Sign in with
Sign up | Sign in
Your question

Zalman CNPS9700-NT on Q6600 - strange behavior

Last response: in Overclocking
Share
May 26, 2007 1:31:04 PM

Hi guys,

I am rather new to all that stuff, so please forgive my ignorance: I have build a system with Zalman CNPS9700-NT on Intel Q6600 and almost died from scare when I turned on PC and system fans started working fine but bloody Zalman did not - but it was showing signs of live by lighting up from time to time and moving fan a bit, but not working as one would expect.

However after a period of time, perhaps 40-50 secs it will start working - when looking at BIOS page showing system health it seemed that cooler fan was starting working when CPU reached temp of 36C. After fan started it seemed to keep going, however after a while on idle OS it stops again - CPU temp is reported at below 30C.

Is this normal behavior? I really freak out at thought of a quad core CPU being on a cooler fan that may stop beause it thinks CPU is cold enough or am I overreacting here? 8O
May 26, 2007 3:21:32 PM

8O

The motherboard is Gigabyte GA_965P_DS3P.

Zalman is connected to the place on mobo where it should be - and cooler works but only if temp reaches around 36C, right now the box is idle and CPU temp is reported to be 23 C, cooler on it is not running 8O

I can confirm that cooler connection on mobo is 4 pins just like on Zalmans.

Is this normal at all? I always thought that maybe revs will drop down on idle, but surely not to 0?
May 26, 2007 3:36:05 PM

Quad is idle in this situation and case temp is reported as around 33C, I can only guess that good airflow going through cooler makes it sufficient for idle speed - I actually have got wattage used by the system reported and it is less than 100 watts - this is only mobo, CPU, PSU, couple of drives and that's about it - so actual power usage is very low, all on idle though.

I tried switching off smart fan control in BIOS, but this made Zalman run at top speed as if it was trying to take off! 8O

Not ideal too - perhaps there are software utils that allow to specify min fan speed to override bios?
Related resources
May 26, 2007 3:49:36 PM

Ok thanks for that.

It is really strange even though all seems to work - the power usage gadget that I have and I know it works correctly would show very low power usage on idle - less than 100 watts, and it was 130 watts whole system usage when installing Windoze. :?
May 26, 2007 4:02:56 PM

The main worry I have is that when I Start loading server 24/7 the cooler might think cpu is cooler than it is :?

So, can it be said definitively that when computer is turned on (not sleeping etc) then Zalman cooler should be working even if it is at very low revs - it does start revving at 500-600 rpm after a minute or so now, other 2 system fans are constant at around 1400 rpm.
a b K Overclocking
May 26, 2007 4:38:05 PM

why not try speedfan to slow it down(with smart fan off)

I know my AC Freezer sounds like its gonna blow at full speed, yet its almost silent @ 80%

SpeedFan download
never hurts to try, and it can check your cpu temps too :) 
May 26, 2007 4:50:14 PM

I second the Speedfan idea. It is a great program and very customizable. I've found it works much better than the motherboard's bios fan control.
May 26, 2007 5:16:02 PM

Is your CPU running at the proper speed? It could be that your CPU is running so cool because it's detected at a much lower speed than it should be running.

Check your CPU multiplier and FSB speed to see if that's where it should be. Also check your voltage, it could be undervolted.
May 26, 2007 5:46:21 PM

Thanks for the idea guys: changed BIOS to disable Smart Fan so it starts spinning it pretty quickly at max speed, this ought to cool the damn thing even if temp values are all lies - system is reported to be at 32 C, but CPU is at 11 C 8O

The CPU is detected correctly as Quad running at 2.4 Ghz - Windows XP Pro 64 bit sees it as 4 cores, memory runs at 667 - 2 x 2 GB.

What really cheeses me off now is that when I added more disks to the system it became unstable - windows crash etc, now thatI unplugged them all it seems okay again - I've been programming these things for good 15 years, but they still can't make them fool proof that would work all the time :x

Voltages reported by SpeedFan:

Vcore1: 1.14V
Vcore2: 1.84V
May 26, 2007 5:57:14 PM

Neither of those Speedfan voltages are likely correct. Try CPU-Z to check the voltage.
May 26, 2007 6:39:56 PM

CPU-Z reads CPU voltages inaccurately. Check to see that you set the vCore to be in the BIOS, that's what your CPU voltage is. Your voltage should be at least 1.3 for a quad core.

Tacos is right, a thermalprobe is a good way to check the actual temperature. Here's a good guide for installing one. You don't have to mount it to your case if you don't want to, the most important part of the process is where you place the thermistor.
May 26, 2007 7:00:13 PM

I have the vcore set to 1.35v, and CPU-Z is detecting it at 1.264v.

Quote:
Not quite true, a lot of the time it's not cpuz, but the motherboard, many asus motherboards have huge vdrop and vdroop problems :twisted:


That may be, but a drop of over a tenth of a volt? I'd RMA the board if it was that much, that doesn't speak well for the rest of the board's reliability. I prefer to take the CMOS's word for it that I'm running at 1.35 than believe that my CPU is running 500MHz over stock, stable at 1.264.
May 26, 2007 7:15:12 PM

But not to the extent of putting out 0.8v at all times even when set to 1.55V in BIOS and overclocking :p 

CPU-Z sucks for voltage measurement. Infact, the only thing I would rely on for this is a good multimeter :p 
May 26, 2007 7:39:32 PM

I have the zalman 9700 led and it came with a fan mate 2 controller that allowed me to select the speed of the fan with a knob.
here is a link for it on newegg its only 4 bucks and is included with the 9700 led so i imagine it would work fine on your 9700

http://www.newegg.com/Product/Product.aspx?Item=N82E168...

hook it up, turn fan control off in bios, select your speed and then it just works.
May 26, 2007 7:40:49 PM

You guys are way more hardcore than me 8O

Had more instability traced to yet another 2 GB memory chip that seems faulty - 3 out of 4 (2x 2 gB OCZ memory), darn my luck - and one of the 1 TB drives was recognised as 30 GB one and then some crashes happening until I unplugged that drive 8O :x

But now it seems stable with 2 GB RAM and 4 drives - idle system power usage is just 130 watts, I was expecting more - when running benchmark loading fully all 4 cores it jumpst to around 180 watts, so, assuming quad core itself is 120 watts when fully loaded, then during idle time it was using 120-50=70 watts, perhaps that's why it was reasonably cool - note that Zalman's cooler is big and it is placed right on the way of airflow blowing into it sucking air from case, and back fan takes it out - so perhaps this airflow was sufficient to cool most of idle CPU requirements without having cpu cooler to spin around more? :?

Out of interest I enabled Intel's management of smart cooler in Award BIOS and it started spinning cooler on average of 1200 rpm, however that mode seemed unstable, but at the time I still had faulty chip and bad drive, but bad experience made me pull back to default settings :cry: 
May 26, 2007 8:47:43 PM

I think I will leave things as is at defaults - fan starts kicking in if CPU is used, so at least this works: much bigger problem is faulty memory that I am currently memtesting and one failed disk that somehow was causing additional system instability until I switched it off 8O

Thanks for trying to help guys, looks like it is all just my "luck" 8O
May 26, 2007 10:28:18 PM

I am running memtest right now: 2 x 2 GB chips from OCZ on their own tested fine, now running memtest of them together - so far so good, but the comp won't boot if other 2 chips are used, the funny thing is that 2 chips that work are from different pairs, memory is this: http://www.aria.co.uk/SuperSpecials/Other+products/+OCZ...(2+x+2GB)?productId=26604

Now one drive seems a failure - have not tested on another PC yet, will do tomorrow, however 5 other drives are fine, so I am inclined to think this is a genuine drive fault, either way when it is switched off things were okay so long as I used one memory chip, now testing 2 if all works will try tomorrow.

The motherboard is supposed to be good - power supply is good too - 500 Watt, so power wise all should be good. :?

The CPU itsels seems to work fine, at least memtest runs very nicely on those modules with which I was able to boot that is. 8O
!