Clamp something on to the wire, needlenose pliers maybe, will act as a heatsink!
Good suggestions! I made a template, I’m gonna give it a try. I’m not sure I will have the right combination of patience and free time before the contest ends though. It caught me at a time I don’t have my workshop setup so I’m kind of doing it on the kitchen table.
I’ll give it a shot but if I wire it up and it doesn’t boot up or have more problems getting it soldered it may have to be put aside for a later day.
Status Update: 3 wires completed
Status Update more than 3 wires:
I didn’t leave enough room for the battery now I gotta move the oscillator
I had also wired it up wrong too anyways.
This thing better boot up when I put on the power or I will be a sad panda.
Welp, I knew I was playing with fire. Battery ran out of charge after a few minutes and this apparently was long enough for an undervoltage situation to corrupt the flash.
A lot of work for medium levels of magic blue smoke. Maybe I could endup on hackaday fail. LOL
Time to put the tools away…
But I guess this kinda means we can have an Arduboy with 8 parts counting the circuit board, maybe I should make a few bare essential Arduboys to commemorate the passing of this art project. On flex. Oh wait the battery is still a problem.
EDIT: The more I think about it, I don’t think it was a straight undervoltage problem because there is a brownout. I think it was intermittent connection or maybe even a short circuit.
Yeah the only way having a low input voltage could faff things up possibly would be if you were flashing a new hex with low voltage. I’d bet on something simple like Murphy’s hiding a cold joint somewhere.
Cold joint is possible, it’s kind if too difficult / not worth it to troubleshoot. I’m fairly sure the flash is corrupted though. I’ve killed these chips in a similar situation when trying to connect them to a coin cell battery. The atmega chips as a whole tend to be pretty robust but any chip is susceptible to being fried when you flick on and off it’s power supply on and off rapidly.
I’ve been guilty of thinking it would be fine to just hold an icsp programming header onto a dev board I made only to accidentally move a bit during flashing which corrupted the fuse bits preventing me from reflashing without a high voltage programmer before (the only way to reset fuse bits once they are set to disable spi flashing over icsp). I’m sure your chip is recoverable but like you said it will be tedious to fix it. It’s just a shame that you obviously put a lot of work into making this and now it’s a paperweight.
I do this kind of thing more often than I should, so far I’ve only killed (temporarily) 1 Arduboy.
@bateske Out of curiosity what’s the origin of the chip was it pulled from an Arduboy?
I had an Arduboy someone sent in with a busted reset button so I just took it off that with a hot air tool.
And it is a paperweight now, I learned a little bit when building it. Mostly I remembered how difficult it is to work with coin cell batteries (you really need some kind of mechanical force).
To be honest, I was mostly doing this to try and get some attention for Arduboy on the hackaday sculpture contest, but @rv6502 did better than I ever could have getting Starduino featured on the front page! So, I don’t feel so bad actually. Thanks Stephane!
if you have some mini grabber test clips to hand it may be worth trying to reflash it.
I once had an arduboy go garbled like that I forget what I was doing but its been fine since reflashing the bootloader.
The actual reflashing wouldn’t be the problem, the issue is getting a reliable connection to the battery.
I had planned on using a liberal application of conductive ink to help make the connection, but it had dried up.
But even these aren’t reliable connections to the coin cell battery. It’s something I struggled with immensely when trying to develop the original prototype when trying to move away from the battery with solder tabs. I tried virtually every method of interface with a coin cell battery, but only a very strong mechanical connection is reliable.
From what I understand this is because you need the mechanical force to overcome whatever oxides are on the surface of the metals. Even with the best conductive media, you still will have a poor connection without a lot of pressure or a laser weld.
If I had used one of these maybe it would have worked.
I’m really kicking around the idea of what it would mean to build this same circuit on flex. MAYBE Z-tape is enough. I figure adding a coin cell connector with a PCB that could all be hidden behind the screen is another sorta-close approximation, but using a PCB kind of defeats the purpose in the first place.
I took a more practical approach on my ATtiny85 setup
You could fashion a battery holder out of copper or a wooden clothes peg with copper pads.
I’ve read about using paperclips and have also used just bits of wire. You could use a small rubber band to add tension.
Edit: it could be disastrous but clamping the battery between two magnets just popped into my head.
You can see that the paper binder is a good use here, that’s the kind of force you need.
Build it for next year!
And put a coil, an integrated bridge rectifier and a 4V Zener to do wireless charging while encased in epoxy. It won’t be a great protection circuit but it should be good enough.
You can also use the coil as the data transfer method with a modified bootloader and another MCU controlling the charging coil. (I haven’t tried that method yet.)
I tested with a couple of weak magnets, I had to wrap each one in some conductive kitchen foil to get a connection.
I think if you was to stick some copper tape to some earth magnets you could get a strong connection. Possibly use one magnet if it’s strong enough
Edit again: actually you may want to avoid copper tape you could end up with a levitating battery
Shame the 32u4 isn’t available in a DIP like this ATtiny, could be done with a 328p though!
I mean, it worked, the flash just got corrupted and I was too frustrated to keep it going. There is the TQFP package which would be loads easier as it actually has leads instead of just being tiny pads.
Also I did actually design a bare minimum PCB that fits everything behind the screen but I’m not totally happy with it yet.