Good to know, MTK358.

I've had a few computer with the same sort of 'ducting to direct casefan airflow over the heatsinks' setups as well. I dont count ducting as 'fanless' myself, I call it 'semi-passive' (and if you wwant to count semi-passive setups, I've run a lot, up Phenom II X2 550 (only a short test, not extended running).

There is a stage between- BIOS temp monitoring and shutdown. I've had it work for me (long story) and seen it work, and fail, for others.

72 watts TDP? Not for any athlon 64 model I've seen, and all the early models were rated at 89 watts TDP as far as I know. I'm pretty sure that all the athlon 64s have thermal protection. Never seen it used though, all the athlons 64s I've seen have run cool.

BTW, you might find this interesting-

http://www.dansdata.com/goop.htm

Only 4 wire 'PWM' setups pulse the the power as far as I know.

Out of sheer curiosity, how much down can the speed go for a modern microprocessor. I bet it can't be run at 0 Hz. That is, imagine a switch going to the clock pin of the CPU, and you manually pulse the switch on and off. Would the CPU still run your program?

As a matter of fact, this could be done with one of the earliest micros, the Zilog Z80 despite the fact the Intel said in the manual, "Although static by design, testing guaranties tW(phiH) of 200us maximum". Here tW(phiH) is Clock Pulse Width High. This micro, the Z80A, could be run at the impressive speed of 4MHz.

Depends on the CPU/Chipset.

Getting down to the minimum clock sperd that cpu-freq will downclock to is easy, getting under that number (e.g. 800MHz for AMDs) is more tricky.

So that means modern micros, contrary to Z80, are dynamic by design, as dynamic memories vs static ones.

Also, I see the manufacturer gives not only an upper limit but a lower one.

That's what I always thought, but then howcome it works for the OP?

I have checked that TDP figure in Wikipedia and you are right about the TDP: 89W. My bad memory.

I am not sure about thermal protection. I remember blowing up an AMD Athlon XP a few years ago. That was the first generation after Slot A, socket 754 I believe. I was so unbelievably stupid as to forget to remove the protective cover on the heatsink before installing it during maintenance. The only sound I heard from the speaker was a sort of "buuuurp" and that was the end.

I think I found the best solution regarding fans: none at all. This months I started to replace the ordinary computers at home with Shuttle PCs with Atom D510 installed. Totally fanless, small heatsinks which conduct the heat to a small radiator at the top of the case. The graphic chips is an Nvidia Ion. I am running Debian + KDE4 with all graphic effects effortlessly.

jlinkels

Well, the fans not working :P You wont find any P3/Celeron motherboard with a PWM fan control anyway.

Sorry, silly jokes aside, to say what I said in post #8 again- the way that 3-wire fans vary the rotation speed is by changing the voltage to the 'power' (12v) line.

No problem, its just a detail anyway.

The socket after 'slot a' was 'socket a'. Socket 754 never got 'athlon XP' CPUs, they were all athlon 64 based (though the early versions were 'crippled' to be 32bit only).

With socket a, the early boards had no overtemp BIOS protection. The early models with overtemp BIOS protection were varing in quality (depending on chipset, BIOS and manufacturer). IIRC soemtimes even if the board had decent, working overtemp protection, sometimes the BIOS could come from the factory with the overtemp set high enough that the CPU could be damaged, or even blow up.

*edit- I should have mentioned that most socket a boards didnt read the internal thermal diode, they used a probe under the socket. There could be huge difference between the diode and probe readings, with the probe beign slow to react and generally giving lower temps when it had reacted. So even with overtemp protection, if some really nasty happened (eg heatsink falls off) the CPU could blow before the probe had read a temp high enough to shut down. Bad. Thank the gods they've fix it with more modern CPUs.

Yes, I've spent far to much time working with and checking AMD CPUs.

I prefer 'normal' desktop CPUs, with quiet 120mm case fans, and decent power supply (better quality fans), installed in a heavy, good quality case to avoid noise myself.

I used to play with fanless setup often, but these days I dont bother. With an antec 'solo' case (taped up seams, some air holes taped up, a few other minor air/flow noise mods), 120mm case fan, corsair HX520 power supply, a Phenom II X2 550 (stock cooling!), and a GT8600 (passive cooled) my computer is pretty much 'whisper quiet'.

At least it is in the cooler part of teh year, in summer I bump the case fan speed from 'minimum' to 'medium/max' (depending on the temp that day). With the ambient temp hitting up to 40-42C here on a bad summer day, I'd prefer not to cook anything.

stf92, I take it you've had a close look at the fan blades for 'crud', dust build up.
I have a Intel Celeron 2.6Ghz processor. About once a year I unscrew the fan, four little screws on each corner, and give the blades a thorough clean. Build up of crud can make the thing weigh double what it should, obviously not what it was designed to run at.

Anyway depending on your room temp, having the case of, suppose all's o.k. You say your getting a repair tomorrow, personally I'd shut down the machine till then. Go read a book or something

1 members found this post helpful.

You've taken it wrong. Every time I look to my right (did you say its safer to have the machine cover off?) I see the poor fan. And just looking at it,it occurred to me what would happen if I gently push the blades. I immediately felt a force on my finger! So I keep pushing and... low and behold: the fan started spinning, ejecting a jet of dust in the process.

I had been changing the power source a day before, and putting the chassis upside down. As this machine, which was my sister's, is incredibly filled with dust, part of it has gone to the blades which already had a lot on it.

I feel like putting this in the Success Stories section: about how a finger solved a terrible situation.

It would be better to really clean that machine up.

Depends on the case airflow and fan setup. I prefer to never have the side panel off, its more noisy and lets more dust into the case (I have air filters where the air comes into my case)

The fan can sieze like that, from the bearings locking, the lube on the fan turning into sludge due to time, or even from dust.

Just because the fan started to spin when you pushed it doesnt mean that the fan will spin without a push next time. Even if the fan does spin, it might not be spinning as fast as it should be, and not giving the cooling it would if the fan was spinnign freely.

Clean it out. Seriously.

Don't worry, cascade9. It's been a long time since I have the idea of taking the machine to a friend's who has an air compressor and clean it all up.

Are you telling me the dust gets into the bearings?

It can do. Or it can stick to the fan blades, making it harder for the motor to spin the blade, and if there was already a problem with the bearings or lube, that can be enough to stop the fan spinning.

In really bad cases, the dust and bits of hair, etc. can stick the blades to the heatisnk, or actually make a solid block of dust and crap, stopping the blades from spinning.

Huh, what's going on here ?
08-19-11, 06:14 PM
Post #1.stf92
Uluru Yesterday, 12:12 AM :
stf92 @ 10:27 PM
.
Doh, I'm not sure what you mean.
1.)Please read the post you first wrote, explaining problem.
2.) Please read the highlighted pieces of my post.
3.) Please consider staring at a stopped Fan, used to cool a CPU, is harmful, and in all situations if you want to remain using your machine in the future, and the data on your computer, maintain and regularly check inside the box.
3.a) First shut down the Operating System, turning the Computer OFF, and disconnect the power to the machine, before doing anything inside the machine.

As I had said: check for build up of dust on the Fan blades, so common.
You gave it a push, the added weight needed 'a push' to get it spinning, it's Mass was larger than it was designed to run at.You do not have a 'Magic finger'
It is no miracle that the Fan blades spun. It's called the direct approach.
Us Technicians love tackling the simple, then speculating rationally if the obvious doesn't solve it.
What I've read here is a little pitiful I'm afraid to says; sincerely with the best of intentions my Linux friend.

stf92, you turn the Machine off, if it is in danger of overheating. Dust is a Computer's No.1 enemy. We live and learn.
Removing the fan over the Heat sink and ever so gently wiping with a tissue, or whatever 'nano appliance' you want, till clean would have solved the problem for you.
As a Tech I get the box off, and skillfully blow, or vacuum (not recommended for novices) the dust out off that MB, case, and all. These machines don't like dust.
Also, if using a pressurised Air 'Spray Pack', do be careful not to 'freeze' any components. Happy Computering.