres 3 frequency bands on the right antenna

why are there gaps between the frequencies it can see

[dry pelican]

Antennas are weird

[idle python]

true

[dry pelican]

Well they can be

[idle python]

ykw im not even gonna bother with antennas i think im just gonna buy one from quectel and hope the best

[dry pelican]

Probably just get the one with the frequency bands that you need

[idle python]

thats what im gonna do

[dry pelican]

And if you don't know get the one that supports more bands

[idle python]

this whole project is going to be fit inside a tictac box

[dry pelican]

That is extremely small

[idle python]

no, not the smaller one

the bigger size

[dry pelican]

What is the project anyways

[idle python]

surveillance cam with 4g

i live in an apartment and someone keeps stealing my solar powered lamps, which is why i tried to make it small and un noticeable and as wireless as possible because i live on the 2nd story im not routing cables down to the entrance

plus i always wanted to experiment with lte connectivity and cellular stuff so i thought this project would be a good way to learn about this stuff

[formal sand]

Nuts

Hate when that happens

Regarding antennas, they are built in such a way that their resonant frequencies land in specific lte bands

[idle python]

ill buy one with the biggest frequency range, but i want ti to be a pcb because you know

i have to put it in a tic tac case

i dont want an antenna that sticks out like those older phones

i think i might cut the lid of the tic tac box so the antenna would have better range since it would have worser range if it had to pass through plastic

[formal sand]

Anything metal around the antenna will have an effect on it

Keep that in mind

[idle python]

could i isolate it somehow from the rest of the circuit boards?

like would electrical tape work?

ill worry about the antenna last

now i gotta design the traces and placement for these components

btw what width should i make these traces by default? and traces not for 5v supply, antenna traces or usb, i mean traces for things that dont need controlled impedance like SPI or UART or LED, etc

[worldly schooner]

Nah, most boards just avoid having ground planes directly beneath any kind of antenna.

[unreal flax]

Personally I default to 10mil unless I'm trying to be especially compact.

[idle python]

anyone ever use an online ruler?

ik it seems silly but it actually works pretty well

[dry pelican]

EasyEDA has a measurement tool built in

[idle python]

ik but i like to see what my board would look like in real life, that way whenever i expand the board i never think that i made it too big

finished 1 part of it, do i finish placing the rest of the parts, or do i make the traces for this voltage stepdown then place the components along the way?

[hushed smelt]

Which some of us are bad at using sometimes. 😅

It works very well, when you remember to use it in the correct units.

[idle python]

here comes the part where i gotta choose between multiple component placements

[dry pelican]

That looks good

but can you route it

once you finish more PCBs, you get better at routing

when laying out the components, you have to think about how to route

and with that tight component placement, you might need a 4L PCB

yeah 2L is going to be very difficult

and you have some chips towards the edge

if you can't get enough vias under the chip, you need to move stuff around

[idle python]

should i put them in the center?

like this ft220

[dry pelican]

that needs to move

imagine how you will get the traces to other components

[idle python]

true

i need to route usb differential pair from the sim module to that cihp

chip

then the spi lines to these resistors

then the pis

pins

[dry pelican]

and spi is pretty high speed so put that chip on top near the sim module and complementary components

idk why spi defaults to 10MHz (or maybe it's 5. idk) or something

[idle python]

the spi clock speed will be set by the esp32

ill set it to whatever the max is

[dry pelican]

don't set it to max

low speed signals are more stable

unless you need high speed

[idle python]

oh

[dry pelican]

but past a certain speed, the signals need special routing

[idle python]

15mhz works? sim module max up/download speeds are 5/10 mbps

[dry pelican]

although 15mhz may be fine without any special routing if you just keep traces short

I think it's mostly determined by inductance

the max speed

but where are you getting the USB signal from?

[idle python]

sim module

[idle python]

im definitely relocating that chip

[dry pelican]

usb signals need to run along each other and be short

they should be parallel

[idle python]

same length

[supple pollen]

We'll just ignore the Redboard which ran USB traces most of the length of an ATX mobo

[idle python]

in this case, its better if theyre short because if the usb traces are long they start acting like radio antennas

and i think it would screw with the sim module

[supple pollen]

The other way around is much more likely.

[idle python]

oh

this feels like im playing a parked car puzzle game

[idle python]

new design

[unique patio]

are you doing any routing yet? Instead of trying place components firmly now, consider connecting the pins per the schematic, and then look at the unrouted connections, seeing if rotating or moving components will possibly simplify the routing later.

[dry pelican]

Pcb routing in a nutshell

And @unique patio has a good tip. Try routing some components or sections of the circuit and change things later if you run into a problem

[supple pollen]

I do tend to start with specific connections (like antenna or USB) that should be short, and arranging the parts accordingly

[idle python]

i fell asleep so no routing done

[idle python]

does 15mhz spi need terminating resistors?

[spice turtle]

No, Ive never used terminating resistors with SPI in general.

[idle python]

oh

idk why the datasheet has them tho

[spice turtle]

that seems like a proprietary bus. Those arent terminating resistors though, they are pull ups

[idle python]

proprietary?

[spice turtle]

a thing that isnt open for all. Or its something they came up with

[idle python]

i guess i dont need the resistors then

[spice turtle]

If it shows them, Id add them, at least put the footprint in. Im not familar with that protocol though. It does look like SPI which doesnt need pull ups..

[idle python]

should i do 90 deg angles if theres a via between

?

or is that not good

nvm

i found out that 90 degree angles were more of a manufacturing problem than electrical problem

but that was a very old problem

[supple pollen]

90° angles are fine with vias

[idle python]

i love the 3d feature

[sage hill]

Hi, this is my first time here in the Adafruit discord is this the right spot to ask for hardware design advice? I had a quick question regarding switching power sources on feather boards..

[unreal flax]

Yup, exactly the right place!

[sage hill]

Hope I’m not interrupting an ongoing discussion.. I wanted to know if its advisable to have two diodes or the mosfet solution. To my understanding the mosfet solution seems to be better as it shuts off the load to the battery when vbus goes high but I wanted to make sure I wasn’t missing anything.

[idle python]

self bomb? that sounds scary

[supple pollen]

The main advantage to the MOSFET is a lower voltage drop, but the diodes do work.

[distant raven]

Some SPI devices specifically want pull ups if they’re expected to run at high speeds. Not sure how fast the ft1248 communicates over spi though

[sage hill]

Thank you!

[spice turtle]

https://ww1.microchip.com/downloads/en/DeviceDoc/51746a.pdf
its basically variation on this, except on its on the input.

[distant raven]

Oh it can support 8 bit parallel transfer up to 30MByte per second

Impressive

I can absolutely see why they want pull ups on the spi interface

[sage hill]

Oh nice, thanks!

[supple pollen]

I can see why they'd want terminators, but pull-ups for a bipolar drive signal? That just seems pointless to me.

[idle python]

max 80 mhz i think?

most im using is 17mhz (+2 mhz for good measure) and thats if i use both download and upload bandwidth

[distant raven]

Might be that they we’re saving money on the die costs and pull ups are just a side effect of that

[supple pollen]

That seems like a much less assertive statement than "I can absolutely see why"

[distant raven]

Well, some higher speed spi chips do want pull ups but I’m offering a possible alternative

Though TI E2E forum suggests that only MISO tends to need pull ups 🤷‍♂️

[idle python]

anyone know a sim slot without those

without these plastic bits

id have to add 2 drill holes to the pcb

[marsh nest]

Is there a reason not to? I've used microSD cards both with and without locating pins in the past, and honestly I much prefer the ones with them.

A lot easier to hand solder

[supple pollen]

Less likely to tear the socket/pads off the board too.

[idle python]

im not the one soldering them. jlcpcb or pcbway (whichever one is cheaper in the end) will be assembling all the components

i dont have any other place where i can put the sim slot, one drill hole goes through the resistor pad and the other drills through 2 pads

[idle python]

i can move the resistor but idk what to do about the other hole

[marsh nest]

Mm, yeah, I see the issue. Having to completely redesign a board because of one component is really annoying.

[idle python]

not anymore

i found a replacement part

sadly theres no 3d model for it

but it is much more smaller than the other one its a bit bigger than a normal nano sim card

[elder peak]

I have a vectorized image from a past photoshoot with my friend Anna where I used a prior PCB design and now I'm putting the image on the back of a new PCB and so kicad has me making a local footprint library directory 'anna.pretty' and I'm like.. Well, she is pretty, thanks KiCad.

[idle python]

lol

why does kicad name their footprints .pretty tho

man when i submit this pcb to pcbway theyre going to wonder what i was thinking when i made this

im gonna lie im the most organised person ever

[idle python]

i had to make some traces 6 mil instead of 10mil so i could get them through the crevices of the sim footprint

[supple pollen]

Straight between 100 and 53 would route better

[idle python]

after that screenshot i did reroute it between those points

[idle python]

i have no idea how to fix this

[supple pollen]

It looks like the via is too close to the pad under the "81". I would probably move the via upward, probably the same height as the SIM1_RST via, and then have the trace on the blue layer angle back down to the pad. Doing things like that, going out of your way a short distance and then going back, often become necessary on densely populated boards like this.

You can see five such backtracking examples in this board I designed a while back: two at the upper right (KEY1 and WIFI_IRQ) that go down slightly to a via, then up, a third in the middle right (GPS_ANT_ON) which goes right and then back left, and two more (M_EPD_BUSY and M_TEMP_SEN) at the lower left that go right to vias and then back left.

I might take out that extra zigzag in the +5V trace in the lower left of your board.

[idle python]

i fixed everything

that worked btw

track width point to these traces which arent really a problem

and the incomplete connection are the usb pins i was supposed to add but i ran out of space

[woven bluff]

why ZVS need two big chokes? If I use push-pull with floating channel gate driver, can I avoid those chokes?

[supple pollen]

The voltage has to come from somewhere, when you say push-pull, are you thinking an H bridge?

[idle python]

the way jlcpcb designed the footprint pads angers me

[distant raven]

Why?

[idle python]

theyre too big

[distant raven]

It’s probably normal for modules like that to have longer pads going out for mechanical security

[supple pollen]

More often than not, I'll just design my own footprints.

[idle python]

how long can an usb trace be without having its impedance regulated, just normal 6/6mil traces

[supple pollen]

It depends on which USB protocol you're using, if it's length matched, and how tolerant you are to USB errors. In reality, you can probably get away with a lot (especially if you include damping resistors), but you might have problems in some cases.

[idle python]

2.0

[distant raven]

2.0 is pretty resilient

6/6 trace/spacing I wouldn’t do for more than a few millimeters without termination resistors (usually 27ohm), otherwise I’d use at least 10mil traces, running as parallel as possible. Length matched isn’t nearly as important here. But you can if you want.

[supple pollen]

USB 2.0 "full speed" (12Mbps) is more tolerant than USB 2.0 "high speed" (480Mbps)