dicky96 wrote on 2020-01-05, 12:56:

However I notice now that the CPU will fit in the socket any way round. It is so long since I was building 486 computers I had forgotten they do that. I have inserted it the way I think it should go back in. See pics. Is that right?

The dot on the CPU matches with the hole separated from other holes, so that's right:

Is that a decent motherboard / CPU combo?

Define "decent".
It's a typical late 486 platform, I think from 1996, when Pentium was already mainstream, and 486 was low-end.
So, nothing to be excited about.
But it can be very useful for DOS games - it's fast enough for Doom 1, Doom 2, and thereabouts, but it can be also made slow enough for 286 and 386 games - just set it to de-turbo, and disable caches.
Everybody interested in PC games from about 1985..95 should have such a machine.

Thanks Gryzb, it's good to know I have the CPU the right way round

The BIOS sticker says 1995 so you are spot on.

I already have a similar working 486DLC 40MHz machine I picked up out of a trash skip (dumpster) last year for DOS games, plus a SS7 with K6 CPU and ISA AWE32 that can also play DOS games so I will probably sell this one once I get it working, but only to buy something nice like a Voodoo 3 3500 for my SS7 Rig which I have been wanting for some time. I had a look around on ebay, so I have an idea of what this is worth now. Seems like the two items could have a similar value.

best regards

OK so it's not working but I think it can be resurrected

The AT PSU (I used an old IDE HDD as a load instead of the motherboard seeing as I didn't know it's condition) powers up but makes a hissing noise and generates low voltages (about 0.7V on 5V and 1.8V on 12V)
I cleaned all the dust out, the fan has seized but all the electrolytic capacitors look good to the eye. That's not to say that they are, I'll have to desolder them and check on my ESR meter and capacitance meter. I'll probably just change them. It could be an optocoupler/feedback type fault also but due to the age I suspect capacitors.

The CPU cooler fan is also seized. The CMOS battery was very low 0.8V so I replaced it.

Using a known good AT PSU (from my working 486DLC rig) The motherboard makes a continuous beep-beep-beep-beep immediately on power up. There is only one EDO sim fitted. it makes no difference which slot I put it in or if I remove it altogether I get the beep-beep-beep just the same. I have +12V, -12V, +5V, -5V present plus an orange lead from the AT PSU with 4.7V on it. Googling says EDO RAM runs on 5V and I can't see a VRM anywhere so I guess I can't have a Vram fault on this motherboard. There is a TIP127 PNP transistor near to the SIS 85C497 chip which appears to be generating 3.4V on the collector so I guess that is acting as a series regulator. Anyway whatever it is for, 3.4V looks good to me.

Possilbly I have a duff EDO RAM but I don't have any more here to test with. I have 4 known good ones back in the UK and I can get my wife to pick them up for me when she goes over for a few days at the end of Jan.

According to my POST diagnostic card I have code C6 which the manual says is 'External Cache Size Detection'.

So is it possible I actually have faulty cache and not the EDO RAM? There are some jumpers on the motherboard but I couldn't find the right manual for this motherboard when I googled it. Does anyone have the manual for this motherboard (SIS-496 PCI rev 1.0) or know if the external cache can be disabled using any of the jumpers, to eliminate that possibility?

Other than possibly eliminating the cache I guess I will have to put this on one side until the end of the month and try some known good simms. I can take a look at the PSU in the meantime

Rich

Have you tried FPM RAM too?
And I'm wondering where the RTC/CMOS chip is located...

With a system of that old, it's pretty certain that it only takes fast-page RAM (FPM). AFAIK EDO RAM was more of a Pentium thing

Ahh OK I just thought it looked like a stick of EDO RAM (looks like same size at a glance) The one I have is the original one that came with the motherboard so I guess it is the right sort.

You are probably right it is not EDO RAM. In that case I won't have any known good ram to test the board

I'll look for the CMOS chip tomorrow and post a pic of the SIMM

Although 486 chipsets generally don't support EDO, SIS 496 actually does. Not just accepts modules, but actually supports EDO timings with slightly reduced burst latencies.

My SIS 496 board's BIOS (ASUS PVI-486SP3) boots into recovery boot block when populated with EDO DRAM.

Could be a motherboard board limitation or perhaps chipset revision? But the chipset does have EDO support as per datasheet.

EDIT: According to this thread EDO is supported from revision B5:
ASUS PVI-486SP3 problems with L2 (external) cache

Yes it is NV (SIS 496 B4). So my PVI-486SP3 does not support EDO.

Mine is SIS496 NV Also

I am assuming the SIMM that came with it is the correct type. I have no idea as I got it all 'as is'.

Not just accepts modules, but actually supports EDO timings with slightly reduced burst latencies.

AFAIK only UMC can do that, both on VLB an PCI. SIS can only support EDO in FPM mode, and that's it.

You might want to try re-seating the cache chips. Oxidation can make stuff act funny. Maybe re-seat the BIOS chip and the CPU as well.

Oh yeah.. clean the dust off of the board as well.

The Serpent Rider wrote on 2020-01-07, 18:24:

Not just accepts modules, but actually supports EDO timings with slightly reduced burst latencies.

AFAIK only UMC can do that, both on VLB an PCI. SIS can only support EDO in FPM mode, and that's it.

In the datasheet they provide bus cycles for both page mode and EDO and there is a difference . But again this could be revision specific.

OK
The RTC is built into the SIS85C497 chip

I figured that out by looking with a multimeter to see where the 3V from the CMOS battery went to. I then checked the datasheet for that chip and it does contain the RTC. I assume the CMOS is also in that chip

I've cleaned the board. I tried reseating the cache RAM, Tag RAM and BIOS - no difference. The DIL sockets and pins looked clean anyway. To be honest Gran Canaria is renowned for it's climate (best in the world apparently) and in the south we get average 88mm of rain a year. So nothing corrodes much here.

As an aside why did the manufacturers of these boards go to the extra expense of putting the cache ram on sockets. Did they expect them to fail regularly, or was the cache size meant to be upgradeable?

I tried reseating the CPU. I had already done that anyway as I had removed it to replace the old heatsink compound. To be sure I also tried the CPU in the socket without the little plastic bracket and heatsink, as it sits further in the socket that way. No difference.

There are no identifying marking on the SIMM. The chips are Panasonic MN414400BSJ-7. I looked up the datasheet for that chip and it is Fast Page RAM

I tried the SIMM in all four sockets (which the SIS496/497 chipset datasheet says it supports. No difference. I tried putting a little pressure on the SIMM when it is in the socket. No difference.

I can occasionally get the board not to beep when the SIMM is fitted but it still stops on the same POST code C6 even if it does not beep.

I guess the next thing now is I need to buy some known good FPM to see if I simply have a faulty SIMM here. Unless anyone can think of anything other to try first.

Looking on ebay there are single sided and double sided FPM SIMMS and some with 9 chips (parity bit?) while others have 8 chips. The double sided seem to be a bit cheaper than the single sided. And all the ones I can see listed are in Germany for some reason! The one I have how has 8 chips and is single sided, though I don't actually know it is the correct type.. Should I get the same eg.

https://www.ebay.co.uk/itm/8MB-Ps-2-Fpm-Simm- … 1MAAOSw5wFcSJUB

best regards

On the plus side I fixed the AT PSU. It was capacitors as I guessed. It runs sweet now and no longer hisses like a snake. It just needs a fan but that's OK I have some good spare used ones lying around.

Should I be buying any particular type of FPM to test this motherboard? With Parity, without parity? Single or double sided....

In the datasheet they provide bus cycles for both page mode and EDO and there is a difference

From my experience with SIS496 PR, the performance between FPM and EDO sticks is literally identical. So even if there are some EDO specific tweaks, most vendors probably didn't bothered to implement them in BIOS at all. Practically, PR revision provides only some minor convenience.

Should I be buying any particular type of FPM to test this motherboard? With Parity, without parity? Single or double sided....

Any non-parity FPM in 60-70ns range.

HI guys
I finally got round to looking a this again

I bought two Mb sticks of FPM RAM from ebay £8.83 inc postage. These ones
https://www.ebay.co.uk/itm/Pair-of-Smart-sm53 … 872.m2749.l2649

I inserted those and it boots. as the simm that came with the board was also 4Mb FPM I can only assume it was faulty.

However I have another problem with it now. If I put the Cirrus SVGA card in the PCI slot nearest the ISA slots it will boot up to POST screen

If I put the SVGA card in either of the other PCI slots it bleeps......... beee-bip-bip-bip which Ithink is saying no graphics card present. All the PCI slots look clean. is this a fault or a 'feature' of these old motherboards?

Cheers