This board uses an "acer" BIOS, which if I had to guess might be based on Phoenix BIOS. I don't think there are alternate BIOSes available for me to play with.

However, I also have two SiS 406 based VL/EISA 486 that don't like Tsengs either. I was able to change these from AMI to Award, but the behaviour was the same. I suspect chipset incompatibility, or there is just something wrong with my two cards.

Bad news. I finally physically got ahold of my Powerleap PL 54C/MMX. It does not appear to function properly on the J3 motherboard with any CPU I tried. With a P166 classic, it posts but crashes before boot. With a P200MMX, it functions but L2 cache is automatically disabled by the BIOS. My Tillamook is the same, but for some reason runs at 133MHz when set to 4X mode. K6-2-500 and K6-2+ 500 prevent the system from posting. This module does support the correct voltages and multipliers to run these CPUs. I will check the capacitors, but I suspect it's just a good old fashioned incompatibility. Quite frankly, I'm not really surprised.

I ordered a second POD5V133 to try SMP mode in NT4. Will probably be a few months before I get it. Both my POD5V133s have broken heatsinks and no fans, so I will need to rig something up.

Did you try the classic Pentium at 133 MHz? What about a regular K6? Cyrix 6x86MX or MII? Winchip or Rise? If a classic Pentium isn't functional, it is unlikely these other chips will function, but it would make for a more complete test.

For the K6-2+ test, did you try physically removing your L2 motherboard cache?

I think testing the classic Pentium at 133MHz is a good idea. I didn't think of that. Infact, come to think of it I think I accidentally the multiplier set at 3X too.

I do not have an original K6, any Cyrix socket7 chips, winchip or rise. I don't really have any interest in running them on this build either.

I didn't try physically removing the cache when I tested K6+/K6-2 because the system didn't even post. Not a good sign.

Correction with the P166 classic. I made a booboo. It was being undervolted and the multiplier was set wrong. At 2.5X and 3.5V it works the same as the Pentium MMX chips....no L2 cache. I tried 133MHz too, no difference. I have a feeling that the BIOS is detecting the CPU type and automatically disabling the cache when it doesn't find what it wants. The Socket 4 CPUs are all Family 5 Model 1, but the socket 7 chips have different model numbers in their CPUIDs. I have a feeling that's throwing something off.

About the K6-2 and K6+ chips. I could remove the L2 cache from the board, but that's one hell of a lot of ICs to take out. Since the system doesn't even post I think it's unlikely to help.

*edit* I removed the cache chips. It had no effect. I tried all the multiplier options too. No dice. I wonder if the guy that adds k6+ support would like to have a shot at my decrepit pentium bios.

I tried tillamook again. I was able to get it up to 233MHz, but the 4X (and higher) multiplier isn't working. It's like pin BF2 isn't connected...but it is.

If you got AWARD Bios you can Patch CPU Support yourself.

With BP Bios Patcher.

Sadly no. It seems to be an Acer BIOS. Yes, I checked the BIOS in a hex editor, and it does indeed appear to be an Acer homebrew, not something licensed from another company like AMI, Award or Phoenix.

Damn, so no way to add microcode support 😒

I asked Jan Steunebrink to look into the BIOS issue for me some time ago. I did not hear back from him, so I can only assume there wasn't much that could be done. Anyway, I went ahead and bought a second Intel Overdrive 133:

Unfortunately both of my PODP5V133 chips have broken heatsinks (it was the only way to get them cheaply), so I need to remove those and epoxy on some new ones so that I can mount 5cm fans.

I will probably also bump the cache from 512kb to 1mb once I get the heatsink issue taken care of.

This system is currently not cased, but after my big move in June I will consider doing so. Probably another ATX conversion, since AT stuff is too hard to come by.

I see what looks like a solder pad for connecting power directly from the PSU's cabling. Under what circumstance do you think this is needed?

There is also a jumper for an 80MHz FSB. My guess is that Acer anticipated that Intel would release faster 5V Pentiums, thus requiring more power. Thankfully the POD5V133s (12W) use less power than the P60s (15W).

That is an awesome board! I've never saw one of those back in the days. I like your setup! 😎

dear lord, how did I miss this thread...

I've somehow missed this thread too. Good to hear, that this board works, though I'm a bit disappointed, that it's performance is crap - I like all these high-end PC chipsets which gave you a slight edge over the competition.

ramiro77, not too surprising you've never seen one of these before. Back in late 1993 only the King of Town would have had something like this.

It shouldn't be a big surprise that the memory performance is crap, as it was well documented at the time. This particular consumer oriented motherboard was from the second wave of Pentiums. The first wave in the first half of 1993 was expensive server type stuff. Old copies of PC Magazine have wealth of information available on this topic.

I'm just happy that this board works, and that it survived the trip from Poland To China. The customs nazis almost kept it.

Sure. I remember 1995 in my country. Socket 7 Pentiums were really expensive. I can't imagine how expensive was something like your rig. It looks really awesome.

This is a P60/66 without a built-in local APIC. I wonder if this is an Intel Multiprocessor Specification compliant system with external APIC or something else (it is not in the NT4 HAL list).

yuhong wrote:

This is a P60/66 without a built-in local APIC. I wonder if this is an Intel Multiprocessor Specification compliant system with external APIC or something else (it is not in the NT4 HAL list).

Shrugs shoulders, wouldn't be surprised that it is not compliant seeing that it is so rare and likely very early. Pity the performance isn't so great but gems like that are great learning tools.

It should be supported by NT 3.1 and 3.5 at the very least (does this imply 3.51 is okay as well?).

I could not find the NT4 HCL. Where did you find it?

Yea, no reason why NT 3.51 and NT 4.0 would not support it too AFAIK. Do you know if it uses the MPS HAL or something else?