Delays are Inevitable

Published on 2017-11-03 in µGame.

Both of my prototypes stopped working. Just like that. And that happened just as I was going to show them off in Adafruit’s show-and- tell. Oh well, bad luck I guess. I decided it must be the flex ribbon connecting the displays to the PCBs, and I re-did the connection with my soldering iron, inspecting it carefully for shorts, etc. They started to work again. For a moment. Then the newer one stopped responding completely, and no amount of re-soldering helped. So I de- soldered the display from both, and put the one from the old prototype on the new one. It worked for two hours, and then stopped again. This time I decided it must be broken traces on the flex ribbon, because I bent it too much. Since I don’t have any more displays with 14 pins, but I have a couple of modules with the same display but with a different connector, I decided to continue work by using one of those modules and a Trinket M0.

The displays stopped working with the Trinket too, even though I didn’t touch the flex ribbon or the connections at all.

So I began to poke at the non-working display, trying slower SPI speeds, different commands, etc. At some point I got it to work — I was sure then that the problem is not with the display module, but with my code.

Turns out that the display working or not was a completely random thing. The problem is that some of the SPI commands, such as reset or wake up from sleep, require a lot of time for the display to perform, and delays are needed. I added those, and also lowered the SPI speed during the initialization sequence, and it works on the Trinket every time. So I compiled the firmware for the “dead” prototype, uploaded it — and it works too! I bet that when I solder the other display it will also work.

One more obstacle out of the way.