Looks bad. My Voodoo 1 behaved similarly (of course, no 2d, but it passed-through fine until 3d lockup).

There's video corruption present before the card locks up. Some things you can test to determine if anything can be gotten out of it, but except for power, the other checks will only prove there's a serious problem:

Have you tried a dedicated power cable directly to the 5500 in either of the systems?
Have you tried downclocking the card from 166mhz?
Have you tried disabling one of the VSA-100 chips?
Did you look over the card for any physical damage?

Your card most likely has issues creating stable voltages under load. You should better not use it until it has been repaired. My guess is bad caps but it could also be one of the VRM circuits has failed and the other can deliver enough current for 2D but not for 3D.

I can highly recommend this thread here (unfortunately in German):

https://www.voodooalert.de/board/forum/index. … deren-ursachen/

The main problem of the Voodoo 5 is the voltage transformer. Stupidly, there is no protection circuit, i.e. if it dies, it takes the VSAs with it. So it's best not to put the card under voltage and change all electrolytic capacitors.

PC-Engineer wrote on 2022-02-25, 08:05:

I can highly recommend this thread here (unfortunately in German):

https://www.voodooalert.de/board/forum/index. … deren-ursachen/

The main problem of the Voodoo 5 is the voltage transformer. Stupidly, there is no protection circuit, i.e. if it dies, it takes the VSAs with it. So it's best not to put the card under voltage and change all electrolytic capacitors.

Exactly. And check wether one of the VRMs has already failed open so the other would be overloaded even after a recap.

@murrayman,

Got a weird issue with another Voodoo 5 I bought. Runs stable in 2D, but the second you try to render in 3D - be it Glide, D3D, OpenGL, anything - it freezes. Tried using reference 1.04 drivers like normal on two separate rigs, one with a clean install of Win98SE. See the video below for demonstration that I sent to the seller:

https://youtu.be/UioL4sQw6QU

Has anyone had this issue before? Is this VSA death or something else?

I see artefacts when it locked up- It looks as memory problem. Please, run same games but in single mode and let me know if card works ok or it continue freezing and showing artefacts. It should be run ok since seems as problem is related slave memory and worse case on slave vsa100.

Regards,
Oscar.

Uh 😅 Okay. Already tried a dedicated power cable and there’s no visible damage that I saw, so those are checked off. Should I now test SLI off and downclocking, or should I replace capacitors / check VRM first? Getting mixed messages on what’s safe and what’s not.

I would highly recommend to change capacitors (in power circuit) first before proceed with your experiments. Bad capacitors can cause extreme stress to the FETs, so that these can die and some Miliseconds later the VSA will die.

PC-Engineer wrote on 2022-02-26, 10:33:

I would highly recommend to change capacitors (in power circuit) first before proceed with your experiments. Bad capacitors can cause extreme stress to the FETs, so that these can die and some Miliseconds later the VSA will die.

This is a warning stronger than on cigarettes' packs...

"Take care of your caps before it is too late for your FETs"

@murrayman,

If you have enough knowledge in reworks, problem is on sdram, but not try it if you are not sure since you can damage more the V5. If you need help, you can contact me, I can fix it wihtout problems.

Not reemplace caps, not reemplace nothing. Just test the card as I told you. Once you do it you will not where is the problem or at least where the problem can be.

Regards,
Oscar.

Physics cannot be tricked, regardless of the words chosen @nexxen, but thanks for the tip. Maybe you also have a substantive contribution to this thread.

It is a fact that the FET of the voltage converters of the V5 already get their problems with comparatively few aged capacitors, compared to many other cards. There is no downstream protection circuit. Since the capacitors inevitably age, unless the V5 was in the fridge for the last 20 years, a replacement of the capacitors is really highly recommended. Especially since an exchange of the 9 relevant capacitors is not very demanding on the user and his equipment and costs very little (compared to V5).

And while you are at it, you can also exchange the 12 other electrolytic capacitors. A little tip for removing the SMD capacitors: carefully twist them off with pliers, then clean the pads with SMD flux and desoldering wire.

The SDRAM change has higher demands on Skil and equipment and also means more stress and risk for the card.

If the benchmark does not start at all, I would rather suspect the power supply or the stable voltage levels than a defective RAM. Especially since the capacitors will have to be replaced at some point anyway.

All is ok, but main question is - at least it is my guess - fix the problem. Not start to reemplace component, if not find where is the problem and later we have time for reemplacing and upgrade the card.

Once find the problem and fix it tunes up the card and reemplace caps, internal power circuit and so on, yes, it is an option and a safe option - yes, I recommend it but not start to reemplace components without know where is the porblem- I always work so and I have have fixed a lot 3dfx cards since Rampage, V5-6000, Daytonas, V5, Aalchemy and a lot if 3dfx prototypes.

Regards,
Oscar.