So, the BIOS POST doesn't show that "table of information" - like a rectangular box, with info about the system's resources, such as serial and com ports, etc etc? I noticed that one of my 486 boards hangs after displaying that table of info, but it only does this if I use an ISA IO controller. If your mobo works with another 5V CPU, such as any old SX or DX, then I'm not sure what the problem is. But if you can't get the mobo to work at all, you could try -

Changing IO card, and also the ISA slot that it goes in to.
Similarly, changing the VGA card, and also the ISA slot that it goes in to.

That's correct, it doesn't get to the info table. It doesn't go past the first POST screen with memory test, but I can go into the bios and fiddle with as many options as I like.

The board works perfectly with a DX2-66, looks like a forgot to mention that, oops 😒

This board has 2 VLB slots and I am using an ET4000 VLB video card and a VLB IDE/IO card. I've also got a Gravis Ultrasound and an SB16 CT2800 in there.
It has 8x 30 pin SIMM slots, currently using 4x 4MB SIMMs.

Its the only working VLB board I have at the moment... I guess I can try pulling cards and maybe an ISA I/O controller card instead of VLB.

Have you tried disabling the L2 cache in BIOS?

No, I guess that's another thing to try. A bit disappointing if thats what it takes to make it work. If thats the case maybe it will work with only 128kb or 64kb cache instead.

I have the IH4077D and it is one quirky little board. I have been able to get it to work with a DX2/66, but nothing faster and not a P24D, which stason lists as supported. It has also been cranky with different cache combinations, but eventually worked with 256kb.

The IH4077D uses an EFAR chipset and not the UMC chipset that yours has, so it's not a straight comparison, but based on our combined experiences, I'm going to say these boards have issues. In fact, if I can get another setup going, the IH4077D is going back on the shelf.

Good luck.

Hmmm. I have managed to get the Pentium Overdrive and the DX4 Overdrive to work, by changing the jumpers on the board to 128kb.

I have swapped the 20ns cache chips with 15ns chips off a dead board but this doesn't seem to help... if the jumpers are set to 256kb and a DX4/POD is installed, it will hang at the end of the first POST screen.

Even with 128kb, and the new cache chips, the system is unstable at the fastest wait states/burst settings. However, a DX2-66 works perfectly at these settings.

I've been fiddling for a while but haven't quite found a stable combination.
I wonder if there is some limitation of this board or chipset, that if an overdrive is installed you are limited to 128kb or certain minimum RAM/cache wait states.

I've been searching for information regarding that overdrive dx4, and found that there are 2 types. dx4ODP100 and dx4ODPR100, for regular and overdrive sockets. However, here it says that both of them are 3.3v cpus, as opposed to the Pentium overdrive which works at 5v. I wonder if that is true, and if it may affect anything.

Well that's strange. The CPU I'm using is an SZ959, with integrated heatsink and VRM on the surface of the processor. I though the whole point of the overdrive DX4's was that they would work on a 5V board?

I suppose if it has a VRM it's ok. I've been reading some POD datasheets, and it said that it works at 3.3v, but it has also a built in VRM so that it can be placed on 5v sockets. Now, what I've been thinking, in what mode do those overdrive cpus work. I mean Write-Back or Right-Through. I believe it's the first. While reading that datasheet, it mentioned something about some boards that cannot handle write-back correctly, so Intel had another thing that you place between the cpu and the socket, so it would disable write-back mode. I don't know if that's your case though...

As far as I know, the SZ959 is write-through. My understanding is that write-back DX4's had &EW stamped on top of them; the DX4 Overdrive has no such markings. I'm not sure how you can tell, is there some util that can check this? There is no option in the BIOS for write thru/write back cache mode.

I've just had a bit of a brain wave. The socket in the motherboard is clearly marked 'Socket 3'. That implies it actually supports 3.3V CPU's doesn't it? There certainly isn't any jumper options to change voltage (only CPU selection - DX, SX and P23N/T) and I can't see a voltage regulator on the board. Do all Socket 3 boards support 3.3V?

DonutKing wrote:

As far as I know, the SZ959 is write-through. My understanding is that write-back DX4's had &EW stamped on top of them; the DX4 Overdrive has no such markings. I'm not sure how you can tell, is there some util that can check this? There is no option in the BIOS for write thru/write back cache mode.

I've just had a bit of a brain wave. The socket in the motherboard is clearly marked 'Socket 3'. That implies it actually supports 3.3V CPU's doesn't it? There certainly isn't any jumper options to change voltage (only CPU selection - DX, SX and P23N/T) and I can't see a voltage regulator on the board. Do all Socket 3 boards support 3.3V?

Most Intel DX4's were write through. The WB variant is more uncommon (I had to look for them specifically before I finally managed to get a hold of one).
There were 2 different overdrives made also. One was for plain Socket 3 (or older) and the other specifically to fit that overdrive socket Intel came up with.

Iirc the Overdrives ALL worked in 5V-only boards. It was supposed to "overdrive" your old non-upgradeable 5V-only motherboard.
Iirc the ceramic Intel DX4's were oem only and not for sale in retail channels.
And iirc none of Intel's retail DX4's supported WB, only WT.

Because of the price of those 5V DX4's many companies started offering those upgrade chips like Evergreen and Turbochip and such.

And not all Socket 3 board work with 3.3V parts. I got an Acer/AOpen VI15G(?) which supposedly only works with 5V parts.
I think the Socket 3 specification had more to do with the pin layout in order to support the Pentium Overdrives.
I once found a Socket 2 (the real while socket with the lever and all that) ISA board YEARS ago and still regret not having it brought home 🙁

Thanks for that, I pretty much suspected that was the case.

I looked on the internet and even a couple of books I've got here that all say that Socket 3 supports 3.3V and 5V but it definitely appears this is not always the case. Just goes to show- can't believe everything you read.

OK here's something interesting:

I took the plunge and put in an AMD DX4-100, which clearly had 3 VOLT written on the surface.
Stuck a heatsink/fan on it, didn't change any jumpers from the 486DX2 setting, and fired it up.

It booted first go, at maximum performance settings. Didn't need to change a thing and it worked fine.

Intrigued, I pulled out the AMD DX4 and put in an Intel DX4- a standard 3 volt one, not an Overdrive.
The same problems I had with the overdrives occurred.

Interestingly, the AMD chip was picked up by the BIOS as an 80486DX2 while the Intel chip was detected as a DX4-S. Both were picked up at 100MHz. I confirmed this with cachechk and sysinfo.

The AMD DX4 only has 8KB L1 cache while the Intel DX4 has 16KB. I presume this is part of the issue- perhaps my board has trouble with CPU's with 16kb of L1 cache?
It's very odd.

I'm not sure how long the AMD DX4 will last running at 5V but I'll see how it goes... with the heatsink/fan on there it doesn't get very hot at all (certainly much colder than a DX2 with no fan).

DonutKing wrote:

Intrigued, I pulled out the AMD DX4 and put in an Intel DX4- a standard 3 volt one, not an Overdrive. The same problems I had wi […]
Show full quote

Intrigued, I pulled out the AMD DX4 and put in an Intel DX4- a standard 3 volt one, not an Overdrive.
The same problems I had with the overdrives occurred.

Interestingly, the AMD chip was picked up by the BIOS as an 80486DX2 while the Intel chip was detected as a DX4-S. Both were picked up at 100MHz. I confirmed this with cachechk and sysinfo.

The AMD DX4 only has 8KB L1 cache while the Intel DX4 has 16KB. I presume this is part of the issue- perhaps my board has trouble with CPU's with 16kb of L1 cache?
It's very odd.

I'm not sure how long the AMD DX4 will last running at 5V but I'll see how it goes... with the heatsink/fan on there it doesn't get very hot at all (certainly much colder than a DX2 with no fan).

The AMD ones (at least the older 8kb WT models) needed to be jumpered differently.

Seems to be working fine regardless. I've had it looping PCPBENCH for about 20 minutes now and it hasn't crashed.
Besides, I don't have any options apart from the 3 already mentioned.

Interestingly, I found this page: http://www.pcguide.com/ref/cpu/fam/g4I486DX4-c.html
Which suggests that the 8KB AMD DX4's can handle taking 5V.... whether its true or not, I guess I'll find out soon enough 😜

DonutKing wrote:

Seems to be working fine regardless. I've had it looping PCPBENCH for about 20 minutes now and it hasn't crashed. Besides, I don […]
Show full quote

Seems to be working fine regardless. I've had it looping PCPBENCH for about 20 minutes now and it hasn't crashed.
Besides, I don't have any options apart from the 3 already mentioned.

Interestingly, I found this page: http://www.pcguide.com/ref/cpu/fam/g4I486DX4-c.html
Which suggests that the 8KB AMD DX4's can handle taking 5V.... whether its true or not, I guess I'll find out soon enough 😜

Those chips are quite common and cheap to replace 😉

For completeness, the AMD DX4's were in reality 2 different chips:
The older 8T chips (8kb Write Through) and the 8B and 16B chips (Write Back).
I suspect the 8B chips were the exact same chips as the 16B chips but with half of it's cache disabled.
Also the 8B/16B chips are typically jumpered similarly to the Intel ones, whilst the 8T ones are jumpered differently.

Tetrium wrote:

DonutKing wrote:

Seems to be working fine regardless. I've had it looping PCPBENCH for about 20 minutes now and it hasn't crashed. Besides, I don […]
Show full quote

Seems to be working fine regardless. I've had it looping PCPBENCH for about 20 minutes now and it hasn't crashed.
Besides, I don't have any options apart from the 3 already mentioned.

Interestingly, I found this page: http://www.pcguide.com/ref/cpu/fam/g4I486DX4-c.html
Which suggests that the 8KB AMD DX4's can handle taking 5V.... whether its true or not, I guess I'll find out soon enough 😜

Those chips are quite common and cheap to replace 😉

I think that's a good way of looking at this situation. 5V might eventually kill the CPU at some point in the future. Hopefully it'll take a long time. But if you have a small pile of cheap back ups, you can keep this system running for ages! 😀

But otoh, there are plenty of alternatives for 5v-only boards.
You could get a DX4 overdrive or one of those 5x86 upgrade chips which use a voltage regulator. The POD might work, but iirc there were some older boards that had problems with them (though the same could be said about any of the other chips!).

You could get a DX4 overdrive or one of those 5x86 upgrade chips which use a voltage regulator. The POD might work,

The DX4 overdrive and pentium overdrive don't work in this board 😜 that's what prompted me to make this thread in the first place 😀

The 5x86 upgrade chips you are referring to, do they have a voltage regulator built into the CPU like the Overdrives do, or are you referring to an interposer?

I've been looking for a socket 2/3 interposer for a while but haven't found one.

If it is something to do with the 16kb L1 cache that stops the Intel DX4's from working, then the 5x86 won't work in this board either.