CSPPC not recognising on-board memory

[Boot_WB]

The title says it.

I have a CSPPC (200/060) on which the memory is not being recognised.  At first I thought it could be a motherboard fault, but after getting hold of a second a4k motherboard I still have the same problem.

I have tried 4 x 32MB simms (tested working FPM modules) and 4 x 4MB simms (tested working FPM modules) with nothing being recognised by the board.

I have replaced both the motherboard fast ram (and tried removing it comlpetely) and the chip-ram.  I have checked my jumper settings a thousand times.  I have read the CSPPC manual midway to the back cover (I didn't bother reading the first half, which is in German).

As a consequence I am unable to test the PPC side of things as it requires min 16MB on board the CSPPC to run the PPC (afaik).

The 68060 is recognised in showconfig, although Whichammiga crashes with lots of brightly coloured flashes across the screen - this crash sequence seems identical every time (I think this is probably as a result of trying to initialise ppc.library, although I'll run snoopdos to find out what is happening there). If I uninstall the 68k and PPC libraries Whichamiga just hangs.

I've tried 3 different psu's - one of which is a 550W new ATX unit.  Two of them measure 4.65-4.7 volts on the 5v line, the 550W atx measures 4.8V on the 5V line.
OK - I could understand the ppc not working, or crashing with low voltage, but to just not register the memeory each time? I think this is very strange.  

Also the 550 W is a brand new ATx unit - I've measured it at 4.99V under load on a different system.  Also, one of the other psus runs my BPPC for hours on end with no complaints/crashes whatsoever.

What the hell am I supposed to do - keep buying ATX psus until I randomly find one that outputs 5V as it should?  Do I need to pay £50-ish for a psu with dashpot-adjustable voltages - all that on the off chance that it is just a bizzare (yet stable, predictable) side-effect of the low voltage?

The CSPPC is a phase 5 model, with only 2 oscillators installed.  It looks to be in mint condition - not even  scratch, or a speck of dust.

One thought - is the memory dependent on one or the other oscillators, or is it just run by the clock of the cpu which is accessing it.  What I mean is - if the PPC oscillator was not working would the 68k processor still be able to access the memory - or does the ppc oscillator also drive the memory?

Any help would be most appreciated - at this oint I am seriously considering sending it to Jean Jacques in France.

Anyway, I'm off to cook my tea.

Regards



Rich

[X-ray]

Has this board recognised memory before?

[Erol]

Have you checked the memory settings on bios on the card?

hold down the ESC key and enter a configuration menu for the card.

[Argus]

I certainly hope it isn't this, but if those memory controller chips located between the first and second simm slot are fried, it might be curtains for that board.  Afaik, even JJ can't replace them.  Did you buy the board or was it working until just recently?  If it's a recent purchase I'd ask/demand my money back.

[Boot_WB]

I bought the board recently, and have not had it recognising memory thus far.  I'm also hoping it is not the memory chips - I would have thought that would cause the board not to boot at all though - presumably there is SOME sort of self-test run by the board before booting.
There is no physical damage at all to the board - I've inspected the board (especially the memory chips) with a high-powered magnifying glass and can see no sign of any damage/shorting.

I'm totally at a loss.

Ideally if there was someone in the Sheffield area with a CSPPC setup who would try the board on their syste, at least that would show that it did/did not work in a known working setup.

Thanks for the replies.

[Erol]

I know that holding the M key down disables onboard memory, but that should reset on powercycle..  worth mentioning though on here.

[Boot_WB]

Hi Erol (are you Erol in Israel btw?), I thought that too, but I've tried with the keyboard unplugged, and I've reflashed the ROM just in case something was 'stuck'.
btw - I've got the memory set to the most basic config - 70ns.  Have also tried with 60ns and with freeconfig.  No change.

[Erol]

im not mate, im in london

[Karlos]

My BPPC does this every time the flash rom goes mental. Nothing, other than the basic 68040, is recognised until it is reflashed.

[Boot_WB]

@Karlos
Hmmm.. may try some different versions of the flashrom then, see if that helps.  There may be hope yet.
When I do I'll output Snoopdos to a log file and see exactly what happens when WhichAmiga crashes.

[Argus]

Any luck with the reflash?

[Boot_WB]

The earliest I can find through Google in the 23-03-99 flash update from Aminet. No luck with this.

I've found that removing the ppc.library allows whichamiga to return, showing the 68060, fpu, mmu, and only the 8mb fastram on the motherboard.  

Having done a thorough search a lot of the memory I have which I thought was FPM is actually EDO.  
I've found two 4mb simms which are confirmed FPM (checking using the chip part numbers), however they are still not recognised (inserted in slots 1 and 3 - ie bank '0').  They are not matching part numbers, so I'll carry on searching see if I can find an identical pair.

I would like to try some earlier flash updates, so if anyone knows where these are archived that would be a great help.

I've just e-mailed Jean-Jacques to ask his advice also.

On one side - at least the WhichAmiga mystery is solved, on the other hand it is still not recognising memory.