I have just got one of the new Vostro 430's with a Core i7 860.
When the machine first boots up, the fans spin upto full speed, which is loud but ok, as once Windows loads, the fans drop to their idle state which is 1500rpm. I'm one of these people who leaves their PC on 24x7. Anyway, during the night, I can hear the PC noise increasing considerably and after a few days it's becoming unbearable.
After investigating i've discovered a design fault with the temperature sensor under to CPU socket.
When the PC is not under load and using HWMONITOR, the power usage of the CPU drops to about 24W and the fan speed is 1500rpm. The sensor TMPIN0 will eventually drop to 0c, when this happens the fan speeds increase to 4500rpm for the CPU fan and 2400rpm for the case fan which generates an excessive amount of noise. It will take a while for the sensor to drop to 0c, but once it does happen, the excessive cooling causes the sensor to drop to -2c and therefore the problem with the noise will not stop.
During the day the sensor will go above 0c and when that happens the fan speed drops back to 1500rpm for the CPU and 1200rpm for the case fan, which is quiet.
I have also noticed that the temperature sensor under the CPU socket also fluctuates a lot, going from 5c to 16c in a second and back again, so the problem could be any of the following :
Solutions attempted so far :
A little background about myself, IT Consultant building PC's for 20 years since the old 286 / 386.
I too have just taken delivery of a Vostro 430 running Windows 7 and have exactly the same problem. When I turn the computer on the fan starts up at full speed and then periodically turns on and off.
Now, more by luck than good judgement, I rebotted the computer and pressed F2 to get into the BIOS. I wanted to see if there was anything that may affect this fan speed. Just by luck I turned off the "NUM LOCK" setting and pressed the Save to exit and continue the reboot. Much to my surprise the fan did not ever go back up to full speed. The following day, the same thing happened again. When the computer first started up the fan was doing the same thing again, periodically full speed and off. So I rebooted and again went into the BIOS with F2 and this time just did a save without changing anything. Lo and behold, the fan has now stoppped this on/off behaviour. More tests will continue as to whether this is a real fix and the exact sequence of events required.
My BIOS version is still at V1.0.1
What I have discovered is that the issue only occurrs if you have very low CPU usage and the speedstep kicks in to lower the power usage (24W). When the issue occurrs, if I run SuperPI to force the CPU to max speed (115w power usage), the problem stops immediately.
You can download HWMONITOR to confirm this if you want.
Spoken to Dell Support, and it seems this is a known issue and at this time there is no fix.
Dell are collecting my PC and refunding. They wouldn't replace with a same spec Studio XPS 8000.
Little bit saddened that I am having to now order a Studio XPS 8000, and it's costing me an extra £100 for the same spec machine, especially as I order about 20 dells a year and they couldn't at least offer to replace it for the same cost.
I have found that it does not seem to matter what the computer is up to it either does it or it doesn't. The impression I get is the following:
When the computer starts up something, presumably the BIOS, takes a temperature mesurement and uses this as the normal setting for controlling the CPU. So if you start the computer and the ambient temperature is 15degC, after about a minute the temperature has risen and the control mechanism kicks in to get the temperature back to 15degC. And so the cycle continues.
So what I have found that seems to work is the following:
Start the computer and press F2 to get into the BIOS setting.
Wait for about a minute for things to warm up.
Press "Save & Exit" to continue.
The fan speed now remains at a low speed regardless of CPU usage.
So I get the impression that when you exit the BIOS setup, a new temperature measurement is taken and becomes the new referance point (which of course is now much hotter).
Am I deluding myself or is this a possible explanation and solution until Dell get their act together.
Download HWMONITOR from http://www.cpuid.com/hwmonitor.php
When the fan speeds up, notice that the temp reads 0c for TMPIN1 and the power usage will read 24W. Then run anything that uses a lot of CPU, you'll see the temperature jump above 0c and the fans will reduce speed.
Downloaded the HWMONITOR program and whet I get here is the following:
TMPIN1 is constantly at about 19degC
TMPIN0 hovers at about 1 or 2 degC occasionally jumping to somthing higher.
Everything is fine (FANIN0 is about 1500 rpm)
Current thinking is this:
As soon as TMPIN0 goes negative the the temperature control mechanism goes haywire. Can the software hanlde negative numbers or is it treating it as a very larger positive number? The fan would obviously ramp up and try to cool the processor. The temperature gets even colder now at about -2, so you cant get out of the loop until you make the processor do some work and generate some heat.
Does this make sense?
I am now convinced that my last post was correct. The temperature control system can't handle the temperature going negativ and treats the reading as a very large positive value.
So to cut a long story short, my solution, which I shall be monitoring over the next few days, is just to turn the fan upside down. It is easy to do as it just unclips. No need to undo those 4 screws. Using the HWMONITOR program to monitor the value of TMPIN0, originally, under minimum CPU usage, this was hovering around 1-2degC. Reverse the fan and, under the same conditions the temperature hovers around 4-6degC, well away from the 0degC problem point.
As an additional test, I disconnected the fan completely and monitored the temperature. As it turns out, the combination of the large heatsink and the other fan do a pretty good job of keeping everything cool. However, probably not a good idea for long term usage. Also it does give a message when you reboot about the fan not working.
Glad you have confirmed my analysis of the issue, but as Dell have agreed that this is a known issue, they have also agreed to collect and refund my machine.
But I would be extremely cautious of reversing the fan to reduce the amount of cooling and stopping the temperature reaching 0c. From my other analysis relating to the correlation of temperature to power usage of the CPU. When the power usage is at 24w and Temp at 0c, fan runs at 4500rpm, when I start something that causes more CPU usage (eg. SuperPI) the instance the power usage jumps to 115w, then Temp jumps to 20c and fan speed drops to 1800rpm. I suspect that the temperature sensor hardware is in fact faulty and I would never expect to see a temp of 0c on ANY pc, also the ambient room temperature here is 15c, therefore I suspect the temperature is only being read correctly under CPU load. Running the PC with the fan in reverse may actually cause excessive heat generation and lower the lifespan of the CPU. I highly recommend you DO NOT run your fan in reverse!
You can disable the speedstep in the bios, which would in theory resolve the issue due to the increase heat being generated from the CPU, but as I run the machine 24x7, this would lead to a very large increase in running cost of the machine at the processor would use the full 115w instead of the lower 11w~24w under low power modes.
As such this solution is not viable, especially not in a business enviroment where we were discussing purchasing 30 of these machines.