Dumbest Question of the Day

[Braegnok]

Did I cook my chip? A few days ago I started my rig with my waterblock pump shut off, twice I had F@H set in start menu to start both SMP and GPU WUs at start up @100% CPU.

The system blue screened within seconds of start up, then second time I started, I saw pump lights were not on and as I was reaching for the power switch got blue screen again. I fixed the bad Molex connector and restarted, every thing started OK . I ran a sensor test and seems my movements changed from 8-8-8-7 to 21-19-12-9 Then I checked core temps, and noticed core 0 and core 1 run 16 c hotter then before the [no pump issue] , I checked online about sensor movement and found " typical movement is between 3-5, A difference of + - 2 between two cores is typical, Grater than 3 is a warning sign. That the sensor is stuck". So it seems my sensors are stuck, how do I unstick them?, or is being stuck just causing inaccurate temp. readings and no big deal?.

This chip running at 3.2 GHz always ran cool 27c idle and 50o load on all 4 cores. Now at 3.0 GHz it's running 70c on cores 0 and 2.

Ambient air temp @ screenshot 28c Thanks for any help or suggestions. Edit: also my core MHz during testing bounce all over from 800MHz to 3125.25MHz under load, see cpuz MHz on screenshot.

Edited September 6, 2012 by Braegnok

[Coors]

Odds are that it is still good. They're usually pretty good about shutting down when they get to hot before any major damage is done. If you can still run a stability test without failing I wouldn't worry too much about it.

[Braegnok]

Odds are that it is still good. They're usually pretty good about shutting down when they get to hot before any major damage is done. If you can still run a stability test without failing I wouldn't worry too much about it.

Thank You, It will run Prime 95 Blend, or F@H all day but runs hotter then before on two cores, and CPUz shows MHz constantly bounce around and change from 800MHz to 3125MHz every few seconds like a strobe light.

[dr_bowtie]

question is why would you have your pump switched and not wired to turn on when the rig is powered on? but what Coors said too

 

you prolly cooked the TIM so a re-apply and re-seat is in order.... its prolly got a contact issue now

[Braegnok]

question is why would you have your pump switched and not wired to turn on when the rig is powered on? but what Coors said too

 

you prolly cooked the TIM so a re-apply and re-seat is in order.... its prolly got a contact issue now

Thank You, I pulled my QX9770 chip and put in my Q6600 chip three days ago, when I did the swap out, I unplugged the H100's fan plugs and the power connector at the backside of case, for more room to work, then when I plugged the Molex connectors together, one of the pins pushed up out of the Molex connector enough to Not make contact. But looked good and seemed properly connected.

To re-apply and re-seat, Do I just remove chip and reinstall? Thanks That's Great News I thought I Killed it.

[sword_fish]

you don't have to remove the chip from the socket anyways, just wipe off the old TIM with a cotton dampened with isopropyl alcohol, reapply TIM then you're good to go.

[Braegnok]

Thanks Everyone!, I re-applied the thermal compound, my chip is back. And running just like before,... idle temps all 4 cores @ 27c and load 45c with sensor movement @ 8-8-8-7, :woo:

 

You guys are Great!, I'm back to folding now with 42c all four cores. I thought for Shure it was a goner.

[ir_cow]

I had the same problem with the h100 , the molex connector came half broken and you had o reseat the pins . Beware other people

[Braegnok]

I had the same problem with the h100 , the molex connector came half broken and you had o reseat the pins . Beware other people

Yes, From now on, I will always take a vary small flat head screw driver, and re-set my pins after connecting the cheep Molex connector. Seems like they can be vary easily pushed out 1/8 to 3/16 inch when connected, and looking normal. I hope my Dumb question helps someone.