|
Some dash cams use super caps instead of batteries for their acceleration triggered recording. Keeps the RTC and acceleration sensor running, when triggered it records like 30 seconds after the hit. Not super useful, but at least tells you when it happened and you might be able to get the make/model/color.
|
# ? Apr 13, 2024 07:54 |
|
|
# ? May 28, 2024 16:34 |
|
ryanrs posted:I would also have accepted "We used a supercap to avoid scary lithium battery shipping labels". Just use one of those 50 year lifetime, radioactive Betavolt batteries in its place and the scary lithium warning sticker goes away! https://www.batterytechonline.com/materials/betavolt-battery-50-year-charge-cellphone-powered-for-decades These actually seem pretty cool, there is some nickel-63 inside that sprays its radioactive decay products into a diamond crystal semiconductor where they ionize electron-hole pairs that are then swept apart by the internal field of a PN junction similar to how a solar panel works. That's what I could glean from a pop culture article anyway. Low power output but you'll never have to reset that RTC again.
|
# ? Apr 13, 2024 14:51 |
|
ryanrs posted:Actually seen them being used? I think ElectroBOOM did a video on it.
|
# ? Apr 13, 2024 15:38 |
|
PDP-1 posted:Just use one of those 50 year lifetime, radioactive Betavolt batteries in its place and the scary lithium warning sticker goes away! The little I've read on these things indicates they effectively store something like 3,300 MWh but only actually generate 100uW @ 3V. That's not even enough to power a cell phone.
|
# ? Apr 13, 2024 16:06 |
|
Just received my aliexpress hotplate I ordered during a sale, it's got to be a record at 10 days from order: https://www.aliexpress.com/item/1005006053783156.html Seems like there's a relatively small heating element in the middle, but it does spread out evenly pretty quickly. It might be about 5-10 degrees C off at the top end, but I haven't really calibrated the camera for emissivity so who knows. Not difficult to compensate for it anway.
IMO worth it over the smaller one like (like https://www.aliexpress.com/item/1005006380068665.html) and it's a good price which goes up quickly for larger ones. Also received a bunch of WS2812 LEDs, this is my 3rd order because I don't pay attention/think things through First I missed that the LEDs came on little individual boards with a cap and resistor on each, I guess to let you solder them together with wires? Anyway, no good for me. Then I ordered a bunch of surface-mount LEDs which is what I thought I needed, but then I realized that I wanted them to be about level with the top of a 7-segment dispaly, which would mean makeing a riser board or something So the solution it turns out is WS2812D-F5 (or F8) They're supposed to be like the one on the right so you can tell which side is +/- but one of them was instead lol PS. also they work just fine with 3.3v Vdd and data so that's gonna make my life much easier mobby_6kl fucked around with this message at 17:12 on Apr 13, 2024 |
# ? Apr 13, 2024 16:55 |
|
AlternateNu posted:The little I've read on these things indicates they effectively store something like 3,300 MWh but only actually generate 100uW @ 3V. That's not even enough to power a cell phone. Yeah, the ones they're coming to market with now are basically RTC backup batteries. The 1W+ version is a future goal and I haven't heard if they are even able to make prototype versions of it.
|
# ? Apr 13, 2024 18:01 |
|
I don't think you can even buy them without an NRC license (or equivalent wherever you live) either which sucks cuz I would absolutely buy one just to add to my collection of Radioactive Things e: Okay the one being sold has two loving curies of tritium in it which is a heckin' lot of tritium, way more than any exempt quantity anywhere. Thing is, every article about it says this: quote:The company reports that it has just been granted a Product Regulatory General License to manufacture, sell, and distribute its NanoTritium battery, making its P100a the first betavoltaic power source to be made available to customers who don't have a radiation license, haven't obtained regulatory approval or undergone special training. Except I think that's actually just them straight-up lying?? Because googling "Product Regulatory General License" just turns up articles talking about this product and not what the hell a "Product Regulatory General License" actually is, what entity granted it, or what it allows you to do. To sell something with radioactive sources in it in the US anyway you either need to show it's an NRC exempt quantity made into an approved product type (like a smoke detector), or you need an NRC license to buy it. Tritium has large quantity exemptions specifically for self-luminous elements for clocks, watches and compasses (which this is certainly not) but the highest you can go even for that is 750 millicuries, and that's only for marine compasses manufactured before 2007, sooo Shame Boy fucked around with this message at 21:46 on Apr 13, 2024 |
# ? Apr 13, 2024 21:35 |
|
"We used a nuclear battery because the lithium shipping labels weren't scary enough."
|
# ? Apr 13, 2024 21:40 |
|
I wonder if it would be enough to power something like a wireless keyboard or mouse. Would be nice to never have to charge that again.
|
# ? Apr 14, 2024 14:05 |
|
Collateral Damage posted:I wonder if it would be enough to power something like a wireless keyboard or mouse. Would be nice to never have to charge that again. Oh man. That gives me a flashback. I bought Microsoft's first gen wireless keyboard and mouse. Either one took 2 AAs. The batteries in the mouse lasted 1 week. The keyboard lasted a month.
|
# ? Apr 14, 2024 15:39 |
|
I built a custom wireless board a while back using a nice!nano controller (basically an Arduino Micro with a built in BLE module). It's powered by a 130mAh battery and a charge is good for about two weeks of daily use.
|
# ? Apr 14, 2024 16:25 |
|
OK, so this isn't really micro power, but it's related if you power your circuits from a USB power bank. The Anker PowerCore Slim 10000 has low quiescent power draw. Specifically, the cheap $16, 3 port model (USB-C input only, micro USB input, standard USB A output). You need to pull a current pulse of 100-150 mA every <120 seconds to keep the power bank turned on. But if you do this, it will run for over 2 weeks. Most power banks have atrocious idle power draw, because it doesn't matter for their standard use case. But this specific Anker model is cheap and good, and the engineers did pay attention to quiescent draw. e: Hmm, maybe you could operate off a supercap and wake up the power bank every hour/day/week for a minute to recharge it? That could extend the time to months maybe. I should try this! ryanrs fucked around with this message at 17:23 on Apr 14, 2024 |
# ? Apr 14, 2024 17:19 |
|
kid sinister posted:Oh man. That gives me a flashback. I bought Microsoft's first gen wireless keyboard and mouse. Either one took 2 AAs. The batteries in the mouse lasted 1 week. The keyboard lasted a month. (NB. this is using the dedicated USB radio receiver, I assume using bluetooth would make it drain a lot faster.)
|
# ? Apr 14, 2024 17:56 |
|
ryanrs posted:e: Hmm, maybe you could operate off a supercap and wake up the power bank every hour/day/week for a minute to recharge it? That could extend the time to months maybe. I should try this! So this was the idea: 5V USB in, 3.3V out to your circuit. USB > linear-charger > super-cap > linear-regulator > 3.3V When the supercap voltage hits 3.6V, wake up the USB power bank by simulating a plug event, charge up the supercap, then disconnect the power bank and let it turn itself off. Efficiency will be identical to a 5V to 3.3V linear regulator. All the power savings come from using the power bank in burst mode. Quiescent current will be under 100uA, hopefully significantly so. But I'm not going to expend great effort on reducing it to <10uA, because there is a more efficient version of this design that uses a switching regulator to charge and discharge the super cap. This linear version is designed to be simple. With a 3F super cap and a 1mA load, it will turn on the power bank every ~1 hr. Goal is 3-12 months runtime at 1mA using a $20 Anker USB power bank. I really, really wanted this thing to be based on a 555, but even the CMOS versions use 200uA. It's that 3 resistor chain and the Control pin. I think that circuit architecture makes ultra low power fundamentally difficult. So instead I have to use a discrete comparator and 74LVC74 flip flop. But I really wanted to use a 555.
|
# ? Apr 15, 2024 06:31 |
|
PDP-1 posted:I experienced an interesting bug this morning while messing around with a dual core STM32H745 microcontroller. As someone who is planning on using this exact chip later this year I appreciate the warning! Any other experiences worth sharing? I'm planning on using the big core as a sort of DSP pumping audio through I2S/SAI interfaces and basically running all the business logic and serial ports on the M4 core. -- Bit back but the answer to why you float DC return in devices (but not at the source) is that if you chassis ground DC return in devices you're setting up ground loops. This can be an EMC issue, but the main problem is that now anything that is grounded is part of the power supply return. nbd when it's in the steel cabinet on a metal rail, but you could easily end up in the case where mounting screws was the only DC return path, so you pull the screws and now your device is missing ground. This either causes the thing to float at 24 V (not great, not terrible), or the entire power return could flow through e.g. a tiny ground line in a signal connector, which might just catch fire. Bonding power supply return to ground (PE) is mandatory in at least some industrial DC IEC standards but I can't remember the exact rationale. Leaving the secondary floating makes the voltage vs. chassis potentially undefined, so e.g. an insulation failure in the power supply could make your DC bus connect directly to the AC mains, that's one reason to do it at least. longview fucked around with this message at 17:06 on Apr 16, 2024 |
# ? Apr 16, 2024 17:03 |
|
At my previous job we had a cabinet full of fancy arduino clones for all the usual bit banging/interface noise. I miss that so I designed some DIP32-sized RL78 boards. I'm not even sure 16-bit microcontrollers need to exist with so many options in ARM and AVR, but it's a thing you can buy. Now to see if I can get the thing to talk to my UART at all.
|
# ? Apr 16, 2024 17:14 |
|
Stack Machine posted:At my previous job we had a cabinet full of fancy arduino clones for all the usual bit banging/interface noise. I miss that so I designed some DIP32-sized RL78 boards. I still feel weirdly dirty every time I have something that's just a teeny bit too intense for what a lil' 8-bit fella can do and the next level up solution for basically the same fuckin' price is some dual-core ARM thing that runs at 200MHz+. Like there's absolutely no real reason for this it just feels wasteful even though as far as I know the resources that go into making them are basically the same. Maybe it's a craftsmanship thing? Using a constrained chip makes me have to be very careful and clever with my design and programming, whereas throwing a raspberry pi pico at it means I can be a lot lazier, idk.
|
# ? Apr 16, 2024 17:30 |
|
Extra complexity is free until you have to start releasing security patches. For example, I've been looking at how to IoT-ize my various microcontroller projects. Until now, they've all talked over proprietary wireless or USB, where the security threat wasn't such an issue. But once wifi and ethernet is in play, I need to actually think about proper encrypted protocols. For my USB/Internet protocol translators, I decided to use second-hand wifi access points running OpenWRT. They are plentiful and dirt cheap on ebay / craigslist / facebook. Many models have USB ports. And if it runs vanilla OpenWRT, then I am likely to receive security patches and improvements for a long time.
|
# ? Apr 16, 2024 19:01 |
|
Shame Boy posted:I still feel weirdly dirty every time I have something that's just a teeny bit too intense for what a lil' 8-bit fella can do and the next level up solution for basically the same fuckin' price is some dual-core ARM thing that runs at 200MHz+. This one is in the $2 range and tops out at 32MHz. You can get a 250MHz STM32 for that, but the RL78 had a better supply range and I kind of wanted to play with a new architecture. Stack Machine posted:Now to see if I can get the thing to talk to my UART at all. Welp, I only had to cut one trace so I'll mark it as a W.
|
# ? Apr 17, 2024 00:52 |
|
250mhz, what are you, the king of France?
|
# ? Apr 17, 2024 01:06 |
|
longview posted:As someone who is planning on using this exact chip later this year I appreciate the warning! Oh cool! I pretty much picked the chip for the same reason - the M7 will be living in an almost interrupt-less ivory tower doing very deterministic DSP-ish stuff on a stream of data while the M4 handles all of the communication to the outside world and general housekeeping. The cores talk to each other by passing messages through a pair of FIFO queues in shared memory with hardware semaphores used to coordinate access. I'm only a bit over a week into programming this thing so I'm no expert yet, but be aware that a lot of stuff works different than it did on the old F4xx chips so plan for some extra dev time or get an eval board and start early. For example, SPI ports now have their own local Tx/Rx FIFOs and you can specify things like a wait time between when the chip select line goes low and when the clock/data lines start transmitting. e: Nerd pride on getting to blinky: LED1 is controlled by the M4, LED2 by the M7, and LED3 is controlled by the M4 passing commands through the shared memory FIFO that tell the M7 to toggle a pin. Cojawfee posted:250mhz, what are you, the king of France? PDP-1 fucked around with this message at 02:38 on Apr 17, 2024 |
# ? Apr 17, 2024 02:32 |
|
multi-core superscalar microcontrollers, lol
|
# ? Apr 17, 2024 03:04 |
|
PJRC Teensy took it too far, imo. I had to abandon them because their boards run at 600 MHz and use 100 mA. Low power consumption was a low priority for development, so out of the box, that's all you get. It's totally unworkable for anything that isn't plugged in.
|
# ? Apr 17, 2024 04:59 |
|
USB Super Capacitor BOM highlights: LTC3127 Buck-boost switcher MIC841L comparators + reference SIL2308 Dual n-/p-ch mosfet Turns out the switching version is smaller and has fewer components, although one of them is an expensive Analog Devices switching regulator. The MIC841 is a cool little comparator chip. It has all the guts of a 555, except the resistor divider. It even includes the flip-flop! This circuit is really just a big relaxation oscillator. When the supercap hits 3.7V, the MIC841 turns on the USB power bank and starts charging the cap. When it hits 5.1V, it turns off the charger. It does this on 3uA, 1/100th the power of a CMOS 555. Enough signals are brought out to pins to let you change the threshold voltages, explicitly request USB power, and get notified when the USB power bank is dead, but the supercap still has a little juice.
|
# ? Apr 17, 2024 06:42 |
|
Today in the lab there was a loud BANG and the smell of burning wood. I was across the room when it happened, so I didn't see which device it came from. The smell was strongest around my Tek scope (no!), but the smell was more like the building was smoldering, not a modern(-ish) digital scope on fire. Nope! A capacitor had blown up in my 30+ year old HP 6236B triple output power supply. It was a 0.1uF RIFA cap across the AC main. It smelled like burning wood because it's a paper capacitor. I will be replacing it with a modern metallized polypropylene film X2 cap, even though Digikey still sells paper RIFA caps.
|
# ? Apr 21, 2024 06:13 |
|
Some electrolytic caps smell strongly of fish sauce
|
# ? Apr 21, 2024 16:29 |
|
So why are RIFA caps still being manufactured? What's so good about them to offset the large physical size, high price, and explosive failures?
|
# ? Apr 21, 2024 17:41 |
|
ryanrs posted:So why are RIFA caps still being manufactured? What's so good about them to offset the large physical size, high price, and explosive failures? Rifa is one of those brands known to fail, isn't it?
|
# ? Apr 21, 2024 20:33 |
|
ryanrs posted:So why are RIFA caps still being manufactured? What's so good about them to offset the large physical size, high price, and explosive failures? They lose less capacity than those DAIM or MEX x2 capacitors. So for their normal life span, their performance can be slightly better. However, by now there should be better x2 capacitors that don't lose so much capacity due to self healing. Fun fact: you can even get those Rifa/Kemet paper caps in SMD versions!
|
# ? Apr 21, 2024 20:54 |
|
I had an 80s rectifying stick welder do that to me a few years ago, made an incredible amount of smoke. I had a suitable RIFA replacement though and was up and running in 20 minutes. I find lots of RIFAs in old equipment, I replaced them all pre-emptively in another welder.
|
# ? Apr 25, 2024 08:17 |
|
quote:[Senator Collins:] Well, there are … regulations governing the materials they can be made of
|
# ? Apr 25, 2024 08:25 |
|
I was thinking back now and it was indeed RIFA caps in old Mac computers that are known to go bad. I said some wrong 4 letter brand previously.
|
# ? Apr 25, 2024 18:56 |
|
Fun fact: you can actually get brand new Rifa paper caps in SMD versions, for all your modern, miniturized smoke generating needs! https://nl.farnell.com/kemet/smp253ma4470mtv24/cap-film-suppression-4700pf-class/dp/3527162 This one's out of stock but there are others from both Rifa and Kemet. They're actually surprisingly expensive. I want to buy a bunch and see how the newest, most modern paper caps fare. Hook them up to mains with a fast acting fuse in series and leave them for the next 1 or 2 decades to see what happens. But they're a bit too expensive just for a silly test.
|
# ? Apr 25, 2024 21:06 |
|
Paper caps have always been junk. Back in the 40s and 50s it was the acids in the paper would throw off the values over the decades, eventually shorting. Edit: and they were covered in wax, so they're gross and sticky.
|
# ? Apr 25, 2024 21:27 |
|
Decided to roll the dice on a Chinese PCB assembler (PCBWay) sourcing my more complicated components, specifically a SmartFusion2. They found some, but the PCBs are not usable, they've already been programmed (once, I suspect using Microsemi's own "we'll program your chips so your manufacturer can't steal your firmware and knock out knock offs after hours" service, as they have an extra part number on them they don't normally have), and the security key is turned on and so that's that, pretty much (the "factory key" is still enabled but I doubt I can get Microchip to give me that, either, being "some guy" and not a major customer). Time to finally learn desoldering a QFP144 and replacing it with a fresh FPGA, I guess!
|
# ? Apr 26, 2024 19:00 |
|
Charles Ford posted:Decided to roll the dice on a Chinese PCB assembler (PCBWay) sourcing my more complicated components, specifically a SmartFusion2. They found some, but the PCBs are not usable, they've already been programmed (once, I suspect using Microsemi's own "we'll program your chips so your manufacturer can't steal your firmware and knock out knock offs after hours" service, as they have an extra part number on them they don't normally have), and the security key is turned on and so that's that, pretty much (the "factory key" is still enabled but I doubt I can get Microchip to give me that, either, being "some guy" and not a major customer). Youre telling us you have suspected counterfeits with PCB Way pretty casually here. Is that what youre telling us?
|
# ? Apr 26, 2024 20:18 |
|
CarForumPoster posted:Youre telling us you have suspected counterfeits with PCB Way pretty casually here. Is that what youre telling us? No, the chips are 100% legit - they're just preloaded with some firmware intended for some other PCB. It's kind of the opposite of a counterfeit, they're super-legit, they're only good for one application and it's not mine (short of me getting the encryption key to be able to run the erase command). Basically, either they're stolen or (in my opinion, more likely) they came from some liquidation sale of a company's assets and the chip recycling place didn't realise they were programmed. You'd only want to steal pre-programmed chips if you intended to make more units containing that specific firmware. Sadly, I can't identify the original owners, the SmartFusion2 software will tell you it's programmed (1 cycle and it doesn't know how it was programmed, a sign it was done with a more serious tool than the JTAG programmer) and that the name of the design was "A1_top". The name of the design is just your project name, so their project name in Libero SoC was literally just "A1_top", not much to go on.
|
# ? Apr 26, 2024 20:26 |
|
Charles Ford posted:No, the chips are 100% legit - they're just preloaded with some firmware intended for some other PCB. It's kind of the opposite of a counterfeit, they're super-legit, they're only good for one application and it's not mine (short of me getting the encryption key to be able to run the erase command). That sounds like recycling components that are marketed as new, a form of counterfeiting.
|
# ? Apr 26, 2024 20:28 |
|
I suspect they weren't "recycled" (like they weren't pulled off a board or anything), but programmed at the Microsemi factory (as a service they offer), placed on the reel, then at some point they fell into the hands of PCBWay's suppliers. If you want to count just selling pre-owned, unused otherwise original parts as counterfeit then sure, of course. Those suppliers definitely do find specific, fancy ICs like FPGAs under sofas and things though, past runs with PCBWay where I used a Lattice Ice40 had the ICs branded "SiliconBlue", the original manufacturer that was bought out by Lattice a long time ago, and a date code saying they were manufactured 10 years before my order.
|
# ? Apr 26, 2024 20:40 |
|
|
# ? May 28, 2024 16:34 |
|
Quick question. I lost the IR remote for this RGB LED string: https://ibb.co/3kTDXxR The string itself is black and made of a thicker coated flex material, in contrast to those cheaper white strip ones. I've got a few of those as well, but none of their controllers work. I naively figured they all used the same chip, but evidently not.
|
# ? Apr 26, 2024 20:59 |