Did you check beforehand if that motherboard is compatible with standard ATX pinouts? You might've fried the motherboard if it isn't. Proprietary (mostly from OEMs like HP, Dell, IBM, etc) server parts tend to have this problem.

Safest buy would be a Intel, SuperMicro or Tyan motherboard since those are designed to work with off-the-shelf parts mostly.

I did check it before, but i didn't find any info on it so I assumed that it is standard connectors. If i knew it, i wouldn't have bought it 😁 I knew that Dell uses proprietary connectors, but i thought that HP uses standard atx. The light on motherboard does light up though! When i hook it up the fan spins up for a second but thats it.

Maybe the CPU model require a BIOS update? I have a pair of dual 1366 SuperMicros that only boot up with E55xx Xeons (there's also the fact that one of them "died" after a BIOS update that went without warnings or errors, so, no dual X5690s for me until I replace that chip 🙁 )

The connectors are standard, they are not PCIe however.
They are like the 4-pin "P4 connector" only with more pins to handle multiple CPUs.
Sometimes there is an additional 4,6 or 8 PIN specifically for RAM. (Which has it's own VRM.)

PCIe is the proprietary one. - The others came first.
PCIe connectors have the polarity reversed from standard. (They are wired backwards.)

These systems don't use standard ATX PSUs. They use EPS12V PSUs. (The SSI standard.)
There is little difference except for the connectors available.

There are many combined standard ATX/EPS12V PSUs around if you look close enough.
(Because some are but don't say so.)

If in your tinkering you connected something with reversed polarity you may have fried the PSU, the motherboard or both.
.

The xw8600 only officially supports E54xx, X54xx, E52xx, X52xx CPUs.
While technically the chipset should support other chips HP may not have provided BIOS support for them.
.
The HP xw____ systems are known caps blowers. Heat issues.
I know most of the caps on that model are solid polymer but you should closely inspect those few that aren't.
.

Sorry for your loss.
I did find that the connections different so i rewired my connector by finding which pins on board are ground with meter and double checked by tracing the other pins to nearby 16v caps for +12v. Could the main connector be different? Maybe by just one pin? My xeons are 3ghz (or maybe 2.66,but most likely not) quad cores from 2008 mac pro so I think they should be supported (cant see their model because their fronts are ruined by apples thermal glue or whatever that crap is) and I don't have any other lga771 chips. Could it be the ram? The xeons in mac pro 2008 are 5400 series as far as i can find

Caps appear to be fine

varanoidas wrote:

Sorry for your loss.

What loss?

varanoidas wrote:

Could it be the ram?

Yes. What are you using?

Snayperskaya said that he killed a xeon after bios update. That loss

varanoidas wrote:

Snayperskaya said that he killed a xeon after bios update. That loss

Gotcha.

These systems require FB-DIMM. - Fully Buffered DIMM.
.

Two sticks of ram same as this http://www.ebay.co.uk/itm/1GB-PC2-4200F-44-11 … A-/111233330934
It is fully buffered. Other sticks dont even fit.

varanoidas wrote:

Two sticks of ram same as this http://www.ebay.co.uk/itm/1GB-PC2-4200F-44-11 … A-/111233330934
It is fully buffered. Other sticks dont even fit.

That is FB-DIMM but it's only PC2-4200F.
The system calls for PC2-5300F. (DDR2-667 Fully Buffered)
And socket 771 boards are very picky about RAM type/speed.
.
Fortunately 1Gb and 2Gb sticks are quite cheap these days.
.
You can ignore the ECC part of the description. All FB-DIMM is ECC.
.

varanoidas wrote:

Other sticks dont even fit.

Not supposed to. I've seen people manage it.
I think some people install RAM with hammers.
.

Could it really be that ram is too slow? In that case shouldn't it display an error code with its led or something? Now it just spins up the fan for a second and powers off.

Guess I am going to have to return the ram back to the shop and ask if they have PC2-5300F. Thank you your your help. Btw, do you know if it can take a single dimm or do they have to be in pairs?

varanoidas wrote:

Snayperskaya said that he killed a xeon after bios update. That loss

🤣

On 10+ years tinkering with hardware that was the first time I've ever seen a BIOS dying after a successful update. IT never ceases to amuse me 😵

varanoidas wrote:

Guess I am going to have to return the ram back to the shop and ask if they have PC2-5300F. Thank you your your help. Btw, do you know if it can take a single dimm or do they have to be in pairs?

You will need at least one DIMM for each CPU installed.

Oh, try to edit your posts to avoid double (triple, etc) posting.

varanoidas wrote:

Could it really be that ram is too slow? In that case shouldn't it display an error code with its led or something? Now it just spins up the fan for a second and powers off.

Yes I believe the wrong RAM in a socket 771 board could result in the old "fan kick" scenario.
I'm on a dual skt 771 system right now. (Kind of my thing.)
I could tell you for sure if I had any PC2-4200F around but I don't think I do.

From the discussion so far the possible problems are bad or incompatible PSU, bad motherboard or wrong RAM.
The RAM is probably the cheapest to explore unless you have a better (for this) PSU on hand.
.

You can test the board in single CPU mode.
That's probably a good idea because it's less load on the PSU.
.

Snayperskaya wrote:

You will need at least one DIMM for each CPU installed.

Not all socket 771 boards are like that but given this one is an HP this one probably is.
The next problem is figuring out which RAM slots go to which CPU.
Testing in single CPU mode is just easier.