View Full Version : CS MKII bug

10th September 2012, 10:13

I have a CSMKII here with the DMA issue. According to Amiga Ressources, some parts have to be changed. Do we know what are the parts affected or only phase 5 knows that ?

Thanks for your help !

10th September 2012, 11:45
There's a high res photo of an apparently bug-free CS MK II on amiga.resource, maybe you can attempt a visual comparison. I tried to find some info when pondering getting a CS MK II some time ago but got nothing, but I'd guess it has to do with the board's interface logic so reworking would be from hard to near impossible (save maybe using donor parts).
There's also the off-chance that S/N stickers have been swapped on occasion - have you noticed the actual problems?

10th September 2012, 11:58
Thanks for your answer BLTCON0.

Yes I thought the same concerning the differences and, if it concerns the programmed chips, it will be impossible to fix it as only phase 5 owns the files to program them.

The symptoms : every single expansion board or scsi controlers that uses DMA access is not recognized or is corrupt : In my A3k, the internal SCSI controller does not work when fast ram is installed (the HDD is not recogniezd even in the early), in my A4K the Deneb has to be set in PIO mode. DMA mode results in corrupts files and files on usb sticks are not shown.

Well it is not a big problem as many solutions exist now to bypass it : Buddha, DENEB in PIO mode, Fastata, using Zorram and do not put ram on the accelerator (tested and very fast)...

But it could be interesting to find what the differences are and to fix the board.

Do you think it is possible to reach someone who worked for phase 5 or it is useless ?

10th September 2012, 16:22
I have no idea if someone ex-Phase 5 could reprogram the chips (if that's the culprit), but considering the rights have been bought by DCE it might even be formally illegal for such a person to do so if said rights include the CS MK II.

It's good indeed there are acceptable compromises - I guess the caches on the 060 make slower access to ZIII RAM less of an issue.
Have you tried using the DMA-less Buster 7? (although it may still allow ZII DMA so no real solution in that case). Deneb would be plenty even in PIO mode anyway.

10th September 2012, 17:12
Have you tried using the DMA-less Buster 7? (although it may still allow ZII DMA so no real solution in that case). Deneb would be plenty even in PIO mode anyway.

No chance (at least, not of getting the A3000's onboard SCSI back)

I also have to have my deneb set to Z3 PIO

10th September 2012, 19:45
Thanks guys.

So I emailed DCE but I am pretty sure that I will receive no answer...

I will stick with the CSMKII in my A4k and will plug the CSMKIII that was inside in the A3k, this way it will solve the problem.

11th September 2012, 21:28
There is a guy named Carsten Schlote that used to work at Phase 5, in fact he did an unofficial Cyberstorm MKI/MKII firmware replacement.

No personal web available nowadays but thanks to the archive you can find it here:


I've no idea if he can help you and I don't know this guy, but maybe it deserves a try...

11th September 2012, 21:31
You can contact Thomas Knabel too, he's currently working at bPlan (you know Pegasos motherboards etc); he might help.

11th September 2012, 23:21
Many thanks guys! Will try ASAP ;)

27th September 2012, 08:16
Ok so I had some answers from another source :

To fix the board it seems that a solder shunt is needed (I have the exact position), a trace has to be cut (my source cannot remember where but it is near the solder shunt) and unfortunateley, the XILINX chip has to be reprogrammed... In other words it is almost impossible for us today to fix them :picard :(

No answer from DCE as expected...