I have a Voodoo 3000 with an S-Video port and it works fine on a cheap MVP3 board I have

Can it be that not enough Amp gets to the slot for the voodoo3? The cyrix iirc uses quite some power, by this draining so much the voodoo3 lacks on this so it can't power on. 😐

I was thinking along the same lines as Nvm1.
Perhaps the card with s-video draws slightly more power and it's just enough more to give you the problem on that particular board (and PSU).
.

feipoa wrote:

I just received a Voodoo3 3000 AGP card containing an S-Video port from eBay. The card turns on fine in a 440BX, VIA Appolo Pro 266, and an i815EP board. However, when I tried it in my SS7 VIA MVP3 board (containing a Cyrix MII-433), I receive some beeps and no display. What is strange is that my this same MVP3 w/MII-433 runs fine with my Voodoo3 3000 AGP card without the S-Video port.

And the beeps are?

All my voodoo 3 cards work fine in my MVP3 boards, including my Voodoo 3 3500TV

Nvm1 wrote:

Can it be that not enough Amp gets to the slot for the voodoo3? The cyrix iirc uses quite some power, by this draining so much the voodoo3 lacks on this so it can't power on. 😐

The MII-433 doesn't use that much power actually. According to this webpage it only uses 12W max http://pclinks.xtreemhost.com/elec_pentium.htm#cyrix
The amount of heat produced by CPU's depends greatly on the size of the die. The MII had several die-shrinks which helped cut down powerusage.

And I don't know what the problem could be. Nvm, I didn't read correctly *blush*

Didn't we have a case here where the voodoo was sensitive to spread spectrum?

h-a-l-9000 wrote:

Didn't we have a case here where the voodoo was sensitive to spread spectrum?

Even prior to POST?

To respond to another poster, the beeps are 1 long followed by 3 short.

saturn wrote:

And the beeps are?

Single tone audible sounds that in this case emanate from the system speaker.

feipoa wrote:

the beeps are 1 long followed by 3 short.

Beep codes are BIOS type and sometimes board manufacturer dependent.
For generic AMI: 1-long 3-short = Conventional/Extended memory failure.
For generic Award: 1-long 3-short = Video card not detected or bad video card.
By generic I mean one where the board manufacturer did not customize the beep codes.
.

Indeed - Bad video card. I'm not sure why this particular Voodoo3 card comes up as "bad" while another Voodoo3 card works fine. This is the first time I've come across this problem for the same graphics card series. As noted previously, the Voodoo3 w/S-Video works fine in other systems.

feipoa wrote:

Indeed - Bad video card. I'm not sure why this particular Voodoo3 card comes up as "bad" while another Voodoo3 card works fine. This is the first time I've come across this problem for the same graphics card series. As noted previously, the Voodoo3 w/S-Video works fine in other systems.

I have had Award throw that code when there were bad caps on the card and also when the video RAM was bad.

Are all the caps present? I've seen them get knocked off in handling and just not be there.
Also look closely at the IC chips. Especially VR's or MOSFETs. Are any cracked?

As (I believe) you have a card to compare it to look for any missing SMD parts. Those can get knocked off too.
.

Those are all good points; I will take a closer physical examination of the card. It is possible that the AWARD BIOS in the MVP3 board is more particular about bad caps or RAM. How would I test for non-bulging bad caps and video memory while these components are still soldered on?

feipoa wrote:

Those are all good points; I will take a closer physical examination of the card. It is possible that the AWARD BIOS in the MVP3 board is more particular about bad caps or RAM. How would I test for non-bulging bad caps and video memory while these components are still soldered on?

Despite what some people say you can't really test most caps in-circuit. Least not without knowing the schematic and even then it won't always work.
The problem comes from parallel paths through other components that will muck up the readings.

For video RAM there are some softwares that will test that but at the moment an example isn't coming to mind.
And of course it has to be in a system that will at least boot with the card in.

If you want to be thorough testing caps you really need three test-equipment tools.
An ANALOG multimeter. (Digital models are too hard to read for the test needed.)
A Capacitance meter.
An ESR meter. (With a 100kHz test signal.)
- Is not common but there are situations where if you don't use all three you can miss a bad cap.
There are more advanced test gears that combined some or all of those into one unit. Beware: The hand held models, even from good companies like BK Precision, rarely have appropriate ranges for caps used in PC equipment. They are more for radio and communications gear than computer gear.
A popular combined unit that can do this job is the Sencore Z Meter. They are a bit pricey for hobby use.
You can work around using the meters with a Sig-Gen an O'Scope and math but I've always been too lazy to do it that way so I dunno how without some study.

And if you want to go old-school and play with old cool looking test gear there are Leak-Down testers.
Leakage rate is not an important/common issue in PC equipment so one of those would just be a cool toy.
.

Not discounting any of the previous replies, but SS7 boards (VIA chipsets in particular) are pretty well known to be picky/flakey with AGP. Back in the day, I ran a SS7 Gainward 5VPA (VIA MVP3 chipset) and it would NOT work with a brand new Voodoo3 3000 (I ended up buying the PCI version of the card). Sadly, that system is long gone, but I too have been working on a new SS7 build and have some recent experience to offer! With the following boards, I have tried 3 different AGP Voodoo3 cards: a "Compaq" Voodoo3 3000, a typical 3Dfx Voodoo3 3000, and a Voodoo3 2000. All three are from eBay.

Board: FIC PA2013 (MVP3 chipset). Would not work with either of the 3000 cards. The 2000 AGP seemed to work ok.
Board: Gigabyte 5AX (rev. 3.x, ALI chipset). Worked fine with the typical 3000 and 2000 cards. Did not like the Compaq Voodoo3.
Board: ASUS P5A (rev. 1.04, ALI chipset). Same as the Gigabyte board.

Oddly enough, the Compaq Voodoo3 seemed to work fine in an Intel P4 motherboard.

Visual inspection did not reveal any bad caps, cracked VRM, or anything else out of the ordinary. The six caps of any size are surface mount and it is not worth the trouble to replace or even test them at this point. I think the issue is most likely related some something specific in my Tyan Trinity S1598 motherboard (VIA MVP3-based) or BIOS which does not like the S-Video variant of this graphics card. I only bought this recent Voodoo3 card for testing purposes across various platforms so that I don't need to continually canabolise my systems.

I also have a FIC PA-2013 sitting in a box. Perhaps when I get my 486SXL off the test bed I will try this card on that board. I was under the impression that since the Voodoo3 did not use any AGP features, the crummy VIA MVP3 implementation of AGP would not be an issue.

Besides checking for signs of damage, cleaning the card might also help. And another possibility might be something with the slot (dirt perhaps?) on the motherboard itself? Though I don't think this is likely, it's easily and cheaply fixed with some cleaning alcohol so it might be worth a shot 😀

Tetrium wrote:

Besides checking for signs of damage, cleaning the card might also help. And another possibility might be something with the slot (dirt perhaps?) on the motherboard itself? Though I don't think this is likely, it's easily and cheaply fixed with some cleaning alcohol so it might be worth a shot 😀

Yes.
I've seen RAM modules test bad but then test good after cleaning away 'stuck' dust around the pin area of the RAM chips.
.

feipoa wrote:

I also have a FIC PA-2013 sitting in a box.

I would be very interested to hear your results. This is the board I originally wanted to build a system with because of the 2MB cache.