i just put together a system that enables me to have separate systems, securely separated from each other, using a multiple physical drives and removable drive bay.

COMPONENTS
[list]motherboard - Asus M3A78-EM
CPU - AMD Phenom II X4 940 (125W)
memory - Crucial CT2KIT25672AA800 - 2x2GB unbuffered ECC DDR2-800 1.8V CL=6
hard disk - Hitachi 320GB 7200rpm 2.5â€

Quite a nice clean build.
Good job !!

Nice. I have an almost identical board; mine's the nVidia version: M3N78-VM. I also have fairly cool chipset temperatures. If the reported sensor value is accurate, it's usually about 38C. Same chipset cooler as yours.

Given that you don't have a discreet graphics card, nor any hard drives in the upper chamber, why on earth do you have all those fans in there? It strikes me as a bit of overkill. You have two fans in close proximity to your Ninja, and I'm wondering if both are necessary. I'd think you could run just as well with two fans as with three.

I encourage you to play around with fewer fans; start by taking out the orange Nexus intake fan and see what that does to your temps. Then maybe try the rear or upper exhaust fan. Your machine already has pretty good potential for cooling, given the lack of additional power-hungry parts besides your CPU.

Why aren't you using the CPU fan header? I'm also cooling a 125 Watt CPU + GPU with Ninja passive using only 2 anemic Noctua fans. I would not have been able to do that without thermostatically controlled fans. The Asus BIOS q-fan thingy ramps up the fans only when needed, but under normal use they run ~800 rpm which makes for a near-silent machine. I'm also running micro-ATX with 2 2.5" drives, so the two systems should be comparable in terms of noise and heat.

http://www.silentpcreview.com/forums/vi ... highlight=

i always try to build redundancy into my cooling systems, since i've had several system failures over the years caused by fan failures. with my strategy of "all components passively cooled; multiple low-speed case fans," any single fan failure should not stop or damage my system (i hope). i guess i would not disagree with the assessment of "overkill." call it intentional overkill.



i'm sure all my fans are well under 800 rpm now. still, i may take your advice and try using Q-fan to reduce noise even further, while keeping the same number of fans.

one concern: CPU and northbridge temps are fairly independent of each other, but Q-fan only looks at CPU temp. so if you had a graphics-intensive program that used the CPU only lightly, Q-fan wouldn't speed up the fans, since the CPU would remain cool, even as the northbridge heated up from the graphics load. for this reason, i think i only want to let Q-fan regulate the rear-facing 120mm fan, which i think is the least necessary of the 3 fans in my system.

question: if you have fans plugged into the "CPU fan" and "chassis fan" connectors, does Q-fan regulate both of them, or only the CPU fan?

The 200mm blowhole fan on top connected to the CPU fan header will suck up a storm through that case. That should be sufficient -- a one-fan machine! If that creates too much vacuum, then connect the 2 120mm fans in tandem as intakes to equalize the pressure. CHA_FAN header is not regulated on that mobo.

I think you worry too much about the northbridge, mobo manufacturers use adequate heatsinks, otherwise they'd be replacing them by the thousands under warranty, so I'd leave that up to Asus. If airflow is sufficient to cool a 125W CPU then it's plenty good for the northbridge.

well, i decided to experiment with Q-fan, but i was unable to scare up a pwm fan, and the MB doesn't do any voltage control on the CPU_FAN header (i tried).

interestingly, when i turn on Q-fan, the BIOS undervolts the CHA_FAN connector. not sure why it does that, but it is just a constant undervolt: it runs about half speed, and it doesn't vary with temp.

anyway, with just the static fan controls, this system really is acceptably quiet, especially since its ultimate place is going to be under my desk. i think i'm already at the point of diminishing returns; i'm not sure i'm going bother ordering a pwm fan just to get a little further out on the noise asymptote.

CPU header works fine with 3-pin fans. Set BIOS q-fan to:
auto mode start voltage = 5 v or lower
auto mode start speed = 30 deg
auto mode full speed = whatever AMD says is the max core temp for your cpu

Do use q-fan, by running fans at constant speed you're forfeiting a lot of technology that's built into your mobo, and you risk frying your cpu.

If you do connect the 2 120s in tandem as intakes, I'd love to have a reading of your temps with/without them connected. Also, in the end you'll find that the top chamber is cooler and has more airflow, so I'd move the hard disks upstairs too, and suspend them while you're at it.

//

using a 3-pin Nexus D12SL-12 fan, connected to the CPU_FAN header, with start voltage 5V, i tried:

(1) Q-fan disabled
(2) Q-fan enabled, start @ 40C, full speed @ 66C, cpu temp = 36C (fan should be slow or off)
(3) Q-fan enabled, start @ 25C, full speed @ 51C, cpu temp ramped from 36C to 52C (fan should range from from med to hi speed)

in all cases, fan speed constantly stayed at full speed (very slight variations around 1140 rpm). there was never any change, at any time.

i'd assumed that Q-fan only worked with 4-pin pwm fans. if, in fact, it is supposed to work with 3-pin fans, then i don't know why i'm not seeing any fan speed changes.

It should work. If the fan you're referring to is the orange one in the front, then it won't be able to affect cpu temp from that location and will run at full speed after cpu reaches 50+. Disconnect all fans, and plug just the 200mm to the cpu fan header, and see what happens.

did a search thru the 600 reviews at newegg for the M3A78-EM. confirmed the behavior i'm seeing. need a PWM fan to get Q-fan to work.

i'm done polishing this apple. it's already a great system in its current form. i'm declaring victory and moving on...

Hmm RMA time lol, that's unusual, 3-pin fans worked on all Asus mobos that I've used. You can try fan control with SpeedFan free software, it can adjust fan speeds based on CPU/GPU/case temps.

I have a similar board Asus M3N78-VM. I can control 3 pin fans connected to the chasis fan header with speedfan or q-fan but not a 3 pin fan connected to CPU header. I also assumed that it's because the CPU fan header needed a 4 pin fan with PWM.

On another note, my northbridge is blazing hot. My mother board is sitting on a desktop (antec fusion case) with a mini ninja passively cooling the CPU. I have a tuner card that is practically sitting on top of the northbridge heatsink. I am reading the temps of my northbridge using speedfan. I am idling at 75 C, and with watching blueray temp goes to 90 C. The north bridge heat sink is not getting any air flow. You say that you cannot install a thermalright HR-05. Have you tried? Are you sure? How about a thermalright HR-05 SLI or xigmatek PTR-N881 HDT.

http://www.overclockersonline.net/index ... es&num=641

http://www.newegg.com/Product/Product.a ... 6835233021

i did not actually remove the stock northbridge heat sink, but i did place an HR-05 IFX directly above the stock heat sink, to see where it would lie on the board. i'm quite certain that the heat pipes conflict with the capacitors that are tucked in under the fins of the stock heat sink (see the first photo on my original post above).

i'm farily sure, but not 100% certain, that the other versions of the HR-05 and the xigmatek would conflict with the same capacitors.

it appears that your motherboard does not have capacitors tucked in so tightly, so there's probably a good chance that the HR-05 will work.

This is a known problem with newer Asus motherboards, same behaviour has been observed here.. a bad design choice, Asus, boo!

I have the same M3A78-EM board.
The CPU header is PWM only, which I tried using AC's "PWM sharing" technology with the exhaust AC 12 PWM. The result is, no matter what temperature startup in the bios, either the fan would spin too fast, or the cpu idle/ load temperature would be too high.

After playing around with Q fan for a while, I scrapped the idea of using automatic fan controls, and hardwired resistors on my fans to undervolt them. The AC 120mm runs @ 700RPM, the AC freezer's fan runs @ 900RPM, and the Seasonic S12 II is left to ramp up on its own. Idle CPU: 34, MOBO: 41, GPU: 59, HD: 38. Load CPU: 52, MOBO: 42, GPU: 90, HD: 38 *C.

All on an x2 5600, with an 8600gt silent, stuffed in a SOLO

gpu's a little warm there.
maybe time to install a 92mm in the lower front fan mount of the solo?

I've completely removed the front fan cage, filter, and metal 5.25 bay covers to allow for a larger air intake and decrease the whooshing noise.

For a stock Asus 8600GT silent, I thought 90C tops is great, compared to what the SPCR review got of the card. Its the PSU's ramping up what gets the air really moving through the case under load that keeps the card (and CPU) cool.

The mobo temp sensor I bet is somewhere near the RAM or VRM's on the board, because I did notice an increase from 33C to ~41C idle when I changed my top down stock cooler to the Freezer 64.

Hey folks,

I found this thread while searching for info on the fan contol on the ASUS M3A78-EM. The 4pin CPU header does not speed control a 3pin fan. However, I've found that the 3pin chassis fan header works fine for a 3pin CPU fan as QFan does change its speed according to cpu temp (or mobo temp which is related closely enough IMO).