VOGONS


First post, by Linoleum

User metadata
Rank Member
Rank
Member

I've been working on my P2 setup for few weeks...It uses a HP Condor 1 (P2B-VE) motherboard. The moment I was mostly done with the project (software only), it wouldn't post.

The first time could smell burning followed by a long beep, pause, and then 3 or 4 beeps. I tried again and fans are spinning without any other beeps.

I tried:
- removing all card and all drives
- moving the ram, replacing the ram
- reseating the cpu, replacing the cpu
- reseating the psu, replacing the psu
- visual inspection: nothing
- nose inspection: nothing

Now, with the POST diagnostic card I get these two codes (at the end):
3A: Auto Size Cache
28: Auto Size DRAM

Any ideas are more than welcomed!

Athlon64 3200+, 1Gb, HD3650 AGP, SB Audigy 2ZS
P4 1.8Ghz, 512Mb, Voodoo3, SBLive
P3 866Mhz, 384Mb, Riva TNT2, Audigy
P2 266Mhz, 256Mb, RageIIc, V2, SBLive
P233 MMX, 64Mb, Mystique 220, SB 32
P100, 32Mb , S3 Virge GX, AWE64, WavetablePi & PicoGus v2

Reply 1 of 46, by Trashbytes

User metadata
Rank Oldbie
Rank
Oldbie

Check caps and SMDs around the VRM and then around the Ram slots, sounds and smells like something has let the angry pixies lose. Hopefully it didn't send voltage the wrong way and take something important with it.

Reply 2 of 46, by Linoleum

User metadata
Rank Member
Rank
Member
Trashbytes wrote on 2024-11-01, 01:15:

Check caps and SMDs around the VRM and then around the Ram slots, sounds and smells like something has let the angry pixies lose. Hopefully it didn't send voltage the wrong way and take something important with it.

All caps are rated at 1000uf 6.3v. I checked capacitance for all them caps... There are 10 caps in a row beside VRM where 9 of them seem to be within their rated range (900uf) while 1 is in the range of 4700uf. However, i am measuring them with a multimeter on pcb; which I don't believe is very reliable... Even if this cap is not bulged, shouldn't it be closer to the 9 if they are (and seem) to be in the same circuit?

Athlon64 3200+, 1Gb, HD3650 AGP, SB Audigy 2ZS
P4 1.8Ghz, 512Mb, Voodoo3, SBLive
P3 866Mhz, 384Mb, Riva TNT2, Audigy
P2 266Mhz, 256Mb, RageIIc, V2, SBLive
P233 MMX, 64Mb, Mystique 220, SB 32
P100, 32Mb , S3 Virge GX, AWE64, WavetablePi & PicoGus v2

Reply 4 of 46, by Linoleum

User metadata
Rank Member
Rank
Member
akimmet wrote on 2024-11-01, 02:55:

The burning smell was likely a shorted tantalum. That is where I would concentrate on first.

I assume it's these two orange-yellow caps (CT-25 & CT-26) you are referring to... Any specific way I can test them before proceeding?

Attachments

  • IMG_5529.jpg
    Filename
    IMG_5529.jpg
    File size
    1.68 MiB
    Views
    934 views
    File license
    Fair use/fair dealing exception

Athlon64 3200+, 1Gb, HD3650 AGP, SB Audigy 2ZS
P4 1.8Ghz, 512Mb, Voodoo3, SBLive
P3 866Mhz, 384Mb, Riva TNT2, Audigy
P2 266Mhz, 256Mb, RageIIc, V2, SBLive
P233 MMX, 64Mb, Mystique 220, SB 32
P100, 32Mb , S3 Virge GX, AWE64, WavetablePi & PicoGus v2

Reply 5 of 46, by Unknown_K

User metadata
Rank Oldbie
Rank
Oldbie

I had a marginal power supply blow up southbridge chips on a P2 board and a P4 board before.

Collector of old computers, hardware, and software

Reply 6 of 46, by momaka

User metadata
Rank Member
Rank
Member

Since you are getting POST codes as high as the board checking cache and DRAM, that should (in theory) rule out anything to be majorly wrong with any of the hardware. Unfortunately, there's still a lot that can be wrong.
For example, it could be that a MOSFET or voltage regulator has failed, letting in higher or lower voltage than it's supposed to. In such case, the board could still pass POST in regards to the voltage rails checks.
It could also be a random part failure causing the issue - anything from an IC going shorted or "wonky" to a resistor going open-circuit. I actually (and coincidentally) also had an ASUS board fail on me recently - a CUSL2. It worked fine the first two times I had it running (configuring CMOS.) Then I powered off to connect some HDDs and... nada, would power up for a second and turn off. After a few tries, it powered On... only to not POST at all. I put my finger on various components (ICs) on the board and noticed the clockgen IC was burning hot... and that's where I've left it off. There's nothing in the vicinity of the clockgen IC that appears to be bad, so it looks like the clockgen IC itself has just failed. Why? - I have no clue. I suppose we will know when I replace it (need to make an order for it still.) But old parts can fail out of the blue.
In my case, I'm not sure if the PSU contributed to the issue, too. It's an old PSU, but I've recapped it completely. But it's an old CWT design I'm a bit skeptical about, and I've had a few issues with various PCs with that PSU in it. Coincidence maybe? In any case, it's more of a reminder that I will never ever test any unknown PC/motherboard without a 100% known-good PSU anymore - that means choosing one of my Delta, LiteOn, Chicony/HiPro, Astec, or recapped Bestec units, as these are OEM PSUs that I know for sure were designed well.

Linoleum wrote on 2024-11-01, 02:38:

However, i am measuring them with a multimeter on pcb; which I don't believe is very reliable...

Indeed it isn't.
In fact, it's usually pointless most of the time, as motherboards often have multiple caps in parallel. So you can never measure just one single capacitor. Also, the capacitance is one parameter. The impedance / ESR is equally important, which is something you cannot measure with a multimeter... and even less so with the cap attached on the motherboard.

That being said, this appears to be an ASUS-built board (which is typical to most HP's). As such, rest assured that the electrolytic capacitors are usually of good quality and unlikely to be bad. Based on the picture you provided, I see black capacitors with "K" -style vents on top, which suggests Rubycon... and probably their YXG series, which was more or less the best low ESR they had at the time. These are very reliable electrolytic capacitors, so unlikely they would be bad.

akimmet wrote on 2024-11-01, 02:55:

The burning smell was likely a shorted tantalum. That is where I would concentrate on first.

Unlikely. Tantalums almost never burn out without leaving a (serious) mark on the board.

Reply 7 of 46, by akimmet

User metadata
Rank Member
Rank
Member
momaka wrote on 2024-11-01, 13:43:
akimmet wrote on 2024-11-01, 02:55:

The burning smell was likely a shorted tantalum. That is where I would concentrate on first.

Unlikely. Tantalums almost never burn out without leaving a (serious) mark on the board.

The most common failure mode for tantalum capacitors is to fail shorted.
They don't always burn up. Sometimes a fuse or the power supply protection trips first.

Reply 8 of 46, by momaka

User metadata
Rank Member
Rank
Member
akimmet wrote on 2024-11-01, 19:15:

The most common failure mode for tantalum capacitors is to fail shorted.

Indeed

akimmet wrote on 2024-11-01, 19:15:

They don't always burn up. Sometimes a fuse or the power supply protection trips first.

Aside from the USB and PS/2 ports, mostly everything else is unfused on motherboards. So good luck looking for blown fuses.
The power supply may trip... or it may not. If it doesn't, then the shorted Tantalum cap will surely burn. However, notice O/P says board powers on (fans come on), yet no POST. If PSU was tripping due to a shorted Tantalum (or shorted any component, really), then the fans wouldn't stay running - you'd get a "blip" of power with most fans twitching for a second or two and that's it. So that's one way to know whether you have a "hard" short-circuit somewhere or not. In rare cases, if the short is on a buck-regulated rail (VRM), the controller for the associated buck VRM may detect an overload and shut down without making the PSU trip. But I find this to be an exception rather than the rule, at least with retro hardware.

Reply 9 of 46, by the3dfxdude

User metadata
Rank Oldbie
Rank
Oldbie

It's not clear that you have a short due to a capacitor to me. If you are truly getting post codes, then that means part of your board is working. That means a short is unlikely, because normally a short across a cap would trip the protection on the power supply.

A burning smell can be many things. ICs can get damaged, burn up... I've seen VRMs go. My son burned up a PNP not too long ago.

I'd look at the original power supply and see if something went bad in there. I'm guessing all the swapped in parts are known working? You can get a meter and check voltages on the power supply and check resistance on rails on the motherboard.

Reply 10 of 46, by rasz_pl

User metadata
Rank l33t
Rank
l33t

Step one is to find what is getting hot and emanating that burnt smell.

https://github.com/raszpl/FIC-486-GAC-2-Cache-Module for AT&T Globalyst
https://github.com/raszpl/386RC-16 memory board
https://github.com/raszpl/440BX Reference Design adapted to Kicad
https://github.com/raszpl/Zenith_ZBIOS MFM-300 Monitor

Reply 11 of 46, by Linoleum

User metadata
Rank Member
Rank
Member

Yes, the parts I swapped in are all known working parts.

Bit more info: when removing the RAM I get beeping errors (on top of fans spinning)... But still, no display.

Unfortunately, the smoke and beeping with everything in place only happened once when it blew up... Now, no smell nor anything's getting hot.

Athlon64 3200+, 1Gb, HD3650 AGP, SB Audigy 2ZS
P4 1.8Ghz, 512Mb, Voodoo3, SBLive
P3 866Mhz, 384Mb, Riva TNT2, Audigy
P2 266Mhz, 256Mb, RageIIc, V2, SBLive
P233 MMX, 64Mb, Mystique 220, SB 32
P100, 32Mb , S3 Virge GX, AWE64, WavetablePi & PicoGus v2

Reply 12 of 46, by rasz_pl

User metadata
Rank l33t
Rank
l33t
Linoleum wrote on 2024-11-06, 00:25:

Now, no smell nor anything's getting hot.

pull it out, get a magnifying glass and go over front and back looking for fried tracks or discolored components.
Last two dimm pins 84/168 (https://allpinouts.org/pinouts/connectors/mem … -sdram-168-pin/) are 3V power, check if there is still 3V there.

https://github.com/raszpl/FIC-486-GAC-2-Cache-Module for AT&T Globalyst
https://github.com/raszpl/386RC-16 memory board
https://github.com/raszpl/440BX Reference Design adapted to Kicad
https://github.com/raszpl/Zenith_ZBIOS MFM-300 Monitor

Reply 13 of 46, by tanis

User metadata
Rank Newbie
Rank
Newbie
Linoleum wrote on 2024-11-01, 01:09:
I've been working on my P2 setup for few weeks...It uses a HP Condor 1 (P2B-VE) motherboard. The moment I was mostly done with t […]
Show full quote

I've been working on my P2 setup for few weeks...It uses a HP Condor 1 (P2B-VE) motherboard. The moment I was mostly done with the project (software only), it wouldn't post.

The first time could smell burning followed by a long beep, pause, and then 3 or 4 beeps. I tried again and fans are spinning without any other beeps.

I tried:
- removing all card and all drives
- moving the ram, replacing the ram
- reseating the cpu, replacing the cpu
- reseating the psu, replacing the psu
- visual inspection: nothing
- nose inspection: nothing

Now, with the POST diagnostic card I get these two codes (at the end):
3A: Auto Size Cache
28: Auto Size DRAM

Any ideas are more than welcomed!

The two codes, if I'm reading them correctly, mean that the last successful POST test was Auto Size DRAM, and the one that failed was Auto Size Cache. Have you looked up what that failure means from the BIOS reference?

Strangely enough I had the same symptoms on a Pentium 4 I'm working on . Booted fine after cleanup, added an IDE HD, fans all spin up, no POST beeps at all. I've got a POST card otw, but none of my efforts so far have managed to recover it. We'll see what my codes say.

Reply 15 of 46, by Linoleum

User metadata
Rank Member
Rank
Member
MikeSG wrote on 2024-11-06, 10:51:

Any gold coloured pins on the cache? Or nearby?

Isn’t cache on the cpu for pentium 2?

Athlon64 3200+, 1Gb, HD3650 AGP, SB Audigy 2ZS
P4 1.8Ghz, 512Mb, Voodoo3, SBLive
P3 866Mhz, 384Mb, Riva TNT2, Audigy
P2 266Mhz, 256Mb, RageIIc, V2, SBLive
P233 MMX, 64Mb, Mystique 220, SB 32
P100, 32Mb , S3 Virge GX, AWE64, WavetablePi & PicoGus v2

Reply 16 of 46, by Linoleum

User metadata
Rank Member
Rank
Member

Additional clue: the cpu fan header blew up in the process… I just noticed since I’ve used the easiest fan header to reach during my tests.

Athlon64 3200+, 1Gb, HD3650 AGP, SB Audigy 2ZS
P4 1.8Ghz, 512Mb, Voodoo3, SBLive
P3 866Mhz, 384Mb, Riva TNT2, Audigy
P2 266Mhz, 256Mb, RageIIc, V2, SBLive
P233 MMX, 64Mb, Mystique 220, SB 32
P100, 32Mb , S3 Virge GX, AWE64, WavetablePi & PicoGus v2

Reply 17 of 46, by rasz_pl

User metadata
Rank l33t
Rank
l33t

as in melted?
Shorted fan header would do it, 12V line or with any luck a fuse will be fried somewhere between FAN and ATX connector.
Still this would only break 12V and nothing on this board uses 12V apart from Fan headers and RS232 translators + Asus monitoring chip measuring voltage on all rails.

https://github.com/raszpl/FIC-486-GAC-2-Cache-Module for AT&T Globalyst
https://github.com/raszpl/386RC-16 memory board
https://github.com/raszpl/440BX Reference Design adapted to Kicad
https://github.com/raszpl/Zenith_ZBIOS MFM-300 Monitor

Reply 18 of 46, by Linoleum

User metadata
Rank Member
Rank
Member
rasz_pl wrote on 2024-11-08, 09:49:

as in melted?
Shorted fan header would do it, 12V line or with any luck a fuse will be fried somewhere between FAN and ATX connector.
Still this would only break 12V and nothing on this board uses 12V apart from Fan headers and RS232 translators + Asus monitoring chip measuring voltage on all rails.

No, sorry! It's in good condition visually. It just doesn't work anymore... Also, I checked all solderings and could not find anything suspicious...

Something also bothers me: there was no heatsink on the chipset. However, I'm pretty sure these things don't burn to the point of going up in smoke... Ugh. Anyways, I am giving up. I've already replaced it...

Athlon64 3200+, 1Gb, HD3650 AGP, SB Audigy 2ZS
P4 1.8Ghz, 512Mb, Voodoo3, SBLive
P3 866Mhz, 384Mb, Riva TNT2, Audigy
P2 266Mhz, 256Mb, RageIIc, V2, SBLive
P233 MMX, 64Mb, Mystique 220, SB 32
P100, 32Mb , S3 Virge GX, AWE64, WavetablePi & PicoGus v2