Anyone ever had their RTC stop, and then start again?
 

As the title suggests. I'm not long after repairing my A501 Rev C and putting a new NI-mh battery on it etc last week, and setting the clock and proving that it maintains it again. Everything seemed fine until last night and then weirdly enough, i found the RTC was 1hr behind the real time, and then with closer inspection, i realised the seconds weren't moving, suggesting the crystal had stopped pulsing. Then after faffing around with it between WB and Amiga Tools, it suddenly kicked into action again, and has been working all night with the Amiga turned off (i looked at it again this morning). Really weird. The only thing i can add is that I've been messing around with keyboard(s) plugging and unplugging them from the motherboard, but seeing as the RTC circuit is localised inside the expansion card, in my mind, this should have no impact on the RTC.

Anyone had anything similar? It's almost as if it stopped because the Amiga was cold, but then started up again as it warmed up through being powered up?

How were you noticing that the seconds weren't moving? If it's with the normal clock program from Workbench, then the problem lies elsewhere. The OS reads the RTC just once on boot to get the time from it, and after that the time is counted by the CIA. So if the Workbench clock has stopped, there's an issue related to the CIAs (either hardware or software). When you reboot, it will read the time from the RTC again (which was probably ticking along fine in the background), making it look like it's suddenly recovered.

Amiga test kit CIA timer test is a good place to start with.

If it fails, see whether you have vsync/hsync on the CIA timer pins.

If only one of them fails, power down, swap the CIAs around, see if the problem follows the CIA.

It was in Amiga Test Kit i noticed the time was 1hr behind, and frozen. I went into this after booting into WB and noticing that during booting, it gave the message to say there was no battery backed clock available. In Amiga Test Kit, F8 told me that the chip was recognised for the RTC, and it showed me the time, but the seconds were not oscillating. I thought "that's weird, i'm sure this shows the seconds moving!?" I set the time to the correct time, and still no seconds movement. So it was at this point i was doubting myself, and tried booting into WB again. This time, it came up with the correct time and date during booting sequence rather than the mesaage saying no battery backed clock. Within WB, looking at the clock widget, the time was indeed working, and you could watch the seconds move. So then i rebooted again back into Amiga Test kit, pressed F8, and right enough now the seconds were oscillating there too! So basically, it looks to me that the time time was being kept, in 2021, even the correct date and only 1 hr out, but it was frozen there, unable to move....and then it started moving again after a short time of the Amiga having power on it.

I checked it again this morning after being powered down all night, and even switched off at the wall, and once again, both WB and Amiga Test Kit was fine. Time was correct and seconds oscillating...

Weird eh.

Ah yes, I've seen this with Amiga Test Kit too - it will actually read directly from the RTC chip, but it only does it for a while before it somehow locks up the RTC (or the RTC reading code), needing a reset. I meant to report it as a bug but never got around to it.

Of course, there's always the possibility that the RTC genuinely did stop for a while and start up again, but I feel that's probably pretty unlikely.

The RTC chip has a stop bit. The bit can be set by software and causes the RTC to stop moving. If the RTC does not move it might not be recognised by the OS (depending on OS version). If you clear the bit, the RTC starts moving again.

Many years ago, when I didn't know the details of the clock chip yet, I wrote this brute force program: http://thomas-rapp.homepage.t-online...startClock.lha
It just overwrites the entire address range of the clock with zeroes. By that it also clears the stop bit.

This all might be the case, but tonight, i was having repeated stops and was struggling to get it to start again each time. Eventually i took the shield off the expansion card again, and so far, that has fixed it. I'm pretty sure nothing was shorting out on the shield casing. I might just leave it off now...

I tried this last night. The test (at least the non dangerous test) comes back "Ok" for all tests of the CIA.

Ok, so the operating system should be happy then. Now you can do the RTC test with Amiga Test Kit or by the methods suggested by others.

Yeah i've gone into both F7 and F8 options within Amiga Test kit multiple times.
They all look healthy and check out .....even when the seconds are showing as frozen within Amiga Test kit!?

Typically, if the seconds are frozen in Amiga Test Kit, then trying WB shows "battery backed clock not present" (presumably because it needs to see increments from the clock seconds to say its ok and display the time on the chip!?) and if Amiga test kit shows the seconds rolling, then WN displays the current date and time fine.

I've still no idea what was causing it to randomly stop. I'll keep monitoring it and see if it does it again now with the expansion memory shielding removed.

The seconds are frozen if you are in set mode, but they should be running when you are not.

Sounds like you need to do some rework near the RTC chip, in case there are bad solder joints.

Exactly.

Oh god, i'm hoping not. I was so careful with that board, and put a lot of attention and time into it, testing it fully with my meter and high intensity magnifying headset and then caked a load of solder mask all over the exposed areas. I'd probably have to remove the solder mask now if i wanted to do anything else with it. I'll go over it again with my meter while it's cold, and see what that brings up.

Try adding jump wires for the crystal? :-)

....and connect it to the car battery you mean? :laughing

I think i'd perhaps get better results if i used a defibrillator! :spin:guru

Hehheh.. Context is king, eh.

In this context just additional wires in case the traces between the RTC and xtal are broken.

Ah right! You were being serious! :) I thought you were kidding around. :laughing

So i'm resurrecting this old chestnut. I've been having my clock stopping and starting on me, seemingly randomly. Except it wasn't random. I suspected it anyway, but have confirmed that when the Amiga was warm, the clock would work fine. When the Amiga was cold, like sitting over night in a cold room, the clock would stop. Except, as I previously tried to explain, although the clock was not moving, it was retaining the clock time in which it actually stopped at.

I suspected the quartz oscillator crystal to be at fault, pehaps through age. So last week I ordered a couple replacement 32768Hz oscillators from Ebay, and they arrived this week. I've replaced the original crystal with one of the new Ebay ones, and hey presto, not only is my clock working again, it's working consistently now without stopping even when cold.

Maybe this thread will help another in the future...?