Welcome, Guest. Please login or register.

Author Topic: A4000T CSMII hassle  (Read 4940 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline Castellen

Re: A4000T CSMII hassle
« on: June 02, 2003, 03:07:47 AM »
There is nothing on the A4000T's audio/video module which it needs to boot, the system will work without it.
I have a CSmkII here in my A4000T as well and it has never given any problems.

I repaired an A4000T a few months back which the 22µF capacitors onthe audio/video module had leaked electrolyte onto the PCB, that's the brown stuff you see.
I'd carefully inspect all the SMD electrolytic capacitors on the motherboard for signs of leakage.

Things to watch for is the silver legs of the capacitors appear black or tarnished instead of smooth and silver.
A distinct fishy smell, especially when you solder it.
Discolouration of the PCB and surrounding components.

The leaking capacitors can semi-bridge PCB tracks, which can introduce crosstalk, meaning you can get data corruption.  Perhaps this is what you're seeing with the A3640 in place?

I'd highly recommend replacing all the 22µF caps on the motherboard anyway.

Also you really need to test the CSmkII in another machine to make sure it's working correctly.  You've got no easy way of knowing otherwise.

If you need schematics or more info, feel free to give me a yell.

Anthony.
 

Offline Castellen

Re: A4000T CSMII hassle
« Reply #1 on: June 03, 2003, 12:16:33 PM »
Condenser is the "old fashioned" name for capacitor, same thing :)

As for the clock source jumpers, the CSmkII requires these to be in the "internal" position, this is where the clock signal generated by the motherboard's crystal oscillator is used by the rest of the motherboard for timing purposes.  Perhaps the CSmkII does not generate this itself for some reason, most of the other CPU boards with built in clock require the motherboard jumpers to be set to external clock.

It could be that there is something wrong with either the motherboard oscillator, frequency dividers, or delay line (U102).
If you have a counter, I think you should have 25MHz on both clock jumpers, from memory.
A lot of those oscillators were fitted into machine sockets, which can become loose and cause an intermittent connection on the oscillator module.  Would be worth checking.

If you have an EC030 card (A3630?), you can try that, as they require motherboard clock.

Don't forget the CSmkII uses a few additional lines to the motherboard that the A3640 doesn't, so it can be possible where one card can work where the other one won't.
However, unless you can check the CSmkII on a known working machine, you're only speculating.

Perhaps you can send it to a friend somewhere else to test?  Would only cost freight.
 

Offline Castellen

Re: A4000T CSMII hassle
« Reply #2 on: June 07, 2003, 12:01:14 AM »
Did you know that Vesalia Computer in Germany have A4000T modules for £5 each?

But anyway, the fact that the power LED is blinking, shows that the CPU is actually doing something.  There are custom logic devices on the Cyberstorm board, so these may have failed?  Possibly died due to the extra demands put on them from overclocking.  Remember that clock speeds are what the devices are rated for.  By exceeding them, you're using the devices outside their specifications.

See how the A3640 board performs in your A4000T in the meantime.  Don't forget to remove/rename the Cyberstorm libraries and use the correct 68040.library or you'll get werid problems.

Sending the Cyberstorm to France (or Hyperion?) will probably be the next thing to do.