A2000 RTC clock not advancing

[amigasith]

Cheers, folks! I recently got a pretty nice A2000 that seems to have at least one problem (maybe there are more that I haven't discovered yet ): I can successfully set the time on this A2000, but the RTC does not seem to advance. So for example, when I turn the A2000 off and then on again, it still has the same time when I last did a "setclock save". And even while the A2000 is up and running, a "setclock load" will restore the time of the last "setclock save". Just as if the time stored in the RTC is not advancing...

The battery was replaced by the previous owner and it hasn't leaked yet. Measuring its voltage also shows nice ~4V. I also checked pin 9 and 18 of the RTC chip - it actually gets the 4V from the battery. Therefore, I think this problem is not related to the battery.

The A2000 has a Rev 4.3 board, but it seems that one of the previous owners did all the upgrades in order to turn it into a Rev 4.5. One thing that caught my attention is a resistor soldered to diode D800, which is part of the RTC circuit. All the guides I found in the Internet that describe the upgrade say

NOTE: On rev. 4.3 and above this resistor is on PCB as R1000, located to the left of Q302

So this makes me think that on my Rev 4.3 board, this resistor is not needed, right

Or maybe I am looking at the wrong spot and somebody has a simple explanation? Any hint is greatly appreciated!

[thomas]

Try "setclock reset".

[hooverphonique]

maybe the little 32KHz oscillator can has packed it in?

[amigasith]

@thomas: Thanks for your tip I just tried "setclock reset", but unfortunately, the RTC clock still does not seem to advance Here's what I did: "setclock reset" -> "setclock load" -> set the current date / time by hand -> "setclock save" -> wait a few seconds -> "setclock load" and the time is again set to the time of the last "setclock save".

@hooverphonique: That was my thought, too, but somewhere else in this forum I read that those oscillators hardly ever break. But it is for sure worth a try in my opinion! What do you think - would this be an adequate replacement?

[bebek]

Originally Posted by amigasith

@thomas: Thanks for your tip I just tried "setclock reset", but unfortunately, the RTC clock still does not seem to advance Here's what I did: "setclock reset" -> "setclock load" -> set the current date / time by hand -> "setclock save" -> wait a few seconds -> "setclock load" and the time is again set to the time of the last "setclock save".

@hooverphonique: That was my thought, too, but somewhere else in this forum I read that those oscillators hardly ever break. But it is for sure worth a try in my opinion! What do you think - would this be an adequate replacement?

I would say oscillator and it happens quite often. Otherwise ic.

[amigasith]

Yay, I found it In the end, it was very simple actually, so hopefully the following description is of use for other people who might have the same issue

I really thought that the oscillator was the issue, since the IC at least stored the time correctly and the battery was also okay as described above. So I inspected the RTC circuit once again and read about the small sub-circuit for the oscillator crystal that consists of one fixed capacitor (20pF) and a trimmer capacitor (VC800, 4.5 - 47pF). If this small extra circuit has wrong capacitances, the quartz crystal does not start to oscillate. See this part of the A2000 schematics to see what I mean:



So basically, wrong capacitances disable the oscillation effect if I got it correctly. Since the A2000 is quite a few years old, I thought "why not check the trimmer capacitor". And indeed - I had to use some force to turn the small screw around initially - it was stuck a little. I turned the screw around a bit to the left and a bit to the right and finally back to its original position. Then I stored the time to the RTC with "setclock save" and voila, it now advances time like it should

So my guess (or poor man's explanation ) is that maybe the capacitance of the trimmer capacitor was either close to zero or close to infinity before turning the screw around a bit. At least this would explain why the RTC circuit behaved as if there was no oscillator at all.

[hooverphonique]

that would be a suitable replacement, yes.. glad you got it working..

[Hewitson]

Originally Posted by bebek

I would say oscillator and it happens quite often.

It is actually very rare for a crystal or crystal oscillator to fail.

[pandy71]

Battery leakage? Oscillator (trimmer?) affected?

[amigasith]

Unfortunately, the "turn a little left and right" solution for the trimmer capacitor didn't last very long... The A2000 kept its time for a couple of days and then the clock stopped advancing again. But it was the trimmer capacitor for sure, since the "turn a little left and right" trick made the clock going again for another couple of days every time I tried it.

Anyway, I wanted a proper solution for this and therefore I removed the old trimmer capacitor and the other 20pF cap (C810) and soldered 2 x 22pF capacitors instead. Why 2 x 22pF? Well, (1) because the manual of the OKI M6242B RTC chip says so, and (2) because most A500 512K memory expansions with the same RTC chip also have 2 x 22pF capacitors soldered on the board.

So here is how it looks like now:



And it works very nicely The A2000 clock hasn't stopped advancing since I made this mod a couple of weeks ago. And the stored time doesn't drift away, too. So all in all, I think it's a viable option for people who are not looking at fiddling around with the capacitance values of the trimmer cap

[K.C.Lee]

Crystals do age over time, so their electrical characteristic could change.
As to the cap values, they (2 * "load capacitance") really should be what the crystal manufacturer specified and it is related to the accuracy. Also one would need to account for the board layout and package parasitic capacitance.

On the other hand, some trim cap electrode (e.g. silver paint based) could be corroded over time, so the parasitic (e.g. internal resistance) could change.