Bev Clock

[wotnot]

..last installment ; white cotton glove day (don't go near clear perspex without them =)



Assembly the enclosure is a fickle exercise -- the nuts aren't captured in the slots, so you have to do this...



...pass the screws through the perspex, and wind on nuts 3~4mm, then assemble to sides. Use pointy tweezers to hold nuts, and do not overtighten...(clearances are tight, too tight and you'll crack the perspex...found that out first try =)



PCB assembly just freely sits in enclosure (mounting holes in PCB not used)...damn, missed an LED straighten...I'll remove side panel later and fix =)

Finally, a small potato cam video of it working...

[wotnot]

Oh!!... forgot to unravel the LED 'ghosting' evident with GT250's kit...remember, I figured this was something to do with the multiplexing between LEDs and 7-segment display...

...the video was captured @ 30fps, and as mentioned the immediately obvious difference between the two firmwares was the blinking rate of the 7-segment colon...

...even at that framerate, you can pick up what the eyes can't see...

frame @ 26.043



frames @ 26.075, 26.111, 26.148, 26.179, 26.215



frame @ 26.251



Your eyes can't/don't see the 'blanking' frames.

You can't have everything happening 'at the same time' on these MCUs. so I was suspicious the 1Hz event turning LEDs on, was aligning with the moment the 7-segment colon turning on...ie; in GT250's kit to colon flashes at 1Hz.

So..in my head m'kay, I saw the colon flashing at a higher rate on this kit, and if you were trying to avoid this conflict, the logical choice would be to toggle the colon at around 3Hz (3 times per second), so I checked the colon timing events in the video (time in ms);

26.975
27.323
27.667
27.979
28.327
28.671

That's about what I expected... you'd need to toggle the colon on a prime number, and either 331ms or 337ms would avoid the conflict (and you catch/correct the under-run with blanking time)...I don't quite get the accuracy I need @30fps, but it's close enough =)

If one can be bothered, a click frame by frame will show the ghosting isn't gone altogether ~ it's been moved to other segments of the 7-seg display, but due to the way the firmware is written, it's too brief for the naked eye to see.

@GT250 ~ after these tests (and me knowing for sure your PCB wiring is correct), I'd reckon there's nothing wrong with your clock kit now at a hardware level --- D55 not fully lighting & ghosting on LEDs D56 thru D59 I feel is entirely the fault of the code/firmware being used (presuming the MCU itself is good electrically).

I consider that *now*, I've actually finished "looking at this"... and I've got a pressy for grand-daughter