(Ice Tube Clock) Time not being kept while powered under battery backup

[rgollub]

Have built kit thru completion and ALL works as advertised with a minor exception as per the topic title.

So far what I did was to confirm that:
a. battery voltage is sound (it is; measured in the range 2.9-3.0 V while inserted)
b. processor clock is running (it is as per oscilloscope)
c. voltage of "batt sense" pin is as designed (it is; approx 1.2 V on external power; approx 0.7 V on battery).

The description is straightforward: when powering the clock back onto the 9 V source, the time displayed back starts off exactly from the time it was when the external power was removed! In other words, time "was not kept" while under battery power.

It could be either a hardware (faulty/marginal processor IC) or a software problem (untriggerred interrupt?). Unsure as yet.

Would appreciate any instructions and hints as to how I should proceed from here in order to troubleshoot my clock, which otherwise is superb.

Thanks in advance for attention.

[adafruit_support_bill]

b. processor clock is running (it is as per oscilloscope)

Is it still running when the 9v supply is removed?

[rgollub]

Tks for quick reply.

I wasn't clear enough in my point (b).

What I should have made unambiguous is that "the processor oscillator is orderly running while also under battery power".

Hence, replying to the query objectively: Yes, once the external power is removed, the oscillator continues to operate. I checked by observing the pin waveform with an oscilloscope. Didn't bother checking frequency, as I just wanted to make sure that the processor was still alive under the voltage supplied by the battery. Thus, as far as the oscillator is concerned, _it_ was "alive". However, if other areas of the processor weren't, that still remains to be tested.

Hope now the whole scenario is complete.

TIA for attn.

[adafruit_support_bill]

That is unusual. If the oscillator is running, it should still update the time while in sleep mode. Check the oscillator frequency. Is the time exactly the same as when the power is removed, or is it possibly just running very slowly?

[rgollub]

That is unusual.

Yes, I know... That's why I am consulting you.

If the oscillator is running, it should still update the time while in sleep mode.

Exactly. But, as it isn't updating then what must be assessed is whether "while in sleep mode" is true. Or, more to the point, "in what state the processor actually is after ext power is removed?".

Check the oscillator frequency.

Will do, but the probability that it is running at the crystal frequency is very close to 100%. See following comment.

Is the time exactly the same as when the power is removed, or is it possibly just running very slowly?

One experiment I have executed was to record the time being displayed at the time of removing the power. Then, roughly one hour later I checked what had happened to the internal time once I reinstated the external power. Needless to say that time started off from the recorded value (give or take a few seconds due to measurement imprecision). Had it running been slow, I imagine I would see a significant discernible difference: unfortunately that was not the case.

(For the record, this type of experiment was conducted at least thrice. The last one it is reported hereinbefore. Surely the outcome is reproducible as many times as required.)

Hope you are now convinced that the condition requires a strategy of attack, as, I am almost sure, the answer to the riddle won't be trivial, and I count on you to help me out. TIA.

[adafruit]

can you post a clear photo of your main pcb?

[rgollub]

can you post a clear photo of your main pcb?

Sure. No problem. Will revert in due course, probably tomorrow.

For my guidance, could you explain me your rationale behind the request? Maybe I could enhance your thinking with my own...

TIA.

[rgollub]

Check the oscillator frequency.

Will do, but the probability that it is running at the crystal frequency is very close to 100%. See following comment.

OK, here is the data regarding the oscillator:
Attached picture shows waveform, measured at pin 10 (resolution of V=.5 V/div and H=20 usec/div) giving a roughly period of 30 usec, or, roughly 33 kHz for the oscillator frequency.

Amplitude of roughly 1 Vpp with battery voltage of 2.82 V, which delivers (after diode) 2.66 V as VCC.

Finally, voltage at pin 13 (BATTSENS) is 0.70 V, or roughly 26% of VCC.

Waveform at pin 10

Oscillator_pin_10.JPG (51.66 KiB) Viewed 1368 times

can you post a clear photo of your main pcb?

Sure. No problem. Will revert in due course, probably tomorrow.

I have the picture but, due to its high resolution, your restriction of 1 MB per attachment disallows me from attaching to this post. Please instruct how/where I should send it to and thus have it reach you.

For my guidance, could you explain me your rationale behind the request? Maybe I could enhance your thinking with my own...

Your eloquent silence is well understood: you didn't want to hurt my feelings...

[adafruit_support_bill]

I have the picture but, due to its high resolution, your restriction of 1 MB per attachment disallows me from attaching to this post. Please instruct how/where I should send it to and thus have it reach you.

You can re-size it (800x600 should be fine). Or you can put it up on one of the many photo sites and post a link here.

[rgollub]

You can re-size it (800x600 should be fine). Or you can put it up on one of the many photo sites and post a link here.

OK. Re-sized pictured to within allowed limit.

Overview-cropped.jpg (570.75 KiB) Viewed 1366 times

PCB-cropped.JPG (749.79 KiB) Viewed 1366 times

[adafruit_support_bill]

That looks like a good clean build. Not sure what would cause the loss of time other than a malfunctioning processor. We can try replacing it.
Contact support@adafruit.com with a link to this thread and ask for a replacement Ice Tube processor.