#help-with-hw-design

1 messages · Page 6 of 1

knotty tiger
#

oh, 22µF on both input and output of your 5V regulator? that's going to have a kind of large inrush current. you might want to think about ways of managing that

#

your 12V power comes in on EXP1? where is the ground from your 12V attached?

distant raven
low anchor
#

The 12v power should normally come in on the right hand side connector

#

But in this case yes I connected it on EXP1

knotty tiger
low anchor
#

Yup that sounds like what happened. From what I remember I connected 12v to EXP1 and ground to the top left of EXP2

#

In that order :/

#

What can I do to decrease the inrush current?

distant raven
knotty tiger
low anchor
#

I meant top right, my bad

low anchor
distant raven
#

Or 500mA?

knotty tiger
distant raven
#

But that’s the 500mA device

#

I didn’t look at the 1A device

#

Even the 1A device says 10uF/50V for input

knotty tiger
#

oh, the first search hit i got was for the -2000R3, which recommends 22µF/50V for input

distant raven
#

Not sure which they are using though

#

So 2000r3 might be right

hushed smelt
#

Is there a diode that will only allow power one way for 4 pinouts?

#

I can use 4 separate diodes but it would be far more convenient and space saving to have 1 diode chip handle all that.

vast flume
#

Yes

#

I've only seen them for esd-protection diodes but I'm sure they exist for other applications

#

Its a zener diode so you can't use it in general applications I dont think; but I'm sure theres other diode chips in the same package

spice turtle
hushed smelt
#

Thank you. I’ll probably just stick with 4 cheapy individual little diodes then.

fervent lance
spice turtle
#

Yes, its an analog switch, so it just matches the input. Think of it as if it were a relay, but an electronic one.

merry grail
#

I have a problem with a usb-c project. It did work, I even got a blinky running at one time. But now when I plug it in all other usb devices disconnect. I've looked for shorts, can't find any. What else should I look for?

merry grail
#

Turns out it was the jack itself.

supple pollen
#

Thanks for sharing what you found, I was curious.

woven bluff
#

@supple pollen seems icestudio support for ECP5 Evaluation Board is on paper only..... There is not even a single input clock.

#

now I have a paperweight sent to lab corner collecting dust

distant raven
#

There are a few ecp5 boards in the supported boards directory

woven bluff
#

that's all there is

distant raven
# woven bluff

Looking at the docs, it bases the clock speed of the fpga off the external clock. So it does make sense you can’t select or add it

#

Have you tried looking at the documentation for IceStudio?

woven bluff
#

the board has internal 12M and 200M

#

the nwjs GUI is not working properly for some reason

#

is IceStudio production ready?

distant raven
#

It’s open source, last release was April

#

It appears to work for a lot of people. Looking further at an issue about external clocking; icestudio takes care of it. I don’t think there’s any need to specify it

spice turtle
#

Has anyone used those cheap 433Mhz Modules in a 3D print successfully? Just wondering If I can get away with a thin wall and still be able to use them to get reception. Im not pushing out much data either

unreal flax
#

You should be able to without trouble... plastic doesn't interfere with radio that much, especially at lower frequencies.

spice turtle
#

Thats what I was thinking too. I dont need to go far, literally 20ft would be OK

#

I know not many people know about them, but I have these modules I bought due to their ease of use; E43-433T13S,. They are small but wayyyy to OP for what I need.

spice turtle
#

I should do a write up but Im using the "H5V4D" receiver and it seems to be running OK. I tied the data line high because Ive found it floats and gives crappy data a lot. I am beginning each "packet" with 0xFF and 0xAA. These seem to work OK. Baud Rate is 2400 BPS. Im only at a foot or two of distance, LOS. Transmitter is standing up, and the receiver is laying down.

As long as both are still, it works. I think this will work 🙂 I forgot where I got these though

supple pollen
#

I think the RadioHead library supports those simple radios, and there's another library optimized for them (I forget what it's called)

#

It may have been the VirtualWire library, which has been superseded by the RadioHead library anyway. There's also the RFTransmitter and RFReceiver libraries by zeitgeist87.

hushed smelt
#

The RFM95 and RFM69 modules that Adafruit use are compatible with radiohead, there’s mention of it in the library which I had to dive into for my rfm project. I have no idea what Radiohead is but seemed compatible out of the box. Just make sure you get a module with compatible mhz 400 or 900 for your needs.

#

If you only need to go 20 ft that should be within WiFi range. Radio will work line of sight. Putting a brick wall between you wouldn’t be advised but a pane of glass shouldn’t be a problem.

spice turtle
#

You know, I did forget to check if my complier has support for the nrf series line of RF transceivers and apparently, it does. It seems that in much the same way there are arduino boards, another company has kind of made the same thing for PIC micro called "Click boards". Well, turns out, there is a nRF24L01P board and it looks like there is software for it pre-written. These are also just as small.
ex:
https://www.mikroe.com/nrf-c-click

But..I wonder as well: how hard would it be to just DIY it myself?
https://www.mikroe.com/ook-tx-click

Ie, what if I just...remade the schematic? I dont exactly have rf testing equipment, but, how hard can it be?

unreal flax
dry pelican
#

This is the datasheet for the 61 ferrite material that I'm planning to use in a gate drive transformer for my Tesla Coil. I'm not sure what size toroid/ring I should get. Basically I want it to be "good" for 700kHz to 2MHz. The coil's current resonant frequency is around 1.2MHz, but I may want to change that to something lower that the driving components are better rated for. I'm driving a half bridge (probably 2 x IRF460P). I assume that I want a transformer that can work at high frequencies, has a high coupling coefficient, and does not emit much radiation (EMI). Would a bigger or smaller toroid be better? Or is there a better material? I've never read a ferrite core's datasheet before and I don't know what most of the things mean because I am a hobbyist and not an EE.

#

I think smaller is better, but I'm not sure.

#

The biggest toroid that I would use (because of space) would be around 22mm in diameter and 10mm in height, but smaller is preferred.

#

about the diameter of a quarter

#

coin

#

TLDR: Is 61 ferrite a good Tesla coil GDT material for around 1.2MHz driving frequency? What size should the toroid be?

#

80 material seems good as well. It's rated for power applications from 1-5MHz, which is where the coil is operating. Although there's not much margin on the lower side. 61 just says "inductive applications up to 25MHz"

verbal moss
#

Hello! UberGuide says to use 300-500 Ohm resistor close to 1st NeoPixel's DIN.

Does it matter if Thin Film vs Thick Film resistor for NeoPixel DIN?

Wondering if Thin Film should is recommended? My first PCB attempt has Thick Film resistors, they result in sloped/ramped signal in the pic. Doesn't look good to me, at least that's what the scope is telling me, which I'm also learning how to use.

Note, do NOT use BSS138 for high frequency signals (>400khz), they won't be level shifted. Something else I've learnt.

I read and searched through NeoPixel UberGuide https://learn.adafruit.com/adafruit-neopixel-uberguide/neopixel-matrices (pdf) didn't see specifics on resistor type, maybe doesn't matter?

Already posted to https://forums.adafruit.com/viewtopic.php?t=195153 but figure folks here will know

Yellow = signal from QT PY pin
Blue = signal at 1st Pixel DIN, just after 470 Ohm Thick film resistor.

Measuring signal before resistor looks identical to Yellow, indicating slope introduced by resistor, and that the BSS138 isn't shifting squat.

Cheers!

scope_after_470ohm.jpg qtpy_bss138.png sch_ws2812b-mini-470ohm.png
Adafruit Learning System
Adafruit NeoPixel Überguide

Everything you always wanted to know about Adafruit NeoPixels but were afraid to ask

distant raven
#

It’s not going to be drawing huge amounts of current

#

As for using the bss138 for level shifting, how is your level shifter set up?

#

I use the bss138 and my signal level shifts for neopixel. BSS138 is more than sufficient for 1MHz and below.

verbal moss
distant raven
#

Wait

#

Ignore me

#

G1 is probably the gate pin

#

seems right, this is what I do

verbal moss
#

Thanks for sharing your schematic, it's reassuring to see I got that ok.

Curious what BSS138 you're using? Maybe I picked wrong BSS138, used https://lcsc.com/product-detail/MOSFETs_MCC-Micro-Commercial-Components-BSS138AKDW-TP_C779245.html. For mine, I see 3V3 boosted to 5V if I set the pin to constant high. But I don't see any boost at NeoPixel 400khz or 800khz frequencies. I haven't investigated at what frequency the level shift starts to drop off. I didn't see an explicit frequency range in the datasheet, maybe I need to infer/compute based on the switch on delay (3.8ns). Currently digging through misc BSS138 datasheets to eyeball for differences...

BSS138AKDW-TP MCC(Micro Commercial Components) | C779245 - LCSC Ele...

BSS138AKDW-TP MCC(Micro Commercial Components) US$0.0848 - 50V 220mA 1.9Ω@10V,500mA 350mW 1V@250uA 2.9pF@25V 2 N-Channel 22.8pF@25V -55℃~+150℃@(Tj) SOT-363 MOSFETs ROHS datasheet, price, inventory C779245

distant raven
supple pollen
#

Würth Elecktronik publishes some good books on the subject (ABCs of Inductors and Trilogy of Magnetics). Art of Electronics also covers it. However, it's often quicker to eyeball similar circuits or commercial gate drive transformer offerings and use them as a guide. For many circuits in that range, their little 10mm dia by 3.8mm thick toroid is entirely sufficient.

#

The 80 material has much higher permeability than the 61, but I doubt you need much permeability at that frequency.

distant raven
#

I’m joking of course hehe

#

Toroidal cores are lots of fun. I’ve considered getting some bare cores and trying to wrap my own

#

There are some P200 ferrite cores that are just a few $ a piece that seem like they’d be a great starting point

supple pollen
#

Yeah, I don't think Fair-Rite offers a 200 mix. But there are some amazingly high permeability mixes out there. I've been playing with core memory built out of toroids, but the usual mixes available today don't generally offer the squarish B-H curve that's best for it.

distant raven
#

Most P200 cores are washer shaped. Like a squished donut from what I can tell

#

Toroids are not my strongest area so my knowledge of making inductors with them is limited

supple pollen
#

I did eventually find out that the Fair-Rite 85 mix seems like it would be good for core memory. They don't list it any more, but Mouser has some stock.

distant raven
#

I’ve watched a few videos but I’ve not done it

supple pollen
#

My early experiments are encouraging. You want a "hard" ferrite with a large magnetic remanence for good performance.

core-demo.jpg
distant raven
#

That’s really cool

#

Have you made a toroidal transformer at all?

#

I had thought it would be fun to make one

supple pollen
#

Yeah, a few hand-wound ones. I may make more, although my last homemade transformer used E cores and a bobbin.

transformer-secondary-layer2.jpg
#

Heh, I found a pic of my last toroid project: just wound a few turns on a "Flexiformer" to provide filament power for an argon laser. Ugly, but it worked.

flexiformer.jpg
#

@knotty tiger would be amused

scarlet totem
#

Is there somewhere I could read about thunderbolt and displayport conversions?

plush vector
#

I need help please guide

#

I'm consufed

#

Is the implemented of diodes and capacitor correct?

supple pollen
#

The diagram doesn't show what's connected to D2, nor where the lower connection to the battery nor C1 goes (presumably a 0V reference?). However, the breadboard shows largely the same thing: the cathode bands of the diode point toward the capacitor lead (which appears not to be the negative one). It's unclear where the capacitor's other lead goes (best guess is a black jumper to the red rail on the left, which is presumably the 0V reference?).

untold leaf
#

sorry, might be missing something... bss138 works great for me in level shifting ws281x data signals. If it works up to 2MHz, that's ok for 800KHz, no? (2000KHz)

jlcpcb assembly error sounds completely plausible to me. I have found the ws2812 5050 and 3535s to be very heat sensitive in reflow... you can actually hear the little twangs as the die wires break if you get it wrong. a damaged LED will likely have a break in its lens. there's a trick where you sharpie the lens to expose its structure better when troubleshooting (does damage appearance). you can also jump around a failed led by shorting its DI and DO pins with a bodge wire.

weirdly, the ws2812E (economic version) seems a little more reflow tolerant. but I have found all of these to be total pains and I prefer individual LEDs with dedicated LED drivers like IS31FL3733.

distant raven
#

I prefer the sk6812, it’s nearly identical to the ws2812 but it has a constant current driver. Colors are more consistent and I’ve had no issues reflowing them.

#

The IS31FL3733 is a great chip though. I thought about using it to make a tiny RGB display with the UHD1110 tiny RGB LEDs I have

#

Or there’s a version that has tons of pins for doing large RGB LED displays row/column counts

supple pollen
#

I wonder if those are the same tiny RGB LEDs used in the AdaFruit LED glasses

distant raven
#

I don’t think so, these are like 0404 sized LEDs

supple pollen
#

Wow, that is small. Yeah, the glasses ones are like 1.5x3mm or so, much "bigger"

#

Have any pictures of that display? It sounds amazing.

distant raven
#

The data sheet lists “displays” as a use so I want to try it out haha

#

I need a powerful enough RGB led display driver though 😅

supple pollen
#

I'd probably go with something like the HUB75 style displays use, with shift registers, drivers, and multiplexing

distant raven
#

If I’m going to make it, it’s gonna be at least 1.5”x3”

#

Which would be like.. easily 400+ LEDs

#

I could probably do 20 per inch, so it would be like 30x60 pixels

supple pollen
#

Wow, I want one. In that sweet spot between 2mm HUB75 displays and OLED.

#

Come to think of it, an OLED controller might work too

distant raven
#

Placing 1800 LEDs would be fun

supple pollen
#

That's a job for a machine or a fanatically dedicated human

distant raven
#

At that rate I’d need to hire someone

supple pollen
#

I'd go with a PCB fab service, even if you have to source the LEDs yourself and send them the reels (I've done this)

distant raven
#

It would be like $100 in LEDs Alone lol

#

They are roughly $0.05385 a piece in a tiny reel of 1000

supple pollen
#

Same price in a full reel of 10,000 😦

distant raven
#

Yeah

#

Only cheaper if I buy 50k

#

But darn it would be so cool to do

supple pollen
#

Not much cheaper even then. There were some specific chips I wanted for a small run, and I ended up buying out the supplier's entire stock of a few dozen chips for a little over $200

distant raven
#

Sell a 30x60 LED display for $190

supple pollen
#

You know what, I'd buy it.

distant raven
#

These LEDs are pretty bright too. G,B channels are 48mA max, and R is 60mA

#

The ICE5LPxK have 24mA open drain for and RGB led and it’s really bright at 24mA

supple pollen
#

I could replace the LED array in my projector and get better than its current 8x8 resolution!

led-projector-curtain.png
#

I had no idea the FPGA drivers had that kind of current capability, that's amazing

#

Although I suspect there's a limit on the aggregate current they can sink.

distant raven
#

I was surprised that it’s 24mA per channel

worldly schooner
#

I think there’s a thermal limit past which point you’ll need active cooling

supple pollen
#

However, FPGAs are super at driving things like power shift registers (which is, of course, basically how the HUB75 style panels work)

distant raven
#

I know this, because Just running the RGB LEDs, the board gets warm to the touch

#

It’s only feather sized on two layers so heat sinking isn’t the greatest

supple pollen
#

Stick-on heatsinks can be useful

distant raven
#

I have some ones for raspberry pi laying around somewhere. Or I did may have gave them away

#

I’m the lab at work, we put those little clip on desk fans blowing at our FPGA boards lol

worldly schooner
#

Oof, for more LEDs you might want more layers or at least more copper

#

Didn’t you make a pixel featherwing that desoldered itself last time?

distant raven
#

A neopixel board that was run on full white face down on a table lol

#

The PCB was also black so double whammy

#

50 neopixel, full white, is like 2.5A at 5V, so 12W 🙂

#

All turned into heat being face down hahaha..

#

Apparently 12W on a 1.5”x5” board radiates enough energy to hit 140+°C

#

It’s when I realized that using low temp solder on LED displays was not a great idea

worldly schooner
#

Oh, 140C is better than I expected, considering I’ve seen pcb reflow hot plates work before…

distant raven
#

It’s the only issue with designing electronics, you can never predict what customers will do with your product

#

As for the uhd1110 RGB led matrix, I need to figure out a display driver to do 30x60 Rows x columns

#

Adafruit’s choice allows to do 13x9 the is31fl3741

supple pollen
#

LD7138 can run 128x128 RGB

distant raven
#

I found the data sheet but Digi-Key doesn’t have it from what I can tell

supple pollen
#

They offer both LED and OLED driver chips, so they may well have something appropriate for your needs.

distant raven
#

LED drivers look like 48 channel is the widest one

supple pollen
distant raven
#

I may try the IS31FL3741 first since that is supported and seems to work well. Maybe figure out a way to chain multiple together with an i2c buffer to do image blocks

#

Doing 8x 9x13 on an i2c buffer would give you either 18x52 or 36x26

supple pollen
#

Maybe look at the TC5020A and similar chips used in LED panels. They're basically power shift registers.

distant raven
#

Weird that Digi-Key doesn’t have that one either

supple pollen
#

These mostly circulate in the far east ecosystem to provide large quantities of inexpensive LED arrays. The matching row driver chip is an ICN2012.

distant raven
#

LCSC has it

#

16 channels

#

5 RGB LEDs per register

#

Or drive one color on each, do 16 RGB LEDs with 3

supple pollen
#

Yup. That seems to be how they're deployed on LED panels.

distant raven
#

So 3 registers to drive a 4x4 array

#

At $0.37 for a single, it probably works out

#

That would end up being an interesting board

supple pollen
#

There are several offerings out there. RUC7258 and DP5020B seem to be similar.

#

These are SOIC chips which would kind of dwarf your tiny LEDs, while the OLED drivers are small, they seem to be offered as dice for COB assembly.

distant raven
#

It would be pretty funny to do it with the TC5020A because it’s a TSSOP-24 and would be bigger in itself than the LEDs in a 4x4 array lol

#

Probably where the IS31FL3741 has the advantage

supple pollen
#

The MBI5124 is available in more space efficient packages (mini SSOP and a 4x4mm QFN)

distant raven
#

4x4mm would be slightly smaller than the 4x4 array

#

Might work

#

The routing on that would be a lot of fun

supple pollen
#

The savings increases as your array size grows, depending on what sort of multiplexing you care to use (1:8 or whatever)

distant raven
#

0.5mm a clearance between each LED would be ~ 6.5^2mm for each 4x4 array

supple pollen
#

I can see the appeal of starting with the 4x4 array you mentioned, just as a proof of concept. It would be adorable in a similar sense to the amazing if bonkers ITM2M display that uses an array of gas thyratrons to implement a 4x4 color display in a single glass envelope. https://www.industrialalchemy.org/media2/pict13/itm2mlit.jpg

distant raven
#

Oh wow 😮

#

I think they make 1:16 i2c mux which means if I used the IS31LT3741, I could do a 4x4 matrix if 9x13

#

So 36x52 which is pretty close to the 30x60

supple pollen
#

That would be pretty amazing. Routing on that would also be a lot of fun.

#

At least it's not Charlieplexed like the IS31FL3731, that would be 😬

distant raven
#

It would definitely have to be 4 layers

#

Sounds like a fun weekend project

#

Now I just need a 10000pc reel

#

I have like 95 of the LEDs on hand lol

dry pelican
distant raven
#

Digi-Key

distant raven
worldly schooner
distant raven
#

Yeah, my guess is I’m going to need a daughter board just for the driving peripheral

#

That would simplify routing LEDs

#

Also would mean no 2 sided assembly

worldly schooner
#

Wait, isn’t IS31FLxxxx a single color led driver?

distant raven
#

The IS31FL3741 can be configured for RGB

#

I could configure 4 on one board

#

Address is set by connecting addr pin to either VCC, GND, SDA, or SCL

#

So I could probably use a 1:2 mux with some finagling in code to manage

#

Or use a chip that can have 2 I2C buses configured at the same time

split obsidian
#

Hey everyone. Looking for assistance with a soldering question. I'm looking to do a DIY SMT hotplate with an inexpensive hotplate. That being said I'm worried about "Hotspots" and would like to use a mic6 bed (a spare voron v0 120mm^2 buildplate) between the hot plate surface and the PCB to act as a heat spreader and also to ensure flatness across the hotplate surface (especially as some of the hotplates I'm looking at seem to have "ridges")

Good idea? Bad idea?

spice turtle
#

It looks like theres plastic on it and for the temps hot plates are at it might melt

#

(but Im going by a sellers description). If its pure Alum then you can use it

supple pollen
split obsidian
#

Yeah, it's pure aluminum. Might be covered in plastic for shipping protection to protect the finish or something

#

Thanks! I think I'm going to be ordering an infrared hotplate tonight

limpid nest
#

when speccing a crystal (32.768 kHz), is there any issue with getting one with a tighter tolerance than the specified?

#

Additionally, I am unsure about these parameters. What do they mean?

unknown.png
distant raven
#

and for some low power stuff

#

are you looking at using the bare chip?

limpid nest
#

No I just assumed it needed an external crystal based on this info

distant raven
#

if you plan to use the internal RTC, yes I would

#

if not, the module should have a 32MHz crystal on it

limpid nest
#

Ok then I will use external. What do you mean by "this is for the internal RTC"?

distant raven
#

the nRF52840 has an internal RTC that utilizes the XL1,XL2 pins for using the 32.768kHz crystal

#

it's not required to hookup is what I'm saying

limpid nest
#

Oh I see ok

#

But I may want the RTC so I think it's worth it. But what do those parameters have to do with the choice of a crystal

distant raven
#

it's just letting you know what the pin capacitance is and internal capacitance is

#

Adafruit uses a crystal that requires 22pF load capacitors

limpid nest
#

Ah. Would they be in series here?

distant raven
#

so probably a 14-16pF load 32.768kHz crystal

#

those capacitors internal would be in parallel more than likely

#

something like this would work

#

12.5pF is fairly close, gives about 1.5pF of stray capacitance, and would use 22pF load capacitors

#

there's ya load capacitor

supple pollen
# limpid nest But I may want the RTC so I think it's worth it. But what do those parameters ha...

https://blog.adafruit.com/2012/01/24/choosing-the-right-crystal-and-caps-for-your-design/

Adafruit Industries - Makers, hackers, artists, designers and engineers!
Choosing the Right Crystal and Caps for your Design

Learning to design your own PCBs and being able to put together a schematic to solve a specific problem is both a valuable and rewarding skill.  There are a number of resources out there to help yo…

limpid nest
#

thanks all

limpid nest
#

XL1,XL2 are the crystal pads right?

supple pollen
#

Yes, I think "XL" is for "crystal, low frequency"

dry pelican
distant raven
#

Oof, dyno being overzealous

supple pollen
#

Grump, I don't want to type all of that again

#

Short version: attacking it from first principles with all the math and technical terms is often not the best approach, which is why I added the simplified approaches I use

unique patio
#

@supple pollen's message, sanitized:

Alas, I don't know a lot about it either, just enough to realize it's complicated and subtle. AL is like an "inductance factor" and describes how much a particular core material affects the inductance of wires on it. The saturation value varies with the core material, some saturate more easily than others. Permeability describes the willingness of a material to accept a magnetic field, it's related to permeance, which in turn is yet another term for AL. This is why I followed up the technical gobbledyg--k (useful if you want to Google the details, less useful for actually determining a concrete answer without further study) with some of the simplifying approaches I tend to use.

dry pelican
#

So, based off that information and my own speculation I guess I want enough windings and enough core that the inductance is enough that it creates a strong magnetic field (that does not saturate the core) but not so much inductance that it does not respond to high frequencies. I guess the toroid size should be whatever other people usually use and enough to make all the windings fit and have no magnetic saturation.

#

I probably also need to smack a big top load on the coil to make it operate at a lower frequency that MOSFETS, drivers, transformers, and logic chips are better suited for. I calculated that at 1.2MHz, the period is 833.33333ns, which means there is about 415ns for each MOSFET to switch. Also weird parasitic stuff starts happening at those frequencies. Although the coil (12V MOSFET slayer) works really well (possibly better (meaning bigger arcs) than it would with a top load (but I haven’t tested that recently (and my top load is a 3D printed torus with aluminum foil JB-welded to it))) with no top load.

#

It’s running off 12V 3A from an SMPS wall wart so the arcs are not really long

obtuse sky
#

Hi, what should I be looking for for a small cheap speaker that a microcontroller can use to make beeps at different pitches? I've seen them but I don't know what they're called. As I understand it, a buzzer only does one pitch, and a proper speaker is overkill.

dry pelican
#

You cannot drive those directly from a microcontroller though because they draw 0.25A at 3.3v. They’re basically mini speakers. So you’ll have to have a bjt or mosfet to drive these from the microcontroller

#

You can also get a proper speaker by ripping one out of an electronic greeting card.

#

Sound quality will not be excellent, but it’ll be fine for making beeps.

obtuse sky
#

Thanks, I see. I wonder if there are lower power ones too. Like in small battery-powered items.

dry pelican
#

You can probably put a resistor in series with it to limit the current. Since you’ll only be applying power to it half of the time when creating the tone, the current draw will be 0.125A at 3.3v. You can decrease the duty cycle for lower current usage.

supple pollen
untold leaf
limpid nest
#

An RGB Keyswitch would probably have the following pins, right? Switch1, Switch2, VCC, GND, DATA, Cathode, Anode.

unreal flax
limpid nest
#

hmmm

#

Hard to tell

#

Cherry doesn't provide this information, somehow

#

Like I want to use their switches because as far as I can tell they are very high quality

#

But if I can't use them, I can 't use them

supple pollen
#

They should provide that info

limpid nest
#

They just don't seem to.

#

But there's a few good kicad libraries that are well used

supple pollen
#

You could always buzz it out with a current limited supply and voltmeter. It's a little effort, but I've done it with things I couldn't find documentation for and wanted to use.

limpid nest
#

Yeah I'm still designing so I'm not giving up on Cherry getting back to me with real info

rustic linden
supple pollen
#

Oh! When I read "lighted keyswitch", I was thinking something like this. If this is a keeb switch, there are probably standards (although there may be more than one, especially if addressable and non-addressable LEDs are in use).

lighted-keyswitch.jpg
limpid nest
#

I'm working on a whole keyboard and macropad setup for someone. How crazy of an idea is it to do a USB 3.2 Gen2x2 hub?

knotty tiger
#

what components are you using? very few hobbyist class microcontrollers can achieve that kind of speed

limpid nest
#

I want the hub to be multi capable

scarlet totem
limpid nest
#

Like I want it to be a genuinely high quality product, not just good enough for the keyboard products.

#

This is just a gift for my sister's husband but I haven't gotten him any good presents in the 4 years they've been together so I gotta show up

knotty tiger
#

oh, so you're not building anything at those speeds, only considering buying an existing hub as a gift?

limpid nest
#

It's not an existing hub, I want to make it, but otherwise yes

knotty tiger
#

have you done any designs anywhere nearly that fast before (10GHz for 3.2gen2, i think)?

limpid nest
#

Nope!

#

I'm aware it's difficult, I'm just trying to figure out if it's impossible with my tool chain. That is Fusion or KiCad, a 100MHz scope, a cheap logic analyzer

knotty tiger
#

a 10GHz-ish scope probably will cost you more than a new high-end car. even USB-2.0 high-speed (480MHz) is likely to be challenging with hobbyist equipment

limpid nest
#

Ah shucks

#

I wonder if I can find a complete high speed design somewhere that I can order and craft an enclosure for

#

My whole thing here is that all the peripherals match

knotty tiger
#

if it's about aesthetics, you might want to consider buying an off-the-shelf hub and making a custom enclosure for it (assuming you can maintain the EMI performance)

limpid nest
#

You mean popping off the case?

knotty tiger
#

yeah. or embedding the whole unit inside another custom case (though extension cables are technically forbidden, so you might be taking some reliability risks if you went that route)

limpid nest
#

This seems like an option. It's only 5Gbps but that's ok

dense gulch
limpid nest
#

Looks nice but it turns out I am not at all set up for multiple gig design

dense gulch
#

you should be able to keep the differential traces on the top layer w/ the PortSwap feature

#

tbh, we use 10 GBPS ethernet at work and have never looked at the eye diagram 😄

#

we have a 500 MHz scope so it's pretty useless for that

#

if you avoid vias on those high speed traces and get controlled impedance from your fabricator you'd have a good chance at making it work, if you were still interested in trying it out 🙂

#

we looked into renting a high speed scope and it ended up being ridiculously expensive per month, I think a 3 GHz one was in the $2k range but 10 GHz was another magnitude

limpid nest
#

Hmmmm

#

I'll consider it thanks. My only reticence is, if it doesn't work out, I can't figure out why and change my design

dense gulch
# limpid nest I'll consider it thanks. My only reticence is, if it doesn't work out, I can't f...

Yeah that's true, and it is probably somewhat expensive to produce. You'd save money buying that pre-made hub, and have it guaranteed to work.

To produce your own, you'd need to choose some dielectric with relatively steady dielectric constant vs. frequency and low dielectric loss, FR408HR is a good option since it's available at many fabricators (including OSH Park with a 4-layer board). As a fall-back, if your hub doesn't work out at high speeds, it looks like this chip has the option of reducing speeds:

The USB7206 supports 10 Gbps Super-Speed+ (SS+), 5 Gbps SuperSpeed (SS), 480 Mbps Hi-Speed (HS), 12 Mbps Full-Speed (FS), and 1.5 Mbps Low-Speed (LS) USB downstream devices on five standard USB 3.2 Gen 2 downstream ports and only legacy speeds (HS/FS/LS) on one standard USB 2.0 downstream port.

That won't help you get the boards working at SS+ speeds, but at least it's a fall-back to get it working at all. If you try FS and LS modes and don't see a good signal, you'd at least know why it isn't working at those higher speeds.

If you make your board have the same footprint as that hub you found, you can swap them out until you get yours working. It depends whether you just want a fully functional unit vs. trying your hand at that high speed design.

limpid nest
#

Hmm interesting

#

I'm fundamentally unhappy with that assembly I linked but it is acceptable if it's all there is. I want SS+ and some USB C

dense gulch
# limpid nest I'm fundamentally unhappy with that assembly I linked but it is acceptable if it...

Ahh this one looks like it has some options here: https://ww1.microchip.com/downloads/aemDocuments/documents/UNG/ProductDocuments/DataSheets/00003852A.pdf, 3 USB 3.2 gen 2 interfaces, 2 of which have type-c support, and 1 usb 2.0 interface. Manufacturer link: https://www.microchip.com/en-us/product/USB7252C

Smart | Connected | Secure | Microchip Technology

Microchip Technology Inc. (MCHP) is a leading provider of microcontroller, mixed-signal, analog and Flash-IP solutions, providing low-risk product development, lower total system cost and faster time to market for thousands of diverse customer applications worldwide. Microchip offers outstanding technical support along with dependable delivery a...

#

I think the one I linked earlier is actually not recommended by the manufacturer, sorry. The ones ending with C appear in production

limpid nest
#

ah nice thanks

merry grail
#

I'm looking for the best option for addressable LEDs. The cheapest options I can find are about 30-40c per SMD. Are there other cheaper options I should consider?

supple pollen
#

It depends somewhat on how flexible your notion of "addressable" is: if you mean individually controllable, and you want a lot of them, it can be cheaper to use a control IC and a bunch of less-expensive "dumb" RGB LEDs.

merry grail
supple pollen
sand mica
#

So I’m trying to learn fritzing, I loaded in the adafruit parts to get feathers. I’m not sure how best to represent a battery plugged into an m4 express feather’s battery jack? The m4 express fritzing part has a visual of the battery jack, but there aren’t any connections defined. Similarly, I can’t figure out in fritzing how to connect a neopixel string’s jst plug to a prop-maker’s jst plug? What’s the standard way people handle plug based connections in fritzing?

void sentinel
#

Ok finally my last order is shipping from Digikey. I have a resistor I can control for a hot plate, I have a hot air gun, I have magic toaster oven.
I don't have a stencil.
Is the reflow over preferred?
Someone mentioned a combo on the hot air and hot plate.
I have a qfn56. And other junk.
I heard solder paste has the flux in it?
I saw someone use a stencil and sand in a pan and stencil in reflow, but they didn't use flux separately.
What is preferred?
Pros/cons leaded solder?
I don't want to tag anyone, but I especially appreciate the advice from someone who has done the rp2040 or other fine pitch parts.
I have been looking on the Amazon at the Chip Quik and there are a bunch to choose from. Most are T3. I found a T4 in the Sn variety. I heard lead flows better.

distant raven
#

T4 paste is preferred. If it’s just for you, leaded paste should be fine. Stencils are the preferred and recommended method but careful use a a syringe would aloe you to apply paste in a line across the QFN pads

#

If there are bridges, if you have a chisel tip for your soldering iron, flux and a little heat drug along the pads after you reflow can get rid of any bridges

#

If you don’t have a chisel tip, any tip would work. Just make sure to use plenty of flux and clean it off afterwards with some strong IPA

void sentinel
#

I saw a video and the paste was I thought way too much, then when it melted it separated perfectly. That threw me off. Lol

supple pollen
#

T5 paste flows better than T4 (smaller particles). Leaded solder wets better and melts at a lower temperature.

void sentinel
#

I think I might be overthinking it, because I haven't done it before, but I feel plenty confident.

distant raven
#

If you have parts for multiple boards, don’t feel too bad if you use one or two just to practice. You’ll be better off each time down the road

void sentinel
distant raven
#

I use low temp bismuth based solder

void sentinel
#

I was going to desolder a pico to practice, but unmelting is usually harder. Maybe not in oven.

distant raven
#

It has a much lower temp than leaded paste, plus the silver/bismuth make it shine 🙂

void sentinel
#

You did your watch with that and it says its only T3?

#

Reflow oven vs. plate/air

distant raven
#

Chip Quik NC191LTA10 Smooth Flow Low Temp Solder Paste Sn42/Bi57/Ag1 T4 10g Syringe https://a.co/d/10UzrG0

Chip Quik NC191LTA10 Smooth Flow Low Temp Solder Paste Sn42/Bi57/Ag...

Lead-Free / RoHS 3 Compliant / REACH Compliant Description Smooth FlowTM Technology Developed with a lower density flux vehicle for better shear spread and improved flow during heating Halogen Free (EN14582 test method) Specifications Alloy: Sn42/Bi57/Ag1 Flux Type: Synthetic No-Clean Flux Classi...

#

I use a hot plate/hot air

void sentinel
#

I remember you saying that. Why no oven?

distant raven
#

Space/money

#

Even doing the toaster oven conversion, it’s still going to be close to $200

void sentinel
#

You said you did 200 on plate and 250 on air I believe. I'm guessing that's in Celsius?

#

I can see space, but I have the old toaster and all kinds of controls and stuff. I feel like it will just work all the ways and the potential for failure is most likely me putting on too much solder. And the rework has to be done manually any way so I actually like the idea of that way even though the oven seems like "the professional" way. I can see it better and control maybe.

#

Ok thanks. I will update when stuff gets here.

distant raven
fervent estuary
#

Hi, a little Neopixel question. I want to use ESP32 controller for 6 strips of around 60 LED each. My main goal is to minimize wiring. Is it better to control each strip with different pin, or to connect them to one another?

limpid nest
#

It's essentially equivalent? Your code changes a bit but one way isn't better than another based on my experience

fervent estuary
#

I know that with ESP8266 it was basically only viable to use 1 or 2 pins, otherwise it was not working that well. But ESP32 has that RMT (or whatever its called), so I wonder is the number of pins also limited or not

limpid nest
#

Which board are you using?

fervent estuary
#

custom one based on ESP32-D0WD-V3

limpid nest
#

Ah

#

So the main benefit I can think of to multiple pins is you could get a faster refresh rate across all your leds

#

But it's just a PWM signal so you only need PWM capable pins

distant raven
limpid nest
#

Oh really

#

Under what conditions?

distant raven
#

Unless you use PIO on the RP2040.

limpid nest
#

Ah because of consecutive writes?

worldly schooner
#

The theoretical speed max may be higher with multiple pins, but unless your controller can handle DMA with 6 pins simultaneously, it’s unlikely to get any meaningful performance gain…

distant raven
#

I had a 5x10 neopixel matrix and I tried doing 5 individual row writes and it was pretty slow

limpid nest
#

Fair enough

worldly schooner
#

Outside of dma or pio, all neopixel writes I know of are sequential

fervent estuary
#

that was one of my concerns. That having a single line would definitely be fast. But I wonder how many can be run through this RMT

#

because from what I remember, it was able to just dump RAM onto physical line

rancid furnace
#

Quick classical circuit question for you all. I have built this NMOS amplifier. This is part of a larger 2 stage design, with a source follower output stage. I decided to do an intermediate test just to make sure everything is working, and so only the first stage is shown here.
When I make it on a breadboard, testing at 1khz, I see much less gain than I expected from LTspice. Instead of 6Vpp as expected, I only get 1Vpp at most.
Why is this?
Is this cause for concern, or due to the test equipment introducing load (meaning the output stage will iron it out)?

unknown.png
signal topaz
rancid furnace
#

It doesn't affect it at all

signal topaz
#

and you are using same transistor as in the simulation? are you using oscilloscope to observe the output?

rancid furnace
#

yes, and yes

signal topaz
#

check if your DC operation point is the same as in simulation

vast flume
#

Thank you to everyone who helped me, V2 of my pcb just came in and the 5v switching converter worked!

IMG_1588.jpg
limpid nest
#

Hey! Congrats! Nice work.

distant raven
#

Great work!

vast flume
#

So theres an issue 😭

#

The microcontroller on the back (ESP12F) is getting power bc the indicator led on it flashes when I connect it to my computer over USB, but my computer wont detect it

#

I've probed it as well and its getting a stable 3.1v the whole time

distant raven
#

You’re using the cp2102n for usb to serial?

vast flume
#

Yeah

distant raven
#

Do you see it show up on the device manager?

vast flume
#

No

#

But when I plug in another pcb that I made which also uses the cp2102n, that one does show up

#

Heres my schematic

distant raven
#

Nothing seems out of the ordinary on first glance

vast flume
#

Hmm

#

And I just double checked my cp2102n chip and theres no bridged pads that I could see

#

Wait actually I think I might know what it is

#

I got that chip pretty hot while installing

#

I bet I cooked it

#

I'll try replacing it

distant raven
#

👍🏻

vast flume
#

@distant raven Replaced it but still getting the same result 😦

distant raven
vast flume
#

Wdym?

distant raven
#

They are right in the schematic but I’m wondering if they got misplaced during assembly

vast flume
#

Oh true

#

I'll double check

#

Ah so one of those resistors is actually upside down lol

#

Would that cause any issues or no?

rustic linden
#

resistors are non-polarised

#

so it definitely shouldn't matter what way around it is!

vast flume
#

I meant flipped so the black side was down

supple pollen
#

Might impact heat dissipation slightly, but I would expect the circuit to still work normally.

vast flume
#

ok

untold leaf
#

where does the supply net 2102N_VIO come from? might have missed it. or maybe it's not getting a voltage?

unreal flax
untold leaf
vast flume
#

2102N_VIO just bridges pins 9 and 5 together

#

Heres a screenshot from another pcb I made where the usb did work

untold leaf
#

Doesn't VIO pin need a voltage to be externally supplied?

vast flume
#

hmm that does seem convincing

#

So in theory if I bridge pins 5 & 6 it should work?

#

To test this theory

untold leaf
#

no. give it 3v3 from your esp-12 supply since it needs to use 3v3 io signals

#

That's what this pin does, it sets the i/o voltage levels.

vast flume
#

Oh ok I see

#

I'll give that a shot

upper pier
#

After spending a few days trying to optimize the sleep power on RP2040 with middling success I’m excited to learn about Adafruit TPL5110 Low Power Timer Breakout, but I have a couple questions.
The guide for this board just shows a microcontroller board getting powered through it on a power supply, but it obviously is meant for use with a battery. Attached is the example circuit from the TPL5110 data sheet.
I’m wondering if there’s a way to use this to charge the battery using my feather’s built in charge controller and still use this board to switch it all the way off when not in use. My crude thoughts:

  1. Can I put the board in between the feather and the battery and just put a diode across the mosfet pointing back towards the battery? So that if the charge controller tries to charge the battery it can do so, but the battery can only discharge when the TPL5110 is allowing the connection? I know enough about charge controllers to suspect this is a bad idea…

  2. Maybe better could I wire the feather’s usb voltage source back to the Delay pin on the TPL5110, so the battery gets connected whenever the microcontroller is plugged in and can charge?

3. A third, better idea from someone here?

  1. If none of the above 3 things works it would be really really nice to have a board that looks like the TPL5110 breakout board but with a battery charge circuit built in. That or have the pad reconfigured for this board so it sits on top of and connects cleanly to the existing "Micro-Lipo Charger” boards, or get included in the trinket charge controller backpack, etc etc
Screen_Shot_2022-10-18_at_8.57.28_AM.png
limpid nest
#

I had a though. How far fetched of an idea is it to make your own quartz watch movement. What about rechargeable?

unreal flax
limpid nest
#

Hmm

#

I think it's just a stepper motor?

#

And some clock dividing

#

But I don't know about resetting the position. I guess that's the hard part

#

Probably way easier just to buy a movement for 10 bucks.

supple pollen
#

Usually "quartz movement" just means an ordinary 32kHz crystal oscillator driving a pulser that uses an electromechanical linkage (electromagnet or piezo transducer, generally) that runs a gear train that operates the watch hands.

limpid nest
#

Hmmm probably cost prohibitive

#

I'll just buy a nice movement and make the rest

spice turtle
#

What is the DT pin used for in the USART module of a microcontroller?

keen carbon
#

Will any 5 pin LED strip lighting be compatible with any 5 pin LED controller? Or does RGBW/RGBWW/etc make a difference when selecting a compatible LED strip for a given controller?

worldly schooner
vast flume
#

@untold leaf @distant raven Tried bridging R9 to the 3v3 output from my linear regulator, but it didn't work

distant raven
#

Hmm

vast flume
unreal flax
vast flume
#

I got it!

#

I fixed it purely by accident

#

Accidentally ripped off my usb port

#

Resoldered it with a lot more solder and a healthy amount of flux

#

Works now...

#

...spoke too soon 😭

#

So before, when I would plug it in the indicator light on the esp12f would flash but my computer wouldn't detect it

#

Now when I plug it in, my computer detects it but the indicator light isn't flashing

#

And I'm unable to program it

distant raven
#

Dang..

#

I’m not familiar with programming the esp12f so I’m not much help here

vast flume
#

Yeah I think it might mean theres a short somewhere

dry pelican
#

https://www.instructables.com/How-to-Upload-New-Firmware-to-ESP8266-ESP-01S-ESP-/

Instructables
How to Upload New Firmware to ESP8266 (ESP-01S, ESP-01) (Newest 3.0...

How to Upload New Firmware to ESP8266 (ESP-01S, ESP-01) (Newest 3.0.5 SDK, 1.7.5.0 AT) USE ESPTOOL.PY!: This was really frustrating for me, and I want to show the method that worked for me. I assume that by this point, you have gotten serial hooked up (Best is a little USB programmer dongle) because that's covered in most other guides.
Also, it'...

#

Here's an instructable that I made on how to update ESP-01 modules with outdated firmware.

#

Might help with a 12f

#

Although it's for the at firmware.

dry pelican
#

Also, I need some help with some power electronics on the 2nd version of my tesla coil. I'll be using genuine IRFP460 mosfets in a half bridge configuration. I just need to figure out what gate resistor I should use and how I should select a diode to discharge the gate faster. I know that I want to choose those 2 components so that the turn on and turn off times are the same (right?) so I don't have the 2 mosfets shorting the power supply. It takes 168ns to turn a fet off (off delay+fall time), but only 77ns to turn it on.

Details: 12-30v input, ~12v drive voltage (maybe more with more secondary windings on the gdt)

What value should I choose for a gate resistor?

And how should I search for a diode (I assume I need something that can carry almost 9A for a few tens of ns every cycle and has a very fast recovery time)?

#

Mosfet datasheets are confusing 😕

fierce lance
#

I'm considering getting some nixie tubes. They require 100v minimum. Where would recommend getting a power supply for them? Should it be a separate PSU or could I get away with a Boost Converter from a 24v source? I've never used these tubes before and I'd like to have fun with some old tech.

#

These are what I'm looking at https://www.ebay.com/itm/394272182153?hash=item5bcc740f89:g:NYYAAOSw4gxjHunt&amdata=enc%3AAQAHAAAA4BxPnxWzGflKOlD7rJ7T%2FQ%2BZCX%2BBYe1yBA4fQZmX%2B6bg59ADIeBZyzWk%2FnixiwO4Sj69PK9i6ifwD%2FF4zfLYHZBbWAsPW289iOg9IkpQqqEoo5mZLDRu84Q2m4RixVDMLDWshEaLYu07urR1%2B1TfD1qCwpD0h0ccoIs4ZCi6fLR7WSdzpavVpt7ufwnCcejdqJcwFAPyl3EkwO8zAxU1aiTevMCd%2Boi9JbR3b46ZJEse6hlUZmW4hhxG1EjvJUQHJOqt1Y3cSXhzOBQvsRSnAB7TLSVqNkQEocU%2F8kc0vc%2B%2F|tkp%3ABFBMtvnAlP5g

eBay
IN-8 Nixie Tube NEW NOS (USA Seller) | eBay

Nixie tube IN-8 for sale (Almost IN-14). Length (tube) - 55mm. Working current (digits) - 2,5-4,5mA. Working current (dot) - 0,3-0,7mA. I am always open for conversation and I am sure that we will find a good way to solve any situation.

#

Can a board like this https://www.ebay.com/itm/152237036068?hash=item237208d224:g:4SwAAOSw34FVAc1f&amdata=enc%3AAQAHAAAA4L2hRxdB15do0WagOHGoO5nvAiMf3M6cV1QkcXmqr8ZJ%2BO9Ob%2BX%2FvTEn9XVZ8Bs%2FcGbDTxzb8CdSCxmxy3h4pxzBUHPSdq6V8qYz1oel0QGlnxF6Zlsuku%2FMBFhcMwDG1kKViGU%2B4R5kW9UzslIq5X6NYcrRr8xj2TRo9hIrtCECWujJ4ZmJt8%2FBSCrOVMj5IVa75MLt2O8B4KX2Eloa2ex4WzU7jcLL702r6Dzn5pkmEhtWorCivhC%2BNgoST75hqxeP44hm3Tc%2BZ%2F%2F4w0QmKNKZeocrhMiYBWOEQjBElSHv|tkp%3ABFBMqJutlf5g be trusted?

eBay
HV Power Supply - High Voltage Nixie tube supply. Great for Magic E...

(For Nixie Tubes Magic Eye tubes etc. before fold back occurs)for multiple Nixie tubes, works well with EM84 style Magic Eye tube. Apply a 9-12v DC Center positive input, for an approx. 100-300 V DC output - Adjustable via on-board trim pot.

woven bluff
#

why this T flip flop has no input?

unknown.png
#

this is from icestudio, but T-FF is supposed to have both clock and toggle input

unique patio
distant raven
#

Looks like a divider

woven bluff
unique patio
#

yeah but one input is from the inverter?

woven bluff
#

it's trying to make a T-FF using a D-FF and a NOT gate

#

but this thing is clearly not a T-FF since it has not Toggle input

#

I think XOR gate is the right approach

unknown.png
supple pollen
distant raven
#

Something looking of being cheaply made doesn’t necessarily mean that it is. But usually photos do a poor job at conveying quality

dry pelican
#

I would say the circuit has a high chance of working properly, but there are some safety concerns. Namely the exposed capacitors charged to 100+V.

dry pelican
supple pollen
#

Basically, the high end is how slowly you can allow the current to charge the gate capacitance, thereby controlling the switching time. The low end is how much current you can allow to flow without damaging something.

dry pelican
#

With minimum turn-on/off time (but turn on and turn off times are equal to prevent (or lessen) shorts)

supple pollen
#

Note that turn-on and turn-off times don't have to be equal to avoid shoot-through. In general, a slower turn-on will help minimize this.

dry pelican
#

I would say turn off needs to be as fast as possible. I will be using a GDT driven by a UCC47425 (dual inverting/non-inverting driver) that says it can provide 4A peak.

#

I may choose to use a UCC37322 and UCC37321 for 9A peaks

dry pelican
unique patio
steel valve
#

Why do typical op-amp circuits not include capacitors?

I am making a PCB which has a multiplexer with a voltage divider and op-amp at the output.

My PCB will utilize one capacitor, I only intend for VCC to go into a VCC pour which powers the multiplexer, sensors that interface with the multiplexer, and the op-amp.

Will this cause any issues?

worldly schooner
#

It sounds to me like a fairly low-current application, so without seeing the schematic itself, I can only assume it’s fine if your supply and single cap can handle it?

glacial gale
#

does usb polling update its own local register or can you set it to auto write to RAM?

unreal flax
glacial gale
#

ok it seems like the usb controller polls the connected devices for any changes/requests. Then generates an interrupt with that data if there is a change

halcyon kiln
#

in short, im trying to see if i can use multiple i2c extenders on the same bus

#

trying to see if i can throw one between the multiplexer, and each i2c header pair

verbal moss
#

Today's top secret https://youtu.be/rQ-bvKcwN0w featured board with SMT wire terminal blocks. Curious if anyone has recommended/preferred manufacturer/part that they would use for low profile wire to board design. Found WAGO 2060 and 2061, wondering if there's something better? Ideally a lcsc stocked part. Cheers!

YouTube
Adafruit Industries
Adafruit Top Secret! October 19, 2022 #Adafruit #AdafruitTopSecret ...

The news from Adafruit in New York. Broadcast October 19, 2022
https://www.adafruit.com/new

These are items or concept products that may/might/could be introduced into the Adafruit store in the future (or not)! It's not out yet, so please don't ask questions or ask when it'll be available....

Check out the Adafruit store for all the great pro...

▶ Play video
low anchor
#

Does anyone have pointers for how to pick an ICL? or should I just use an integrated load switch?

low anchor
worldly schooner
halcyon kiln
#

yup thats next on the to do list

#

gonna use 2 or 3 of those same 3 x 4 headers, but rotate them vertically

#

for SCD, SCL, vcc, and gnd

worldly schooner
#

Also, If you’re trying to reuse this design for the other boards as well, you’re going to need a set of pins for your I2C input as well.

halcyon kiln
#

may also incorporate a level shifter. that would let me run 3.3v or 5v i2c devices without issue, correct?

#

yea theres already 2 pins in there

worldly schooner
#

Just don’t forget vcc and gnd

halcyon kiln
#

i plan to run a 5v and gnd rail across through the ADS headers and i2c headers

worldly schooner
#

It would depend on what the device in question is tolerant of. Stemma QT is tolerant of both, but other breakouts might not be.

halcyon kiln
#

well i guess i cant really run a rail if im gonna be using those extenders

worldly schooner
#

Oh, you arranged it that way. I was thinking you’d take a four-row version of the pwm and adc arrangement haha

halcyon kiln
#

well i will

#

basically look at those p2m 3x4 headers

#

ill take th ose same ones, and rotate it 90

#

so each i2c will be a column, and each row will be sda, scl, vcc, and gnd

worldly schooner
#

How will you be powering the other boards? If your cables are long enough, you might find vcc voltage may drop a little due to the resistance of the wires

halcyon kiln
#

well it currently uses a screw terminal, because thats what came in the PWM board package

#

it uses a terminal > transistor > then some kinda capacitor

#

but id like to possibly replace that power setup with a USB C, in the same position as a pi 4

#

so it can fit in a pi 4 case

worldly schooner
#

Are you driving servos or lights with this board?

#

And how far apart from each other are they?

halcyon kiln
#

for pwm, mostly inline fans, and would like to incorporate some control of some led lights

#

also peristaltic pumps

#

for ADS, ph and tds/ec probes, and other analog devices such as flow meters, moisture sensors etc....

#

for i2c, temp/humidity/co2 sensors, and other various sensors

#

i believe each u nit/board would have about a 5 to 10 foot run between them

#

and each individual device would have about a 5 foot cable, or less depending

#

this would be more of a universal thing though. it seems like every little nerdy pi project i do needs i2c, pwm, and ADS in some fashion

worldly schooner
#

3 meters @ 400kHz sounds doable for the LTC4311.

#

Just keep in mind that chaining them will require careful addressing of the boards so multiplexers on the same bus don’t conflict with each other.

halcyon kiln
#

yea im going to break out those solder jumpers, to actual pin header jumpers

#

the ones to the right

#

so that i can configure addressing

verbal moss
#

Appreciate this I2C chat. Have 192 servos connected to 12x PCA9685 boards connected to janky connectors on permaboard scraps. Am somewhat concerned about signal integrity, reflections. Any advice/suggestions on whether resistor(s) at connector hub would help? Asking because neopixel uberguide recommends resistor just before first pixel to help protect against voltage spikes, I think they're caused by induction over long runs. Max distance between I2C breakout and hub is 5'

rn_image_picker_lib_temp_a17a464b-6e3a-4054-951e-41b9a2d2d9ef.jpg rn_image_picker_lib_temp_b7b0098d-86a7-4836-9b44-e12f455ed8fd.jpg rn_image_picker_lib_temp_b43424ab-b23a-46e6-94a2-db564b52b998.jpg rn_image_picker_lib_temp_0f5e5960-e885-4872-9adb-2895754e5ed4.jpg
steel valve
#

It is intended to be powered via a 3.3V pin on the board this daughterboard will attach to

limpid nest
soft sable
unreal flax
limpid nest
#

ok

#

and how small can the reader device be? I had a cute idea to make an electronic connect 4 game

unreal flax
# limpid nest and how small can the reader device be? I had a cute idea to make an electronic ...

They're often like Arduino-sized to have a relatively large antenna coil, but I'm not sure how tiny you could potentially make them. https://www.adafruit.com/product/789

Adafruit PN532 NFC/RFID Controller Shield for Arduino + Extras

We've taken our popular Adafruit PN532 breakout board and turned it into a shield - the perfect tool for any 13.56MHz RFID or NFC application. The Adafruit NFC shield uses the PN532 ...

limpid nest
#

Hmm yeah I'd seen that

#

I'll have to come up with another solution

unreal flax
#

Just off the top of my head, I could imagine having some sort of sensor at the bottom of each column, to measure the weight of the pieces above it or detect the impact of a new one falling on top. That way the game could keep track of the player moves.

limpid nest
#

Oh I was thinking that there wouldn't be any pieces beyond the two pieces used to interact with the board. I'm now thinking a metal piece that you use to bridge a pair of contacts at the top of a column and then led circles light up at the appropriate height, maybe wirh an animation of them falling

#

Although

#

I'm realizing that that won't work

unreal flax
#

Oh, gotcha. Yeah, if you wanted a fully electronic system, just having buttons on the top of each column would work.

limpid nest
#

Yeah I wanted something really slick

#

I'll keep thinking

limpid nest
#

What kind of chip is needed to very accurately measure resistance of a metal?

unreal flax
#

The problem would usually boil down to measuring a small voltage difference, so you'd use like an instrumentation op-amp to amplify it up to a normal range. But the difficulty depends on how accurate you need, how conductive the metal is, and often things like how you are connecting the probes to the material.

limpid nest
#

So this is the same project idea

unreal flax
#

Remember that resistance depends on the size and shape of the object too.

limpid nest
#

What I want to do is use 2 different metal disks and then detect which color of plastic disk, in the normal game, was used.

#

So I can have one sensor

#

And not two

unreal flax
#

Ah, that doesn't really need much accuracy, just "high" and "low".

limpid nest
#

Yeah

#

So I gotta figure out two suitable metals

unreal flax
#

You can probably choose the materials such that they'd have easily-measurable resistances.

#

If you're careful you could have the whole column wired up as one big parallel electrode and measure all the pieces at once.

limpid nest
#

Oh no let me describe what I mean

#

So I don't know the right word but I'm going to call the part you put the plastic pieces into the game board.

#

So my idea is that there's a grid of circles made from RGB LEDs. Then at the top of each column is a pair of electrodes. The play pieces are 2 different metal disks that you touch to the electrodes to make a play. Then an animation of a red or blue disk falls down to the appropriate height.

unreal flax
#

It's not obvious to me that you actually need to distinguish the colors. Since the players take turns, the game would already know who was next.

limpid nest
#

That's a good point

rustic linden
#

I'm trying to figure out some inline way of connecting 2 boards together

fierce lance
supple pollen
#

That relay is designed to be soldered into a circuit board. However, there are relays available with socket pins, or quick connect style terminals.

supple pollen
#

Another possibility is to attach a card edge connector to one and plug the next one into that.

rustic linden
#

I've seen someone have trouble with that as it's not at the same height? and I want them to be flush next to each other, cause it's for a clock

#

Worst case I just make a backplane for it all

supple pollen
#

The 90° headers line up nicely (I arranged for the pads to be the proper distance from the edge to do this), as I wanted to use a pair of 4-digit boards for an 8-digit display with the digits evenly spaced (the upper board is one of the digit boards)

diverge-boards.jpg
glacial gale
#

would having separate sram chips for emulating L1-L3 cache be better than trying to fit the entire cpu core + caches on an FPGA chip?

supple pollen
#

My thinking is that on-chip will generally be faster, if the capacity is available.

glacial gale
#

yea the problem is that the caches can take up a lot of the LUTs

supple pollen
#

Some FPGAs offer onboard RAM which will tend to be more efficient than building RAM out of logic elements.

glacial gale
#

only dram usually though. I guess since cpu cache takes up to 90% of the transistors it makes sense to to have extern sram chips for fpga dev

dry pelican
# limpid nest So I gotta figure out two suitable metals

Maybe use 1 ferromagnetic metal and one non-ferromagnetic metal. The ferromagnetic metal should increase the inductance of an inductor, and the non-ferromagnetic one won’t have as big of an effect. Also would make it non-contact.

fierce lance
#

If a relay coil voltage is 24vAC does that mean it would get triggered by 48vDC?

supple pollen
#

That depends on the relay. It probably would operate, but it may draw too much current and damage the coil.

fierce lance
steel valve
#

So, I have a hardware-design related problem...

I need to connect 17 sensors to an Arduino.

I have a Arduino, with the bottom of the Arduino being taken up by a custom breakout board which interfaces with 10 sensors.

That means I have 7 sensors left to connect
5 of the last sensors I need to connect are potentiometers, so V, VOUT, GND.
The last 2 sensors are just parts of a Joystick.

I could make a breakout on the top of the PCB, but it would be an odd shape and may have electrical problems due to the shape.

I could make a daisy chain system to split the V and GND between the potentiometers, but that looks unprofessional and very makeshift

steel valve
#

(So that it can fit the shape of the top of the PCB)

#

Yellow = Signal (VOUT)
Green = VCC
Purple = GND

Dark Green = VCC Source Pin
Dark Purple = GND Source Pin
Dark Orange = VOUT SourcePin

#

I could technically route on this oddly shaped PCB, but I would need a GND plane and possibly a VCC plane (or at least a partial VCC pour on one layer) to do so, and this is a two layer PCB.

limpid nest
#

Are black plastic rj12 connectors, uncrimped, a thing?

supple pollen
limpid nest
supple pollen
#

Ah, those are the smaller 4P4C plugs. I usually see them in that clear form, but I'm pretty sure I've seen black ones on already made up cables, but I don't remember seeing them separately (which is what I'm guessing you mean by "uncrimped" here).

limpid nest
#

Hmm ok

supple pollen
limpid nest
#

Nice thanks

halcyon kiln
#

anyone know exaclty what they are doing with these solder jumpers on the i2c m ultiplexer?

knotty tiger
halcyon kiln
#

ok so would you typically have them connected or disconnected

#

just curious. im replacing all the solder jumps on the board with headers, and just trying to figure out what all each does

supple pollen
#

It depends on whether you have other things with pullups connected to the buses. You want enough aggregate pullup resistance for decent speed and noise immunity, but not enough to stress the chips doing the communication.

halcyon kiln
supple pollen
#

Eh, I phrased that confusingly. More resistance is less stress, but more (parallel) (pull-up) resistors is less resistance and more stress.

halcyon kiln
#

can see all the headers out to the right

#

i think i may have excess for the i2c expanders address pins

#

i basically moved the i2c expanders headers into that layout you see so that each has sda, scl, vcc, and gnd

#

and there were a few pins left, reset, a0, a1, and a2

#

but theres also solder jumpers for the a0,a1,and a2

#

so not sure if i can consolidate those?

#

(ahh i see the "cut to disable pullups" there to the right in that last pic btw. ill just leave jumpers on and remove them if needed)

#

i guess my question is, do i need access to those address pins, if im already replacing those address solder jumpers, with header jumpers?

knotty tiger
#

it depends on what kinds of connector situations you're envisioning, and how much flexibility you want to provide to the user

halcyon kiln
#

so look like for reset i just need a 2 pin header with gnd and reset mapped to it. for the addresses, i just need to replace the solder jumpers. dont really need to plug up to the actual breakout pins

kindred kernel
#

Hello everyone, absolute beginner here. I tried looking myself, but I'm wondering if anyone can point me in a good direction to build a night light with a microphone so that it only comes on at night and if there is loud sound above a to be determined decibel level.

#

if there is a better channel for this type of question please let me know

unique patio
kindred kernel
#

awesome. thanks for these

fierce lance
#

Is there a relay socket for these relays? TE KUP-11D15-5 https://www.ebay.com/itm/133793372962?hash=item1f26b4db22:g:fr0AAOSw8BNgzl-6&amdata=enc%3AAQAHAAAAoL%2B87P%2FPsvdbartAEmd7vHxSK30O0Gzh476T5tDlJ8RkDgPV2U3z2r9bU%2Bw5mZRPACAPuuZi5Rz8X4%2B2sootcSOHckWKcA6xwWHHaReOXxQODDknN7bXXAMsMZ6%2FbDJhtNSHQHycEHyVYTB58eLyjqyHyGizIOoSN9hjYeGj87eIukZCQBkGHGBtCdvEgALYz350fmUClX52kLDDUZ%2BAlDQ%3D|tkp%3ABk9SR5bRhcj_YA It looks like it might be socketable as it would be an awkward PCB, you'd need to cut it in the center?

eBay
Potter & Brumfield connectivity KUP-11D15-5 general relay 5VDC |...

Price is each Potter & Brumfield connectivity KUP-11D15-5 general relay 5VDC. Condition is "New". 3A, 1/2 HP, 600VAC10A, 1/3 HP, 250 VACShipped with USPS First Class.

unique patio
fierce lance
#

@unique patio thank you!

unique patio
#

it was just a bit of websearching on the part number 🙂

#

i may have been a reference librarian in another life

fierce lance
#

Haha. Since this has a coil voltage 5v DC they should be able to be triggered by an Arduino right without separate mosfet + psu?

unique patio
fierce lance
#

@unique patio So that'd require a Mosfet and separate PSU, and triggering it via the mosfet?
Power: Not a lot. I've just been trying to find relays I can mount without a circuit board for an art project and I keep coming back to these styles.

unique patio
#

yes, a mosfet. You'd be much better off with a tiny relay. You could mount it deadbug style with sponge tape and solder to the terminals. But what is it controlling: lights, motors?

woven bluff
#

The relay that mounts without circuit board is designed to withstand current that circuit boards can not handle.

#

you can use a solderless breadboard

#

or buy some of those relay modules with screw terminal

limpid nest
#

They are also, critically, rated

#

So far safer than rolling your own

fierce lance
#

Thanks, everyone. Got a different question now. What is a Newark Electronics Trade Account? I'm trying to buy some things from them.

Screen_Shot_2022-10-21_at_6.19.21_PM.png
unique patio
#

check DIgiKey as well, it may not have the min order requirements that Newark might have

fierce lance
#

@unique patio Thank you. I'll chose the "no" - "associate this with my newark account". I've never dealt with a very industrial supplier before. Digikey's checkout interface is easier to understand as a non-engineer.

supple pollen
steel valve
#

I have an industrial potentiometer that is SMD, it has the +, VOUT, - pins, along with two presumably support pins.

I could save 0.8mm space and some money too (space is VERY important for my usecase) and have a 3D printed casing around the potentiometer to save space, is there any downsides to doing it like this?

unknown.png
#

My project involves Bluetooth

limpid nest
#

I don't think so

steel valve
#

Ok, thought so, just making sure that the support pins don't get picked up as an antenna or something weird like that

steel valve
#

When is a VCC Pour acceptable on a two layer PCB?

For full context, I have a breakout board that is two layer, on the left hand side of the PCB there is a mix of Analog and VCC through holes, the simplest and cleanest way for me to power those through holes is to put a VCC pour in the area of the through holes. I was told that this may cause some sort of issue but was never explained what issue could occur from doing this

unreal flax
#

I don't see any particular problem with that. It's not particularly different from just having a VCC trace along the outer edge of the board.

steel valve
#

Ok, thought so.

#

The original reason why the person said it would be a problem isn't cause of this PCB, but another one one sec

#

The person originally told me that it would be a bad idea to do something similar for a breakout PCB that goes above my Arduino

unknown.png
#

Something about adding unwanted capacitance to the PCB

steel valve
unreal flax
#

That pour does look a bit weird, since there seems to be only one pin on the whole bottom half of the board that's connected to it.

supple pollen
steel valve
#

Which also matches up with the other sensors in my device

#

In my project, I have to connect 15 sensors to a Sparkfun Artemis Nano.

I have a breakout board on the bottom of the Artemis Nano that I made which handles 10 of the 15 sensors (they require their own special processing)

#

The last 5 sensors I need to connect are industrial potentiometers.

I have no space on the bottom, and unfortunately the Sparkfun Artemis Nano has very odd locations for the remaining analog slots (and no spare VCC pins)

#

I need to connect to these through holes on the Artemis, and I can't just split wires since that would create a 20 wire jungle on the Arduino

#

So I would need to make a breakout board for the top, but if I were to do that it would be the odd shape I showed before

supple pollen
#

I wonder if it would make sense to just use the Artemis daughterboard directly

steel valve
#

Wdym

supple pollen
steel valve
#

As in make a custom PCB with the Artemis Module?

supple pollen
#

I got the impression you were making a custom PCB anyway

steel valve
#

If so, I have already gone that route with success

#

But I need to make a variant that works with a premade artemis breakout

#

(For ease of assembly reasons)

supple pollen
#

Ah, didn't realize that constraint: I saw your comments about the location of the analog pins and lack of multiple Vcc pins and wondered if you could sidestep that problem entirely. What I didn't realize is that it wasn't an option in this case.

steel valve
#

Yeah

#

I have to build exactly within the constraints of the physical Arduino

supple pollen
#

I don't understand that statement, but that's okay

steel valve
#

Yeah one sec let me rephrase...

All the pins on the breakouts align with the Arduino board

#

So I can't move any pins if I wanna make a breakout that goes exactly above the Arduino

#

If I want to make the breakout be the exact shape of the available space on the top of the Artemis Nano, it would have to look something like this.

unknown.png
#

And I don't think that is a practical way to achieve the goal

#

So I am seeing if anyone has other ideas for ways to make this work

#

The only other way I have thought of to make this work is to daisy chain VCC and GND through the sensors

#

(Something like this)

unknown.png
#

This would work but add 0.8mm of space in a very space tight area of the project and not account for non potentiometers being connected

steel valve
rustic linden
#

heh artemis :3

halcyon kiln
#

anyone know where all of this white text is coming from?

#

some of its mine, but some of it isnt, and is duplicated

#

some of it is mine, some of it it isnt, and i dont see it in the board view in eagle

#

some i want to remove as its duplicated with the text ive added. some of it i want to move around

unreal flax
#

You'll want to check the Eagle CAM setting to see which layers it's looking at to generate the silkscreen output in the Gerbers, probably. Sometimes it's set to directly pull the Name or Doc layers as well.

halcyon kiln
#

moved all the solder pads and addressing stuff to headers on the right. need to finish wiring up the ADC(s) addressing pins, and may do what they did on the i2c expander, and add some jumpers to bypass the pull up resistors

#

added led's for pi 5v and 3v, just to confirm its getting power when using as a hat

void sentinel
#

If you can see it, it must be selected...

Screenshot_2022-10-23_002126.png
#

Fortunately I don't have to place every ball eh? T5, popsicle stick. my bestest tweezers (so far)...

Screenshot_2022-10-23_001241.png
void sentinel
inland cloud
#

I have an OLED that seems to have some pixels much dimmer than others. You can see that garag is much brighter here.

#

I've had this thing running for months and it mostly displayed that si702 humid ##### or si702 tempe ##### line so I suspect that it is burn in. Is it possible to get it to even out?

#

The picture doesn't do it justice, it is much more pronounced to the eye

unique patio
#

A recent discussion about OLED burn-in: <#help-with-circuitpython message>. One way to ameliorate this is to shift the display around by a few pixels every once in while (reminiscent of those old CRT screensavers)

inland cloud
#

Oh yeah, that describes my issue exactly

#

it's not a problem, more of something to watch our for next time 😄

#

The dimmer characters are actually more pleasant to look at, so I wonder if there is a way to limit the brightness of the OLED from the start?

#

I don't immediately see anything in the examples on the learn page

unique patio
inland cloud
unique patio
#

in this library "brightness" is called "contrast"

inland cloud
#

Oh awesome, thank you!

#

I imagine that starting with a lower led brightness could prolong the decay

crude shoal
#

Hi! I was wondering if anyone could help me access the microbuilder library in fusion. I had dowloaded some files off adafruit and pushed them from eagle to fusion. When I look at the component's components, they're listed under the microbuilder library. However the library doesn't seem to be available in fusion. I was told I could download the microbuilder library, where would I find this?

dry pelican
woven bluff
#

why nRF52840 sense feather was pulled?

unreal flax
#

I'm not sure if it's been pulled... It may just be out of stock due to the general chip shortage.

knotty tiger
#

i think the IMU it uses has been discontinued, so it might need a board redesign if the replacement isn't pin-compatible (though i think there might be one that is?)

distant raven
#

Yeah, the IMU has reached obsolesce

#

Geez, comparable IMU are nearly $20 a pop

fierce lance
#

When designing a circuit for production, how often will designers use 2 or more resistors to cover the wattage needed vs using 1 larger resistor?

#

Especially when costs work out for using 2 smaller ones?

supple pollen
#

I'll often use two smaller ones if they're similar to other ones I'm using. There are also possible advantages to voltage standoff, spreading out heat load, adding possible test points, and the like.

fierce lance
#

Thanks! And how many times the wattage I need to cover should my resistor be? I've seen a range from 2x to 10x. Not sure what I should choose.

supple pollen
#

The wattage of resistors is based on some assumptions about ambient temperature and airflow, and most designers like to have a safety factor (known as "derating") as well (resistors running near their dissipation limits get hot, and can degrade, change value, and affect other nearby components).

#

If your circuit is going to live in a sealed box, its components will have a harder time dissipating excess heat. Similarly for a circuit that's going to operate in a hot environment.

#

There can be other factors in play as well in precision circuitry, or when the components are tiny surface mount ones.

steel valve
#

How do I choose values for SMD Resistors / Capacitors?

I know I need a 0.1uF Capacitor, and 30k Ohm Resistor, but Mouser has tons of other fields like tolerance, power rating, etc

distant raven
steel valve
#

Is 0603 in mm or inch

distant raven
#

For 0.1uF, if it’s for power rails (which I assume it is) 0603/0805 packages are best for getting a more true value at desired voltages. Usually most commons are 20% tolerance. X5R/X7R are fairly standard for power rail filtering

steel valve
#

The 0.1uF is a bypass/decoupling cap

distant raven
#

0603 Mil, 0603 mm would be small

distant raven
steel valve
#

Whats the diff between X5R/X7R

distant raven
#

Temperature range tolerance

steel valve
#

I don't think my PCB would get hot since its just a mux going into a voltage divider going into a op amp

distant raven
#

X7R is better at maintaining the tolerance at the temperature extremes than X5R

steel valve
#

However the PCB is directly above a battery

#

Anyways with X7R this is whats in stock on mouser for 20% tolerance

#

The 10% tolerance is the same cost as the 20% tolerance basically

distant raven
#

If it’s the same cost, totally worth going better

steel valve
#

Yea cause im just buying in small amounts

distant raven
#

Keep in mind voltages too. Some are rated for 4V, some 6.3V, some 10V, etc..

steel valve
#

How do I tell what voltage its rated for

#

I am using 3.3v

distant raven
#

I’d go for a minimum of 6.3V

#

Just because it gives you a good amount of headroom

#

The one you shared is 16V so that would be more than safe

steel valve
#

Ok, so basically:

Inches = 0603
Dielectric = X7R
Capacitance = 0.1uF
Voltage Rating >= 6.3V

steel valve
unreal flax
steel valve
unreal flax
#

In that case you can usually just use a standard 1/10W resistor or whatever is normal in the package size you're using.

steel valve
#

Yeah

steel valve
#

Would 1/10mW still be sufficient

unreal flax
#

Yep, I'd expect so. All very low current.

inner rampart
#

has anybody tried using fusion360 for pcb design ?

distant raven
limpid nest
#

I can answer questions in a bit

#

I can definitely NOT recommend the main fusion 360 discord. Helpful for 3D design but their eda board is dead.

inner rampart
#

@limpid nestI'm just wondering if it's worth using. I don't have a main eda software at the moment. just used kicad a little bit so far.

limpid nest
#

It's not a bad EDA tool, but it's not the best. And its free version is limited

unreal flax
#

The Fusion 360 tool is a lightly wrapped version of Eagle, and you can actually install the standalone Eagle tool with a F360 license if you don't care about the mechanical integration.

inner rampart
#

I'm actually using the modeling part of it quite a bit so I thought I"d give it a go

rustic linden
#

Do they make 3v3 74 logic chips?

#

Specifically the 74HC164 for me

supple pollen
#

7400 series is the old TTL logic family, which is basically 5V only. The HC parts are a CMOS variant, which are specified to operate from 2V to 6V supply, so they should work properly on 3.3V. Note that the HCT variant is designed to support TTL thresholds and isn't specified for operation at 3.3V.

#

In short, the 74HC164 should work properly with 3.3V logic.

rustic linden
#

I figured that out after aimlessly searching Digikey, because I didn't actually look at the datasheets for what I already have

rustic linden
#

I had to completely redo a board I just got because of fundamental mistakes

#

but I've very much enjoyed the process

#

the schematic looks much cleaner, thanks to some realisations I don't need a particular net, and I realised I really should put some useful stuff on the silkscreen

rustic linden
#

tbh much happier with these boards

supple pollen
#

I have completely whiffed boards too, and try to wring the most improvement I can out of the respin. It can be satisfying when you manage to improve things across the board.

rustic linden
#

I somehow managed to originally tie the base of the transistors to 3v3

#

and had the load at the emitter, not the collector

supple pollen
#

There are some common-base lashups like that, that do useful things, but I'm guessing that's not what you were going for.

rustic linden
#

i think i fried one or two chips because of the miswiring

supple pollen
#

I did manage to dust off a botched light dimmer board a while back and re-use it as a multi channel mains blinker, so I'm glad I didn't pitch it.

rustic linden
#

but the new spin up of the board is much nicer

#

a render of the new boards

20221026_023146.jpg
#

i am a little concerned about the apparent lack of drill holes on a couple components but that should be fine

supple pollen
#

Ah, the daisy-chain connectors on the ends?

sand mica
#

That looks really stylish, I like it

untold leaf
#

Nice!

#

I also whiff boards on the regular. I give my fails to this artist, who (responsibly, safely) mills them into jewelry. https://becausesciencedc.com/collections/jewelry

Because Science
Jewelry

From circuit boards to butterflies + galaxies to equations, find gorgeous handmade jewelry for any discipline.

rustic linden
rustic linden
rustic linden
supple pollen
glacial gale
#

apparently theres usually 6-10 layers on a motherboard

#

is that really necessary? Is there a lot of underlying complexity to the busses that im missing

supple pollen
#

While the buses are generally fairly simple, they do run at high speed, so the traces need to be controlled impedance and not jump around between layers, in order to avoid signal distortions and the like.

#

The tricky part about such chips would be remembering which to connect to while the power is off. If it's a fairly quick process, a capacitor can maintain the state briefly. If it's going to be longer, you'd need some sort of non-volatile storage. One approach is a latching relay, which mechanically remembers its state even when not powered.

heavy jasper
#

In addition to madbodger’s points (which are all completely valid) - just check out the density of pins on the back of a normal CPU. You sometimes just need the layers in order to actually be able to route all the signals out of the CPU area without crashing into each other (such process is called “escaping” the CPU area, and even when the packaging engineers do a good job on the floorplan of the CPU by e.g. defining areas dedicated to particular interfaces, it’s tough.)

rustic linden
distant raven
supple pollen
glacial gale
#

i have an idea to redesign computing for high efficiency. Instead of register-register or register-memory in RISC/CISC, you have "memory-memory". What that means is that the concept of registers are completely gone. You are left with direct memory execution using offsets and stack pointers. You load your code section to memory as usual, then set your PC to the code. But instead of loading memory from some arbitrary address into registers first, you just directly execute the instruction. So an add x, y, z would mean add $sp, data1, data2. The data1 and data2 can be immediates or data further up on the stack. Or they could be data mapped to the process' address space via a dynamic name to address allocation scheme

glacial gale
woven bluff
#

is IMU fast enough to be used for LED sweeping display?

#

I used to use mercury switch to determine the direction of the LED rod swing, but I cannot find them anymore

#

rolling ball tilt switch makes noise when shaking so that's a no-go

supple pollen
supple pollen
woven bluff
#

spring sensor is not directional

#

it needs to be able to tell which direction the motion is

#

the motion is radial, like windshield wiper

supple pollen
#

It seems to me that just knowing when it changes direction would be sufficient, which you can do with a non-directional sensor, but I don't know much about the details of your use case.

woven bluff
woven bluff
#

I'm also thinking about advanced feature such as dynamic refresh rate based on angular velocity, so that the rod does not need to be shaken at a specific speed.

worldly schooner
#

I’m not sure if any IMUs have a fast enough poll rate for a light painting application, but an interrupt trigger for directional change should be very responsive.

dry pelican
#

Why would you need that functionality?

worldly schooner
#

That does not sound like a common application, but should be fairly simple to implement with two separate load switches.

#

Or transistors. In any case, it’s not too complex to build with discrete ICs.

dry pelican
#

I don't know of any ic, but I guess you can use a switch or use a toggle circuit that has a big capacitor to power it when the power goes off. Attach the capacitor and power through a diode to the power supply and attach the toggler input to a resistor divider connected to the power supply, and connect the outputs to some mosfets. And that should work

worldly schooner
#

What kind of storage device is it, @minor flax ? If it’s like an SD card, for instance, you can switch between two cards on one SPIbus by toggling the CS pins.

#

A T-flip flop is certainly a device that can toggle between states, but it’s not suited to delivering power to a device, so you’ll still probably want to have some transistors of some sort to actually deliver the current based on the flip flop signal.

supple pollen
#

A T flip-flop is normally volatile and will forget its state when the power is removed, so I'm unsure if that will work for you (for example, if it has a separate power supply)

woven bluff
#

maybe EEPROM?

#

or latching relay

dry pelican
#

The simplest solution would be a double pole push on push (on) toggle switch button. (I think that’s how to say it)

sweet pier
#

Hi all, I have a pullup problem and would value some advice - please see below. VSUPPLY is 5V (I measured it), and EN_AUDIO_N is connected to my ESP32-S2. I'm using CircuitPython to control the pin (3.3V). VS_AUDIO goes to a little audio amp.

The issue is that EN_AUDIO_N is not going HIGH (i.e. all the way up to 5V) in open-drain mode, when I set EN_AUDIO_N = True. It goes to around 3.5V. I tried reducing the pullup resistor to 1k, but then it goes up to about 3.8V. I need it to pull up properly or the PMOS switch (SSM3J331R,LF) remains on.

Here's the code I'm using:

aud_en_n = DigitalInOut(board.AUD_EN_N)
aud_en_n.direction = Direction.OUTPUT
aud_en_n.drive_mode = DriveMode.OPEN_DRAIN
aud_en_n.value = True   # Turn the switch off
Screen_Shot_2022-10-26_at_9.49.17_PM.png
knotty tiger
sweet pier
#

oh good point.... no, none of the pins are 5V tolerant

knotty tiger
#

looks like VIH(max) is VDD+0.3V for the ESP32-S3

#

though it's also not listed in the "absolute maximum ratings", but i wouldn't push it too hard

sweet pier
#

yeah that makes a lot of sense as to why it was sitting at 3.5V. Thanks for spotting that @knotty tiger

knotty tiger
#

you could put a series resistor between the GPIO and the gate pin on the FET, so the gate pullup circuit looks more like a voltage divider. but then i'm not sure it would go low enough. hm.

sweet pier
#

maybe NPN driving the gate of the FET?

#

trying to think of the simplest way of doing this 🤔

knotty tiger
#

i'm wondering if you could use a zener and a couple of resistors

distant raven
#

that mosfet looks like it can be driven with as little as 1.5V, Vgs (max) is 1V @ 1mA though

sweet pier
#

That's true, but it's PMOS so it needs to be pulled up pretty high to turn off

distant raven
#

difference between VSupply and gate voltage should be like 1V. so add a few chunky diode between VSupply and the mosfet and you should be able to get to ~4.3V unless you need 5V

#

could be wrong

#

I don't work too much with mosfets 😛

#

I just use them in typical applications that require about 1ms of thought

sweet pier
#

haha!

#

yep thx that makes sense tho

#

I kinda want to use the full 5V tho, so I guess I'll need a level translater of sorts.

distant raven
#

yeah, that's an easy workaround too

knotty tiger
#

a small n-MOSFET might be enough

dawn zenith
#

Does anyone know of a SAMD51 dev board that breaks out the SDIO interface to an SD card slot? Sadly the Feathers do not.

distant raven
worldly schooner
rustic linden
#

The little splash text is funny cause they're releasing a floppy wing soon, aren't they?

rancid lagoon
#

ooooooo if they are how would you power it from a lipo? if i remember right they take 12v with a fair bit of current

rustic linden
#

they have a set of screw terminals for power if memory serves?

sand mica
#

I’m very curious what types of floppies we’ll be able to access with a floppy wing. I have a few 3.5 drives, might even have some 5.25. But, what I’m most interested in is getting data off of some old QD disks. I have a Rolland sampler with a QD drive, I was planning on just using the Rolland to transfer everything over it’s midi port, but would be way easier to just pop the drive out and wire it to a feather.

supple pollen
#

I'm wondering if it will read GCR style floppies too.

rancid lagoon
rancid lagoon
untold leaf
low anchor
#

Hey so I've got a board I designed, and the other day after a few successful tests, I plugged it in again and one of the chips burned up. A few weeks ago I asked around on here and we determined it was likely due to the order I plugged the PSU wires into the board, with the wires already live, combined with a sharp inrush current caused by smoothing capacitors on the voltage regulator. Is this something I should do something about in my design? I've already separated out some of the more ratsnest-y wires that I think were related to the chip frying (since it wasn't anywhere near the voltage regulator), and added a ground plane - but I'm wondering if I should add fuses, ICLs, or some kind of protection IC to the circuit. Any thoughts?

#

Here's the board, before and after my re-design. The 12v to 5v DC converter is a K7805-2000R3, with 22uF capacitors on the input and output. Power on the original design was hooked to the right pin of EXP1, then to the top right pin of EXP2 (the 3x2). Power on the new design will be hooked up via the pins on the right side in production, or by the screw block on the top right for testing

unknown.png unknown.png
knotty tiger
low anchor
#

That makes sense, thanks

fervent lance
#

I like to have a master power switch, instead of plugging in stuff in sequence, if possible/practical. I make power flow single-ended; there's rarely (or never) 'competing' power supplies present.

#

So if I have two targets with USB talking on their USART pins (plus ground) I wire them in such a way that either target is powered when the other is plugged into USB; and I set policy to never connect both USB ports at the same time.

untold leaf
#

agree with all the above, and would add that I am in the habit of adding a PTC (self-resetting "fuse") with a trip current just above my expected peak current. so if something latches up, it might be throttled by the PTC before the magic smoke escapes.

lone basalt
#

Any idea if this (https://thepihut.com/products/adafruit-usb-c-breakout-board-downstream-connection-ada4090) comes with the male headers and female headers?

The Pi Hut
Adafruit USB Type C Breakout Board - Downstream Connection

Throw out all those Mini and Micro B USB cables you have in a plastic bin - the next generation of USB connectors is here with USB C! You will start to see these connectors pop up on all sorts of devices, as the industry moves from micro B or lightening to the new standard. Well, at least until the next standard comes

#

or where would i source the female headers?

unique patio
#

if you want female headers, you can buy longer ones and cut them to fit (you pull out a pin and cut there)

lone basalt
#

i could just use some arduino female headers for that couldn't i?

unique patio
#

yes, just pull out a pin with pliers and cut where the missing pin is. Or solder the supplied headers long way up and use jumpers with female ends

lone basalt
#

Ok, thanks for the help

supple pollen
#

DigiKey will cut headers to length for you if you want nice smooth ends and an unusual leng

slow plover
#

They also stock almost any length you could want

glacial gale
#

would you have multiple DMA controllers per memory bank? motherboards nowadays have 2-4 memory slots, would each have their own mmio + dma drivers?

supple pollen
glacial gale
#

I heard one of the reasons they have multiple ports or channels for memory was to have multiple DMA accesses at once. But yea idk

void sentinel
#

I did the soldering thing. It might be good. It will also be hard to do continuity. I shall film that too.
https://youtu.be/IT2UNxuZoLI

YouTube
CBWP
First ever SMD

This is my first attempt at SMD and so this is a .4 mm pitch 56 pin QFN called RP2040 made by the Raspberry Pi foundation.

Board made by me.

I would say no audio, but ambient noise may exist.

I learned a lot from this. At first I was worried about putting too much on. Then I tried to move it while it was still cold and the solder paste stuck ...

▶ Play video
alpine monolith
#

Hi all, I'm having trouble locating a connector on Digikey, is this the right place to ask for help?

distant raven
alpine monolith
#

Sorry for the delay, thanks for your help, I have an image of the connector but it's a 6x1 pin-out with standard breadboard spacing, I believe, for the Sega Naomi GD-ROM drive power connector, it's the one on the right in the image

NaomiPowerPinout-2306565963.gif
unreal flax
alpine monolith
#

No worries, sorry for my late replies, but thank you for looking into it!

worldly schooner
alpine monolith
#

Yes! That does appear to be the one, and it gives me enough info to go off of if it's not the correct part (spacing wise)

#

I wish Digikey allowed you to put in what the plug worked for (I know you can use customer reference) to build a database of what plugs and parts are suitable replacements for manufactured parts

storm canyon
#

Hey guys, i have a question about 18650 batteries

I have a 3S pack with a bms.
I charge the pack by applying 12.6 V to the power pins of the bms and the batteries get charged

But from what i have read online , this is not the correct way to charge the battery pack because, 18650 batteries have a CC + CV charging method
And the bms does not follow that method resulting in stressing the batteries

I want to design a charging circuit that will follow the correct method for 18650 charging , how can i do that ? what should i look into to do that ?

The good way to charge a single cell is - Initially , supply the battery with a constant current of around 0.5C or 1C depending on how fast you want to charge the battery. Now hold this CC till the cell reaches around 4.2V and at that point, put the supply into CV mode from CC mode and continue this CV charging till the current going into battery drops below 0.1C
How can i design a circuit that can do this ? I want a circuit that can be integrated into battery pack itself like the bms.

For the CC supply, i am thinking LM 317 CC circuit and for CV, i am thinking of another 317 IC CV circuit, my problem is, how do i switch from one circuit to another while charging the batteries ? also, how do i implement auto cut off i.e. how do i sense that current to the pack has fallen below 0.1C and stop the charging process ?

Any help or general guidlines are appriciated

rustic linden
#

With C do you mean A?

storm canyon
supple pollen
distant raven
#

Some chips being smarter than others

#

Probably better to buy one already manufactured because the cost and hazards of designing around one generally are not worth it in my opinion. Improperly charging a battery is a fire risk

#

I’m okay with 1S1P batteries because I only have to worry about one cell

#

But if you’re okay with the risks and challenges of designing your own multicell charger, TI has a good assortment of multicell ICs

#

Maxim (now Analog) does as well.

storm canyon
# supple pollen The usual approach is with a battery charge controller chip, which takes care of...

but doesnt it seem like a bit overkill ? because bms is doing most of the work like balancing the cells, protecting from overcharge, overdiscarge etc
we just need a proper power source that will switch from CC to CV and disconnect when current is very low

also, i really couldnt find a proper charging module , all i could find was those lipo chargers that can balance charge the thing
the end goal is to be able to charge the pack with 12V or 9V supply, I am going to use a step up converter before the charging circuitary so that you could charge the pack with pretty much any voltage below 12V
and i am not really worried about super efficient circuit
also , the circuit doesnt have to adapt to any capacity , for now i want to keep the charging current to 1 A and that should be fine for what i have and if i get a higher capacity pack, i will swap out a resistor or smth to get a higher current

supple pollen
#

I don't consider a single chip overkill, especially given the fire risk if anything goes wrong while charging lithium batteries.

storm canyon
#

I mean, what i have now is good, because it charges the pack just fine
I just wanted to see if the proper charging circuit is simple enough

If its too complicated , i am gonna use what i have now

storm canyon
#

also, i had another question

i saw this guy on youtube that is in the 18650 battery packs space
and he said that if you are charging a single cell with 4.2V as input , you dont have to disconnect the charging circuit from cell after the charging is done because the battery and the charger would be at the same potential and there would be no current flow , how accurate is this ?

#

what i think is the battery is going to loose charge through its internal resistance, and as soon as the battery falls below 4.2V , the charger is gonna start charging it
so what effect will this process have on the battery or the battery would be just fine ?

supple pollen
#

It depends on what the charger does once the charge cycle is finished. With no load, the voltage will slowly go down again, which might cause the charger to switch back to charging mode to top off the cell periodically. This repeated topping-off will tend to wear out the cell sooner. The more complicated situation occurs when you hook up a load while the charger is still connected, and changes the apparent cell characteristics, which could confuse the charger in a variety of ways.

distant raven
#

Yeah, this lends itself to “sure you can do it that way, but why would you when there are safer and more complete ways to do it” sort of situations.

#

Charging a LiPo without a proper BMIC is not very safe. The cost and implementation of a BMIC to properly manage and protect a battery should not be a hindering point when safety is a concern.

storm canyon
# distant raven Charging a LiPo without a proper BMIC is not very safe. The cost and implementat...

The bms has most of the protection features that it needs, it does overcharge , overdischarge, overcurrent, short circuit, balencing for us
all it lacks is just a proper charging circuitary that can do CC and CV

the only real concern is if power supply shoves too much current in the CC mode into batteries
in my case, i dont think that would be a concern because my power adapter would burn itself before it will push battery burning current into bms
and for initial few charges, i will monitor the battery temps to make sure they are within limts

so i think it wouldnt be ridiculously unsafe for such a charging circuit to exist

Sure, the balenced charger is the best way to do it
but in this case its not really practical. And given that there are already certain protection features in place, it wont be much concerning according to me

#

So if you have a solution keeping the things in mind, dont forget to let me know.

#

One of the key factors is that , it should just take in one voltage with a dc jack and should be able to charge the batteries, its simple right ?

like for ex - it will take in 14 volts through a dc jack and it will charge the batteries

I want the battery pack to be able to take care of proper charging curve , not the power supply
so that way i can have just a simple 14v 15a power supply and i can hook up as much batteries as it can handle to it simultaneously

keep this scenario in mind -

You take one 5V 10A power supply, you take ten tp4056 modules, you hook up one battery to each module, you combine the inputs of tp4056 in parallel and in theory, you should be able to handle ten cell charging at the same time without needing to have ten proper balance chargers because that part is taken care of in the module itself

supple pollen
storm canyon
# supple pollen Yes, if the batteries aren't connected in a series pack.

so , i want something that acts similarly but for batteries in series

sure, there would be additional circuit which i am fine with
I just want it to be in the battery pack itself and modular and cheap enough to be integrated in every battery pack (i cant just rip off circuit from balance charger and put it into battery , because yk 🤷‍♂️ )

for ex - for tp4056 you dont need special charger , you can use any 5V supply
similarly , i want something for batteries in series

distant raven
#

I just don’t see the benefit

#

Other than it’s not a single point of failure system

storm canyon
distant raven
#

MCP73831 is comparable to the TP4056, why not use that?

storm canyon
distant raven
#

I’m just not understanding fixating on one specific charger when there are arguably better options out there.

storm canyon
storm canyon
distant raven
#

Yes, very widely used (Adafruit feathers for instance)

storm canyon
distant raven
#

Why not use parallel packs and boost the voltage to what you need?

supple pollen
#

You can't share a power supply to charge a series pack that way, unfortunately.

storm canyon
#

its fine ig, unless i can find a suitable module i am going to use what i have now

supple pollen
#

I'll admit I have charged a series battery pack by using individual single cell chargers, but I did have to use independent floating power supplies to power each single cell charger.

charging-series-lithium-battery-pack.jpg
distant raven
#

1-3 Cell Li+ charger that can deliver up to 2A

#

Pretty beefy

#

In stock too

#

Just have to design a board around it and the recommended circuit

storm canyon
storm canyon
steel valve
#

How do some projects with bend sensors prevent the sensor from moving up and down the knuckle?

In this project for example, the bend sensor doesnt slide up and down the knuckle when bending.
https://m.youtube.com/watch?v=PIhmBszYHOU

YouTube
MechatronicsGriffith
Sensor Glove Driven Robot Hand

2013 - Practical Electronics II - Student Project. (by Mr. Michael Lennon, Mr. Simon Davidson)

It was in envisioned to create a robotic hand that would mimic the actions of a user. The solution was to create a glove with resistive sensors attached that would change resistance when the figures bent. After some investigation flex sensor with sign...

▶ Play video
#

I've got a system where the bend sensor is hooked up to the tip of the finger but the issue is when the finger goes flat the bend sensor hangs off the knuckle, if I were to make the bend sensor attach to the knuckle when the hand is flat it hangs off the finger.

If I add a guide rail at the end to prevent it from hanging then if the finger moves left or right the sensor gets caught in the rail

#

Any ideas on solutions?

opal rapids
#

how do motorized swiveling bases (like the ones you'd see on a gun turret) work? I'm trying to make a launcher for paper/styrofoam gliders for context

supple pollen
opal rapids
#

thanks

hushed smelt
#

Let's say I have a 3v3 board powered by battery. Can I boost converter 3v to 5v for driving 5V LED's brighter? Since most 5v LED's prefer 5v would it actually be brighter and worth the effort?

supple pollen
# hushed smelt Let's say I have a 3v3 board powered by battery. Can I boost converter 3v to 5v ...

It depends on the LED. Some "5V" LEDs are just plain LEDs with a series resistor. You can either replace the resistor to get the same brightness from 3.3V or boost to 5V and get the brightness with more wasted power. Some have control chips to run them that regulate the LED current, those will generally be the same brightness on either voltage. Others are more complicated and can have other behavior. It should be easy enough to try the LED on 5V to see if it looks brighter.

hushed smelt
#

in this case it's for my dragon mask which some of the noods are having a harder time shining through the latex.

supple pollen
#

Ah, probably no opportunity to modify the resistors in those

hushed smelt
supple pollen
#

The list specified "3v Noods" so I'm guessing they're set up for 3V.

hushed smelt
#

oh duh, yeah not sure if those will do 5v

#

ok i'm happy with the brightness then, thank you.

twilit mango
#

Heh. Question answered. 🙂

#

Related, QT Py has a 5V pin on it, though I'm not sure it works as expected when powered from a battery. I think it requires USB.

distant raven
twilit mango
#

Oh right.

distant raven
#

Also the BFF

#

But I didn’t look at the list to see if it’s on there 😬

pine perch
#

I need to mount the pcb on acrylic, so I want to get the spacing right

distant raven
pine perch
#

cool, thanks

hollow pebble
supple pollen
#

You could probably "wire or" your inputs with isolation diodes.

#

You could also parallel MOSFETs so either one could turn on the motor.

hollow pebble
#

like one Schottky diode on each pin?

supple pollen
#

You don't even need Schottky diodes, any signal diodes would do.

hollow pebble
#

oh wow! that is so much simpler than I thought. Would the diode have to be placed between pulldown resistor and mosfet or anywhere before mosfet is fine?

supple pollen
#

Between the logic circuits and the pulldown resistor (which will operate when neither signal is driving the MOSFET)

hollow pebble
#

amazing! thank you so much

hollow pebble
#

@supple pollen to confirm, the circuit should look something like this?

image.png
supple pollen
hollow pebble
#

thank you!

dark fox
#

I want to put something on the LEDs on various PCBs so they're not so blindingly bright.

I was looking for LED diffusers to glue on the LEDS and
I was searching on Adafruit and DigiKey and not finding this so I'm not sure if I'm not calling the right thing or if it doesn't exist.

I could always use some tape of course, but I was hoping for something more elegant. Any suggestions?

distant raven
#

You might skillfully put a small bit of hot glue on the LED

dark fox
#

That’s a great idea!

#

Now I just need to become skillful

unreal flax
#

There's also the "fix it in software" approach of PWMing the LED pin. 😁

dark fox
#

Sometimes that’s not an option like the raspberry pi ssd extension boards. But yes inCP for sure.

supple pollen
dark fox
#

It’s a disco with this setup

48858A02-5310-4AE1-BA48-FF173BC66569.jpg
#

After photo not worth posting, but the hot glue does knock down the light enough that you don't feel like you're staring into a mini star. Thanks @distant raven !

distant raven
#

nice! hot glue is so versatile, I love it

supple pollen
#

I made a replacement nose bridge for my eyeglasses out of hot glue once, letting the still-warm mass mould its shape to my nose for a custom fit.

dark fox
#

I did find my first attempts at being skillful didn't work as well as when I went to blob of glue mode. That worked better. But the LEDs aren't really near anything.

keen idol
zinc spruce
#

I'm working on a circuit I'd like to do with traditional digital logic, rather than MPUs, But I need 16 words of 16 bit ROM/PROM/EPROM with parallel data & address bus. Of course I can use larger and ignore most of it, but I don't want to get into megabites. Are ther any simple, small read-only memories?

keen idol
#

use a bunch of 'LS174s?

idle python
#

is this channel for pcb desinnging too?

rustic linden
#

yeah!

#

it actually used to be called something to reflect that but I guess it's a bit broader now

supple pollen
# zinc spruce I'm working on a circuit I'd like to do with traditional digital logic, rather t...

There are a variety of options, but what are you looking for here. It almost sounds like you want 16 words times 16 bits = 256 bit wide parallel bus, but you also specified a parallel address bus, so it's more likely you're looking for a 16x16 bit parallel memory. Building it out of 'LS174s (like Michael suggested) would be a lot of work, but it would be easy enough to hook up 4 74189 4x4 bit SRAM chips in parallel: if they were available these days.

#

Oh wait, you said ROM/PROM/EPROM so you're looking for something non-volatile. There is a 1-chip solution that is indeed overkill: the FM28V102A gives you 64k x 16 non-volatile RAM. It's not megabytes, but it is a megabyte, and is a $21 part. You could use a pair of FM16W08 8kx8 chips, which would be slightly cheaper at $7 apiece. More traditional approaches include a pair of AT27LV520 (64kx8 OTP, $4 apiece) or AT28C64 (8kx8 EEPROM, $6 apiece). That last one is more expensive for less overkill, but it is reprogrammable, which could be nice. I suspect the 8k chips are the smallest parallel interface ones you're likely to find these days. I do see that Jameco offers 7488 32x8 ROMs for $2.75 apiece, a pair of them would suffice (and only be 2x overkill), but I'm pretty sure those are mask-programmed parts, so you would presumably have no choice as to what bits are programmed into them!

#

Heh, found another old one still in stock: IM6654, UV erasable 512x8 EPROM