Amiga 2000 clock problem

matt020

Finger lickin' good
AmiBayer
Blogger
Joined
Dec 20, 2009
Posts
689
Country
Australia
Region
Western Australia
I have an Amiga 2000 with rev 6.2 motherboard. The problem is that the clock resets to default date/time when powered off. It has a new coincell battery, and the clock chip is getting 3v power, I have checked that, but something strange happens. This is what I know so far:

I have reset the clock with CLI command "setclock reset", i then set the date and enter command "setclock save". The time is written to the clock and I can verify this by using command "setclock load" and checking the date/time, and its perfect.

A CTRL AMIGA AMIGA reset, and it holds the date/time, no problem there.

It is when I turn the Amiga off, that the clock is reset to default date/time of 01-Jan-1978 00:00

But the weird thing is, it then keeps time from that point on. So if I switch the Amiga off, then switch back on 7 hours later, the clock has the time as 01-Jan-1978 07:00. If I power off and switch back in 2 days later, the date will be 03-Jan-78.

It seems to me, that when switching off the Amiga, there is a momentary loss of power to the clock chip, it resets, then gets power from the coincell battery and keeps time from 01-Jan-78 00:00 onwards...

Where do I start diagnosing this problem?
 
Hello,

first of all, you need to go into the most lower level, and check if the crystal is working and if there is connectivity between crystal and RTC IC, as schematics say... or maybe someone else has a better solution.

Cheers
Stefano
 
Thank you Stefano, and you are 100% right. I do need to go to the most lower level. I have downloaded the schematics and I have spent time studying them. I have zero experience with schematics so this is fun and horrible at the same time :LOL:

I have ordered an Epson 72421 RTC chip which is a direct replacement, and it has it's own oscillator so I can disregard the onboard oscillator all together. I'll post back here with results.
 
Update:

Replacement clock chip fixed the problem. Although I suspect it was the oscillator that was the issue, and here's why.

The problem initially was that the clock would not keep date/time when switching off and on AND when rebooting. I adjusted the trim cap with a small screw driver a few times, and then it was able to hold date and time on reboot, but still not power off/on. It was at this point that I posted here.

A suggestion was made to replace the clock chip with something modern, that has its own internal oscillator and that was the Epson 72421 chip. I removed the clock chip from the 2000's motherboard, installed an 18 pin DIP socket so I could easily replace clock chips if I ever have to again, and then installed the Epson chip. HOWEVER, I bent up pins 16 and 17 on the chip as they are not required since the chip has its own oscillator.

I did not replace the on board oscillator (I don't have to, the new Epson chip is not using the on board oscillator but it's own internal)

It worked, first time. And continues to work.
 
Hi...Simply check the progression between the A3 leg of the chip and A3 somewhere else, for example the CPU. On the off chance that you have a double follow scope, shockingly better - check A3 at hte CPU and the RTC chip simultaneously and ensure the follows are indistinguishable.

pcb manufacturer
 
Last edited:
Back
Top Bottom