Commodore PC-10 III: RAM bank 2 is not being addressed (Faraday FE2010A chipset)

[1ras]

Hi,

Thank you so much for this excellent tool. It was very helpful in repairing a Commodore PC-10 III.

But unfortunately, it cannot control the RAS signal of memory bank 2 (512-640K) of the Faraday FE2010A chipset:

RS 2 of the Faraday FE2010A is providing the RAS signal for memory bank 2. I measured with an oscilloscope and xtramtest does not control it, RS 2 does not change.
With the Commodore BIOS, RS 2 works as intended. Neither the BIOS nor CheckIt finds any RAM errors.

The Faraday FE2010A was a popular single-chip chipset for XT machines.

Faraday FE2010A chipset details: https://www.dosdays.co.uk/topics/faraday_chipset.php
Faraday FE2010A datasheet: https://github.com/monotech/NuXT/blob/master/Faraday-XT_Controller-FE2010.pdf
Commodore PC-10 III schematics: https://www.zimmers.net/anonftp/pub/cbm-pc/schematics/PC10-III_PC20-III_312626C_Schematics.pdf

Thanks

[misterblack1]

Could you give XTRAMTEST a try from DOS? (LOADXTRT.COM and the 8K ROM)

https://github.com/ki3v/xtramtest/releases/tag/v0.2.0

I asked @640k (author of GlaBIOS) for guidance, as yes this ROM would only really work on 100% compatible systems, so it seems that Faraday chipset is doing something funny and without that specific initialization then it strangely is missing that top chunk of RAM.

[ki3v]

Thank you for the report. It sounds like there's something missing at init time. The actual test algorithm should work fine, but it sounds like there's some setup code for that chipset that we need to include. I'll look into it, maybe take a look to see what GlaBIOS does about it.

Nice diagnostic by the way, checking RAS directly.

[1ras]

Good idea to let the Commodore BIOS initialise the FE2010A. I should have thought of trying LOADXTRT.COM myself, now that the PC is working again and can boot into DOS. With LOADXTRT.COM the memory test works: