ASUS P4V8X-X won't POST

[Dan81]

So I recently got an ASUS P4V8X-X that will not POST.

Further checks showed 12v connector a bit burnt on the 12v pins (GND were fine) so I replaced it. I should mention the CPU it came with was a 2.4GHz Pentium 4 Prescott (as opposed to Northwood like I used to always score)

The motherboard will power up but will not POST at all. After a few seconds the mosfets near the CPU and the whole VRM will get very very hot.

The CPU will heat up, and was tested good on another motherboard (along with a lot of other Socket 478 CPUs so as to rule out any dead CPU - all of them POSTed without problems - motherboard used was a Gigabyte 8IPE1000-G)

So any ideas what I can try? I replaced every cap (regardless of brand - they were all dry it seems) with OST RLX (06 datecode,the dried up ones were from 2004),Rubycon MBZ for VRM low and KZE for VRM high.

Should I try replacing the VRM mosfets? They're 06N03LA and 60T03H respectively, and I can find equivalents I can solder in there.

[cpt.charlie]

Re: ASUS P4V8X-X won't POST

Do you get a proper voltage output from the VRM?

[momaka]

Re: ASUS P4V8X-X won't POST

Do you get a proper voltage output from the VRM?

+1

That would be the first thing to check.

Bad caps on the VRM can make the VRM have all kinds of trouble. What brand and series of caps are on your board?

[Dan81]

Re: ASUS P4V8X-X won't POST

+1

That would be the first thing to check.

Bad caps on the VRM can make the VRM have all kinds of trouble. What brand and series of caps are on your board?

Right now? Rubycon MBZ for VRM low and UCC KZE for VRM high.

Before that,the original caps the board came with were OST RLX for VRM low and UCC KZE for VRM high.

As for the voltage, I measured the voltage on the P4 connector with its respective cable attached and it's 12v.

[ChaosLegionnaire]

Re: ASUS P4V8X-X won't POST

he meant to measure the vcore voltage output. usually this is done from one of the legs of the inductors or the positive leg of the vrm output caps if u can probe under the board. this is to check if the vrm controller ic is working also. sometimes when ost caps go short, they can kill the vrm controller as well. i have a dead msi 875p neo2 board to show for it.

[Dan81]

Re: ASUS P4V8X-X won't POST

he meant to measure the vcore voltage output. usually this is done from one of the legs of the inductors or the positive leg of the vrm output caps if u can probe under the board. this is to check if the vrm controller ic is working also. sometimes when ost caps go short, they can kill the vrm controller as well. i have a dead msi 875p neo2 board to show for it.

I tested and I get this:

1200uF 16V KZE: 12.06v
1500uF 6.3v Rubycon MBZ: 1.50V

Tests done using a 2GHz Celeron (Northwood core) and a DT9205A digital multimeter.

Also, one thing to note that I found now: whenever I turn the board on with a CPU, there is an audible hiss (like an capacitor venting). I checked the capacitors and they are all fine. (I used new caps for everything.)

[cpt.charlie]

Re: ASUS P4V8X-X won't POST

The hiss could be the effect called "coil whine". the 1.5V at the output seems OK, do you have this voltage in all the phases? do all mosfets get hot equally?

[momaka]

Re: ASUS P4V8X-X won't POST

The hiss could be the effect called "coil whine".

+1

Could also be something wrong with the VRM if one of the phases is off. But given that the output voltage is stable at 1.5V, it's probably okay. Though cpt.charlie has a valid point - do the MOSFETs and coils all heat up equally?

Besides that, are there any other buck circuits besides the CPU VRM? (i.e. any other VRMs with coils).

[Dan81]

Re: ASUS P4V8X-X won't POST

The hiss could be the effect called "coil whine". the 1.5V at the output seems OK, do you have this voltage in all the phases? do all mosfets get hot equally?

The whole area heats up equally.

As for the whine, it will disappear as the CPU heats up.

momaka - No. The CPU VRM is the only buck circuit on this motherboard.

[Behemot]

Re: ASUS P4V8X-X won't POST

Difficult to believe, there must be VRM for northbridge, southbridge, memory…

[Dan81]

Re: ASUS P4V8X-X won't POST

Difficult to believe, there must be VRM for northbridge, southbridge, memory…

As much as difficult it is to believe, the CPU VRM is the only circuit where any coil is situated, and the only area that heats up insanely. Other stuff such as NB,SB,memory and others are barely warm.

[Behemot]

Re: ASUS P4V8X-X won't POST

Which does not mean there is no VRM. A presence of coil does not mean there is DC-DC module present, and vice versa.

It seems there are only two FETs in each phase but it has independent deriver?

Anyway, only 478, has SATA, but three memory banks, and VIA chipset. Scrap it, it's useless junk.

[Dan81]

Re: ASUS P4V8X-X won't POST

Which does not mean there is no VRM. A presence of coil does not mean there is DC-DC module present, and vice versa.

It seems there are only two FETs in each phase but it has independent deriver?

Anyway, only 478, has SATA, but three memory banks, and VIA chipset. Scrap it, it's useless junk.

In the end I junked the board - one of the pads of an mosfet ripped and I couldn't be arsed to trace it back. Had it been PT800/880, I might've actually saved (e.g ABIT VT7) it but P4X533 is really useless. I replaced it with a recapped Gigabyte GA-8I848P-G. Same thing - 478, SATA, three memory banks. Came with the 0545 datecode HM's I have been talking about but I replaced them with Rubycon MBZ.

The PSU it came with wasn't really good but not bad either. I actually rebuilt it just for fun, as I've replaced nearly all the output caps with Samwha RD 2200uF 16v (which is nearly overkill in some places). The PSU is an Torent Computers LC-B400ATX.

[ChaosLegionnaire]

Re: ASUS P4V8X-X won't POST

haha nice name for a psu. i'd use it as the psu for my bittorrent pirate rig lol!

[RJARRRPCGP]

Re: ASUS P4V8X-X won't POST

I dunno if a Prescott was even supported out-of-the-box with the P4V8X-X...
But, I strongly doubt the BIOS would cause the VRM to blaze like that with a Prescott...

If made in 2003, then it clearly was in the Northwood era and thus, don't expect a Prescott to be supported out-of-the-box...

[RJARRRPCGP]

Re: ASUS P4V8X-X won't POST

but three memory banks

That period was apparently the origin of the seemingly oddball "3 GB" standard.

Because in the single core era, motherboard makers seemed to like having 3 slots instead of 4... I saw the same thing on an nForce 2 motherboard!

[Behemot]

Re: ASUS P4V8X-X won't POST

This most likely has no dual channel so that made some sense at least. nForce 2 already had though.

Anyway, 4 banks are better so you can use two 512MB modules, which are close to useless these days and you can pretty much pick them in the street. So you get rid of two, with two 1gig, and get the same 3 GB. It's OK for XP, you won't waste anything as XP won't address full 4 GB anyway (so that would be some 3.25-3.5 GB). Just all the prosperity and social certainties…

[momaka]

Re: ASUS P4V8X-X won't POST

A presence of coil does not mean there is DC-DC module present, and vice versa.

Actually, it does.

However, just because there are "X" number of coils/inductors doesn't mean that there will be "X" number of buck-regulator circuits. As with CPU VRMs, you can see many coils/inductors, but they can all be a part of a single buck-regulator (i.e. CPU V_core). That said, you will NEVER see a coil/inductor before or after a linear regulator on a motherboard. So if there are no coils/inductors elsewhere on the motherboard, then there are no buck-type DC-DC regulators and only linear ones.

haha nice name for a psu. i'd use it as the psu for my bittorrent pirate rig lol!

Lol, that's what I thought too.
But it could also be that the name was referring to the output ripple coming out of the PSU due to poor filtering.

[Dan81]

Re: ASUS P4V8X-X won't POST

Lol, that's what I thought too.
But it could also be that the name was referring to the output ripple coming out of the PSU due to poor filtering.

I actually replaced the PCB inside it with another one that had thinner heatsinks (I replaced them with some thicker heatsinks, just enough to be a low power 250W PSU) but had complete filtering on both input and output. It had Jun Fu and Don caps. Since I didn't have any more Japanese caps left on my hand, I used some Samwha RD 2200uF 16v caps I bought new on all rails and used some radioactive green 1000uF 16v Yang Chun (Saturn) caps for the 5vSB - another option was to use some KZGs (the caps are of small diameter - about 8mm?) so I thought YC was better.

As for the fan I reused an Xinruilian 80mm silent fan I had on hand and stuck an K-MEX sticker on its back to secure the fan.

As for the name? There's actually an IT shop in Constanta that was (no idea if it still is) named "Torent Computers". All the PSU's I've seen from them were usually L&Cs with more or less complete filtering. (I have seen one with complete filtering and okay heatsinks but only EI-33 transformer)

The easiest place for me to find Deers are usually thrown-away systems that are sold at rag fairs, but I usually find them separately as well. It depends on luck though . I usually found the more gutless L&Cs, but recently I found some of the better (the best being rebranded by Modecom - a Polish(?) manufacturer that rebranded a lot of other PSUs, including our old friend - FSP.) Deers with thick heatsinks, PFC, real ERL-35 transformers (I usually recognize them by having more rounded edges than EI-33 transformers that usually look square-ish) and nearly overkill filtering.

I am thinking of making an archive with each Deer PSU that I own and uploading it here sometime, just so you can look at each. Regardless of being stock or with filtering completed by me. I need to spraypaint the casing of one of them though (an Allied Al-8400BTX) since it somehow got corroded. The innards are fine (no corrosion,double checked) but the lower half of the casing is just awful.

[Behemot]

Re: ASUS P4V8X-X won't POST

It does not, there can be small coils which are not part of any buck regulator, why not? And there are also small-power regulators like those for southbridge or some additional ICs without coils too (quite often actually) so as stated, it goes both ways.