A while back I posted this guide with details on how the jumpers work for setting L1 WB, clock multiplier, voltage detection and other things an ASUS motherboard with the SiS471 chipset. My main references were the SiS471 datasheet (which I assume you already have) and a really comprehensive page by Jan Steunebrink.

With that info I poked around the board with my multimeter to determine how the jumpers were connected to the CPU and chipset. I suppose you could take a similar approach, using the manual as a hint of what each jumper block does.

TheMobRules wrote on 2020-12-08, 22:27:

A while back I posted this guide with details on how the jumpers work for setting L1 WB, clock multiplier, voltage detection and other things an ASUS motherboard with the SiS471 chipset. My main references were the SiS471 datasheet (which I assume you already have) and a really comprehensive page by Jan Steunebrink.

With that info I poked around the board with my multimeter to determine how the jumpers were connected to the CPU and chipset. I suppose you could take a similar approach, using the manual as a hint of what each jumper block does.

Thank you for your answer. Jan's website was especially informative.

I've followed accurately both your thread and Jan's website and by connecting everything as in your thread and as Jan's says it just stops booting MS-DOS. I had to hard-wire WB/WT# CPU to 3.3V, since I couldn't find anything else to connect it to.

Also pretty much no config on the manual uses what I've found, so they must've either been very drunk at the time or I am doing something wrong. The only config which kinda looks like my findings is the Cx486DX2-V80.

I've also tried retracing my steps in finding how I did get WB working a few times, but I can't seem to be able to reproduce it anymore.

I am at my wits end - I suspect I will have to raise the white flag soon and have to deal with having the L1 cache in WT mode.

Yeah, I gave up on trying to enable L1 WB for a DX4 CPU on my Gigabyte GA-486VS, despite confirming all the required connections. The L1 WB works fine for DX2 CPU with write-back support, but with the DX4 setting the WB/WT# jumper results in a black screen. So it seems that CPU requires something else to work that is not there, maybe at BIOS level.

TheMobRules wrote on 2020-12-11, 01:56:

Yeah, I gave up on trying to enable L1 WB for a DX4 CPU on my Gigabyte GA-486VS, despite confirming all the required connections. The L1 WB works fine for DX2 CPU with write-back support, but with the DX4 setting the WB/WT# jumper results in a black screen. So it seems that CPU requires something else to work that is not there, maybe at BIOS level.

The BIOS supports it - of that I am certain, as I did manage to get it to work for a while some days ago, and that is maddening to me. If I did get it to work, with this same CPU, why can't I get it to work anymore?

I was moving jumpers randomly, and something did it. Problem is, there are too many combinations to try, so I am trying to narrow it down a bit...

Anyway, I suspect it might be an issue with my specific motherboard. I fear some of the required connections might have stopped working (solder joints cracking?). Not even by using ctchip34 (which is amazing, by the way!) have I managed to get it working again in WB mode.

Also, since it hangs when setting every jumper correctly, I suspect that it's the right setting for WB, but it's just missing some signal.

I have since tested the X5 in another motherboard (Chaintech 486SPM) and it's working well in WB mode on that one. Since I have managed to reach a quite decent speed with L1 in WT, I guess I will just call it a day.

I might revisit this issue in the future (this time by using a soldering iron) if I were to remove the motherboard because of some other issue that is going to come up.

One detail from my other thread that may be of interest: at least on my ASUS board, setting the jumper that connects the CACHE# line of the CPU to pin 118 of the chipset causes it to hang when booting if using L1 WB. According to the chipset datasheet, pin 118 is supposed to work as CACHE# line when in WB mode, but for some reason I have to set the jumper so that pin 118 connects to a pull-up resistor instead. Jan's page also mentions this connection being optional for L1 WB, so you may try to play around with that connection if you haven't yet.

TheMobRules wrote on 2020-12-13, 00:24:

One detail from my other thread that may be of interest: at least on my ASUS board, setting the jumper that connects the CACHE# line of the CPU to pin 118 of the chipset causes it to hang when booting if using L1 WB. According to the chipset datasheet, pin 118 is supposed to work as CACHE# line when in WB mode, but for some reason I have to set the jumper so that pin 118 connects to a pull-up resistor instead. Jan's page also mentions this connection being optional for L1 WB, so you may try to play around with that connection if you haven't yet.

Yeah, it looks like JP31 Pin3 is pull-up resistor, still hangs. It's okay - win some, lose some. I still can run the board on pretty much everything else (timings, clocks) maxed out. Seems stable enough.

I'm very interested by this topic 😀

I have sold my motherboard to someone, but maybe you can help us improve this page with these infos ? http://www.win3x.org/uh19/motherboard/show/2600

Deksor wrote on 2020-12-13, 18:19:

I'm very interested by this topic 😀

I have sold my motherboard to someone, but maybe you can help us improve this page with these infos ? http://www.win3x.org/uh19/motherboard/show/2600

I don't think I will be able to provide much more than what is provided in the initial post - they are also unconfirmed findings which I just made by probing randomly at stuff at night. I'm not sure it qualifies as reliable enough info to be added to UH19 🙁