fan runs full speed at start up

[scotsilv]

The new CPU seems to have cured the intermittent problem.

The old CPU apparently had a flaky temp sensor shown by SIW at TMPIN1 (probably the internal Northbridge-function components) that, on occasional powerup, gave a false reading (obviously) of 99C and caused the fans to spin to high heaven as a result of the fan control algorithms.

It seems your computer must also have had a defective CPU.  I am told by a salesperson at Micro Center here that Intel checks batches of CPU's by sampling; therefore defects do slip through, and he knows of other CPU defects.

[mpw101]

I am glad it all worked out. 

The mystery is why HP arranged for you to send your computer to the factory for a new CPU and then sent it back with yet another new motherboard and no new CPU.

I would not have liked buying a new CPU myself.  And I'm pretty sure I could not have installed it myself.  Probably would not have tried.  Still think HP should have done that for you while your computer was under warranty.

[scotsilv]

> The mystery is why HP arranged for you to send your computer to the factory for a new CPU and then sent it back with yet another new motherboard and no new CPU.

Because their repair people didn't look at the very detailed Repair Request I wrote out, saying the problem was the CPU, and apparently whoever I'd spoken to on the phone either failed to enter the CPU replacement request into their computer, or the techs didn't look!  (Only they know for sure.)

> Still think HP should have done that for you while your computer was under warranty.

I could have sent it back a third time - I did get a shipping box - to see if I got lucky and they actually replaced the CPU (and would likely have had to repolish the cabinet once more) - but I was fed up.

To save aggravation (which is worth lots) and time , it took 2 days to get a boxed G630 on EBay.  Now I have a reliable, unscuffed, pleasant to use computer.

I'm using it right now.   No more fan speeding.

[scotsilv]

Oh, and installing is really easy.  All you need is a screwdriver and a tube of thermal paste.  Unscrew and remove heatsink/fan. The CPU is socketed and a little latch releases it, and the new one drops right in.  Relatch, clean the thermal paste off the heatsink fan, apply a pea sized  dab of paste to the top of the new CPU (some people spread it around a bit with a plastic edge as I did but it's probably not needed), replace and resecure the heatsink/fan, and you're done.

[scotsilv]

To anyone else experiencing this problem:

I take it back.  It's not a bad CPU.  This morning, on powerup the fans went fullblast.  After a ctrl-alt-del reset before boot started, the fans went back to normal and have been normal on several more powerups.

Now with two motherboard replacements and a new Pentium G630 CPU boxed Intel CPU, there's not much left that could be causing this.  It's not the Power Supply, as I had substituted a good, almost brand new Antec Basiq 350W supply when the problem first started, and that made no difference (I then put the original Bestec supply back in and returned the Antec to the other machine).

I thought maybe one of the fans could intermittently not be sensed at powerup causing the problem, but when I disconnected each fan, there was actually a BIOS warning "CPU fan not sensed" or "Cabinet fan not sensed" ... "To boot hit F1."

When I hit F1 to continue boot, the remaining fan was at normal speed.  (I did not leave the machine on very long in this state, lacking on or the other fan.)

Finally, I did note the terminal of the wire coming out of the left top side of the P7-1222's wireless module, which ash both WIFI and Bluetooth, was shorting to a little shield.  It rotates so I turned it slightly so it was not touching the shield.  I don't know how that could have caused the fan symptoms, and it may have been that way just since the last motherboard replacement, but I did it anyway and checked for proper Bluetooth and Wifi functionality, which both are well.  I usually leave Bluetooth off anyway.

According to another diagnostic, HWINFO, TMPIN1, which reads 99C at faulty fan-fullspeed powerups is a reading from the onboard FINTEK F71858 LPCIO chip, see attached image.  That chip also controls the fan speeds.

My conclusion is the most likely scenario is a BIOS incompatibility between this motherboard and the Pentium G630 stepping, that causes this wierd "glitch."  Something is occasionally not being initialized properly at powerup, I believe, perhaps due to a timing issue compaed to the faster Core I3+ chips usually used on the MB.  The G630 is a low end oddball chip for the lowend P7-1222.

Both BIOS 7.19 (which my first motherboard had) and 7.17 (which the two replacements have) exhibit the problem. (The supposed latest BIOS for the board at the HP site is 7.16A.)

I am just going to use the machine until it dies.

 

[scotsilv]

This all has me thinking...

The MB was replaced twice...the CPU has been replaced...still the intermittent fan racing on powerup persists.

The only electronic component that was not replaced was the wireless Wifi/Bluetooth module that plugs into a little receptacle on the MB.

A burst of RF in the GHz range, especially if going to the wrong places, can mess up digital circuits (as a ham radio operator I've seen it, in different contexts).

Perhaps that wite antenna lead coming from the module that was accidentally touching the module shield, until I rotated it yesterday, was the culprit?

Time will tell.

[scotsilv]

Saw other reports like this, e.g., at Tom's Hardware:

--------------------------

Fan Speeding up to highest speed right after booting

bladehopper

March 31, 2012 11:04:14 AM

Just wanted to add that I just had a similar problem starting this morning.
My HP Phoenix 1060sc system (5 months old) had fans running at max speed from boot.

SpeedFan showed that the TMPIN1 was runnning af 99 deg. celcius / 210 deg. F.
The CPU didn't get warm though so I expected something else was wrong.

Several restarts later I tried running CPUID that showed the internal temp. of all 4 CPU cores (i2600) and they were low (30 deg. C).

TMPIN1 is reading 35 deg. celcius now... much better:

I found a manual on the net for my Motherboard (Petragon - ASUS manufactured?) and found the pin to reset the BIOS. From the next boot everything was good. A nice an quiet computer...

I had to remove graphics card and cabling to get to those jumpers so but it was well worth the work  


yihaar!

/dann

--------------------------

Doing some investigating, I fidn this in the Data Sheet for the Fintek F71858 LPCIO chip ("Hardware KBC with ACPI, Temp. and Fan Control") used on numerous HP/Pegatron motherboards, including the Carmel/Carmel2 and other HP boards that use the same basic circuit layout:

=======================

8. PCB Layout Guide

F71858 adopts Current Mode measure method to do temperature detected. The measure data will not be affected by different process of CPU due to use current mode technology. This technology measures mini-voltage from the remote sensor so a good PCB layout must be cared about noise minimizing. The noises often come from circuit trace which is a track from remote sensor (CPU side) to detect circuit input (F71858 side). The signal on this track will be inducted mini-noises when it passes through a high electromagnetic area. Those effects will result in the mini-noises and show in the detected side. It will be reported a wrong data which you want to measure. Please pay attention and follow up the check list below in order to get an actual and real temperature inside the chip.

...Place the F71858 as close as practical to the remote sensor diode. In noisy environments, such
as a computer main-board, the distance can be 4 to 8 inches. (typ). This length can be
increased if the worst noise sources are avoided. Noise sources generally include clock
generators, CRTs, memory buses and PCI/ISA bus etc

=======================

I think a possible noise source is the internal PCI Express mini-Wifi card - especially considering the MB has been replaced twice, and now the CPU.

From my ham radio experience, such noise can be a tricky thing.   I moved the wireless mini-coaxial cables around.  Let's see if that cures the intermittent fan speed issue that could be from noise getting onto a bus at powerup and affecting the F71858 readings.

[scotsilv]

Followup.

I replaced the wireless PCI mini card in the machine for an identical unit from Ebay.

Fan speedup problem at powerup seems gone.

The wireless card in the machine seems to have been the culprit, perhaps occasionally sending a noise spike on powerup onto the bus, thus affecting the Fintek F71858 LPCIO  TMPIN1 reading (per mfg. as in prior message) and making it 99C, causing fans to race.

I hope this information is useful to someone with a similar perplexing problem like this, where multiple MB replacements and a CPU replacement did nothing.  (I reinstalled the original CPU and am probably going to sell the Intel boxed one).

I also add - the Seagate 500 Gb drive the machine came with has developed 8 reallocated sectors.  So I replaced that with a WD blue drive -- so now the machine internals other than the cables have been entirely replaced!

[scotsilv]

It's been almost two months since I replaced the wireless card and the machine has not had a fan problem.

Case closed.

[mpw101]

I don’t think you have an HP computer anymore.  I think you have a scotsilv computer now.  I can’t even follow the technical talk in some of your posts.

I am so glad I was fortunate enough that HP replaced my computer when I had this problem. I could never have spent the time, money and expertise and come up with a solution myself. 

I know the case managers in Canada who helped me were frustrated when I kept calling because the fan kept racing intermittently.  They have to close the case and can’t keep it open to wait and see a couple weeks, so I kept getting new case managers because sometimes it was that long before the racing fan would happen again on start up.  But it all worked out and HP sent a replacement and I have had no problems at all since then.

I am glad you finally found the problem and were able to fix it yourself.  Perhaps all your efforts will help others, including HP techs.