After the success of Toms first little Chinese clock kit, this week we moved on to the second. This is the SH-E 879 kit. A somewhat bigger PCB than the last one, 6 digits, and battery back-up, plus an on-board 5V regulator.
We spread the build over several days, tying it in to his home schooling. Each part of the build including me explaining what the component is and does, and why. Starting as always, with the least heat sensitive parts. We had a slight problem when one of the oscillator load capacitors pushed its PCB pad up, so that section of track had to be linked out.
Tom knows that I insist on a specific orientation of parts to ensure component values can be easily read. He managed to put two resistors in the wrong way, but electrically this doesnt matter so ive let him off!
And so today, he completed the build, I checked the soldering for dry joints, missing joints, jumps and splashes. All looked good.
And - it didnt work!
We spent ages trying to find anything amiss. In the end, Tom left me to fault find. After quite some time, I discovered that one out of the four 3mm LEDs, actually fits the opposite way to all the others! After changing this around, and repairing the now broken track that it connected to, I had the four LEDs lit. But still no clock function! The microcontroller Vcc pin voltage was correct, the ground pin had continuity to ground, all the bus pins showed expected voltages, even the clock pins looked right! I was about to break out the 'scope, when I thought to check how the chips reset line operated. It turns out reset on this is active high, rather than low as is usual. So for the processor to run, the reset line should be low, which it was. The microcontroller was in fact the only component that Tom didnt install - I fitted it due to the risk of bending the pins...
...I decided to pull the chip - oh, bugger! Guess which one and only leg had got bent under? Yep, the bloody reset pin!
So now, after fixing my own cock-up, its working. We havent yet played with the one and only button to find out how to set it. Two things im unhappy with though - 1, for some reason there is a 10k resistor over the diode that feeds the back-up battery onto the Vcc rail, this looks like some sort of charging attempt, but the coin cell is non-rechargeable. I wasnt going to fit this resistor, and only did so just in case I was missing something in the fault finding! I'll likely remove it later. And 2, the four 3mm LEDs are far too bright! They are current limited by a 470 ohm resistor, I think maybe 1k or so would be more sensible, and reduce the current draw a little as well.
But on the whole, another successful kit build by Tom.
No comments:
Post a Comment