Anyone seen this before? (Video below).
Random boots seem to play the sound in slow motion, and also light the blade white-when-off for no apparent reason. Then the next reboot seems fine.
Video breakdown:
- First boot might be a little slow - not sure, but blade seems correct;
- Second boot definitely wrong - slow audio with random blade lighting;
- Third boot I think is correct (Anakin’s scream sounds awful by its nature);
- Fourth boot seems good.
OS is 5.9, and I’ve also tried the 7.x from Github. Spec is Bluetooth via LED pad (capacitor fitted between 3.3 volt and ground), 3mm red accent LED, smd blue accent LED, and three sub blade accents inside chassis, one of which is the crystal. So admittedly it’s a busy little install.
Off-camera, after I’d switched from 5.9 to 7.x, I booted it and got a clean boot sound, but when I entered the volume menu, the audio was in slow mo again.
I’ve tried a second SD card, formatted with proper SD formatter, and also removed all the Mac ‘ghost’ files from the SD. Battery is fully charged.
I’m not sure what else to try.
Config is here:
And video is here:
OK, I’ve been fiddling, and I think it may have been a power issue. I was doing the tests with the battery + wire, neopixel connector + wire and positive to charge port wire (which splits and goes to the board) all just twisted hard together so I could leave the extra wire length in case I had to open it up again. Even like that, battery level announcements suggested everything was healthy (4.07 volts) but I’ve now soldered that joint and so far it seems better. Maybe it was an amps or resistance thing? I don’t know.
Anyway I’ll keep testing and report back if things go awry again. I think I’m starting to be a victim of my own caution. I try to test, test, test at every step of every build, so that if there’s an issue I can catch it before it means tearing down a completed install. I’m starting to think that approach may cause more problems than it solves. 
Slow sound tends to be associated with the clock circuit not working. This can sometimes be caused by power problems, but it can also be caused by components touching the clock capacitors. (C7/C8 on the part map):
1 Like
Thanks Prof. That’s good to know. I’ll make a note of that for future reference. (My pile of notes is growing! LOL!).
I’ll double-check when I get home that nothing is touching those clock components, but so far it’s still looking like power was the issue. Fingers crossed that’s all it was.
I’ve got this happening on a 3.9 I just installed. It was working great, then I moved one of the accent pixels (so very low draw) to the 5v port so it would shut off with the board. now it’s doing that regardless of if that lead is connected to the 5v or not.
C7/C8 have not been touched.
What’s odd is if i reset it it starts out normally and then slows down as the clip plays.
It could mean that there is a problem with the clock circuitry.
The clock circuitry is quite sensitive, so if something is touching it, it can stop working.
(In my case, the thing that is touching the clock circuitry is usually my finger.)
For reference, the clock circuitry is Y1, C7 and C8:
https://fredrik.hubbe.net/lightsaber/v6/partmap.html#y1
For information, while working on the Proffieboard V3.9 (longboard version with Aleyan) we get the same issue and identified the problem around the clock circuitry (which in our case is different than the original V3.9).
I then made some research and will try a new routing schems pour this component :
Like adding a GND shield to avoid interference, no signal of power line under the quartz that seems to be the recommandation for those kind of component.
I will give it a try when new prototype arrives and let you know.
As a last call, you can just unsoldered the quartz from the board, and run it with the internal quartz of the STM32. It seems to work pretty well. I made some tests with a config including OLED + BLE + 2 blades of 244 pixels (for each blade) and the board works perfectly.
Researhc sources :
- some talks with a frien who is eletronic engeenier (even if for him, the routing should not impact as much the quartz)
That’s some crazy shit. I didn’t know any part of the board was sensitive to touch like that. I left the board unpowered overnight and it seems fine today. Can I assume that this is OK now, or would it need to be damaged in order to exhibit those symptoms? This is a build for a friend, and he mailed it to me to replace the 2.2 with a janky amp (another one, i don’t know why I’ve had so many of these die on me but it’s kind of put me off the 2.2) but I don’t want to send this one if there’s a chance it’s already pre-fried.
The only way to know is prolonged testing.
