VOGONS


First post, by quzar

User metadata
Rank Newbie
Rank
Newbie

Hello all,

I've been working on a build centered around an M571 v3.2a motherboard. It has been working fine with a number of other cards: ATI All-in-Wonder (Rage II+DVD, not the 'pro'), a Cirrus Logic board, and a Diamond Stealth Video 2500. With this one it simply won't post though. The card is exactly as listed here: https://www.evga.com/products/specs/gpu.aspx? … C7-6DC7ED3928EA

At first, on finding the specs of the card I presumed the issue was power. The board was being driven by an old standard 230W AT power supply. I swapped in a 300W ATX PSU with the requisite 18A +12v rail, but this makes no change and the machine still simply won't post (it still does as soon as an older card is swapped in, so it wasn't an issue caused by the newer PSU). I was able to also test the PSU with the graphics card via a P4 board that has no extra power other than the 20pin ATX connector (so it wouldn't be drawing from an additional p4 4 pin or extra molex).

My last recourse is presuming there is some issue here with the compatibility of the board with newer cards or that the board is unable to provide enough power (despite it being made available by the power supply, perhaps due to the board drawing power to feed 3.3v requirements?) . I am aware that this card is vastly overpowered for the system, and the likely solution will be to go with an older card but I would like to determine a specific cause if possible.

Some other diagnostics:
- A POST card in PCI doesn't progress at all when the 6200 is plugged in (sits at 00). It has voltage indicators and all are the same as without the card plugged in.
- I removed all other devices while doing testing, which make no appreciable difference.
- Tried a more modern 400W PSU with dual 12V rails totaling 28A, no difference.
- The motherboard is using the latest BIOS available (5-14-1999 http://cwcyrix.duckdns.org/techpage/html/m1.html). I have also tried (and primarily have been using) the update BIOS listed there that restores a newer onboard VGA BIOS.

Thank you for any input you can provide.

Reply 1 of 3, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

You're forgetting that it's a typical AT motherboard and 3.3v is not drawed from PSU at all. Instead, onboard VRM is used for 3.3v.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 2 of 3, by quzar

User metadata
Rank Newbie
Rank
Newbie
The Serpent Rider wrote on 2020-09-27, 21:16:

You're forgetting that it's a typical AT motherboard and 3.3v is not drawed from PSU at all. Instead, onboard VRM is used for 3.3v.

Thank you for your reply.

That is what I meant with "perhaps due to the board drawing power to feed 3.3v requirements", but I wasn't clear on the ramifications. So I'm guessing this means that there's an unspecified ceiling to the 3.3V draw allowed from PCI slots on this board regardless of the incoming power. I found notes here: http://cwcyrix.duckdns.org/techpage/docs/tf-faq.html#9 stating specifically that the board does not use the 3.3v rail when ATX power is connected.

So I suppose I've got two followup questions:
1) Is there any clear way to distinguish cards that will work on this or similar motherboards? Would I have to stick to single-notch PCI cards that wouldn't expect 3.3V?
2) How dangerous/common would it be to jumper over the ATX 3.3v line to the output of the board's VRM? Or direct to the PCI slot?

Reply 3 of 3, by sdz

User metadata
Rank Member
Rank
Member

1.No. The notch has nothing to do with it, that is related only to the IO signals voltage. 3.3V keyed slots supply two 3.3V rails, one for IO (low current) and another one for powering stuff. 5V keyed slots supply 5V for IO, and 3.3V for powering stuff.
2. You would have to disable the VRM. And there may be power sequencing issues. I wouldn't recommend this.