Hello! Asking for a little advice on some mods to your kit - I am stuck! Basically I have been trying to use some of these in a project,, but decided that I needed to use dip sockets so I can change the software at my leisure down the road. So I got hold of some high quality sockets (machined, gold, solder-tail contacts - don't try and use those crappy radio shack sockets you will regret it!). I soldered it onto the back of the display just as your kit directions, and I have made a few of them thus far. I have changed the original source code quite a bit. I had been testing the code on a breadboard. Now I am running code on the built socketed versions. Here is what I am finding...

On the socketed pieces, my code runs fine for 30 minutes, maybe an hour, eventually it will blink only one or 2 of the segments (randomly), stuck in a loop. Or it will go blank altogether, or even sometimes just lights one random segment continually, very brightly (ouch). It happens at random times (the worst kind of bug!), maybe after 5 minutes operation, but usually within an hour. So I think... OK I hacked the code up TOO much and jacked something up. So I went back and installed the original code you released on the chips. When testing again, I get same freeze results! I had been testing my code on a breadboard for some time, only a couple times did I get the frozen/loop thing. I could run it for hours with no problems.

My question is, did the introduction of the socket add something new to the equation, some kind of resistance or timing problem? Grounding issue? Again, I have made a few such socketed versions, have tested quite bit, and am using high quality sockets. Is there some code I can add to compensate for timing/resistance changes? At this point in testing, nothing else has changed from your kit, not even the code, with the exception of the addition of socket and the addition of an a/c adapter (3V 1A). I have tested the voltage and current of the adapter, it is regulated, appears to be a good quality. Barely over 3V. Could the culprit be some weird random fluctuations of current/voltage? I am going to test more soon with batteries to eliminate that possibility. But because things work pretty well on the breadboard even with A/C adapter, doesn't seem that is it. I am really baffled that the addition of a socket to the circuit could cause code/timing errors.

Are there delays, pin flushing, or any such tricks I that could add somewhere at the end of the loops to try prevent these bugs? Voltage regulator? OR? I am going mad here trying to wrap this up MAD i teELLLLLlllll yOUwwww!

Perhaps posted that just a tad too early. Doing some more testing now of a/c adapt vs. batteries, looks like things are running good using batteries, and all kinds of whack freeze/timing/hiccup segment errors with the a/c. This is especially so with my own code, which is animating segments much faster than original code. So when going direct from a/c to socketed readerboard circuits (no breadboard) using my latest code, I get errors within seconds on the a/c, and no errors after much testing on the batts!

Can anyone tell me what is going on here? Why is the attiny chip getting pissed off at my a/c adapter? I was careful to buy a supposed 'switched', regulated supply from digikey. Very steady 3.18 readings on the a/c. Rated 1A. Strangely, when this same code of mine is run on a breadboard circuit, I don't get these errors very often when using the same adapter. So I can conclude that something with the breadboard itself must be balancing the fluctuations better vs the direct a/c circuit hookup? Do I need some type of voltage regulator thingamabob?

I also wanted to point out to readers that there never is or was anything wrong with the code nor the kit design from evil mad - all these errors are of my own making

Perhaps mystery solved. 3.18 V just seems to piss off chip. I was going with the +- 10% theory, so much for theories. Dropping a/c adapter V to 3.02 solves the problems! All this debug hairpulling and the solution is a f***ing resistor?! The reason it wasn't obvious was because breadboard must have been providing just enough resistance to compensate during prototyping,, thus when I began circuit construction with dip sockets I blamed the addition of those to the design (or my whack code) and assumed voltage was OK. I'm a elec. genius here as you can tell. Thanks for all your help folks!!!

Sorry-- a little late to the discussion here.

Yeah, a socket works fine (and is a good idea if you want to pull the chip out to reprogram it!), but we do recommend running on 3 V DC, no higher.

-Windell