
UNSOLVED
How to get around "Fan Failure" alert in GX110 BIOS?
The fan in one of the GX110 towers I have to work on (I'm doing my community service requirement for school) is going to fail; I know it will fail in the relatively near future because it's making noises like the bearings are going. So, I bought a Sunon KD1208PTB2 80mm 39cfm fan to replace it; unfortunately, the BIOS is convinced that since the SUNON fan spins at 3500rpms instead of whatever the stock fan did, that the fan is dead... Is there any way to get around this?
//Edit
Shoot. Sorry... Didn't think of posting this in the "OptiPlex" section
Message Edited by Geeky1 on 04-11-2003 11:54 PM
Responses (27)
Solutions (0)
Unfortunately, I don't think it is. It was donated to the local Habitat For Humanity a while ago, but I have no idea how old it is. Also, while I'm thinking about it, the system just died a few weeks ago. It powers on, but no POST, nothing. So, I pulled the CPU (Slot 1, P3 600/133) and replaced it with the CPU in their other OptiPlex GX110 (which is also a P3-600/133) and it booted without a problem. Obviously, the CPU is dead. However, I can't figure out why it would just die when it had been working fine. No one had done anything to it (except change the fan) at that point, and it was running fine for a week or two with the SUNON fan in it (except for the BIOS error), so I'm kind of nervous about putting another CPU in it... have there been cases of the voltage regulators on the GX110 boards failing and frying the CPU? I can't think of anything else that would cause it to die... oh, I should mention that I did put the dead cpu in the other GX110, and it refused to boot with that CPU too, so I'm positive it's the CPU...
AHA! I figured out how to get around it.
After doing some reading and some testing, I determined that you guys (Dell) don't use the 3rd lead as a fan speed lead the way all industry-standard boards and fans do. Instead, at least on the GX110, when the fan is spinning, that lead carries +12v (iirc- i did the mod a while back). When the fan is stopped, the lead carries no voltage. So, i simply soldered a jumper wire between the +12v pin on the fan header and the sensor pin. Problem solved.
vivessence
1 Message
940
0
Posted August 18th, 2004 23:00
Hi,
I tried you method and it worked for me as well. Thank-You.
What a pain in the but this error has been.
Best regards,
Pikaso
simonmccaughey
5 Posts
940
0
Posted October 6th, 2004 08:00
Excellent!! This is a great help, thanks for sharing your finding!
Simon
All Brothers of Electricity and Terrabytes,
NO WARRANTIES OR LIABILITIES EXPRESSED OR IMPLIED BY THE CONTENT OF THIS MESSAGE - USE THIS INFO TOTALLY YOUR OWN RISK!
I only tackled this problem because I'm a knothead, and wanted to fix this so that I could use a specialty heatsink/fan combo on a GX110 low-boy case.
It seems that Dell uses a special fan that does not put out a industry-standard tach signal on the third wire, but instead senses that the rotor is stalled and changes the signal, based on fan running/stalled condition, up the third wire (if I am reading the fan specs correctly, it is a 'sinking' circuit, such that when the fan is running, the third wire is sinking its voltage/current close to ground potential, and when it fails and stalls, it releases the 'sink', bringing the third wire up to a positive non-ground voltage).
If you use a regular industry-standard fan that enforces a tach signal in the third lead AND connect that fan's third lead to the mobo third pin, that pulsing of the tach signal on the third wire applied to the third pin on the mobo, is just enough kick (non-ground potential) for the bios to think that the fan is stalled. Then you get the obnoxious 'fan alert' message at boot time.
When running properly, a Dell-branded original fan will sink the third wire near to ground, and when the fan stalls, the fan forces the third wire and mobo pin positive to some level of voltage that the bios interprets to be a stalled fan (looks like a pulsed signal in the specs for a stalled fan). "Free running" the third pin on the mobo (i.e., disconnecting the fan's third wire from the mobo third pin) results in 5vdc appearing on the third pin of the mobo (possibly a leaky design flaw or possibly intentional??), which again is a non-ground signal, and the bios thinks that the fan is stalled. So disconnecting the third wire is not a solution in itself.
My fix - if you replace the fan/heatsink with a fan that is a taching fan, first thing, cut the third wire from the fan to the plug (Not close to the plug, dummy - you will need some wire on the plug-end to finish the fix - read on!). You can't use the taching signal - end of story. Tape up the fan-end of the third wire to insulate it from the world, and hide it away. BUT you also need to re-wire the fan plug so that the third wire from the fan plug is soldered to (spliced into) the black ground wire of the fan plug. This forces the mobo (and subsequently the bios) to always see ground potential at the third pin of the mobo, effectively lying to them that the fan is running, full-bore. Get the splice in place, insulate it good, and replug the fan in.
Once you have this in place, the pc should boot without the obnoxious 'fan alert' messages, and you shouldn't have to ever press f1 to bypass that righteous bios error message. BUT BE AWARE THAT THE FAN IS NOT REPORTING ITS RUNNING STATUS TO THE MOBO, AND THAT YOU ARE RUNNING THE RISK OF A HEATSINK/FAN FAILURE THAT WILL GO UNREPORTED.
AGAIN, IF YOU ARE NOT SKILLED IN THE WAYS OF ELECTRICITY, LET A BROTHER THAT IS REPAIR THIS FOR YOU. THE CHEAPER SOLUTION IS TO BUY A FAN FROM DELL.
Knowledge is power - Use your powers wisely, my sons.................Geeky1 and Milkey 6400,
The fan noise in a Dimension 4300 has been almost unbearable since the day I received the unit. Tried this fix and installed a third party fan. Everything is working fine, post error is fixed and the fan noise has disappeared. The machine is finally quite!
Excellent Post
Thanks for the help.
Morit

DELL-Cody
2 Intern
•
2166 Posts
940
0
Posted April 14th, 2003 16:00