Warlord wrote:

there are bioses for all of those other ones but I'm not sure which one you have.

PBA 631022-103
AA 634019-009

This is the information on the stickers normally used to identify these boards.

1.00.xx.AF1 (Premier/PCI) or 1.00.xx.AF2 (Premier/PCI ED) bios identifiers.

source
http://www.motherboards.org/forums/viewtopic.php?f=2&t=30189

yours may not even be the batmans revenge, just a normal batman. depends if u have AF1 or AF2

Technically, it is a Gateway 586PCI4 motherboard, which originally contained the 1.00.10.AF2T BIOS. The board is made by Intel for Gateway and many other OEMs. Intel also sold the board with their own branding. The Intel branded BIOS is a little newer than the Gateway one, so I am using it - 1.00.13.AF2. Among a few other things, it supports the POD133.

The documentation which comes with the BIOS update states "Batman Revenge":

1OPSD BIOS ENGINEERING
2 
3Intel Corporation, 5200 NE Elam Young Parkway
4Hillsboro, OR  97214
5DATE:	December 13, 1995
6
7 SUBJECT:	Batman Revenge Release 13.0
8
9About This Release
10o	BIOS 1.00.13.AF2 Release 13.0 - Batman Revenge
11
12Features/Errata Fixed In This Release
13
14Release 13.0
15o	Support for P5 Overdrive added to setup. “Pentium Overdrive” is now displayed and
16	CPU frequency now displays 133MHz.
17o	Fixed setup’s CPU frequency display. 66MHz used to display 50MHz. 

What I am curious about is if these CMOS clearing procedure is really necessary when changing graphic cards:

1How should I add a PCI video card into my Advanced system?
2
3You can do the following:
4Perform a CMOS clear. The procedure is as follows:
5Note your current settings
6Turn power off
7Set the CMOS Clear jumper or switch to the CLEAR position (see the product documentation)
8Turn power on
9After approximately 30 seconds turn the power off
10Set the CMOS Clear jumper or switch to the OFF or NORMAL position
11Turn your system on and enter the CMOS setup to change settings as you require (hard drive, etc.)
12Some PCI video boards have issues with PCI burst transfers. Using an ISA video board you can disable 'PCI Burst' in the Advanced Chipset submenu found in the BIOS Setup. You should remove any add-in PCI video board before installing the ISA video board.

feipoa wrote:

I'm also not sure what all those suffixes are for - ED, GX, LC...

these are just different variants/formfactors of intel 430LX boards, an image search will show boards that have graphics/sound chips soldered on and even RAM. by the way, i find it interesting how the batman's revenge (Premiere/PCI) manual talks about how the board needs to have the VRM onboard to reliably run a 66 MHz CPU, yet the batman (Classic/PCI Pentium) did not have the VRM and the manual for that still lists 66 MHz as supported.

i could perhaps test a voodoo1 on my batman's revenge board once i can find an appropriate heatsink to glue on the CPU.

auron wrote:

i find it interesting how the batman's revenge (Premiere/PCI) manual talks about how the board needs to have the VRM onboard to reliably run a 66 MHz CPU, yet the batman (Classic/PCI Pentium) did not have the VRM and the manual for that still lists 66 MHz as supported.

Possibly Intel wasn't aware at the time of publishing the manual that the Pentium 66's needed a bit more voltage for reliable sustained operation. I haven't come across any Intel socket 4 Mercury-based motherboards with the 5.27 V VRM on it. I saw the socket 5 variant with the VRM though. For my system, I'm using the Pentium Overdrive 133, which contains its own VRM so it is a non-issue.

auron wrote:

i could perhaps test a voodoo1 on my batman's revenge board once i can find an appropriate heatsink to glue on the CPU.

Great, thank you! For benchtop testing, I usually just used a heavy copper heatsink w/fan and don't use the clip to tie the CPU down. If you plan to semi-permanently secure a heatsink using thermal epoxy, know that when you go to remove the heatsink, e.g. with a chisel and hammer, the surface print may be removed.

Do you have a photo of your board? And do you know which BIOS version is on it?

Brief update:

I now have a virgin installation of Windows 95c with only device drivers installed, plus GLQuake. I have the Voodoo1 driver version 2.16 installed and was able to run GLQuake using the GLQuake installation's default miniGL. I haven't installed Direct X yet, so most likely DirectX 2.0 was installed.

It will take quite some time to track down which software or installation was causing the Voodoo not to function on my previous installation, and only for the socket 4 system - as noted, the 486 worked with the Voodoo on the same hard drive.

Next step is to install Direct X 5.0, test, then MiniGL v1.40, test, then Direct X 6.1, test, then Unofficial service pack update, test, etc..

feipoa wrote:

Possibly Intel wasn't aware at the time of publishing the manual that the Pentium 66's needed a bit more voltage for reliable sustained operation. I haven't come across any Intel socket 4 Mercury-based motherboards with the 5.27 V VRM on it. I saw the socket 5 variant with the VRM though. For my system, I'm using the Pentium Overdrive 133, which contains its own VRM so it is a non-issue.

i've heard about reliability issues with the 66 MHz chips, but i have a hard time imagining they'd just keep a CPU prone to crashing on the market, also socket 4 boards from other manufacturers didn't use VRMs either and they certainly seemed to support 66 MHz from what i've seen, so this is still pretty curious to me. besides the 5.27 V capability i would think the extra two big capacitors could especially help when using a low quality PSU though.

feipoa wrote:

Great, thank you! For benchtop testing, I usually just used a heavy copper heatsink w/fan and don't use the clip to tie the CPU down. If you plan to semi-permanently secure a heatsink using thermal epoxy, know that when you go to remove the heatsink, e.g. with a chisel and hammer, the surface print may be removed.

Do you have a photo of your board? And do you know which BIOS version is on it?

there are a few pictures of boards with the VRM on the net, here for instance. incidentally, my board also does have the VRM. it really looks very much like the one on Plato boards at a glance. i dont remember the BIOS version offhand but i'll check when i do the testing, also the RTC battery on mine is dry so hopefully that won't impact usability.

and yeah i'm aware of the downsides of gluing, i think this was the standard for socket 4 though and actual socket 4 heatsinks seem to be even rarer than socket 8 ones nowadays.

Oh wow, you found a Batman's Revenge board with socket 4 and the 5.27 V VRM! Yours also has the standard DIN for the keyboard, but doesn't appear to have a PS/2 mouse header. My board has two mini DINs - keyboard and mouse. I will need to convert it to standard keyboard DIN and make some kind of breakout header for the PS/2 mouse.

Yes, Socket 4 heatsinks have all but disappeared. Perhaps some Z-clips will fit with the right heatsink?

I believe I have resolved the problem with the Voodoo1 with Win95c and NT4.

Win95c needs to have either Voodoo1 driver 2.16 or 3.00.01 installed, but not 3.01.00. Not on this motherboard anyway. The newer version was fine on the 486 board. Also, trying to revert to a lower version number of the driver was not straight forward. Windows95 didn't want to do it because it had a newer driver date. There is no uninstaller. So you need to hunt down all the glide, 3dfx, and directx driver for Voodoo and manually delete them. You also must manually delete the Voodoo Graphics entry in your Device Manager. When Windows 95 reboots, it will insist for those deleted files. At this time, it will not accept lower version files. You must hit cancel and wait for Windows 95 to boot up, then you can "update driver" for the Voodoo entry in the Device Manager.

The same is true for NT4. Voodoo1 driver version 2.46 and 3.00.01 work just fine, but not 3.01.00. Although 3.01.00 worked on the 486.

I also needed to find the proper Dell 3D Nitro version of my VirgeGX for the NT4 resolution sit at 1280x1024x16bit at 60Hz. The NT and S3 provided drivers would only go up to 1024x768x16bit. NT4: https://www.dell.com/support/home/ca/en/cabsd … driverid=r18832 , Win95: https://www.dell.com/support/home/ca/en/cabsd … driverid=r18831

This is probably case closed now, but I still need to try it on my Acer V12P motherboard, which is the one in the Acer case to be built.

I was able to get the Voodoo working in the Acer V12P motherboard, so since the V12P comes with an Acer case, it is the V12P that I will use in the build. The GLQuake benchmarks between the Acer and Batman board are the same. The Acer BIOS is a lot more customisable too.

As I don't normally do system build threads, I'll post a few photos here to wrap it up.

Hardware

Acer V12P Motherboard

Inside the Acer Tower

Front view of Acer Tower

View of the Intel Batman's Revenge Premiere/PCI used for troubleshooting the 3dfx issue

The Promise SATA150TX2plus was the only PCI hard drive controller I could get working on the Acer board. Promise Ultra100 or Ultra133 wouldn't detect hard drives when used on the Acer board (Batman board OK though). I could not get any all-in-one PCI 3D cards working on the Acer board, so I opted for a VirgeGX + Voodoo1. The VGA pass-through cable is lousy and I've ordered a thicker one. I'd like to try S3 Quake on the VirgeGX before putting the system away.

Next system I'll be working on is a socket 8, or socket 5 with direct-mapped cache. I'm considering using an AMD K5-200 in the socket 5, or an Evergreen Spectra upgrade module with a K6-III.

That Compudyne case has similar design to the Acer, but is it actually made by Acer too?

I suspect Compudyne made this particular case for ACER. My Acer V20 motherboard came in this case. It had an ACER V20 sticker over the "Compudyne". The bottom of the case has a sticker naming ACER, a model number, mfg date, etc.

You would think that Acer would have made the case for Compudyne. Acer was a fairly big manufacturer of computer cases in the 90s.