Ice Tube boost stopped working?

[Agent24]

Update:
Progress is being made... The boost converter now works again!

I replaced the inductor but didn't really expect it to fix the problem, since it didn't test short or open - and no, it didn't fix the problem.

So then I decided to just start replacing the only parts I hadn't looked at: D3 and D5.

Replacing D5 (60v Zener) didn't cure the problem - but after replacing D3 (MBR160) the boost converter is now producing 60v once again - see attached photo - no longer just a square wave!


Unfortunately, there is still no display after reattaching the MAX6921 and tube, although interestingly enough, the boost output drops only to 59v with the MAX6921 installed and 48v with both it and the tube... so perhaps it's a bad connection, the MAX6921 isn't doing anything\drawing current, and the voltage does not drop.

Will check to see if there's any signal coming out of the MAX6921, and I guess check all its solder joints if not.

[Agent24]

Update:

More progress! I have lights, action, and I used my camera More on that down the page...

Soon after my previous post, I checked the outputs of the MAX6921 (without tube attached). Everything showed a square wave on the 10-100Khz range on my oscilloscope (although no idea of the voltage level, never made any comparison test - was going to do so today, but not sure I need to now) - which blinked on and off every second (which makes sense as the clock apparently does this when it looses power)

Funny thing was that the boost converter output voltage was not dropping enough (seemed no load?) even though the chip was outputting a signal. Made me think something horribly wrong with my soldering (although confusing since it all looks fine etc) but didn't have the time to try fixing it yet.

So because of this, yesterday I tested the tube outside the clock. Probably due to my crappy test setup I wasn't able to light more than 3 digits at a time, but it seemed to prove the tube was indeed OK (which I already thought, but wanted to test anyway, just in case it was a tube fault somehow)

Today, after re-trawling through the forum, I found someone with a very similar problem as me. He got the tube to light by shorting out the collector and emitter of Q4, or in my case, the drain and source. I did this and was greeted with a nice bright display which said "Alarm On" and then started displaying the time (and of course, blinking on and off)

No idea why it chose to display the time it did but I'm glad it started at all! Now I guess I just have to try and find out what's wrong with Q4 and why it's not letting the tube come on. But at least I know the tube and MAX6921 appear to be working OK.

Will post another update when more progress is made...

[Agent24]

Update:

It's alive! without any franken-wires too! Anyone still reading this, you'll probably laugh...

I checked the output at Pin 5 of the ATMega, it was low as it should be, yet Q4 (ZVP3306A) was clearly not working as it should. A closer inspection and re-check of the datasheet revealed that, you guessed it, Q4 was in fact backwards Damn that silly E-line package! (Yeah, yeah, we know it was my fault anyway, but still...)

Added the switches, and they all seem to work as they should too. Looks like all I've got to do now is just clean the flux residue off and stick it in the case! Hopefully it keeps working



And to answer some likely questions about my build:

Yes, I made the boards myself.
Yes, They are single-sided.
Yes, I had fun modifying the layout, and figuring out where all the jumper wires could go.
Yes, There is a heatsink on the regulator (No component side ground-plane to use as one!)
Yes, The PCB is a bit wider than the official one
Yes, You also need to make your own case so it actually fits.
Yes, You're probably better off buying the actual kit!


Also, thanks again to Adafruit support for mistakenly giving help even though I wasn't supposed to get any. I hope that didn't annoy you too much!