dr_st wrote:

You can connect speakers/headphones to the line-out (green jack) of the onboard audio device, and see if the speech POST reporter says anything interesting... Although even if it does, I'm skeptical if it's going to be easy to fix. 🙁

The BIOS doesn't even get the time to initialize. The power is cut just a fraction of a second after starting.

h-a-l-9000 wrote:

Check if any of the power rails are shorted. Don't forget the ATX 12V.

Been there, done that. No problems there 😀

PCBONEZ wrote:

ASDS - Asus Sudden Death Syndrome. . IIRC many of those boards came with Nichicon HM and/or HN series caps out of the defective […]
Show full quote

ASDS - Asus Sudden Death Syndrome.
.
IIRC many of those boards came with Nichicon HM and/or HN series caps out of the defective production runs.
Also Asus liked OST brand and Chemicon KZG & KZJ, all of which have high failure rates and often fail with no visible signs.
.

Yeah, I believe the caps are Chemicon ones. There's no logo or anything else revealing the make on them, seen that before on Chemicon caps. A few of them got KZE printed on them, which definitely is a Chemicon series. I'm pretty convinced this is either a case of bad caps on the mobo, or a result of bad caps in the PSU that computer used to have. Hopefully it's not the latter. Never got time to do ESR measurements yesterday, hopefully getting to it tonight instead.

Got a couple of PSU:s I plan to recap at some point, one AT PSU that's just old (Seventeam ST-230WHF) that is in use, recapping it as a safety measure, and an otherwise seemingly high quality Delta one OEM made for Chieftec (HPC-360-202), with Teapo and Jenpo caps, that would be a perfect match for my Athlon rig when recapped. Might just as well order new caps for the P4P800 while I'm at it, and hope for the best, no matter the outcome of the ESR measurements. It's not impossible that there's some capacitance problem that can't be pinpointed without desoldering the caps.

Interesting that you mention faulty Nichicon production runs, had no idea there were. My boss' old Dell Precision 650 died much the same way as this computer according to him, and he asked me to take a look at it. There were SCSI drives in it, and getting it up and running again was the easiest/cheapest way to get the data off the drives. Couldn't believe my eyes when I found busted Nichicon caps on the mobo, but this explains it. Recapped it successfully anyways.

kaputnik wrote:

The BIOS doesn't even get the time to initialize. The power is cut just a fraction of a second after starting.

That is not a good sign, I think a VRM MOSFET is blown.

Skyscraper wrote:

kaputnik wrote:

The BIOS doesn't even get the time to initialize. The power is cut just a fraction of a second after starting.

That is not a good sign, I think a VRM MOSFET is blown.

I believe it's some protection circuit in the PSU's I've been testing with that cuts out the power. There's a slight difference in the time it takes before the power is cut, depending on which PSU is used.

kaputnik wrote:

Skyscraper wrote:

kaputnik wrote:

The BIOS doesn't even get the time to initialize. The power is cut just a fraction of a second after starting.

That is not a good sign, I think a VRM MOSFET is blown.

I believe it's some protection circuit in the PSU's I've been testing with that cuts out the power. There's a slight difference in the time it takes before the power is cut, depending on which PSU is used.

I think you are right, and the PSU normally cuts the power because of a short. A sudden short in an untouced system is often a blown VRM MOSFET on the motherboard.

If the caps are totally dead I guess they could trip the PSUs protection circuits but in my experience bad caps normally just make the board fail to post.

I got a ASUS P4P800 for free two years ago. It had the exact same problem as you're describing. Only powered on for a second, then died completely down, no POST or anything.I couldn't figure it out. I tried different PSU's, different CPU's, different RAM, reset BIOS, different AGP and PCI video cards with no luck. I took a good visual inspection of the caps and the rest of the board, but I couldn't see any problems. Ended up giving to recycling as I couldn't be bothered switching all the caps with no gurantee that it'll work and Socket478 boards were a dime in a dozen (I'll probably regret this one day 🤣 )

kaputnik wrote:

I believe it's some protection circuit in the PSU's I've been testing with that cuts out the power. There's a slight difference in the time it takes before the power is cut, depending on which PSU is used.

That makes sense.

You may be able to narrow it down to what rail by checking the resistance to ground for the various power pins in the power connectors.
A short will show regardless of parallel paths.

If you have an ESR meter you can do a similar check with that, though it might not as helpful.
.

vetz wrote:

I got a ASUS P4P800 for free two years ago. It had the exact same problem as you're describing. Only powered on for a second, then died completely down, no POST or anything.I couldn't figure it out. I tried different PSU's, different CPU's, different RAM, reset BIOS, different AGP and PCI video cards with no luck. I took a good visual inspection of the caps and the rest of the board, but I couldn't see any problems. Ended up giving to recycling as I couldn't be bothered switching all the caps with no gurantee that it'll work and Socket478 boards were a dime in a dozen (I'll probably regret this one day 🤣 )

Well, then at least one of us will be a bit happier if I don't manage to bring my board back to life 😁

Skyscraper wrote:

I think you are right, and the PSU normally cuts the power because of a short. A sudden short in an untouced system is often a blown VRM MOSFET on the motherboard.

If the caps are totally dead I guess they could trip the PSUs protection circuits but in my experience bad caps normally just make the board fail to post.

PCBONEZ wrote:

That makes sense. […]
Show full quote

kaputnik wrote:

I believe it's some protection circuit in the PSU's I've been testing with that cuts out the power. There's a slight difference in the time it takes before the power is cut, depending on which PSU is used.

That makes sense.

You may be able to narrow it down to what rail by checking the resistance to ground for the various power pins in the power connectors.
A short will show regardless of parallel paths.

If you have an ESR meter you can do a similar check with that, though it might not as helpful.
.

Testing for shorts was one of the earlier things I did, no indication of that.

Redid it now, and took some resistance readings. Maybe you guys got some idea of what to expect, and are able to determine if some value is completely off? The readings are of course very approximate, guess caps etc in the circuits interferes.

+3.3V - GND: ~560Ω
+5V - GND: ~900Ω
+12V - GND: ~35kΩ
+5V SB - GND: ~12kΩ
-5V - GND: Out of limits (>40MΩ) or not used. Probably the latter.
-12V - GND: 17MΩ
ATX12V - GND: 11MΩ

I do have an ESR meter, it's just one of those cheap chinese AVR based ones though, nothing special. Could its AC output reveal something useful that the multimeter's DC won't, or did you just mean as an alternative to a multimeter if I didn't have one of those?

Tried to do ESR tests with the caps in circuit, didn't work too well, couldn't get any readings from the caps in one of the cap batteries. Looks like I'll have to desolder them to test after all...

It's hard to do anything useful with a ESR meter on a motherboard because there are so many caps in parallel.
The combined ESR is inclined to make ESR so low the meter shows zero even if the individual ESR's are a bit high.
(Would look the same with good caps or a shorted one.)
If -all- the caps in parallel on the rail are high the meter might show something higher than zero and that would clue you that the whole lot of caps have issues.
Not a great test but it's one you can do without pulling caps.
.
For ESR in parallel you can use the same equations as for resisters in parallel.
It's the same math wise and there are parallel resister calculators online that work perfectly well for ESR too.
.

PCBONEZ wrote:

It's hard to do anything useful with a ESR meter on a motherboard because there are so many caps in parallel. The combined ESR i […]
Show full quote

It's hard to do anything useful with a ESR meter on a motherboard because there are so many caps in parallel.
The combined ESR is inclined to make ESR so low the meter shows zero even if the individual ESR's are a bit high.
(Would look the same with good caps or a shorted one.)
If -all- the caps in parallel on the rail are high the meter might show something higher than zero and that would clue you that the whole lot of caps have issues.
Not a great test but it's one you can do without pulling caps.
.
For ESR in parallel you can use the same equations as for resisters in parallel.
It's the same math wise and there are parallel resister calculators online that work perfectly well for ESR too.
.

Yeah, it's hard to come to any conclusions, especially when using an inexact low resolution instrument on a bunch of relatively large caps. Didn't realize the caps were parallel coupled at first though 😀

In any case it's better to just take the bull by the horns and check them individually.

Skyscraper wrote:

Asus Socket-478 boards not posting is a common issue.

Its mostly memory related but often who knows related... In many cases the board starts working again if you just forget about it now and try again in a month or two. You can try different memory modules, I would use a double sided 512MB PC3200 module that can accept CL2, CL2.5 and CL3.

I don't have a very good opinion of asus boards regardless of socket. The only asus board I bought new and never game me trouble is my P6T-Deluxe. The latest victim in my asus collection is my P5K64-WS - blown mosfet. Thankfully it wasn't too hard to fix.

That's OK. If you go by personal experience of failing boards, you will very quickly end up with bad opinion on all manufacturers, with no exceptions. 😀

dr_st wrote:

That's OK. If you go by personal experience of failing boards, you will very quickly end up with bad opinion on all manufacturers, with no exceptions. 😀

^This, though I might have a bit more patience as things may always been my own fault without me even knowing.

Just accept the hardware for what it is, there's always exceptions...in both directions!

dr_st wrote:

That's OK. If you go by personal experience of failing boards, you will very quickly end up with bad opinion on all manufacturers, with no exceptions. 😀

My personal experience of failing boards is something over 2500 boards and I think Asus boards made since 2001 are crap.
There are exceptions but for the most part they used cheap caps, too thin of traces, cheap MOSFETs and VRs, and insufficient EMI/protection/stability components (SMD).
And they are picky about RAM. It's not the RAM it's that the board can't keep the voltage to the slot stable and load tolerant.
Their customer service went to hell about the same time.
Before 2001 (vintage) I thought highly of Asus.

And no, i don't have a bad opinion of all manufacturers nor do I have opinions that ignore the vintage.

I used to use this in my signature: "Friends don't let friends buy Asus."
.

The short might be on the output side of the Vcore regulator.
Try checking the DC resistance across the Vcore caps.
If it's short, then start pulling caps, maybe one of them is shorted.
If the short is still there after pulling the caps, then lift legs on the MOSFETs. One of them might be blown short. Bad caps can cause that to happen.
Chemicon KZE caps are good, but KZG have been failure prone. KZG is more likely to be used on a Vcore circuit.

PCBONEZ wrote:

My personal experience of failing boards is something over 2500 boards and I think Asus boards made since 2001 are crap.

That's a lot of boards indeed. Enough to collect meaningful statistics. Did you happen to collect such data? Would you say then, that the percentage of failures among ASUS boards noticeably exceeded their percentage in the "general population"?

Well, I have 3 P4P800 SE boards. One is working in my main retro PC right now.
The 2nd one - 2 swollen caps and the integrated network controller not work (1 pair is shorted, 1 pair is cut even w/o the cable itself). But the MB itself works ok.
The 3rd one - ASDS. Strange but it worked some time after a long storage then ASDS again. So it looks like the bad caps case. Maybe later I will repair it.
All 3 have DIMM slots corrosion but a metal ruler fixed this.
P.S. Why ASUS used so *censored* fixators for DIMMs and Agp slots?

And yes, my personal cemetery is Gateway mostly...

PCBONEZ wrote:

My personal experience of failing boards is something over 2500 boards and I think Asus boards made since 2001 are crap. There a […]
Show full quote

dr_st wrote:

That's OK. If you go by personal experience of failing boards, you will very quickly end up with bad opinion on all manufacturers, with no exceptions. 😀

My personal experience of failing boards is something over 2500 boards and I think Asus boards made since 2001 are crap.
There are exceptions but for the most part they used cheap caps, too thin of traces, cheap MOSFETs and VRs, and insufficient EMI/protection/stability components (SMD).
And they are picky about RAM. It's not the RAM it's that the board can't keep the voltage to the slot stable and load tolerant.
Their customer service went to hell about the same time.
Before 2001 (vintage) I thought highly of Asus.

And no, i don't have a bad opinion of all manufacturers nor do I have opinions that ignore the vintage.

I used to use this in my signature: "Friends don't let friends buy Asus."
.

Asus customer service is ABYSMAL. Just check out the ROG forums.

They also tend to be huge assholes when designing and "user-proofing" their products. As an example - a week ago I took my ROG G751JY apart to clean it and replace the thermal paste since I haven't touched it since I bought it in december 2014 - and guess what I found - one of four bolts holding the heatpipes to the GPU had a stripped head so you cold not remove it with any kind of screwdriver. And it wasn't destroyed with a screwdriver, I've seen lots of those - it looked like the person who affixed the heatpipes to the GPU used a drill to destroy the head so you can't remove the screw. Cool huh? I spent about 30 minutes removing the screw - I actually had to tin solder a thin torx screwdriver to the damaged screw head to remove it - and even that came off two or three times before I was done. Not to mention the screw was only crewed half way in (thankfully) so the cooler wasn't touching the GPU as well as it should - and the thermal paste they use is crap. It turned solid after a year of use - but I guess all manufacturers do that.

Sufficed to say, after cleaning and replacing thermal paste as well as replacing the damaged screw, temps wend down significantly.

*I bought the machine NEW, sealed, so there's no way someone messed with the cooling system before. I put the matter up on ROG forums and the thread was deleted within a day, but not before other users saying they found the same stripped screw bolted either on the GPU or CPU side of the cooling system.

dr_st wrote:

PCBONEZ wrote:

My personal experience of failing boards is something over 2500 boards and I think Asus boards made since 2001 are crap.

That's a lot of boards indeed. Enough to collect meaningful statistics. Did you happen to collect such data? Would you say then, that the percentage of failures among ASUS boards noticeably exceeded their percentage in the "general population"?

Not the kind of statistics you want. - And I don't do any formal studies for that.

I see them after they have problems so what matters is -what- problems they are inclined to have.
Most brands need nothing more than caps but Asus are a cans of worms.
I have a P4P800-something on the back burner now with a burned a trace to fix.
Also a P4C800-something but I don't recall what it's 'extra' problem is.
Probably 1/2 the MOSFETs I've had to replace were on Asus and Asus is no where near 1/2 the boards I see.
Asus are more likely to stay dead after a recap (when caps are the only evident problem) than other brands.

I can also tell you that Asus continued using cheap Ch/Tw caps from 2001-ish all the way until their transition to all polymer.
In my opinion even if they had no other problems that makes Asus a low end brand.
When you see a board with "Premium" or "Deluxe" in it's name and it has OST caps on it someone is lying.

In contrast Abit and Gigabtye were quick to adopt Japanese caps and I think MSI eventually did. (Meaning adopt Jap before all poly.)
Intel, Dell, Supermicro and Tyan did. Jetway, Soyo and EVGA did not. I am not sure about HP, Biostar or Foxconn.

Some of these companies got screwed by using Nichicon HM/HN which are Japanese but were defective from 2001-2004-ish due to a manufacturing problem.
(Or Chemicon KZG/KZJ which are Jap but just plain unreliable.)
IMHO trying to do quality by using Jap caps and then getting screwed by the rare case of a dud Jap part is much different than using cheap Ch/Tw caps.

In the socket 478 era virtually every model Asus made (including the AMD variety) showed up at some point with cheap caps.
(Bear in mind different production runs of the same model can have entirely different cap brands installed.)

If caps were all of it then you would expect Asus to be great after the transition to all polymer.
That has not happened because Asus is still cheaping down everything else.
Not too long ago there was a thread here on Vogons that discussed an Asus (IIRC socket 1150 or 1155) mini-ITX board that had BSOD problems due to insufficient uF in the high side of the VRM. (In other words they used too few caps on the 12v rails.) Checking on line reviled the board had a reputation for it.

I used to frequent badcaps.net where there were at least 4 or 5 people with motherboard experience comparable to mine.
The consensus over there is that Asus sucks.
Interestingly one of those 5 keeps an Asus P4P800E Deluxe going for his XP retro box. (Kind of why I looked at this thread.)
His got hit with the defective Nichicon HMs. (And I think OST. Not sure on those though.)
It's a case of him having the skills to fix it and he likes it's features so he put up with the extra work.
Even he won't buy a new Asus product though.
.

kanecvr wrote:

Asus customer service is ABYSMAL. Just check out the ROG forums. […]
Show full quote

Asus customer service is ABYSMAL. Just check out the ROG forums.

They also tend to be huge assholes when designing and "user-proofing" their products. As an example - a week ago I took my ROG G751JY apart to clean it and replace the thermal paste since I haven't touched it since I bought it in december 2014 - and guess what I found - one of four bolts holding the heatpipes to the GPU had a stripped head so you cold not remove it with any kind of screwdriver. And it wasn't destroyed with a screwdriver, I've seen lots of those - it looked like the person who affixed the heatpipes to the GPU used a drill to destroy the head so you can't remove the screw. Cool huh? I spent about 30 minutes removing the screw - I actually had to tin solder a thin torx screwdriver to the damaged screw head to remove it - and even that came off two or three times before I was done. Not to mention the screw was only crewed half way in (thankfully) so the cooler wasn't touching the GPU as well as it should - and the thermal paste they use is crap. It turned solid after a year of use - but I guess all manufacturers do that.

Sufficed to say, after cleaning and replacing thermal paste as well as replacing the damaged screw, temps wend down significantly.

*I bought the machine NEW, sealed, so there's no way someone messed with the cooling system before. I put the matter up on ROG forums and the thread was deleted within a day, but not before other users saying they found the same stripped screw bolted either on the GPU or CPU side of the cooling system.

This is fairly recent. Last 3 years maybe.
An on-line friend of mine sent his Asus board in for RMA three times.
Each time they waited a few weeks and then sent the same board back to him untouched.
Was verified by the serial number on the board.
.
Before that happened he was a hard core Asus fan.
By the end of telling the story there were more 4 letter words about Asus and their reps than any other size word.
.

dr_st wrote:

PCBONEZ wrote:

My personal experience of failing boards is something over 2500 boards and I think Asus boards made since 2001 are crap.

That's a lot of boards indeed. Enough to collect meaningful statistics. Did you happen to collect such data? Would you say then, that the percentage of failures among ASUS boards noticeably exceeded their percentage in the "general population"?

I think this is more along the lines of what you want to see.
I do this sometimes when looking for something for my own use.
I call it the "Happy Factor". It looks at how happy the buyers of the product were.
What it does is gel the reviews down to one number so it's easier to compare multiple products.
Also (by # of samples-reviews) it gives a rough indication of how many were sold relative to each other.
This one focuses on motherboards with a P35 chipset.
By creative use of google you can find all such boards listed at Newegg and it shows the total number of reviews.
(It works for any site and product but Newegg is easier than others and motherboards are pertinent.)
I chose only those with 75 or more reviews. I usually use 100 but I wanted more than one MSI included.
(FYI, below "#4stars" means "number of 4 star reviews".)
The Happy Factor = (#5stars + #4stars + #3stars/2 ) / ( #3stars/2 + #2stars + #1stars)
- In the chart:
The top portion shows the totals by brand. The individual scores are below that.
The key is at the bottom.
The cutoffs for good/bad etc were determined by inputting motherboards I know well to see their HF score.
.

.