yes becasue that board require atx 2.1 standard that have -5v on pin 18 . Your corsair power supply is EPS v2.92 and pin 18 is a ground. and its based on 2.2 standard which as 24 pin without -5v on pin 18.

Warlord wrote:

yes becasue that board require atx 2.1 standard that have -5v on pin 18 . Your corsair power supply is EPS v2.92 and pin 18 is a ground. and its based on 2.2 standard which as 24 pin without -5v on pin 18.

I'm confused by what you're saying... Modern EPS12v 24 pin connectors use the same pinout as ATX12v 2.x

And as i mentioned, the Corsair is actually lacking a wire on that pin, it is not a ground (which matches the ATX 2.1 and newer spec).

What I'm confused about is why either lacking an unused voltage, or having 30A instead of 40A on the 5v rail would prevent POST.

well, the -5v was rarely used after socket 7. And it would be silly to have a board that requires it when the p4 was about to ditch it entirely.

Tyan did some stupid shit, so maybe they used it somewhere. try to remove the -5v pin from the connector and see if it still posts. Or try a 2nd new psu. if 2 new ones and 2 old ones act the same way, then it may be true that board needs -5v.

Warlord wrote:

https://i.imgur.com/XrQtH2n.jpg […]
Show full quote

Thank you - yes that is exactly what I said - on 2.2 and newer the pin you mentioned is labeled NC. NC means "no connection", which is why there is no wire there - it is not ground.

luckybob wrote:

well, the -5v was rarely used after socket 7. And it would be silly to have a board that requires it when the p4 was about to ditch it entirely.

Tyan did some stupid shit, so maybe they used it somewhere. try to remove the -5v pin from the connector and see if it still posts. Or try a 2nd new psu. if 2 new ones and 2 old ones act the same way, then it may be true that board needs -5v.

Haha thank you, I never heard of Tyan until now.

Quickly tried your suggestion of de-pinning the -5v wire from my vintage PSU *EDIT ... wasn't wake yet*

It boots even with the -5v wire disconnected so that's definitely not it. Good to know.

So what I can see on the labels of each PSU, this vintage 400w VPower unit has 28A on +3.3v, 40A on +5v, 17A on +12v, and 2A on +5V standby, where the newer Corsair unit has 30A on +3.3v, 30A on +5v, 100A on +12v, and 3.5A on +5v standby.

I'm just really confused because it doesn't seem that 30A vs 40A max on +5v should be 'low' enough not to even boot with just the CPUs, 1 RAM stick, and a video card connected.... or do you think that's exactly the problem?

Lack of the -5v line was never a problem on a Slot 1+ era hardware.

I think your modern PSU just don't like low load on +12v line and could act unpredictably because of that. It's not like high-end PSUs are specifically designed to work in such scenarios.

The Serpent Rider wrote:

Lack of the -5v line was never a problem on a Slot 1+ era hardware.

I think your modern PSU just don't like low load on +12v line and could act unpredictably because of that. It's not like high-end PSUs are specifically designed to work in such scenarios.

ooh that makes sense - so maybe if I can load up the +12v rail enough it might be fine?
3
It works properly on the TUSL2-C with one Pentium III so I hadn't even considered that but it does make sense. Do you know which line these CPUs are pulling from? 3.3v or 5v?

I think they all pull power from +5v line.

The Serpent Rider wrote:

I think they all pull power from +5v line.

Perfect thank you. Digging in a bit more I see that for the 20 pin connector, each pin is only rated for 6A so total 5v input through it can only be 24A max, which means I'm definitely not hitting a lack of adequate current.

The only thing that would kind of make sense now is too low of load on the +12v rail so after work I'll try connecting 8 fans from a molex connector and see if that helps.

after work I'll try connecting 8 fans from a molex connector and see if that helps

Typical fans draw +/- 0.25A each. So 8 fans won't help.

The Serpent Rider wrote:

after work I'll try connecting 8 fans from a molex connector and see if that helps

Typical fans draw +/- 0.25A each. So 8 fans won't help.

Wouldn't that depend on if the issue is not meeting minimum load?

As far as I'm seeing online minimum load should be 2A or less...

I'm familiar with the a 1200. It's massive over kill for this setup. Latest units are designed to generate a massive 12v rail. THEN generate 3.3v & 5v from the 12v rail.

In a nutshell, new units are not susceptible to low loads or unbalanced ones.

Have you tried other new and old units?

luckybob wrote:

I'm familiar with the a 1200. It's massive over kill for this setup. Latest units are designed to generate a massive 12v rail. THEN generate 3.3v & 5v from the 12v rail.

In a nutshell, new units are not susceptible to low loads or unbalanced ones.

Have you tried other new and old units?

That's good to hear, thank you!

The only other PSU I've tried is a cheap EVGA 450w bronze unit. The behavior with that one is identical to with the Corsair 1200 - fans / anything else connected to the PSU turn on (and stay on) but no POST happens.

If no RAM is installed, there will be a beep code, and if i install a graphics card that requires additional supplemental power (I'm intending to use a Quadro FX 4000 SDI which needs two 4 pin molex plugs) there is a loud constant pulse if the supplemental power is not connected.

I have only 1 vintage PSU - the VPower 400w unit.

The only other old system i have to test with is an Asus Tusl2-c board, and it will not power up at all with the 450w EVGA unit, but works great with the Corsair 1200w unit.

Definitely not a lack of load issue... I have ~ 8A of fans connected, an old IDE HDD, all 4 RAM slots populated, 4 PCI cards installed, the Quadro AGP card with both 4p connectors on the back using their own dedicated cables to the PSU, USB keyboard and mouse connected to onboard USB ports, and 2 CPU fans connected to onboard headers, still no POSTing - just lots of fans spinning really quickly and the link lights on the PCI NIC are on.

turning it on (as 'on' as it will go with the Corsair) wall power draw is 100 watts / 1A / 144VA so it is definitely doing something.

I'm not really sure how to troubleshoot this any further since all the individual parts are good and work - just not together.

*EDIT*

With the old 400w PSU, booted up (to BIOS) it is pulling 100 watts / 2A / 201VA

Looks like the Machine God inside this motherboard don't like your modern offerings and rather prefer the Dark Age tech =P

The Serpent Rider wrote:

Looks like the Machine God inside this motherboard don't like your modern offerings and rather prefer the Dark Age tech =P

Yeah definitely something like that. I'm just really stumped as to what's going on since from what I understand about how these are pinned and specs.... it should be working perfectly fine. In fact it would be less strange to me if it just did nothing, but the fact that all the fans and cards turn on forever but the system just doesn't try to POST is blowing my mind.

I decided to pick up a 600w Enermax EG701P-VE from ebay since it has 45A on +5v and 36A on +3.3v... going to recap it and transplant it into the Corsair HX1200 case so that I can keep using Corsair's modular cables. Lots of soldering fun to look forward to.

I have a tyan board that wont post... I wonder if the same issue is on mine...

luckybob wrote:

I have a tyan board that wont post... I wonder if the same issue is on mine...

could be... if you have an old PSU laying around give it a try.

I managed to use the old PSU to get a test install of Windows 2000 loaded up last night but after that the machine kept freezing/locking up (power usage through the UPS was around 230w+ when it started to have stability issues) so I think either this 400w unit is not enough for the setup or it's just got bad caps from being 18 years old.

Really hoping that the 600w Enermax unit (recapped of course) will do the trick if I can't get the brand new Corsair 1200w PSU to get the dang thing to POST.

The motherboard might need a recap itself.

The newer power supplies might just be 'too good'.