Why isnt there a "silent" motherboard?

All about them.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
knutinh
Posts: 99
Joined: Fri Nov 14, 2003 5:53 am

Why isnt there a "silent" motherboard?

Post by knutinh » Wed Apr 13, 2005 4:41 am

With all the competing brands and "special edition" out there, I am surprised that no one offers motherboards specially targeted at our community? I am not shure how different they would be from regular motherboards or how many different models would be needed, but some ideas would be:

*mATX form factor (and possibily regular ATX)
*pentium-m, socket 939 and p4
*no chipset fan (of course)

Integrated m-cubed functionality. IE sensors (cpu probe, usually accessible only to the mobo, mobo temp, awareness of hd/powersupply temp). Cool n quiet or dynamic clock/voltage... All embedded in BIOS or a separate chip.

Offering:
*convenience (just connect everything, then set parameters in BIOS like desired cpu working temperature)
*stability (not a running application, but hardware)
*support for extremes. Possible to shut off "family mode" so every fan can be stopped (provided temp is low) or extreme overclocking (meaning that more users would be interested)


All in all offering dynamic performance/noise trade-off, keeping everyone happy. Since most of it is there allready, it shouldnt be too expensive either.

What you guys think? Did I forget anything?

Knut

sthayashi
*Lifetime Patron*
Posts: 3214
Joined: Wed Nov 12, 2003 10:06 am
Location: Pittsburgh, PA

Post by sthayashi » Wed Apr 13, 2005 5:12 am

So far you've described most motherboards.
*convenience - AOpen's SilentTek or Asus Q-Fan, etc.
*Stability - What motherboard will not advertise themselves as being stable?
*support for extremes - Suspend/S3 will stop all noises. Fanless operation is too risky otherwise with no good methods of measuring PSU temps. Most motherboards provide ways to overclock though.

m0002a
Posts: 2831
Joined: Wed Feb 04, 2004 2:12 am
Location: USA

Post by m0002a » Wed Apr 13, 2005 5:36 am

What does mATX have to do with silence? One usually gets better airflow with full size cases and boards.

Mariner
Friend of SPCR
Posts: 260
Joined: Thu Jan 06, 2005 11:25 am

Post by Mariner » Wed Apr 13, 2005 6:27 am

I think it's more a case of people become interested in silent PCs because of HTPCs. With an HTPC, the smaller the better for aesthetic reasons. The problem is, of course, that smaller doesn't mean better when it comes to cooling and therefore silencing!

knutinh
Posts: 99
Joined: Fri Nov 14, 2003 5:53 am

Post by knutinh » Wed Apr 13, 2005 11:16 pm

about mATX/ATX:
I think the PC industry is moving towards smaller. HTPC and cute desktop boxes is what consumers want ant what technology can deliver. But they/we want speed and low noise at the same time.

As opposed to when I was starting out, now IDE controllers, sound, graphics, networking, etc is available integrated on the mobo. And more and more peripherals is moving towards external usb/fw solutions. How many of you have more than 1xAGP, 2xPCI, 2xHD and 1-2 optical drives installed and in regular use? As long as a mATX system has the functionality we need (and the form factor that we want), we can concentrate on silencing it.

sthayashi:
I purchased a asus Q-fan mobo because of what you say. I found that my cnps 7000Cu cpu fan was running at a high speed (and the cpu was VERY cool), no matter what settings. In other words, it can`t go as low as I want. I think that the p4 internal temp is very accurate, so why would it be risky to turn the fan off???



Some off-the-shelf systems get it quite right. By integrating functionality and dimensioning everyting properly that can deliver ok performance and ok noise levels. Many "low noise" or "temperature controlled" solutions for the pc builder are poorly designed, so that fans are spinning no matter how cool components are. Many custom jobs include MANUAL control of fan speeds so that they can be turned up when one is gaming. How 1960s is that??

regards
Knut

sthayashi
*Lifetime Patron*
Posts: 3214
Joined: Wed Nov 12, 2003 10:06 am
Location: Pittsburgh, PA

Post by sthayashi » Thu Apr 14, 2005 5:46 am

knutinh wrote:sthayashi:
I purchased a asus Q-fan mobo because of what you say. I found that my cnps 7000Cu cpu fan was running at a high speed (and the cpu was VERY cool), no matter what settings. In other words, it can`t go as low as I want. I think that the p4 internal temp is very accurate, so why would it be risky to turn the fan off???
I'm sorry that your CPU fan does not go as low as what you want. It wouldn't be risky to turn that fan off, but there are other fans that ARE risky to turn off. Like the ones involved in cooling the power supply. You have NO way of monitoring accurately how hot that power supply is getting. So to shut off ALL fans is a very dangerous suggestion.

mathias
Posts: 2057
Joined: Sun Jul 11, 2004 3:58 pm
Location: Toronto
Contact:

Post by mathias » Thu Apr 14, 2005 3:27 pm

A while ago, I starteda thread like this one, except focusing mostly on more efficient voltage regulators. If any mods would like to move it to this forum, please do.

knutinh
Posts: 99
Joined: Fri Nov 14, 2003 5:53 am

Post by knutinh » Thu Apr 14, 2005 10:26 pm

If you read my post, I was asking for a mobo that did this intelligently and without any user feedback. I am not saying how it should be done or what changes are needed to the ATX specification to accomplish it.

If a powersupply has automatic shutdown on overheat, you would at worst risk loosing your data.

knut

[quote="sthayashiI'm sorry that your CPU fan does not go as low as what you want. It wouldn't be risky to turn that fan off, but there are other fans that ARE risky to turn off. Like the ones involved in cooling the power supply. You have NO way of monitoring accurately how hot that power supply is getting. So to shut off ALL fans is a very dangerous suggestion.[/quote]

sthayashi
*Lifetime Patron*
Posts: 3214
Joined: Wed Nov 12, 2003 10:06 am
Location: Pittsburgh, PA

Post by sthayashi » Fri Apr 15, 2005 8:04 am

knutinh wrote:If you read my post, I was asking for a mobo that did this intelligently and without any user feedback. I am not saying how it should be done or what changes are needed to the ATX specification to accomplish it.

If a powersupply has automatic shutdown on overheat, you would at worst risk loosing your data.
I don't think I understand what you're saying. You want a motherboard to do what is currently impossible, but you don't care how it's done?

knutinh
Posts: 99
Joined: Fri Nov 14, 2003 5:53 am

Post by knutinh » Sat Apr 16, 2005 12:34 am

I am saying that the actual implementation and engineering could be solved in several ways that might or might not be equivalent.

You said that it would be dangerous to shut off the cpu fan, while shutting off the cpu fan would be ok. I am talking about functionality that doesnt force me into considering that, or anything else. It should just silence everything when possible.

Instead of crude schemes used today ("run everything constantly as if the actual workload was equal to peak workload", "use manual settings to decrease power/heat/noise when expected load is low"). Clearly, computer algorithms should be able to do this more efficiently?


I think this is very much possible, but the industry is slowly adjusting from a boutique approach to neat and inexpensive soultions to problems that they didnt think users cared about (noise)

knut

mathias
Posts: 2057
Joined: Sun Jul 11, 2004 3:58 pm
Location: Toronto
Contact:

Post by mathias » Sat Apr 16, 2005 10:45 am

It looks to me like you don't really want a motherboard aimed at the SPCR community, what you're asking for is a motherboard that simply removes the hassle. This is rather self-contradictory, someone who wants practically no hassle in configuring anything is very unlikely to buy a separate motherboard.

I think my ideas about silent-motherboards are much more practical, more efficient voltage regulation means less heat produced by the mosfets, and less strain on the power supply. Video card undervolting seems to have more potential than ever, since at 110nm process, GPUs seem to be getting quite leaky. I don't disagree that getting rid of a bit of hassle and cutting costs by integrating a fan controller on the motherboard is a good idea, but I think it's more important to increase silence potential, rather than just make normal silence potential a bit easier to reach.

sthayashi
*Lifetime Patron*
Posts: 3214
Joined: Wed Nov 12, 2003 10:06 am
Location: Pittsburgh, PA

Post by sthayashi » Sat Apr 16, 2005 2:23 pm

knutinh wrote:You said that it would be dangerous to shut off the cpu fan, while shutting off the cpu fan would be ok. I am talking about functionality that doesnt force me into considering that, or anything else. It should just silence everything when possible.
No, I did not say this. I said that there are other fans that are dangerous to shut off. Namely, the fan(s) in the power supply. Or if you have a passive power supply, then the case fan becomes quite dangerous to shut off. These fans have to be considered when you want the motherboard to turn off "every fan."

I thought there were motherboards that turned off the CPU fan if the temperature dropped down low enough. Certainly there are motherboards that will regulate the speed of the CPU fan based on temperature. To what degree of regulation varies from motherboard to motherboard.

I guess what I'm trying to figure out is, what ARE you looking for that a) a motherboard CAN do and b) that motherboards currently are not doing?

knutinh
Posts: 99
Joined: Fri Nov 14, 2003 5:53 am

Post by knutinh » Sat Apr 16, 2005 4:25 pm

I meant: "You said that it would be dangerous to shut off the PSU fan, while shutting off the cpu fan would be ok. I am talking about functionality that doesnt force me into considering that, or anything else. It should just silence everything when possible."

I still think that any fan can be quite safely stopped as long as no component is over-heated as a consequence.

I am looking for a motherboard that immidiately make every fan-controller obsolete, as it has access to better sensors (internal cpu temp) and integration. I thought that my asus board did this, but it didnt.

mathias:
I have to disagree with you. I know very many musicians who gladly will build their own pc and needs very low noise, but who will not purchase 4 fans and select the one with lowest subjective noise. Equivalently, I think they would rather pay $30 extra than $100 extra for the same result. Wouldnt you? bThat is what is brilliant with this forum and the spcr site.

mathias
Posts: 2057
Joined: Sun Jul 11, 2004 3:58 pm
Location: Toronto
Contact:

Post by mathias » Sat Apr 16, 2005 4:51 pm

knutinh wrote: I know very many musicians who gladly will build their own pc and needs very low noise, but who will not purchase 4 fans and select the one with lowest subjective noise.
Then stick with nexus fans, no sample consistancy issues.
knutinh wrote:Equivalently, I think they would rather pay $30 extra than $100 extra for the same result. Wouldnt you?
How exactly is a silence geared mboard going to save you that money? An integrated high quality fan controller is not going to be free, and it's definitly not going to make getting good fans unneccessary.

Instead of what little it can save me, I'd rather have said efficient mosfets and less case heat, less PSU heat(both intake and internally generated), and a lower electricity bill.

sthayashi
*Lifetime Patron*
Posts: 3214
Joined: Wed Nov 12, 2003 10:06 am
Location: Pittsburgh, PA

Post by sthayashi » Mon Apr 18, 2005 8:52 am

Things that motherboards can't do based on current technology:
-- Read PSU temperatures
-- Control PSU fans

The above is based on knowledge of every single line that connects the power supply to the motherboard.

Things motherboards MAY be able to do, but currently don't:
-- Read Drive temperatures
-- Read other misc temperatures

This is based on lack of appropriate temperature probes and reading hardware.

You're right that any fan can be safely stopped as long as no component is over-heated, but it is impossible to guarantee that no components will be over-heated if you cannot monitor the component. And no motherboard manufacturer will be willing to stop all fans without this knowledge, since it sets them up for HUGE liability issues.

BTW, according to this page, the Asus K8V CAN stop the fan. Unfortunately, it's only listed for the Socket-754 platform and normal ATX at that (not uATX). It's not clear of other variants of the Asus + Via boards support the stopping of CPU fans.

knutinh
Posts: 99
Joined: Fri Nov 14, 2003 5:53 am

Post by knutinh » Mon Apr 18, 2005 9:57 pm

Now we are getting somewhere :-)

This means that the mobo currently should be able to "perfectly" balance load/temperature/noise for every component except PSU. My experience is that is won`t. And I bet the functionality to do so would cost next to nothing.

The motherboard/PSU connection is something that perhaps should be improved in future A-/BTX specs? Nevertheless, it is possible to optimise "locally" by assuming a PSU that is either fan-less or has the proper frequency/load/fan-speed curve so that its fan stops or spins slowly at low loads. A case fan connected to mobo and controlled by "system temperature" and/or the max temperature of any sensor connected to mobo should still be able to remove extra heat generated by internal components.

That is a solution possible using todays specs, using a simple component list of PSU, mobo, case-fan, cpu-fan and gpu-fan. And no matter how noisy it would be running folding@home (depending on maximum performance), it should be able to fade into silence in a continous fashion until conditions for "standby" applies...

Knut

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Mon Apr 18, 2005 11:09 pm

AOpen motherboards often have a couple connectors marked PSU that is meant for use with their own brand PSUs. On the pic of the AOpen i915Ga-PLF on this page, you will see 2 fan header type connectors just below the left side of the main 24 pin ATX connector. The 2-pin one is marked PWR TEMP, the 3-pin one is marked PWR FAN. It seems obvious to me that it's a thermal fan speed control loop for the PSU... which does not appear to be implemented in the BIOS I examined. I am guessing that it may be an option AOpen offers their OEM clients.

Splinter
Posts: 245
Joined: Mon Jul 05, 2004 9:01 pm

Post by Splinter » Wed Apr 20, 2005 4:07 pm

I can confirm that the K8V stops the CPU fan when temp drops low enough - mine is stopped right now.

Still, if you look at a graph of my CPU temp, it's a long sinewave. I'd rather have a steady low RPM than a constant cycle of speedups and slowdowns, it's a lot harder to tune that out.

Post Reply