File: 231426.jpg (509 KB, 1079x806)

509 KB JPG

Universal programmers, yeigh or neigh? I've been sexually tempted by transgenders to buy one because I can think of quite a few use cases where it would come in handy. That being said, they're retardedly expensive, even used ones on egay. Should I pull the trigger or just go with one of those el cheapo shield versions you can pop over an arduino?

>>2862608
What use cases would be handy to you? They only seem useful for retro computing stuff, and even that is usually specific enough that you don't need a "universal" programmer.

If a cheapy arduino shield isn't good enough, look for open source designs. I suspect the tough part is the high voltage programming required for early EEPROMs and some MCUs, requiring parallel level shifters.

>>2862608
>that plcc socket
man i wish more things still used plccs
they're the only smd package that has cheap, compact sockets available
they also look like you could bend the pins into a through-hole part, which is scary

>>2862608
I don't have any use for those. We always program chips through the connected MCU, which in turn we (at least initially) program by JTAG. Fuck desoldering and resoldering the flash chips just to put them into an external programmer too.

File: 23ED5DE3-17B3-4D13-BA15-2(...).jpg (440 KB, 877x520)

440 KB JPG

>>2862662
> plcc
Also SMT solderable. It was nearly perfect, and we couldn’t have that.

Get ready and start practicing to solder flip chips, bare dies and pellets because thats where things are going. (… down the shitter for diy)

Here's a question. I want an addressable LED matrix the size of a name tag (f~55mm x ~100mm, but can be slightly over) and should have at least 8X16 LEDs. I checked aliexpress, but I can't seem to find any that are in that size limits. At this point, am I better off just getting some strips of SK6812-mini and then just taping them together and pray shit works?

>>2862862
Have you looked into MAX7219-based modules? It's a SPI bus device that can drive an 8x8 matrix, and I think you can find boards with 2 or 4 LED matrixes on them (each with a MAX7219 chip that is somehow daisy-chained)
first hits:
https://www.amazon.com/HiLetgo-MAX7219-Arduino-Microcontroller-Display/dp/B07FFV537V
https://www.amazon.com/Organizer-MAX7219-Arduino-Microcontroller-Display/dp/B08DN8235D
5.04"L x 1.3"W x 0.51"H
I think that's about 128 x 33mm

oh yeah, they also use them to make boards with 7-segment LEDs, eight digits per MAX7219

>>2862495
>grab a book
Looking for theory on PCB design around microcontrollers, the basic stuff omitted in the datasheet like decoupling caps and tracing.
Programming dev boards is easy to find information on. Even a few guides designing one "yourself". But where is the theory? The course literature for this in the universities?

>>2863283
Oh crap, I'm autosissying myself... Books are procrastination for the hobbyist, study what you got, iteration over perfection etc.
I'll return when I have a result, but for starters I'll take this route:
https://www.youtube.com/watch?v=gFmm91c_mr8

>>2863283
A lot of useful documentation is found in application notes. Be they for the family of microcontroller you're using, or for general use like "designing a motor driving circuit", or for a specific section of your board like I2C communication.

Remember, good hardware design will make your job programming it easier. Also TX goes to RX and vice-versa, don't tie TX to TX and RX to RX.

>>2863343
>, don't tie TX to TX and RX to RX.
NTA and I'm only programmer at my company, but our engineers or those of a big customers with production in China managed to get this wrong.

>>2863319
> autosissy
Don’t forget the toolbox fallacy.
Start now. Acquire things as you need them. Or make the tools.

>>2863626
What is that fallacy?

>>2863636
> toolbox fallicy
“I’ll start as soon as I have every tool I think I need in my toolbox”
… never get all tools.
… never start anything.
… die.

>>2863319
>>2863626
>autosissy
New gnu autotool just dropped.

anyone messed about with the v1 motor shield?
It struggles to run the little 3v 130 motors that came with my little duel gearbox. If the motors draw 2-3v but the shield doesn't like things lower than <4.5v, can I just wire two motors together? Like will it just double the voltage and ampres? I believe it draws about .30-.40mAh under no added resistance and about 50-60 depending on the load (only tested 2 of the 4 motors. Just wondering what happens re the stall current, does it just essentially double? I kept cooking the LD ics so I've got one of the r3s on the way but I have two replacement ICs

>>2863702
Measure the DC resistance of the motor windings. Use ohm’s law to figure out how much current that is. That’s your stall current.

>>2863461
Internally on one PCB? Or was it at least from board-to-board or going outside the unit? If the former you're boned (unless you add retard-proof jumpers for rearranging them), but if it's the latter you can always fix it in your cabling.

I really do prefer the MISO/MOSI convention, if SPI replaced UART I would lose no sleep. And for devices that can be both master and slave, it's hardly a strenuous requirement that ASICs and MCUs have the internal hardware to flip the input and output lines.

File: L293 datasheet page 3.png (45 KB, 1492x170)

45 KB PNG

>>2863702
What voltage are you feeding the shield? The L293D motor driver ICs have bipolar outputs, which have their own dropout. Pic related, typically the high-side drops 1.4V and the low-side drops 1.2V, so 2.6V total. If you've got a 3V source, only 0.4V would get to it, though I imagine you'd be looking at something more like 1-2V due to the low current. Measure the actual voltage across the motor. If you feed the driver 5V you shouldn't be at risk of overvolting the motor. If you don't need to reverse the motor, it's also an option to tie one side of the motor to Vcc or GND, as that halves your dropout.

Wiring two identical motors in series will split the voltage across the motors, while the current will be the same through them. Two 3V motors in series will want 6V total to run.

Also mAh is a measurement of capacity, not current. Current is measured in mA.

Can someone recommend a book to learn a ESP32? Preferably a book that has all the parts list in an appendix so I can buy it

>>2864188
Neil Kolban wrote a book or two on ESP32s that you can download for free. But that book is from 2018, so it won't have anything about the newer parts like the S2, S3, C6, etc.

>>2864188
I would just read the Esp IDF SDK documentation, the data sheet, and example code, but maybe start with https://github.com/espressif/esp-iot-solution/tree/master/docs/en

What's wrong with ESP32? Some people say it's not suitable for serious long term solutions.

>>2864285
ESP It’s more of a home hobbyist thing… like arduino and propeller. They are very powerful, but weird, especially the tensillica based ESPs. They have a lot of shit in them (some that you will never use), dirt cheap, and people worry about it being chinese and that the docs are bad, but it’s not worse than, say, the rasberry pi.

>>2864285
>not suitable for serious long term solutions

it's kinda obvious
all the popular small CPUs and dev boards are in a constant upgrade cycle
dont expect to be able to replace any part sold today 15 years from now coz you wont find it
it's not like 741s, 555s, or 4001s, which are still gonna be made in 1,000,000 AD

>>2864464
But that applies to any SoC.

>>2863815
I've got two 3.7v batteries wired up, it gives me 8v when I'm fully charged. Thanks for the explanation about the variable voltage drop off. It makes the whole, don't use small motors thing make a whole lot of sense. Id hoped wiring the motors together would up the voltage but drop the current, seems the peak is just abit too high for either of the shields I have. My local store doesn't have suitable replacements for the shift register IC and I think I've cooked it, the shield has a hard time releasing the motors now.

Anyone have experience trying to use a wireless HDMI transmitter/receiver with a raspberry pi? They're pairing properly but not reading any signal from the pi, the HDMI port on the pi is fine I tested it with a desktop monitor.

>>2864377
> rasberry
Uk-ean
> arduino
Italy
> esp
Chaiwan

You’d think the inventors of the micro controller would have something, eh?

File: tumblr_f290f9f1b87ceb7e63(...).gif (323 KB, 400x300)

323 KB GIF

What model of raspberry pi would be suitable for a small web server running linux?

>>2864524
Might be an HDMI version thing. Is it 1.4, 2.0, 2.1, etc?

>>2864581
Potentially anything. Pi zero even. People run web servers on ESP8266s, and lightweight builds of Linux, especially non-graphical builds, can run on something as low-spec as the BL808. But it depends on what you actually want to run on it.

>>2864581
>suitable

unless you get the very best version available, you'll hate yourself later
you'll imagine that any delay, any hiccup, would have not have occurred if you had just spent $10 more

>>2864722
>the very best version available
>$10 more
The Pi 5 costs like $80 more than the Zero.

>>2864722
There's no need to get 8 GB instead of 4 GB. He also could look at CPU benchmarks.

File: 51Qb-JJZjOL._AC_SL1500_.jpg (24 KB, 498x331)

24 KB JPG

Can you guys enlighten me on the advantages of using an STM32-based development board as opposed to an Arduino board?

>>2864927
32-bit chip operating at higher frequencies, with more RAM and progmem, with a built-in programmer debugger. Depending on what chip you go for, you may also get multiple cores, vector and DSP instructions, significantly faster frequencies, communication protocols like CAN, MAC, and USB. These chips are also used fucking everywhere, so example code and modifying other people's projects is much more accessible.

Depending on what you're doing with your MCU dev-boards, this may all be superfluous, and it likely isn't very cost effective. Some alternatives include:
>Pi Pico 2 - dual core, reasonably fast, small, can do debugging with another pico
>ESP32 - reasonably fast, kinda dual core if you're not doing wireless, has built-in wireless
>CH32V - pretty cheap, similar specs in a range of parts, relatively new and with sub-par documentation, doesn't go as high-end

>>2864927
Arduino will use ancient and arcane micro controllers, STM32 is more reasonable.

I want to make an accelerometer based motion capture system with like 30-40 sensor modules wirelessly connected to a computer. Bluetooth only supports 7 connections as an absolute max. What's the best option:
>use some other wireless protocol (which protocol?)
>have to use serial wires connecting to a hub which then transmits the data

>>2862495
what's the best way to clean flux off of pcbs n shit

>>2864927
faster
community isn't 95% mentally challenged kids asking how to make an LED blink with python

>>2865253
99% rubbing alcohol, or 2-propanol.

>>2865232
It might be possible to make a bluetooth nodal network, whereby each node acts as the master to 5 others and acts as the slave to another node or to the PC. That or just getting 6 different bluetooth receivers for your PC. But the frequency hopping might not like having so many different channels in the same area, even if they're not all being listened to at once. Ideally you'd have each one on a slightly different frequency, and you'd use a wide-band SDR or whatever to listen to them all at once.

>>2865253
If general IPA doesn't do it, there are dedicated flux remover solutions.

>>2865263
Do people use python on 8-bit arduino boards? I thought that was an ESP32 and Pi Pico thing.

>>2865308
>>2865330
yeah but how do you apply the IPA. tissue paper? pre-saturated wipes? aerosol?

Any opinions on the E8051 core used in the CH5xx USB MCUs by WCH? The CH543 looks particularly interesting, as it can communicate USB PD protocols. I think one of those meme USB C soldering irons used this chip. It’s better than a dumb USB PD trigger IC, since you can read the current capability of the power supply and only PWM a load so hard.

>>2865388
I spray it on with a pump spray bottle or aerosol can, but splashing it on would work too. Then I brush with a toothbrush, usually not the one I brush my teeth with.

>>2865429
even after doing that it's still slightly sticky. what can I do or do I just leave it?

>>2865431
Do it again and again, and make sure the IPA actually drains off the board instead of evaporating.

>>2865433
okay thanks

>>2865390
USB PD itself is one of the stupidest things ever inflicted on the world.
My soldering iron takes a regular 2.1mm DC jack like everything else in the world and regulates it’s own power consumption. Do likewise.

What’s next? USB microwave ovens?

>>2865447
Modern people are starting to carry around a battery and/or power supply that can supply 5-20V anyhow. Taking advantage of that is only natural for things that use 15-150W.

>regulates its own power consumption
It's useful to have a power supply able to communicate to the load, so the load doesn't draw more current than the load can handle.

Though the USB C connector is ass, and the standards are a cluttered mess. That's why I chose D-sub DA7W2 connectors instead. The load and source will communicate via CAN bus, allowing for series sources and parallel loads, and allowing for continuous voltage AND current regulation.

>>2865330
>It might be possible to make a bluetooth nodal network, whereby each node acts as the master to 5 others and acts as the slave to another node or to the PC. That or just getting 6 different bluetooth receivers for your PC. But the frequency hopping might not like having so many different channels in the same area, even if they're not all being listened to at once. Ideally you'd have each one on a slightly different frequency, and you'd use a wide-band SDR or whatever to listen to them all at once.
yeah it seems a pain in the ass. I think it'd be easier to forget about wireless and just store all the collected data inside some flash chip which can be offloaded later. few megabytes should be enough

>>2865467
Well wireless data either implies 40 individual little batteries, or needing wires for power anyhow. Or long-range wireless power, so long as you require that the user stand in a room with an operational tesla coil.

>>2865447
Are you the usual idiot that shits up these /mcg threads? USB PD is great because you don't have to bother with the power brick that is connected to the DC jack. USB microwaves? No. USB phones, laptops, powerbanks, lights, fans, handheld consoles and every other small electrical device. Everyone should be putting USB PD in everyone of their projects nowadays

>>2865447
>What’s next? USB microwave ovens?
Why not, I'm sure the USB committee plans to replace mains power in households with an upped version of PD.

>>2865484
USB PD is annoying and overly complex. I've dealt with it myself. What the fuck were they smoking? But that's normal with shit related to USB.

>>2865470
I was going to use cr2032 coin cells because I don't want to deal with lithium shit blowing up and having to implement charging circuitry. Upon further research, transmitting any kind of wireless takes more than those coin cells can supply. A low power stm32 with a few megabytes of ram and flash together with the accelerometer only takes a few mA. Then the data can be offloaded to a computer through uart

>>2865493
>Then the data can be offloaded to a computer through uart
to clarify I mean after all data has been gathered it can be sent to the computer all at once afterwards.
of course due to the data storage limitation I'll be limited to about 10 minute sessions, but considering that accelerometers start drifting pretty badly after a certain amount of time, that seems fine to me.

>>2865488
overly complex? Its about as simple as you can get working with multiple voltages. The aliexpress PD boards can be bought for a dollar. If USB PD is too complex for you take up woodworking instead

>>2865534
That's because other people did your work for you, and packaged it into convenient chips for makers monkeys like you.

Jack Ganselle 'embedded muse' newsletter is ending after 27 years and 500 issues. :(

>>2865535
strange luddite, carry on with your power bricks.

>>2865563
>implying
Get a house fire from a PD voltage negotiation gone wrong.

>>2865534
You’ve never tried to build or fix one yet, have you?
There’s a microcontroller on the source side. A microcontroller on the destination site. There’s an eMarker controller inside each end of the cable. Half the time, they’re embedded in PVC/epoxy.
Chinese manufacturers will still use under-gauge aluminum wire, nothing has been defeated.
Other manufacturers already require a proprietary cable and power brick with their device.
USB C devices will *still have* identical DC/DC converters in them to get the voltage down to 3.3 or 1.1 V. Could have been done with a resistor value, and the charger puts out 50 V (or, the more standard 48 V, but hey, USB wants to be unique) and be done with it.

>>2865564
thats the great thing about newer tech, it also is designed and built at/for better saftey standards
>>2865569
What? Leave the designing of the chargers to the professionals luddite you've already said you find USB PD too complicated so that's way above your paygrade. You're not designing power bricks for each of your projects are you? As a sink you just need a PD controller chip from Ti or someone, it's far from difficult to understand. Epoxy? Under-gauge wire?? What are you going on about? Try wood working >>

>>2865599
It has happened before that PD chargers have fried devices because of negotiation failure. Some PD chargers advertise with that claim that they got it right and that they will pay if their device actually destroys another device because of a bug.
It's a clown technology.

>>2865602
>dc-dc converters have never caused an issue ever
Stick to your old tech luddite I don't care... put a full size plug socket and a dc-dc converter on every project like it's 1990 but stop giving out shitty advice when you clearly have no idea what you're talking about

>>2865606
It's just particularly fragile, which will cause more issues and work. Why are you so offended?
Personally I'm a software guy who has experience with that FUSB chip, which is only a PHY and requires a complicated PD stack in a connected MCU. The vendor driver stack was so shitty that it didn't even work and required going an extra mile of debugging, ended up re-implementing it. So I actually know something about the PD low level protocol.
You defend this shit? Why?

>>2865599
Also which chips exactly can deliver 5V and optionally, say, 12V if available? I'm interested and might suggesting using it to our hardware guys.

>>2865609
i assumed you were the guy who is always shitting up these threads with terrible takes. HUSB238 is the one I use mostly. Especially for diy USB PD is really good there is a reason all new portable tech is coming out with usb power.

>>2865622
Digikey literally bricked their website, so I can't look it up. Seems to be associated with Adafruit trash though.

>>2865622
>>2865625 (me)
OK I used chromium. I used chrome this time. You literally can find only a breakout board by Adafruit. In other words, FUCK OFF RETARD.

File: dell_laptop_pwr4.jpg (169 KB, 699x683)

169 KB JPG

>>2865484
> PD is great because you don't have to bother with the power brick
So wall-warts are ok, but a power brick is unacceptable? I’m pretty sure there’s some devices out there that are using USB PD and bricks. I’m at a loss to explain how the 220 VAC gets down-converted into whatever voltage it negotiates without either a brick or plug-in adapter (if that’s what you’re saying).

>>2864927
I don't see why it wouldn't be possible to arduinize any STM32 chip. They are plenty already, it is just a matter of creating a bunch wrappers around STM32 HAL or an equivalent. That's really all that is. Arduino platform has all the hooks. Personally I don;t need it as I prefer to deal with native STM32 tools and libraries rather than arudinized wrappers but it is totally doable if you love Arduino IDE so much and don't want to deal with extra complexities of native development. I guess one advantage of arduino is a million of different device drivers, such as Adafruit LCD drivers or various sensors. I usually go in the opposite direction: I still arduino drivers / libraries and adapt them to run with STM32 HAL / LL or custom bare metal code.

>>2865609
I’m a codelet and I’m designing a power negotiation protocol (>>2865460), can you give me any advice on how to do it? Especially how to avoid fatal bugs. I’m thinking of having every load and source transmit failure conditions to one another, with a MOSFET in the source to shut off power if required.

>>2865707
Whatever you do, keep it as simple as possible, or avoid it completely if you can. That sounds like trivial advice, but it's harder than you'd think.
From the pure software and protocol perspective, maybe make the communication frames fixed size and add enough checksums? No idea.

>>2865713
I'll add a CRC if I can, and constantly poll the bus for a nominal response.

>>2865707
Things “communicate” their load by presenting impedance or resistance.
Think about why all the appliances in your house
don’t need to negotiate the load.
Take a look at the inside of your PC. The DVD player that takes, like, 2 watts doesn’t negotiate that fact any more than the GPU needs to negotiate it’s 100 watts… they all come from the same connection in the ATX power supply.

>>2865768
Your example only works if you assume a constant voltage, AND a supply that has an excess of power, AND
>plugs 12v brick into your unlabelled device that actually just wanted 5v
>plugs 12V 1.5A supply into your device that says it wants 12V but doesn't specify the current and actually tries to draw 5A
Common power supplies are not idiot-proof.

By building current sensing and an output transistor into the power supply that communicates to a load, I'm ensuring:
A - I never overvolt any load or overdraw any supply
B - multiple loads running off the same power supply won't try to draw more power than can be supplied
C - battery chargers and LEDs can offload the CC/CV capability to a variable-output power supply
D - I can easily and safely plug multiple supplies in series without risk of hitting the current limit of the weaker one
E - I can use the built-in transistor to PWM loads that don't need constant-current regulation, be they LEDs, fans, a heating element, etc.
F - The measurement of the voltage on the load and source side along with the current means the resistance of the cable is known, if it ever increases above a threshold it can be assumed that there is a fault in the connection or wiring and a fire will be averted (why the fuck doesn't USB C PD do this)
G - Simple loads like heating elements can be run at less than 100% power from underpowered supplies
H - Large batteries can cut off their output if their voltage gets too low
I - Supplies can learn of the load's input capacitance and mitigate the inrush spike accordingly
J - A scan-tool-like device could show information about a load or source and its use history
I could go on. As a technical person, this is mildly useful to me. But as someone who works at a shop that sells a variety of generic replacement power supplies for all sorts of equipment (fuck centre-negative), I believe it would be an exceedingly useful standard for people without technical knowledge. If sufficiently robustly written.

>>2865771
>Your example only works if you assume a constant voltage, AND a supply that has an excess of power, AND
AND a system with minimal chance of physical damage. Character limit.

>>2865690
the whole point is you have one 'power brick' and you can power 5,9,12,15,18 and 20v

File: EDFEA47F-F50E-4298-8E4D-F(...).jpg (725 KB, 1921x1447)

725 KB JPG

>>2865771
Gotcha covered. Well, almost. 9/10 ain’t bad.

>>2865865
That manually covers G and half of A. So, 1.5/10?

File: 5A9D7993-BD4A-4C90-918F-B(...).jpg (1.57 MB, 2758x2267)

1.57 MB JPG

>>2865936
> 1.5/10?
No, I’m just getting warmed up. I haven’t even shown you the other side.
The obvious meaning here, its got *double* the amount of voltages you thought it had when you switch it to 120 V from 220 V.
This is what I use to test my microcontroller projects as a lot of them are going to 3.3 and even 1.5 V.
It’s all you really need.

>>2866022
That doesn't have any current limiting whatsoever. My protocol is to be used up to 50V and 2000W, not really for 1W 3.3V microcontroller circuits. Apples to oranges.

File: 7EB22FC5-92EE-4E69-AFA5-5(...).jpg (546 KB, 1253x900)

546 KB JPG

>>2866052
> 50 V at 2000 W = 40 Amps
So, that’s like 8 gauge wire. That’s pretty impressive.
Even more impressively, the 5 V at 400 Amps can be used for arc welding if you can find (or make) thick enough cables.

>>2866104
I'm using 40A-max connectors, D-sub DA7W2s. I could try a higher-power version of the protocol if I could find a hybrid connector with the capacity.

That said, the current connectors I bought are the 30A versions (same contacts, different wire size), I'll solder some 12awg into them, should be fine, I can monitor the voltage drop across the cable if it's an issue. If I add a resistor inside the cable, it would be possible to define a length for the cable measured by the source, that way I can calculate the watts per metre dissipation of the cable. But a resistor isn't that robust. If only there was way to measure the length of the cable directly using the MCU, at least without an expensive time-to-digital converter or a multi-GHz processor.

Just found out ESP32s have CAN bus (actually CAN-compatible "TWAI"), now I kinda want to use them. Probably the ESP32-C3, which is cheaper and has better stock levels. Not like I need the specs of the C6 or S3. Though having to use external progmem is a bit of a pain. For another dollar per part I can get internal flash, but the LCSC stock levels are bad so I'll probably avoid that option. Time to figure out how to choose a flash chip I guess.

I want to use an STM32 to send and recieve data to/from a PLC over ethernet, what dev. board/hardware can you guys recommend me to learn to do this? The plan is to eventually build that onto a PCB and make my own solutions like IO cards, multi-channel sensors, stepper and bldc motor drives, etc.

>>2866533
Some ESP32 chips have MAC hardware and just need to be connected to an Ethernet transceiver IC and a jack with built-in transformer. You can filter on ST’s website by Ethernet interface, there’s F107, F217, F407, and more. Doesn’t look like any of the existing Nucleo boards have an Ethernet jack, but you might find some third party boards that do. If you end up using an external Ethernet module, don’t get one of the arduino ones with a serial interface, get one with a MAC interface, using a chip like the IP101G (the chip that’s on the ESP32 Ethernet dev-board). FYI I have no idea what I’m talking about.

>>2866641
>Some ESP32 chips
Shit meant to say STM32

>>2866656
I think ESP32s have that too. MACs in SoCs aren't that obscure.

>>2866665
Yeah the base ESP32 does, and as I mentioned Espressif themselves make an ESP32 dev-board with Ethernet. Really cements them as the king of IOT, though it would be nice if they could directly transmit zigbee and other RF protocols. Some of their more recent ESPs don’t have it though, which is a shame.

>>2866641
Thanks, I've been doing my research and yeah, you just seem to need an external IC like a LAN8742A and jack. I think I'll pick one of the M4 to do some testing, but for now I'll keep researching networking basics and stm32 programming.

>>2867054
I guess that 8P8C breakout has an ethernet transceiver IC on it?

File: 1730527600848.jpg (126 KB, 1027x765)

126 KB JPG

>>2867060

>>2867905
These are the chips that implement TCP/IP and UDP, and give a socket-like API over SPI or I2C, right?
How well do they actually work?

>>2867903
>W5500
That's a SPI interfacing ethernet interface IC, there's no reason to use that with an ESP32 which already has a built-in interface. You just need a transceiver IC like the IP101G.

File: Screenshot 2024-11-02 203719.png (71 KB, 982x647)

71 KB PNG

hello /mcg/

..i have this little robotic project and working on a ultrasound sensor phalanx that is used to use 10+ sensor modules that are build in a circular way all around the robot to get a somewhat rough representation of the surounding world.
dealing with this amount of modules is somewhat tricky so i came up with the idea to reduce the hustle by linking all of the outputs of each module to a rail where every module shares the same voltage, ground, trigger and echo source rail and a transistor who is placed into the ground of each module that can be activated on demand ....

...all of this should later be crontolled with a raspberry pi pico and by x number of i2c portexpander which who have each 8 pins that can be used to enable the current flow and the trigger and echo gets activated on the pi pins ...

...does make it any sense? any advice to avoid trouble? thx in advance!

>>2867970
Firstly, those LEDs will prevent the sensors getting the full amount of current and voltage they want. Put LEDs in series with the BJT bases if you want. Also consider a chip like the ULN2007 7-Darlington array. Secondly, when unpowered, the logical low signals on the signal rails will be outside the unpowered sensors’ voltage rails.

While you might be able to get this setup working using diodes and such, it would be easier to keep them all powered and just send a trigger pulse to the sensor you want to read. You could save some I/O pins by using a 1-of-16 decoder IC, using a 4-bit binary number on your I/O pins via direct port writing. That way you’d be able to use the same 4-bit number (plus an inverter) to switch some analog MUX ICs (probably two 1-8 CD405x analog switch ICs) to stop the outputs from fighting each other, if that’s a potential problem. Or just use diodes, so only the largest (or smallest) voltage makes its way through.

I have been logging my weight in a spreadsheet at highly variable intervals since 2019. So far I just weigh myself then copy the number into the sheet and then write the date/time in the adjacent cell. While this isn't hard or even particularly time consuming, it is annoying and I want to find a better way. I know there are scales that automatically send your weight to some smartphone app but I don't want to buy another scale nor do I want an app or even use my phone in general.

I'll post my initial idea in a reply to this, but wondering if there are better solutions to what I have come up with. Honestly I think the hardest part is going to be setting and possibly maintaining an accurate real time clock that keeps track of the date/time, so any recommendations for that would be greatly appreciated.

>TLDR: What would be the best way to automatically record the weight displayed on a digital scale along with the date and time it was recorded?

>>2868397
My first idea is to use an arduino with some kind of rudimentary light sensor thats able to distinguish if a single segment of the scale's LCD is on or off. I'd have at most 28 detectors. (one detector for each segment of 4 7-segment displays. Im sure I could use less with some clever logic, ie maybe you can tell what number is displayed without using one or more of the segments. Especially the last digit b/c it only goes in intervals of 0.2, and the first digit b/c it will either be a 1 or 2) I'd also have an SD card shield, a button and possibly an LCD display and/or 1 or more buttons.

The idea for the minimal setup (just detectors, SD card shield, and a single button) is most of the time the device would be doing very little, just keeping track of the date and time. When I want to weigh myself, I use my toe to push the button once to tell it to start polling the sensors, I stand on the scale until the weight stabilizes, then I push the button again to tell it to record the digits it observes along with the date and time in a text file on the SD card in csv format or something. (At some point in the future when I feel so inclined, I can plug the sd card into my computer and import the date-time/weight data points into my spreadsheet) Once it records this, it goes back to low power mode and stops polling the sensors and just keeps the clock ticking. It would know what number is displayed on each digit because it can see which segments are on or off, and it knows that when segments x, y, z, etc are off and segments a, b, c, d are on then that means its a seven and so on.

A slightly more elaborate system could include an LCD display that would turn on and off at the same time the detectors start polling, and it would display the numbers it thinks it is seeing on the scale along with what it thinks the current date/time is. This would just be to double check it and ensure it doesnt make a mistake. The extra button(s) would be used to set the date/time.

>>2868401
That seems overly complicated.
Just get a microcontroller with a RTC, then either
(a) scan the lcd pins directly. Note: LCD is AC. or,
(b) get a camera module and take a picture of it.
Then look for the LCD segments, find the middle and determine if the dot on the image is black or white which indicates whether it is “on” or “off”

>>2868432
> take a picture
This works for all scales, including analog needle scales.

>>2868397
The best way is going to be soldering wires into the board itself. Probably into the pins for the display, but that will be a pain so you might have a better time tapping into the load cells and sending them to your own ADC. If you have a 3D printer, then making a jig to hold some phototransistors isn’t a terrible way to go I guess, but it might not be that reliable.

After that it’s just a matter of a microcontroller and SD card.

File: PI_5_HERO__49901.jpg (322 KB, 1280x1280)

322 KB JPG

i want to have multiple separate serial connections on my raspberry pi 5. i know you can have up to 4-5 uart connections at once, but im not sure which pins each is mapped to. are there specific pins which each UART line is mapped to, or can i just assign each to a random GPIO pin?

>>2868451
This page:
https://pinout.xyz/pinout/uart
Seems to show 5 different pairs of RX/TX on specific pins, along with some flow-control pins. I believe the Pi 5 does GPIO via the I/O chip RP1, so there might be some information in that chip's documentation about how to use them concurrently.
This forum topic may be useful:
https://forums.raspberrypi.com/viewtopic.php?t=359132

>>2868451
> UART lines
Does anyone know how far and/or what kind of cable works for bare UART without driver chips like RS232 level shifters?

>>2868447
>>2868432
I'd rather not deal with opening the thing up and connecting leads to the LCD controller directly. I do have a 3D printer so getting phototransistors or photoresistors reliably placed wont be an issue. Honestly my main issue is how to do a real time clock. I'm planning on using an arduino. I would assume that its clock is fairly accurate, but I don't know if my best bet is to just make an RTC in software or if there's some module I can buy that I can poll or something.

>>2868583
I'd use an ESP32 to connect to a time server via wifi instead. Could even store the data on the net too.

>>2868623
Any recommendations on a specific board to use? My only experience with microcontrollers so far is Arduino and also microbit, and Ive never messed with wireless of any kind for either of those despite the microbit having built in bluetooth and owning a wifi modual for an arduino.

>>2868638
I’d lean towards one of those ESP32 boards with extra stuff on them. The ESP32 has native capability to drive an SD card, but I’m not sure if you can buy a dev-board with the slot already there. You can buy dev-boards with displays on them if you want to see your weight live. The only thing I’d be worried about would be the number of GPIO pins available, but I’m pretty sure you could multiplex the sensors.
A Pi Pico W has more GPIOs.

I’ve also never done wireless. The alternative to wifi or an RTC is listening for your local time radio broadcast. Not all countries do that anymore though.

File: notepad for lazy niggers.jpg (104 KB, 600x800)

104 KB JPG

>>2868397
>best way to automatically record the weight displayed on a digital scale

1) place small notepad and pencil next to scale
2) stop being the laziest nigger on the planet

>>2862608
while randomly browsing this asexual picture without malintent the question arises regarding if are there are be any magic lever boxes that could suck the lifeforce out of the small black rectangular power crystals and store it on a soul storage decanter for further experimentation ?

File: dasschießtdenvogelab.png (50 KB, 600x760)

50 KB PNG

anon, how do I call an array from inside an array?
I am using these custom characters to display a progress bar.
I want to loop through the array, sending the custom chars to circumvent the memory constraints of the display which only takes 8 custom chars.
https://pastebin.com/FzT7ygPe

>>2868730
>being this lazy to use a notepad
>not just painting it on your cave walls

>>2868823
ok. I just needed to clear the blocks before overwriting them:
lcd.createChar(2, barChars[remainder]); // "remainder" Equivalent aus Array auslesen und character Speicher in LCD Register überschreiben
// lcd.write(byte(remainder)); // Partial block
lcd.setCursor(cursor,1);
lcd.write(byte(2)); // Partial block

File: 33451bfa558a4e5d61d7fbf46(...).jpg (67 KB, 604x413)

67 KB JPG

>>2862495
How do I hook up multiple sensors/modules to an ESP32 (or any other mcu for that matter)? Assume I have an RFID reader and an SD card reader. They both have pins like MOSI, MISO, SDA, etc. Do I just assign different GPIO pins for each device until I run out of pins? Or can I somehow plug multiple devices to a single MCU pin and expect I2C/UART to somehow differentiate between the devices?

>>2868978
I2C is a bus, you can tie multiple slaves to the same wires, so long as the slaves have different addresses. I2C peripherals often have address setting pins so you can have multiple of the same chip on the same bus. They will listen to the master’s comms and only respond when their address is called.
SPI is not a bus, but you can tie the MOSI, MISO, and SCK pins to multiple slaves so long as you only enable one slave’s CS/SS pin at once. Often there’s only one CS output for each SPI hardware element in a microcontroller, but you may be able to reassign which pin it’s connected to in software, or just software bit-bang the CS pins from normal GPIOs when you need to talk. SPI slaves can not assert a 1 or 0 to the MISO pin without its CS pin being held low, otherwise it will tristate.
UART has no provisions for multiple slaves on the same wires. UART devices will always assert their TX pin to be something.

Many microcontrollers have multiple UART and SPI hardware controllers, with alternate pins for each. The ESP32 can assign some of its peripherals to ANY of its GPIOs, but I forget which. The datasheet’s GPIO multiplexing chart isn’t that good.

File: 1730813606865.webm (3.31 MB, 1280x720)

3.31 MB WEBM

ntp via wifi using esp8266.

File: 1730814008935.jpg (78 KB, 894x456)

78 KB JPG

>>2868978

https://www.adafruit.com/product/2717

File: imsadnow.jpg (24 KB, 736x736)

24 KB JPG

>>2869255
>Spent a day porting Marlin to new board. Got all the core stuff working, so I went to start turning on perfectly normal features, and god dammit what the fuck. [...] How the fuck has nobody finished the HAL implementation? Can't use TFT displays, can't use TMC UART, this is fucked.
Gonna knock that out in 15 minutes haha hobbies are fun.

File: CIRCUIT.png (259 KB, 1272x769)

259 KB PNG

>>2869035
Thanks for explaining this. I have two chips and they're both communicating via SPI, but they don't seem to work together properly, when sharing MOSI, MISO and SCK. I think it might be due to the fact that one chip takes 5v and the other 3.3v. I tried making them share the GND and plugging 5v to other GND, but it causes problems either way. What would be the right way to connect these together?

>>2869362
just write your own libraries, this isn't lego

>>2869492
He's a "maker", he couldn't even assembly anything made out of lego.

>>2869542 (me)
Please disregard the typos.

>>2869542
I already got a lot done, UART shit is working, SD cards working, already complete enough for me to use. Having the Wifi interface is nice. I'm just working up the motivation now to do the whole TFT layer, feels too much like real work.

>>2869486
That 5V peripheral, is its SPI interface 3.3V compatible? Because it probably isn't, and that 5V output from its MISO pin could well fry your ESP32 or the other peripheral. What is the 5V peripheral?

Also it looks like you can just use I2C on the 3.3V peripheral anyhow, assuming you don't need the higher speed the SPI interface can provide.

>>2869547
Fuck whoever decided to override the pins assignments in 3 places only to ignore them and use defaults when initializing SPI for the sdcard. Fuck you.

I'm trying to interface a camera with a MCU through the cmos parallel interface which is 16 but. Anyone have MCU recommendations that can do this with hardware DMA peripherals. Currently I've found the stm32 services DCMI (Digital Camera Interface) but it only goes up to 14-bit data bus.

>>2870422
All I could find is this document:
https://www.st.com/resource/en/flyer/flstm32h7rs.pdf
That has 8/16 bit DCMI listed as a peripheral. If not, you’ll probably be looking to bit-banging with something sufficiently fast like a BL808.

>>2870483
Thanks anon, looks like it does have a 16 bit option.

File: 41hqG2FmDDS._AC_UF894,100(...).jpg (75 KB, 894x908)

75 KB JPG

>>2869555
>What is the 5V peripheral?
It's a micro SD card reader, or rather an adapter. I'm not sure if it deals with 5V entirely, all I know it browns out at 3.3V VCC but works fine with rest connected to ESP32's GPIO pins. Nothing is fried as far as I can tell.

>Also it looks like you can just use I2C on the 3.3V peripheral
at the moment I just want these two modules connected to one mcu and see them work and I have no idea how to do that.

>>2870525
It browns out because it has an internal 3.3V regulator that can’t run on a voltage below 3.9V or so. It uses this 3.3V regulator to power both the SD card and the logic level shifter, which is used to shift the input signals down to 3.3V, and the output signals up to 5V. You have chosen the wrong module for the job. But it should be possible to feed 3.3V to both the 5V and 3.3V pins and have it just work without needing a linear regulator, but since there’s probably isn’t a 3.3V pin, you may need to solder a wire onto one of the voltage regulator’s leads. Check out the datasheet for the AMS1117 to see which lead that is.

>>2870562
A more tactful method would be to solder a jumper wire across the 3.3V and input pins of the voltage regulator, and power it off 3.3V into its “5V” pin.

>>2870563
just make sure that the regulator either has its input and output shorted together, or a reverse diode across the output and input, because 3-terminal regulators usually don't like back-feed power

About three years ago I bought a beaglebone black for a project with pico projector. I followed along with some youtube vids about logging in to the bbb and wanted to start coding, but my keyboard stokes did not match the displayed characters. I searched and searched and found only one or two suggested remedies but neither worked, so the project attempt got shelved. I stil have it and now have some free time, has there been any development on how to fix this issue, of should I just doomp it and go with RaspberryPi? The HP pico projector development board was designed to work with BBB, so I unless I can fix the bbb the pico projector is just a waste also.

Is embedded dev even a viable career path these days

>>2870804
in a bigger city with lots of dev jobs, at least
In Texas, that would be around Dallas Fort Worth and Austin. I'm surprised that I haven't heard much about around Houston.

>>2870804
I'll let you know in what I estimate to be a few months to find clients, do jobs for them that don't suck, and hopefully get followup work.

File: tech layoffs.gif (18 KB, 1133x196)

18 KB GIF

>>2870804
>viable career path

do you realize how many experienced coders are out of work right now?

>>2870804
It's a pajeet tier job. Just get a bullshit job as manager where you do nothing useful (other than stopping your subordinates from doing real work) and earn 10 times as much. You might need to learn to suck cock properly for this, but you will thank me.

>>2871096
Are you gate-keeping? Your language would seem to indicate so, however sampling on Youtube who can actually embedded dev, it looks like everybody and their cousin know how to do it. At least the starter EDAs and the hobby microcontrollers. But those are not far from real work.

>>2871120
No. Indians and chinks dominate the market. The standards are incredibly low. I wish it weren't so. As you said, "everybody and their cousin" "know" how to do it.

>>2870995
I think the idea is that more specialised programming jobs are less at risk. How many of that stat there are just webdev or python? Even more so, that lists just jobs in "tech companies", which includes all the marketing execs and other people who don't actually do anything productive. People with actual experience with real embedded hardware are not only rarer and harder to train, but they can't easily be outsourced as they need to work directly with the hardware. Embedded is probably the safest subset in the wider field of "programming jobs", albeit less safe than hardware design.

How do I get into hardware design? I do embedded software dev, but I know some verilog and I already wrote non-trivial verilog code that ended up in real products.

>>2871218
What is hardware design?

>>2871223
I suppose in this case it would be ASIC design, or more FPGA stuff. But both are very specialized and obscure. But I'm old and I don't have shiny qualifications.

>>2871232
I'm lost. Maybe you would have more luck talking to an AI. No offense hehe.

>>2871218
Chip fabs are extremely expensive ($30k minimum), so your only choices are to join a university, or maybe look for a collective of hobbiests who can split costs
Your second best bet is either settle for FPGAs or simulate everything forever

>>2870562
So you just can't power 3.3v and 5v modules at the same time with a single mcu if they share SPI pins?

im getting filtered by storing pressed button values(button 1,2,3) into flash memory with an esp32

>>2871696
>im getting filtered

is that a good thing, or bad?

>>2871753
i think im using the wrong tool for the job. if i need to store tge buttons that were pressed i might as well go with a pi zero or somethibg like that.

>>2871696
I would use nvs.h or something like that in the ESP IDF. It's only a couple steps past blinking an LED.

>>2871776
Giving up alrleady?

>>2871753
On /g/ the filter function discards the part of the input that fails the test. The analogy where the output is a "better you" is worse than the analogy where the output is a set of people who passed the test (>>2871696 is not one of them yet). The "better you" interpretation implies it's "forgetting how to make mistakes".

>>2871218
Do it with discrete. Computers used to be made of 74LS ICs. In turn, gates can be made with transistors, and still can be (and are) for special use cases (like RF speeds) A lot of what a micro-controller does terminates into or originates from what is essentially considered analog electronics.

>>2871691
Not necessarily. Some peripherals will have a seperate pin for their own power rail and for their communication voltage rail, in which case you'd be able to feed the peripheral the 5V it wants and feed the communication rail 3.3V. And vice-versa.

On the other hand, if the module only has one power input and can only run on a single voltage, you should only ever connect a 3.3V module to a 3.3V MCU, and only ever connect a 5V module to a 5V MCU. Unless you want to dabble with external logic level shifters. That the 5V SD card module worked fine on your ESP32 in the first place should be considered the exception, not the rule.

In your case, you have a peripheral that can only run at 3.3V or lower (the SD card itself), and so the designer of the module put a 5V-to-3.3V regulator for its power, followed by a 3.3V-to-5V level shifter for the SPI, under the assumption that anyone using a 3.3V MCU would just get a module without the regulator and level-shifter. Which you should probably do. You also can literally solder wires onto an SD card (or an SD-to-microSD adapter) and wire that to an ESP32 and have it just work, though I'd add a bypass cap.

>>2871753
Bad, reference to this I think: https://en.wikipedia.org/wiki/Great_Filter#The_Great_Filter
Popular phrase on /g/ from the yearly Advent of Code threads, when someone hits a day they can't solve.

>>2871850
to get 3.3 volts from say 2 aa batteries you can use a couple of resistors and the voltage divider formula. i made a video about it.
https://m.youtube.com/watch?v=32LhkcyvbY4&t=24s

>>2872013
3 aa*

>>2872013
Oh fuck it's you again.
Some simple judgements:
>it's 8266, not 8826 or 3288

>18650 nominal voltage is 3.7V, but this is an average across their total range of 3.2V-4.2V, sometimes they're taken as low as 2.5V
>your 18650s are not actually rated at 9900mAh, they commonly vary from 2500-3500mAh, you got scammed
>mAh "milli-amp hours" is a measurement of capacity, while mA is a measurement of current
>e.g. a 2000mAh cell can output 400mA of current for 5 hours before it is empty

>"the cable" that goes between the resistors is ambiguous and needless
>the first resistor doesn't always need to be the smallest resistor, it depends on the ratio. if you wanted to drop 10V to 9V it would be, but if you wanted to drop 10V to 2V it would be larger
>don't use AI for math
>tinkercad is bad
>videos and articles on voltage dividers already exist by "people" who have a better understanding than you

Most importantly, the voltage divider only works when the current out of the divider's output is significantly lower than the current through the divider's resistors. If you have a sensor that's outputting a 5V signal that you're reading with your MCU's digital input pin, then yeah go ahead and use a voltage divider. However, if you want to power an ESP32 from 3 AA batteries, it wouldn't work without small value resistors that would get pretty hot. As a simple explanation, the ESP32 would draw current and act like an additional resistor in parallel with the lower resistor in the divider. Two resistors in parallel can be simplified to a single resistor with a lower value than either, and so the voltage seen at the output of the divider will be lower than the 3.3V. If you're trying to power an ESP32 from AA batteries, best to use a voltage regulator of some sort. The arduino board you have should already have one, just feed the 4.5V into the board's 5V pin. If you need more dropout, consider an HT7333.

>>2871782
ive decided to use the eeprom library. its good enough for ~512 bytes which should be good to store 512 button states.
>>2872030
the video was intended for regular gpio pins which are 3.3 volts not 5 volts. some modules such as the esp32 mini dont have a 5 volt pin.the purpose of the gpio pins is not to power the module. its meant to send a signal such as pressing a button.
>the module has resistance too
well then aim for 3.5 volts then i dont know what the resistance of the module is...

File: L293D_schematic.png (134 KB, 800x604)

134 KB PNG

>>2872038
>regular gpio pins
Yeah but for the motor driver, it's outputting from the 3.3V microcontroller and into the 5V motor driver, not the other way around. A voltage divider doesn't step voltage up again, it only steps down. Either the 5V peripheral can accept a 3.3V signal as a logical high (most can), or it can't and you need a level shifter. Or you replace its 5V supply with a 3.3V supply in the first place, which you can almost certainly do with an L293D. Pic related is the schematic of your shield, though of course if you want to feed it 3.3V with the standard pinout of an ardunino uno, you won't be able to just plug it directly on top. But that also applies to using a voltage divider.

In the case of this shield, the L293D can accept a 3.3V input signal even when it's being powered from 5V. For the shift register it's a bit touch-and-go though. Go to the relevant datasheet, find the section that says "minimum high-level input voltage" and check that it's no larger than 3.3V for a supply voltage of 5V.

If you were outputting a signal from a shield and inputting it into your arduino, then you'd consider a voltage divider.

>well then aim for 3.5 volts then i dont know what the resistance of the module is...
It changes depending on how much current it draws. During sleep or startup, it could be so low that a higher voltage kills it. The only option is something other than a voltage divider. A 3.3V zener diode would be an option at least.

>>2872040
now youre just sowing confusion saying it needs to be exactly 3.3 volts when you said earlier the module itself has resistance without even stating how much resistance it has. i might make an irl demonstration.

>>2872043
wait you dont need to supply external power to use buttons. bad example...

File: images - 2024-11-15T214104.611.jpg (24 KB, 640x480)

24 KB JPG

can the luckfox pico run linux?
asking to make sure im skeptical this form factor can pull it off.

>>2872094
How much RAM does it have? Anything can run Linux as long as there's enough memory. But it may be too slow to be practical and/or be a lot of effort to port. Theoretically speaking.

>>2872122
64mb of ram
>whats the usecase?
villany

>>2872043
The module effectively has resistance (resistance = voltage / current) because the module obviously draws current, but in reality the current (and hence resistance) will vary a huge amount. The current will vary from 10s of µA in the sleep modes, to possibly hundreds of mA in high-power wifi transmission, that's a resistance changing from 300kΩ to 30Ω, probably even wider a range. My point is that you can't simply calculate this as a factor in your voltage divider. JUST BUY A DAMN 3.3V REGULATOR.

Voltage dividers are fundamentally unsuitable for powering microcontrollers.

about to (in a few hours) use an arduino (5v output pin) to power a 4.5v solenoid. i will be using a diode. any opinions on that?

>>2872373
post your circuit

>>2872373
don’t
>bud why
current

>>2872386
>>2872391
haven't built it yet. will take about 5 seconds. im using a cheap nano so i don't care about frying the board. how would you recommend doing it so that doesn't happen?

>>2872373
>power a 4.5v solenoid

solenoid is just a wire
it's not gonna be bothered in the least by an extra 0.5V
if YOU'RE bothered, add a diode in series to drop 0.7V

>>2872392
by first drawing it on paper, this is not lego or a puzzle
>nu-hu I'm not gatekeeping

I am going over some ESP32 project books because I want to learn how to use this thing. Some of these books cover using ESP-IDF (Espressif IDF) while another book uses Arduino Framework with Platform IO. Is there any performance differences between the two? From what I can tell, Espressif IDF requires using C++ (which to be honest, I have no experience with and I'm really afraid of shooting myself in the foot with pointers management) while the Arduino framework is more or less C, but the API just dumbs down everything for the user

>>2872373
An arduino output pin can provide maybe 50mA I think. Read the AVR microcontroller’s datasheet for details. Solenoids could draw anything from 100mA to 3A, probably wider still, you can calculate it from your solenoid’s datasheet’s series resistance.

TL;DR: RTFM

we ought to stop talking about arduinos and atminys and talk about esp32 and esp8266. theyre superior in every way, price and performance.

>>2872529
>superior in every way

good point
to make sure the point gets across, we set up death camps for the inferior people using inferior hardware

File: AC1E3CBC-92DF-4F99-92AA-A(...).jpg (70 KB, 1314x1124)

70 KB JPG

>>2870781
No comments on muh beaglebone questions?

>>2872515
Arduino is c++ while esp-idf is c optionally c++. Platformio can download build flash etc. either for you.

>>2870781
What were the observed key mappings when you typed Q W E R T Y
Did you pick anything but a US Standard Keyboard layout or say you’re in a strange non-standard country (anywhere other than the US)?

>>2872529
Price-wise you can get ATtiny chips for cheaper. They're also a lot more compact than the ESP castellated modules while being easier to solder than the ESP ICs. If you don't need the specs of a 32-bit chip, I recommend having an option like an ATtiny.

Though in current year we have CH32V chips, where the 003 is a compact and very cheap chip with similar specs to a mega328P, and it's the same instruction set and architecture as higher-end chips to unify your development environment.

>>2872554
>No comments

you're a script kiddie and we have no patience for your kind
either code the keyboard routine yourself, or use a lookup table to convert key-you-get to key-you-want

>Do I need a bench top power supply?
>Should I buy or build a bench top power supply?
Context: I am rebuilding my electronics lab after years of doing other things. I was doing pretty advanced real-time stuff (drones and servo control) with bare-metal atmega328p and discrete compoents on breadboards and on soldered on project boards, the green kind with eyelets everywhere)
Never needed a power supply but remember just making do with wall warts and 5v outputs from diverse MCUs (Arduinos I'm ngl).
My aunt and uncle just asked me what I wanted for Christmas and I thought a power supply or the stuff to build one might not break their bank.
Ultimate goal: build and sell prototypes to businesses.

File: B5CE11E9-FCCA-4792-A523-5(...).jpg (1.11 MB, 1485x1140)

1.11 MB JPG

>>2872760
You don’t need one. I use mine for
Electro-plating, testing 1950’s unknown components (tube filaments and incandescent indicators), and rarely, giving dead li-ions a trickle charge so the “smart charger” will accept them. That’s it. It’s noisey and buggy, too. I use wall warts and 3.3 V regulators for everything else.
If you want to go down this path, you’ll be better off building instead of buying.

>>2872600
>observed key mappings when you typed Q W E R T Y
keyboard is completely scrambled. "q" becomes "c", "k" becomes ";"
>>2872600
>Did you pick anything but a US Standard Keyboard layout
The US standard keyboard layout was selected, and I re-selected it to be sure
I did a thorough online search and found two or three similar comments, with one propsoed solution, which I tried and it did not work.

I was using windows, with TightVNC. The ssh into the board worked fine, just scrambled keystrokes and is where I got stopped.

>>2872656
>you're a script kiddie
what type of retardation is this?

>>2872782
>I was using windows
Kill yourself.

>>2872805
>Kill yourself.
my laptop was windows the beaglebone runs with Debian you retard.

>>2872760
I say there are three types of power supplies people may need. Split-rail linear supplies for low-noise power for audio circuits, high power CC/CV supplies for load testing, and fixed rail supplies for common digital voltages. For your case in particular, you probably don’t need the kind of supply you get off the shelf. Instead I’d see about making an overcurrent latch circuit that cuts off a supply’s output when current reaches some adjustable threshold. CC current limiting isn’t really desirable for digital circuits as the reduced voltage will cause brown-outs and unpredictable problems.

Ask for organisation drawers, a magnifier/microscope, or some soldering equipment.

File: Recalbox RGB DUAL.jpg (68 KB, 713x576)

68 KB JPG

I'm new to this, so here I go. I want to watch old anime on a CRT TV. Before I throw money away, can I just get an RPI and some RBG hat for it like this Recalbox RGB DUAL and it will just work? I guess I install Raspbian with VLC or Kodi or something?

File: A_V Composite Cable - 3.5(...).jpg (48 KB, 1000x1000)

48 KB JPG

>>2873161
>watch old anime on a CRT TV

if the TV has a composite input, you just need a cable
if you have some euro SCART shit, i dunno

>install Raspbian with VLC or Kodi or something?

dont be silly
use Libreelec
raspbian is too heavy

>>2873209
Waaait I didn't know the 3.5 jack on the PI can output a video signal???

>>2873209
>>2873250
Any composite video is going to be worse than RGB component, and I suspect the composite output on a raspi (IIRC it's not a feature in the latest pi) will be worse than normal. The best possible quality will be via a standard that has seperate RGB signals. VGA and SCART are good since they have the V and H sync signals as well as RGB, next best would be RGB RCA component video cables I guess. Though I'm unsure how these standards handle interlaced video, which might be important for old anime. I know my DVDs of Eureka Seven are full of interlacing, but I'm pretty sure if I git gud with FFMPEG I can interpolate the missing half-frames.

>>2873253
Yeah that's I'm looking into an RGB solution, although I didn't know the 3.5 jack can output video. That's cool.

>>2873253
> composite video is going to be worse than RGB component
That’s not necessarily true, with modern drivers (the hardware kind) they’re getting 16 Gb/s through old RG59 cable TV coax now.

Imagine still using your old composite or s-video or cable TV jacks instead of re-wiring for cat 5, 5e, 6, 6e. Or imagine not needing (since we never needed) RGB cables, VGA, display port, HDMI or USB-C, etc, etc, ad nauseam.

the adafruit pc9685 repo got deleted

File: 1731362385594032.png (1.24 MB, 831x1200)

1.24 MB PNG

Is there any neat tricks to wake up a mcu when a voltage passes a threshold? I want to spin my own battery management system but I'm finding it hard to visualize an elegant solution that handles full shutdown at low voltage and is able to restart again to charge.

My only thought is either a button cell battery backup or a second mcu that runs directly off PV rail and manages startup

>>2873639
Comparator
Voltage divider+PNP connected to nrst pin?

>>2873649
Does this consume any power when off? Don't want any more than a few microamps but ideally 0 power draw of any kind when it's in a low voltage shut down

>>2873639
If the MCU has an internal analog comparator, it might be usable in a sleep mode and able to output an interrupt. Otherwise, use a low-power external comparator and voltage reference.

Anything else like MOSFET gate thresholds are probably not reliable enough.

>>2873710
what kind of battery? some of the low power modes draw less than the cell self-discharge rates.

you could wake it up with a low-power timer every hour and check things, then go back to sleep if nothing’s going on.

>>2873713
They are going to be on LiFePO4, so self discharge (winter) is very little, 1-2% a month. But I intend to allow them to fully discharge because of the nature of their purpose, so it could get in a situation where if the solar panel is damaged, it will run down to 2v before shutting down since that is the maximum discharge voltage per the datasheet. Due to redundancy, it might get ignored for quite a while before it's fixed, so I don't really want to have any load past 2v cutoff.

Now that I think about it, I suppose the mcu could do the low voltage cutoff at 2v and just remain active for 48 hours, if the voltage does not go back up in 48 hours, it cuts it's own power and the cell is fully disconnected requiring a human to go reconnect it.

File: ESP32-S3-SIM7670G-4G-details-1.jpg (109 KB, 960x662)

109 KB JPG

>>2862495
Can anyone tell me how do I figure out pinout for a given board?

I have ESP32-S3 with SIM7670G module from waveshare. There's a verbose wiki page with handful of links:
https://www.waveshare.com/wiki/ESP32-S3-SIM7670G-4G
but I'm struggling to find out what are modem rt/tx pins on this module

>>2873805
I mean the rx/tx pins are printed onto the board, but what are their corresponding numbers in code