VOGONS


First post, by stanwebber

User metadata
Rank Member
Rank
Member

someone brought this to my attention while i was considering a post-mortem for a board that suddenly died on me. are there any athlon motherboards that power on, but refuse to post and not even provide any error beep codes when the cpu fan header is disconnected? my cpu fan is powered by a psu molex connector with only the yellow tach wire running to the cpu fan header. when i transferred the heatsink to a replacement system the new bios reads 0 rpm (failed board read actual rpms). when i looked into the issue i discovered it's not a straightforward matter to power a 3-wire fan directly off the psu and expect the motherboard to register the rpm speed by just redirecting the tach wire to a fan header. as best as i can determine it might work with some boards and other boards definitely require a specific wiring setup...my cpu fan only came with a psu molex connector and separate tach wire so it hasn't been modded.

all of this confusion has me second guessing if my original athlon board actually died. i exhaustively tested every component to revive the board you could think of, except this one: cpu fan still works, but stopped reporting rpm signal. (i never pulled the bios chip and read it back in another board to confirm no corruption so there's that too). if someone can tell me that a failed or disconnected cpu fan will present the symptoms that i observed then i might be willing to tear apart my current system for the components to re-test the failed board.

symptoms:
- power button activates psu, case fans start spinning and onboard motherboard led is lit up
- black screen
- no error or beep codes with any combination of peripherals added/removed
- cmos jumper removed or in reset position prevents anything from powering up
- rpm signal on cpu fan tach wire may have failed
- all components (other than cpu fan) tested good when transferred to a replacement motherboard (ram tested in separate ddr board)

i've had systems angrily beep at me when a cpu fan failed and subsequently shut down when exceeding a temp threshold, but i've never had a system just die on me and refuse to repond just because a fan wasn't spinning. did those youtube videos of athlons bursting into flames within seconds of removing a heatsink prompt this reactionary design?

failed motherboard: biostar m7mia-r
replacement mothereboard: iwill kk266-r plus

Reply 2 of 6, by Hoping

User metadata
Rank Oldbie
Rank
Oldbie

In my experience a some boards won't read the rpm signal if only one wire is conented, they want the ground wire also connected to the fan header, it's extrange but I've seen it.
And the Athlon cpus don't die so easilly because of overheating, as long as the have the heatsink atached they will fail/give BSODS and things like that but won't die easilly as long as they have the heatsink atached even if the fan doesn't work at all. That videos of the athlon cpus burning, only showed that motherboard vendors didn't care to implement a correct overheating control for the cpu, even when the Palomino core came out with an integrated thermal diode a lot of boards, old and new didn't use it at all.
And to tell you the truth about that videos, who is so stupid to remove the heatsink when the system is running, or power on the system knowing the heatsink is not installed.
And that was around five of six years after the first cpus that required a heatsink came to the marquet.
You won't try to play on your RTX 3090 withoght the heatsink, or remove it wille it's running, I guess.

Reply 4 of 6, by Cuttoon

User metadata
Rank Oldbie
Rank
Oldbie
Doornkaat wrote on 2022-03-14, 06:34:

I don't think this is the issue here. To receive a tach signal the board would have to start up and spin up the fans.

Yes, the mobo should at least get far enough to say something before it reacts to the fan issue.

General hint on fans:
- Most mobos fail to even register a signal below a certain rpm like 1000. That can happen with really low noise or temp regulated things like 80mm Socket A solutions.
- At the most, the system is entitled to bitch about it via an IBM speaker signal until you deactivate that in the bios hardware or "power saving" section.
- You're perfectly within your rights to attach any fan to the mobo header with all three wires with the help of god, some scotch tape and your teeth. Or even more sophisticated tools and parts.
- missing heatsink will lead to sudden death of Athlons but broken fan isn't half as critical.

Have a Biostar M7MKE on the heap that's dead. They tend to be that.
Don't throw it out right away, oftentimes it's an easy fix like replacing caps or voltage regulators.

I like jumpers.

Reply 5 of 6, by ptr1ck

User metadata
Rank Member
Rank
Member

In my experience, Biostar's aren't built to last. I've had similar experience with Soyo. I always pass on those when I see them available.

Your Iwill replacement is awesome BTW!

"ITXBOX" SFF-Win11
KT133A-NV28-V2 SLI-DOS/WinME

Reply 6 of 6, by stanwebber

User metadata
Rank Member
Rank
Member
Cuttoon wrote on 2022-04-09, 15:25:

Have a Biostar M7MKE on the heap that's dead. They tend to be that.
Don't throw it out right away, oftentimes it's an easy fix like replacing caps or voltage regulators.

oh yeah, i boxed it up. even in its non-working state, i'd value the board at at least $100. it's the only socket 462 ddr board, consumer or industrial, ever produced with an isa slot (and it's the more desirable raid edition plus the last board revision released if that matters). i think the fact that it failed and the psu survived makes it an excellent repair candidate down the road. i just ran across a great deal on a pair of kk266plus/kk266plus-r boards before i did anything about it. everything swapped over except the ram...didn't even reinstall the dual boot xp/98se OSes. ultimately, i did un-raid my drives, but even that wasn't necessary as i was going from hpt370 to hpt372.