Intel had a specification for removable voltage regulator modules that go into sockets like that. They were most common on higher end Pentium Pro and P2 era boards, and most especially with boards that supported multiple CPUs. The latest board I've seen them on were some Dell GX150 P3 Coppermine boards (but later GX150 boards have it soldered permanently instead). I've never seen them on a socket-7 before your picture.

Those VRMs were identified by a revision, but I don't remember what revision corresponds to what generation of CPUs anymore. I've had a little experience using those things but not enough to be certain of how interchangeable they are.
Since it's a socket-7 Pentium board it might be the earliest generation there is of those removable modules. I'm not sure if there's any danger posed by using a VRM that's too new. One issue that I think does have to be considered is that I've seen those VRMs labeled based on what input voltage they expect (all mine say 5v on them). If the input voltage didn't match I'm sure bad things would happen. Lacking any documentation, you'd have to use a multimeter to figure out what input voltage is going to that socket, or given how close it is to the ATX power connector it might be possible to eyeball it.

It's possible the VRM isn't for the CPU but is instead for the onboard cache memory, but I think that's less likely. Some boards had multiple VRM sockets, one for each CPU and one for powering their cache.
There seems to be an onboard linear voltage regulator sitting near the CPU, plus the removable VRM (currently missing). So it's a little confusing what's doing what here. It uses an ATX power supply so it shouldn't need an onboard 3.3v regulator. I'm guessing the removable VRM is for the CPU power, and the onboard regulator is for something else, but not sure what. The only way to figure out the scheme for sure would be to poke around with a multimeter.

I don't think the 'power module' is needed at all. There is a voltage regulator soldered on the motherboard right beside it. I found the manual for this board and it describes a jumper setting to switch between two CPU voltage regulator modes. Willing to bet this was intended to support future CPUs / overdrive if ever required.

http://www.elhvb.com/mobokive/Archive/Micronics/m55hipls.pdf

EDIT:

anecdotal evidence by looking at ebay is the board with pentium cpus do not have it whereas onres with pentium mmx cpus do have it

Yeah, my research showed the same, SquareGuy. So I just now tried powering up the board now that I got the Matrox G450 PCI card from another member here. And it powered up fine without the Voltage Module. In fact, I just ran a full MemText x86 floppy (that I just made for this purpose) and it passed all 40MB of RAM at 100% with 0 errors. Looking pretty good for a late Windows 95 build to me. Now I have to figure out what I want to do about a case, a budget PSU, and storage. I'd like one of those 1GB CF card thingies, but I'm not going to go out of my way to figure this out, if you know what I mean.

Clearer shot of the big thing next to the CPU socket. There's a small transistor screwed onto the side. I'll try to get a better picture of that later.

Thats a heatsink with a voltage regulator, something like a 7805, will be dropping a higher voltage to a lower one like 12v to 5v or 5v to 3.3v or something to feed the cpu / ram / whatever else.

as for a budget power supply if you want to go dirt cheap I’d go for a FSP 250-300w 80+ rated psu theres loads on ebay, or a bit more expensive a corsair CX430 or something similar, nice new and safe.

Here's the codes on the chip on that small heatsink: CS-5206-1

http://www.datasheetarchive.com/CS5206-1-datasheet.html
That CS5206-1 is an adjustable voltage regulator that can output up to 6A. "Adjustable" meaning that the voltage is set by some external resistors. If there is any voltage adjustment jumper on the board then it's probably changing the resistance value that the CS5206 sees.
Assuming that it's providing 3.3v for the CPU then it can handle about 20W. I wonder if it's an isolated circuit for the CPU or if it's tied in with the whole board's 3.3v rail, which might also get power from the ATX connector.

I was surprised to find I already had a datasheet saved for the CS5207, so I guess that must be what I found in the onboard 3.3v regulator of a Baby-AT super-7 board I was working on a while back. Looks like the CS5206 and CS5207 are the same thing just with different amperage ratings. That super-7 board had AT and ATX power connectors. When using AT the CS5207 was providing the whole 3.3v rail for the board, and when using ATX it was still active, tied in parallel with the ATX 3.3v rail.

Well, from what I understand, the Pentium 200MHz is being fed directly from that, but if you had the Pentium MMX that came directly after this, it had a lower voltage requirement, so required that power module to be installed (that I don't have). I'd have to go back to look at all the jumpers, but most of them set features or bus speed or clock multiplier. None of them were marked for voltage, to my knowledge.

Ok, so this past weekend I reconfigured my Windows 98 (1999) PC. I backed up all my files because the case I originally had for this PC only had slots for one hard drive mounted. I wanted to remove the 2nd 40GB hard drive so I can use it on the P200 stuff above later. So I repartitioned the main drive so it has a 32GB C: and an 8GB D: so I can keep the D:\Win98 folder as well as FF7 movie files, downloads, share folder, etc. But while I was at it, I pulled all the expansion cards out (except the AGP card) and reinstalled in this order:

AGP: Diamond Viper V770 nVidia Riva TNT2 32MB
PCI1: STB V2-1000 Voodoo2 12MB
PCI2: STB V2-1000 Voodoo2 12MB
PCI3: PC-MAX 10/100 Ethernet
PCI4/ISA1: BLANK
ISA2: Creative Labs Sound Blaster AWE64 4MB
ISA3: BLANK

Installed Windows 98, then installed what I guess is the wrong Viper V770 drivers because after that the Desk.cpl file wouldn't open. You know, the dreaded Diamond plug-in crash thing. The only way I could get it to work again was to reformat the whole C: partition and re-install Windows. This time, I used the correct ones and it doesn't crash - but then again it also didn't bother trying to put the Diamond tab in there. Nice. I can use InControl Tools 99 to configure those options anyway.

So here I go, trying to install the Voodoo2 drivers right from the God-damned CD-ROM that came with them and they're in SLI. I run the installer. Great. Drivers are installed, I reboot. I go into Control Panel app and sure enough, there's the V2-1000 tab. Go in, see the settings, see Scan Line Interleave: DETECTED!!. Cool, right?

Wrong. After the next reboot, I can't enter that tab. The screen flashes black twice, then I get the Desk.cpl has encountered an error and it crashes back to the desktop. No amount of rebooting will get me back into that. I've tried reinstalling drivers from within Device Manager multiple times, even using older drivers....nada.

When I was just using the single Diamond Monster 3D II 8MB card with the Diamond version of the Voodoo2 drivers, this never happened, but I also never had SLI either. Spent a few hours Googling this issue without any kind of resolution. I think it's just a driver problem.

Any help?

Did you make the sli cable by yourself ? I had the same kind of random issues and I discovered that in fact my cable wasn't made right. Now it doesn't detect the sli in the v2 1000 options but at least games do work in 1024*768 without any issues

I did not make the cable myself. I got it off E-bay for a few bucks. But shouldn't that just break the SLI and not have anything to do with the Voodoo2 Tab in the display Properties CPL?

It did not act that way for me. It detected the SLI but when I was launching a game I had a "scanline" effect because one of the two cards wasn't rendering the picture (or at least did not display it because I took screenshots and they were normal) try to check it with a multimeter ?

I may just pull the SLI bridge out completely and see if the tab works properly after that, but I have a feeling it won't.

So I just pulled out the SLI bridge and booted the system. And the VT-1000 tab works perfectly fine right now, although it does say Scan-Line Interleave: NOT DETECTED.

So now I'm examining the bridge according to the instructions found elsewhere on this site's Wiki.

UPDATE: Ok, so I checked to make sure the SLI bridge was made correctly and it looks correct. Pins 16-19 are reversed and those are in the exact middle. Counting from either side results in 16-19 flipped. So I guess it doesn't matter what orientation it's plugged in as.

So I shut the system down, plugged the SLI bridge back in, and restarted. VT-1000 tab crashed desk.cpl again.

Shut down again, took the SLI bridge back out, started up and I can get into the VT-1000 tab just fine. Reports all 4MB Fame Buffer and 8MB Texture Memory (full 12MB). So why in the hell isn't the SLI bridge or maybe the 2nd card working right?

Check to make sure all the pins on the 3dfx chips are straight. I know alot of times they get crushed due to mishandeling and if they touch together they will cause a short. Although id think there would be different graphical issues, not what you are describing. What came to mind at least.

Nah, that's not it. Those SLI Bridge pins are fairly thick and they're all straight the last time I inspected the cards. It's not the pins at least. Although during the inspection of the SLI thing, I DID break the ever-living-f*** out of the floppy connectors. 🤣 Those things are super-fragile. So whatever. It was only a $8 purchase on E-bay. Not out much.

Ah ok just a thought. Odd, keep your eye out for a cheap Voodoo3 card. They are far less trouble. Or just use a single card and use the second one in another machine!

I already have a Riva TNT2 32MB AGP card in there, which pretty much maxes out the system's AGP bus and the 350MHz Pentium II is bottlenecking the Voodoo cards. The Diamond Monster 3D II worked fine by itself, but I was looking forward to seeing what the SLI could do in this system.

The WinXP system has an ATI Radeon 9550 256MB AGP card which far outstrips the 2x Voodoo2 cards, so there's really no point putting them there.

I might run the Diamond Monster 3D II on the Pentium 200MHz Socket7 system I have if I ever get a case for it. The two cases I have are Micro-ATX cases with only 4 expansion slots. This is a full-ATX motherboard with 7 expansion slots.

You talked about NES Games with GravisGamepadPro on a PII-350 ?
Tell me more 😀 Does this machine have enough power for NES emulation?

dr.zeissler wrote:

You talked about NES Games with GravisGamepadPro on a PII-350 ?
Tell me more 😀 Does this machine have enough power for NES emulation?

Off topic, but we did run Nesticle in school on 486SX-25 machines in the late 90s. And it ran very well, AFAIR. 😀

So, a P2-350 is plenty fast even for SNES emulation.