Ice Tube Clock quit working - 2021 Moderators: adafruit_support_bill, adafruit

[llryan]

My prized Ice Tube Clock quit working. It worked well for years but recently became intermittent: when I would plug in the power it would beep and work for a while and go dead. Now it is simply dead. I want to get it working. So far, I've verified the 5v regulator is working: 9v in 5v out. Reading other posts, looking at the troubleshooting chart and reading the assembly instructions, I checked what I can on the boost converter. Bad results: checking from the cathode end of the zener diode and the center pin of the 5v regulator (7805?) I get 9v. I should be getting about twice that. When I remove power, that 9v measurement goes down slowly indicating that some connected caps are slowly discharging - the caps are ok. I get the same reading when I measure across the diode. I believe the boost converter is not working. The clock is ticking away but there is not a high voltage to light up the tube.

Here is my question: where can I find the schematic and the PCB layout so I can troubleshoot this clock? Please send some links. I hate to trash it even though its an old project, I really like it. I see people selling them on Ebay for over $200 - not what I have in mind to get a working clock back on the shelf!
Thank you for any help,
Larry

[adafruit_support_bill]

The guide is here: https://learn.adafruit.com/ice-tube-clock-kit/overview
You can find links to the schematics on the Downloads page.

Also check the troubleshooting guide here: viewtopic.php?f=41&t=27032

[llryan]

Thank you for the quick response. I was able to troubleshoot my symptoms: no beep and no display down to a bad display tube. When I unplug the tube and plug in the power supply, I get the startup beep, the correct boost voltage and a Vcc of 4.5v measured on the ATMEGA 168V. Also I can check the jack, JP2 and see the voltages for the various tube segments changing in 1 sec intervals indicating that the display would be blinking. This tells me that the circuits are functioning ok. When I plug in the tube then power up, all the voltages on the tube elements are .4v steady and there is no beep indicating that the ATMEGA processor is not starting. The boost voltage measured at the cathode of D3 is between 20 and 14v so its working fine. The tube, not so much.

I measured continuity between all the tube pins and find only the normally grounded pins to be connected internally. Why the tube causes this is a mystery. The tubes are unobtanium so I can't get another to continue troubleshooting. This looks like the end of this fun project. Its place of prominence is now taken by the Animated Eyes project - another favorite Adafruit project.

Thanks again for guiding me toward the schematic diagrams and circuit explanations. For sure, there's enough info there to troubleshoot the clock.

-Larry

[adafruit_support_bill]

That was one of my favorite kits. But we had to discontinue it when the tubes started getting harder to find. You might still see a few on eBay occasionally. But definitely not in stocking quantities. :(

[jarchie]

Sorry for replying so late, but I don't check this forum regularly anymore... If you're still trying to debug the problem, perhaps the following is still of interest.


It is somewhat unusual for a tube to fail as described in your post. They usually just get dimmer and dimmer over time.

What is the part number for Q3? Is it one of the older PN2907A or ZVP3306A transistors? If so, you might try using a ZVP2110A from Digi-Key or Mouser. The PN2907A in particular is known to work for years in the Ice Tube Clock and then cause an apparent display failure similar to what you describe.

[llryan]

Yes, I still want to get my ice clock working!

I believe Q3 is marked 2907. If this transistor is bad, could I still get a boost voltage reading like I mentioned: 20 to 14 v? Seems like it should be higher. DigiKey shows zero stock but I see Newark has some. Its worth a try to replace it with a ZVP2110A. It also makes sense that low boost voltage would result in a dark display. Thanks for the hint. I'll try replacing Q3 and hope for the best.
-Larry

[jarchie]

The original PN2907A PNP transistor acts as a switch for the filament power and VFD driver chip. Normally, there would be a resistor to limit current between the PN2907A and the microcontroller, but this is omitted from the Adafruit design. The microcontroller has a max output of 20 mA per pin, but without a limiting resistor, the PN2907A pulls too much power. Strangely, this configuration can work well for many years, but the microcontroller sometimes starts to act up after a while. Since the ZVP2110A P-FET switches based on charge, it pulls almost no current from the microcontroller. And once the overload condition is corrected, the microcontroller usually works properly again.

By the way, when the display is not attached, the PN2907A does not pull much current from the microcontroller. This is why microcontroller-controlled functions like the startup beep and boost circuit work normally again when you disconnect the display. So--to answer your question--issues with the boost circuit can be explained by the PN2907A. But the PN2907A isn't faulty; it just wasn't the best choice for Q3, which is why Adafruit started using a ZVP2110A for their later kits.

If you cannot find a ZVP2110A, it is probably possible to make the PN2907A work by sticking a resistor on the base pin between the microcontroller and the transistor. Forum user Russell27 has a good post on this approach: viewtopic.php?f=41&t=43908&p=219812

Hope that helps.

[llryan]

Yes, that helps very much! Thank you for the info - it makes sense using a P-ch MOSFET to lighten the load on the microcontroller. I found a source for the ZVP2110A and my order should get here in about 2 weeks. I'll post the results of this change then.
Thanks again, Larry

[llryan]

Ok, the ZVP2110A arrived and I replaced Q3, the 2907 PNP transistor with the MOSFET. ITS WORKING!!!!! I'm glad I didn't throw it away.
Thank you, thank you! I noticed that newer versions of the kit supplied the MOSFET and called it Q4.
jarchie, your support is very much appreciated. Thank you for responding and explaining how the circuit works. It all made sense.

-Larry, happy to have his Ice Tube Clock working again.

[jarchie]

Thank you for the update; I'm delighted to learn that you were able to repair your clock. And thank you also for reminding me how much fun I had building, hacking, modding, and reprogramming this wonderful little kit.

If it were not for your thorough diagnostics and description of the issue, I would have not recognized the problem as the PN2907A. So be sure to give yourself a pat on the back for the excellent work at your end. Well done.

I wish you continued luck with all your future projects.

Happy hacking!

[Limnidytis]

Ok, the ZVP2110A arrived and I replaced Q3, the 2907 PNP transistor with the MOSFET. ITS WORKING!!!!! I'm glad I didn't throw it away.
Thank you, thank you! I noticed that newer versions of the kit supplied the MOSFET and called it Q4.
jarchie, your support is very much appreciated. Thank you for responding and explaining how the circuit works. It all made sense.

-Larry, happy to have his Ice Tube Clock working again.

Larry,
I am in search of ZVP2110A transistors for a tube clock project but am having difficulty in locating a source -- could you tell me your source?

Thanks,
Johnny

[llryan]

I have some. Please contact me by email: llryan@msn.com
-Larry

[Limnidytis]

Larry;
Thanks but I checked eBay and found a supply there. I have ordered some from both sellers in China and Israel - it will be a few weeks before they arrive, but I need to collect quite a few other items before I’m ready to begin construction.

Johnny

[llryan]

No problem. I have 4 bags of 5 each - a surplus for sure.

Good luck getting your clock going.

-Larry

[jarchie]

If you're building an Ice Tube Clock from scratch, it might be worth checking out the board redesign I did a while back. The latest design files and corresponding firmware is on GitHub.