Recommended Motherboards: From a Silence Perspective 1.1

All about them.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Operandi
Posts: 243
Joined: Mon Nov 08, 2004 5:20 pm
Location: Wausau, WI
Contact:

Recommended Motherboards: From a Silence Perspective 1.1

Post by Operandi » Tue Feb 07, 2006 9:21 am

The goal of this thread: To create a list of boards for all modern AMD & Intel boards that have features SPCR readers look for when deciding on a board for a new system. Desired features include good fan control of CPU and system fans, the ability to under-volt the CPU, and most important passive cooling of the chipset.

Common MB Problems for Silencers
- A very hot chipset requiring the use of active cooling or large heatsinks for passive cooling.
- No BIOS fan control for CPU and/or system.
- No support of SpeedFan.
- Lack of CPU under-volting options.
- Poor location of chipset causes problems with large aftermarket passive heatsinks and PCI-E 16x slots.

The above issues are generally what separate a good board from a bad one in terms of building a silent system.

To make the recommended list there are a few important qualities or features the board should have. If anyone feels I've left something out let me know and I'll revise the thread.

Requirements for "Quiet" Motherboards
- Stock passive chipset cooling. Or chipset cooling that can easily be modified from active to passive. If modifying the chipset cooling from active to passive has negative side affects they should be noted, such as interference with PCIE 16x graphic slots on many AMD nForce4 boards.
- Dynamic fan control for CPU and/or system fans either through the BIOS or software. The more details that can be provided here the better. For example; the number of controllable fan headers, what temps each header is tied to, the range of control offered 12-5v, or %, ect.
- SpeedFan support if available for more control over fans in addition to BIOS control would be a welcome bonus.
- CPU core under-volting option in the BIOS for reductions in power consumption and CPU temps.

Note: Most if not all AMD AM2 and Intel socket 775 boards feature 4-pin PWM CPU fan headers. Some of these motherboards will not be able to control the speed of traditional 3-pin fans so you will want to to take special care when selecting your CPU fan.

With that out of the way I'd like to take some recommendations for various platforms. As they come in I'll edit this thread into a list of boards (broken down by chipset) with a description of their features.

---Current Boards--- This section is for modern boards for current generation, still in production CPUs.
AMD Athlon Recommended Boards
  • Socket AM2 AMD Chipsets
  • Gigabyte GA-MA69G-S3H -- mATX / AM2 / AMD 690G -- Passively cooled north and south bridges. -- BIOS control of CPU 4 pin PWM fan (sparse options). -- Speedfan support when BIOS fan control is disabled. -- CPU undervotling options unknown.
  • MSI K9AG Neo2-Digital -- mATX / AM2 / AMD 690G -- Passively cooled north and south bridges -- BIOS can control CPU fan with PWM (Min fan speed adjustable from 0.0% - 87.5% of original fan speed in 12.5% increments) -- Speedfan only works on CPU fan header (if BIOS fan control is disabled) -- CPU Voltage adjustable in BIOS from 0.800V - 1.550V in 0.025V increments -- Other: Integrated DVI and HDMI.

    Socket AM2 nVidia Chipsets
  • Asus M2NPV-VM --mATX / AM2 / nForce 6150 + nForce 430 -- Passive cooling of chipset -- Q Fan control of two system fans, CPU header controls 4 pin PWM fans as well as 3 pin 12v fans. -- Speed fan control possible when BIOS Q Fan is disabled. -- CPU voltage control unkown. -- Other: Integrated DVI & VGA.
  • Abit AN-M2HD -- mATX / AM2 / nVidia GeForce7050PV/nForce 630a -- Passive chipset cooling -- System and CPU controllable; with user defined min and max start/stop speeds, target temp., as well as source temp sensor (nice ;)). -- SpeedFan support unknown. -- CPU voltage options unknown. -- Other: Integrated HDMI.
  • Asus M2N (M2N-E, M2N-SLi Deluxe, M2N32-SLi Deluxe) -- ATX / AM2 / nForce5 (various) -- Passive chipset cooling via heatpipe -- Two to Three fan headers controlled by Q-Fan. Options include; Performance, Optimal or Silent. -- SpeedFan support aprears to be present for three fan headers. -- CPU voltage ranges are 0.8000V to 1.5625V in .0125 increments.
  • Asus M2NPV-VM -- mATX / AM2 / nForce 6150 + nForce 430 -- Passive chipset cooling -- Q-Fan is present in BIOS, settings are only "enable" or "disable" however. -- SpeedFan support appears to be present for three fan headers. -- No CPU voltage options.
  • Gigabyte GA-M55S-S3 -- ATX / AM2 / nForce 550 -- Chipset cooled by a "pretty large" heatsink -- BIOS control for CPU fan header, settings are on or off only. -- SpeedFan not currently supported -- CPU voltages down to 0.8V.
  • Gigabyte GA-M59SLI-S5 -- ATX / AM2 / nForce 590 SLI -- Passive chipset cooling via heatpipe. -- BIOS "Smart Fan Control" for CPU fan. -- SpeedFan support unknown. -- CPU under-volting options unknown.
  • Abit KN9 SLI -- ATX / AM2 / nForce 570 SLI -- Passive chipset cooling via heatpipe. -- BIOS fan control for CPU 4-pin PWM fan header. -- SpeedFan support for at least one 3-pin system fan in addition to the CPU fan header. -- Undervolting not supported through the BIOS but is possible via CrystalCPUID down to 0.8v.

    Socket AM2 ULI Chipsets
  • MSI K9NU Neo-V -- ATX / AM2 / ULI-M1697 -- Passively cooled chipset. -- BIOS CPU SmartFan with up to 5 degree hysteresis and 2-8 steppings. -- Two fan headers controlled by SpeedFan. -- CPU voltage ranges of 1.2 - 1.4V in 0.025V increments.
Intel Core2 Recommended Boards
  • Socket 775 Intel Chipsets
  • MSI X38 Diamond, Platinum -- ATX / 775 / Intel X38 + ICH9R -- Passive heatpipe cooling for chipset. -- Automated CPU control for CPU fan and manual control for two system fans. -- SpeedFan support unknown. -- CPU voltages from 1.3250V to 2.1000V (by 0.7875 increment).
  • Abit IP35 Pro -- ATX / 775 / Intel P35 + Intel ICH9R -- Passively cooled North Bridge, South Bridge -- Control of 5 3-pin fan headers & 1 4 pin cpu header vai uGuru -- Underclockable via BIOS or Windows uGuru
  • Asus P5W DH Deluxe -- ATX / 775 / Intel 975X + Intel ICH7R -- Passive heatpipe based cooling of chipset -- 4 pin PWM CPU fan and two system fan headers controlled by AIQuiet or Q-Fan (performance, optimal, silent). 4 pin PWM CPU fan header also supports 3 pin fans via PWM or DC BIOS option. -- SpeedFan support for CPU fan and two system fan headers. -- CPU voltages range from 1.2000 to 1.7000 in 0.0125 increments. Note: None "Deluxe" version lacks support for 3 pin CPU fans
  • Abit AW9D-MAX -- ATX / 775 / Intel 975X + ICH7R -- Passive heatpipe based cooling of chipset -- Seven fan headers under the control of Abit EQ -- SpeedFan support unknown. -- CPU under-volting options unknown.
  • MSI P965 Platinum -- ATX / 775 / Intel 965 + Intel ICH8R -- Passive cooling of chipset. -- BIOS fan control for CPU and one system fan header, with target temperatures and tolerances for both. -- SpeedFan support is unknown. -- Under-volting options not known.
  • MSI G965MDH -- mATX / 775 / Intel G965 + Intel ICH8DH -- Passive cooling of chipset. -- 4 pin PWM CPU fan and two system fan headers -- Support for "desired CPU temp" in BIOS -- SpeedFan support unknown. -- Fixed CPU voltage

    Socket 775 nVidia Chipsets
  • XFX 650i Ultra -- ATX / 775 / nVidia 650i Ultra -- Passive cooling of chipset. Automatic control of CPU, 1-100 range and manual control for 2 3-pin fan headers. -- CPU voltages down to 0.8V.
  • Asus P5N-E SLI -- ATX / 775 / nVidia 650i -- Passive cooling of chipset. -- CPU and two system fans controllable via AIQuiet or Q-Fan. -- SpeedFan is supported. -- Under-volting options unknown.
---Legacy Boards--- This is reserved for boards that are either no longer in production or are obsolete due to their use of outdated CPUs.
Intel Pentium4 Recommended Boards
  • Socket 775 Intel Chipsets
  • AOpen i945Ga-PLF -- ATX / 775 / Intel 945G -- Passive cooling of chipset. -- CPU and system fan controlled by SilentBIOS -- SpeedFan support unknown -- CPU undervolting avalible in BIOS
    Aditional features: AOpen Power Master allows dynmaic CPU throtteling for all CPUs regardless of Intel's SpeedStep support.
  • Asus P5LD2 775 -- ATX / 775 / Intel 945P -- Passive cooling of chipset. -- 4 pin CPU fan header controlled by Q-Fan. -- SpeedFan support unknown -- CPU core voltage can be adjusted from 1.275V to 1.7V in 0.0125V increments through BIOS controls.

AMD Athlon Recommended Boards
  • Socket 939 nVidia Chipsets
  • Asus A8N-SLI Premium -- ATX / 939 / nForce4 SLI -- Passive cooling via heatpipe. -- Two fan headers controlled by Q-fan, control is completely automated; no selections for temp or speeds are made available. -- SpeedFan support for two fan headers -- CPU under-volting possible via RM Clock utility.
  • Asus A8N32-SLI Deluxe -- ATX / 939 / nForce4 SLI -- Passive cooling via heatpipe -- Two fan headers controlled by Q-fan, control is completely automated; no selections for temp or speeds are made available. -- SpeedFan support for two fan headers -- CPU under-volting possible via RM Clock utility.
  • Asus A8N5X -- ATX / 939 / nForce4 -- Active cooling is stock but ample room is provided for large passive heatsinks. -- Q-Fan control for CPU fan header with selectable target temp. -- SpeedFan supported -- CPU voltage settings from 0.800 to 1.650 Volts in 0.0125 V increments.
  • DFI SLI-DR Expert -- ATX / 939 / nForce4 SLI -- Active cooling of the chipset is stock with room for passive modification without blocking any PCIE slots. -- Three fan headers controlled through the BIOS with detailed controls for high / low temps, and start speed. -- SpeedFan support unknown -- CPU under-volting unknown
  • Gigabyte GA-K8N51GMF-9 -- mATX / 939 / nForce 6100 -- Passive cooling of chipset -- No BIOS controls for fan headers -- SpeedFan support for at least one system fan. CPU fan control would seem likely but is unconfirmed. -- CPU under-volting unknown.

    Socket 939 ATi Chipsets
  • Asus A8R32-MVP Deluxe -- ATX / 939 / ATi Xpress 3200 + ULi M1575 -- Passive cooling of chipset -- CPU and chassis fan headers control by Q-Fan. Fan start temp and full speed temp can be set for both fan headers with a value between 0-100c. -- SpeedFan support unknown -- CPU vCore adjustable down to 0.8v.
  • MSI RD480 Neo2 -- ATX / 939 / ATi Xpress 1600 + ULi M1573 -- Passive cooling of chipset -- CPU and system fan controlled by Core Cell. Temp thresholds are selectable, fan speeds are automated. -- SpeedFan support unknown. -- CPU under-volting unknown.
  • Abit AT8 -- ATX / 939 / ATi Xpress 1600 + ULi M1573 -- Passive cooling of chipset via heatpipe. -- Six fan headers controlled by Abit FanEQ. High and low temp thresholds selectable as well as min and max fan speeds -- SpeedFan support unknown. -- CPU undervolting unknown
  • DFI Infinity RS482 -- mATX / 939 / ATi Xpress 200 + ATi SB450 -- Passive cooling on north and southbridge. -- Good control over fan headers -- SpeedFan support unkown -- supports BIOS undervolting of CPU concurrently with Cool n’ Queit.
    Other: Onboard video features both DVI and VGA as well as S-video, composite, and component output.

    Socket 939 VIA Chipsets
  • Asus A8V -- ATX / 939 / VIA K8T890 -- Active & passive versions exist. There is ample room around the NB for large passive heatsinks for modification however. -- Two fan headers controlled by Q-fan, control is completely automated; no selections for temp or speeds are made available. -- SpeedFan support unkown -- CPU under under-volting options in BIOS

    Socket 754 nVidia Chipsets
  • AOpen n250a-FR -- ATX / 754 / nForce3 -- Passive cooling of chipset -- CPU and System fans controlled by SilentBIOS. Fan start speed and temp thresholds for CPU and system are selectable. -- SpeedFan support unknown -- CPU under-volting unknown

    Socket 754 VIA Chipsets
  • AOpen AK86-L -- ATX / 754 / VIA K8T800 + VT8237 -- Passive cooling of chipset -- CPU and system fan controlled by SilentBIOS -- SpeedFan support for at least the CPU fan header -- CPU under-volting unknown
Last edited by Operandi on Sat Mar 29, 2008 10:44 am, edited 58 times in total.

NeilBlanchard
Moderator
Posts: 7681
Joined: Mon Dec 09, 2002 7:11 pm
Location: Maynard, MA, Eaarth
Contact:

Post by NeilBlanchard » Tue Feb 07, 2006 9:33 am

Greetings,

Several Gigabyte motherboards have passive NB cooling, and they have BIOS control of the CPU fan, and they are undervoltable.

There are the Asus and the Abit models that have the heatpipe-cooled NB, and AFAIK, the Asus BIOS can control two fan headers.

There are a number of mobos that can easily fit a Zalman passive NB heatsink, without any interference with anything. An EPoX model definitely meets this criterea, and it also has control of (at least) the CPU fan header.

Tzeb
Posts: 144
Joined: Sat Apr 30, 2005 4:29 pm

Post by Tzeb » Tue Feb 07, 2006 9:59 am

Asus A8n SLI Premium has passive cooling and works perfectly with speedfan which can control 2 fans.
It also has another fan control feature through bios called ASUS Q-Fan2, but i prefer to disable it and stick with speedfan. :wink:

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Tue Feb 07, 2006 10:10 am

Several Gigabyte motherboards have passive NB cooling, and they have BIOS control of the CPU fan, and they are undervoltable.
Several GB mobos have passive NB, and several are undervoltable, but it is more rare for them to have both passive NB AND undervolting. For example GA-81PE1000 Pro and GA-8KNXP both offer undervolting, but have NB fans.

Operandi
Posts: 243
Joined: Mon Nov 08, 2004 5:20 pm
Location: Wausau, WI
Contact:

Post by Operandi » Tue Feb 07, 2006 10:25 am

NeilBlanchard wrote:Greetings,

Several Gigabyte motherboards have passive NB cooling, and they have BIOS control of the CPU fan, and they are undervoltable.

There are the Asus and the Abit models that have the heatpipe-cooled NB, and AFAIK, the Asus BIOS can control two fan headers.

There are a number of mobos that can easily fit a Zalman passive NB heatsink, without any interference with anything. An EPoX model definitely meets this criterea, and it also has control of (at least) the CPU fan header.
Thanks for the info but if you could be a bit more specific such as providing model numbers it would be appreciated.
Tzeb wrote:Asus A8n SLI Premium has passive cooling and works perfectly with speedfan which can control 2 fans.
It also has another fan control feature through bios called ASUS Q-Fan2, but i prefer to disable it and stick with speedfan. :wink:
This is actually the board I was looking at for my next build. I was under the impression that the A8N didn't have SpeedFan support, it's nice to know that it dose. Do you happen to have any info on undervolting options?

Thanks for the post, this will be the first board to make it into the list when I decide exactly how to format it.

gustavs_a
Posts: 48
Joined: Tue Jul 15, 2003 3:20 am

Post by gustavs_a » Tue Feb 07, 2006 2:00 pm

>Common MB Problems for Silencers

Coil whining should be added to the list. Not sure if capacitors can also make noises. My last 2 MBs have made noises, so I think it's a common issue, but very overlooked.

Erssa
Posts: 1421
Joined: Sat Mar 12, 2005 9:26 pm
Location: Finland

Post by Erssa » Tue Feb 07, 2006 3:07 pm

gustavs_a wrote:>Common MB Problems for Silencers

Coil whining should be added to the list. Not sure if capacitors can also make noises. My last 2 MBs have made noises, so I think it's a common issue, but very overlooked.
My current motherboard whines as well. I guess it's overlooked, since most people don't have computers that are silent enough for them to hear the noises. However getting decent information on the subject may be a bit hard, since motherboards probably have sample variances like any other products. I seriously hope that my motherboards coil whining is just that, sample variance.

Operandi
Posts: 243
Joined: Mon Nov 08, 2004 5:20 pm
Location: Wausau, WI
Contact:

Post by Operandi » Tue Feb 07, 2006 9:29 pm

gustavs_a wrote:>Common MB Problems for Silencers

Coil whining should be added to the list. Not sure if capacitors can also make noises. My last 2 MBs have made noises, so I think it's a common issue, but very overlooked.
Coil whine is pretty rare occurrence in my experience and is often the result of other components interacting in the system. For example Motherboard A + PSU B results may result in coil whine but Motherboard A + PSU C may be coil whine free.

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Tue Feb 07, 2006 11:53 pm

I have an ASUS A8V-E SE and would like to post my experiences with it. Not sure if it belongs on a recommended list, but maybe you'd be willing to use the information in your post as an FYI kind of thing.

The board is based on the Via K8T890 chipset and undervolting is supported in the BIOS. However, there is an extra voltage option that overvolts the CPU either 100mV or 200mV. Unfortunately, there is no was to disable this overvolt in any of the currently available BIOS revisions, so you're always stuck with at least Vcore + 0.1V.

This may present a problem to users wanting to enable CoolnQuiet, as at the CPU will run 0.1V higher at all P-states, which kind of defeats the purpose of CoolnQuiet in the first place. I've been meaning to email ASUS about this and suggest they add an option to disable the undervolt.

Otherwise I have been pretty happy with the board. The northbridge comes with a passive heatsink, but without any airflow from a CPU fan it gets very warm (don't believe the rumors about Via chipsets running so much cooler than nVidia -- they're not true). I've replaced the stock cooling with a slightly larger Zalman NB47J which doesn't seem to help very much. I've measured the chipset heatsink to be over 60*C, so I wouldn't be surprised if the northbridge itself is ~70*C. But I have yet to experience any stability problems, so I'm not too concerned about it.

Operandi
Posts: 243
Joined: Mon Nov 08, 2004 5:20 pm
Location: Wausau, WI
Contact:

Post by Operandi » Wed Feb 08, 2006 5:07 pm

frostedflakes wrote:I have an ASUS A8V-E SE and would like to post my experiences with it. Not sure if it belongs on a recommended list, but maybe you'd be willing to use the information in your post as an FYI kind of thing.

The board is based on the Via K8T890 chipset and undervolting is supported in the BIOS. However, there is an extra voltage option that overvolts the CPU either 100mV or 200mV. Unfortunately, there is no was to disable this overvolt in any of the currently available BIOS revisions, so you're always stuck with at least Vcore + 0.1V.

This may present a problem to users wanting to enable CoolnQuiet, as at the CPU will run 0.1V higher at all P-states, which kind of defeats the purpose of CoolnQuiet in the first place. I've been meaning to email ASUS about this and suggest they add an option to disable the undervolt.

Otherwise I have been pretty happy with the board. The northbridge comes with a passive heatsink, but without any airflow from a CPU fan it gets very warm (don't believe the rumors about Via chipsets running so much cooler than nVidia -- they're not true). I've replaced the stock cooling with a slightly larger Zalman NB47J which doesn't seem to help very much. I've measured the chipset heatsink to be over 60*C, so I wouldn't be surprised if the northbridge itself is ~70*C. But I have yet to experience any stability problems, so I'm not too concerned about it.
I don't think I'd worry too much about a 0.1v change. Do you have any information about fan control and SpeedFan support?

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Wed Feb 08, 2006 6:44 pm

I would consider 0.1V to be significant, as it will increase power consumption / heat output of a processor with default 1.4V by ~12%. :shock:

I've never used this feature, but the board appears to have support for dynamic fan control via Q-fan. Doesn't look like SpeedFan supports it -- at least the ASUS A8V-E SE is not listed in the program. It may support the board's clock generator, but I don't know which one it uses, so I can't confirm this. I've also used CrystalCPUID, but it locked the system up. Not sure about RMClock, haven't tried it yet, although I've been meaning to.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Wed Feb 08, 2006 7:01 pm

I would consider 0.1V to be significant (for a silencer, at least), as it will increase power consumption / heat output of a processor with default 1.4V by ~12%.
Yeah, it doesn't sound much but arguably the supply voltage is the most significant factor in determining how much heat a chip will produce. The relationship between clock speed (frequency) and power dissipation is linear, but the relationship between supply voltage and heat is quadratic:

Calculating CMOS Power
Calculating CMOS Power
A simplified equation for power consumption in a CMOS gate is P = CLV2f + IqV, where CL is the load capacitance, V is the supply voltage, Iq is the leakage current, and f is the switching frequency.

The first thing to notice about the equation P = CLV2f + IqV is that power consumption varies linearly with frequency. If the leakage current Iq is relatively small, reducing the operating frequency of a typical CMOS chip by a factor of two also reduces the power consumption by about a factor of two. More importantly, though, notice that the first voltage term of this equation is squared. This implies that if the supply voltage can be reduced, this also has a significant effect on power consumption. Unfortunately, driving a CMOS chip at lower voltages also means that its maximum operating frequency is reduced, which in turn results in lower performance. Thus, when attempting to optimize battery life, designers must carefully weigh the performance consequences of reduced voltage and chip frequency.

Figure 2 shows the combined effects of frequency and voltage on power consumption in the Analog Devices Blackfin DSP. Notice the quadratic shape of the curve due to the change in voltage.

Tzeb
Posts: 144
Joined: Sat Apr 30, 2005 4:29 pm

Post by Tzeb » Thu Feb 09, 2006 4:08 am

operandi wrote:....Asus A8n SLI Premium ..... Do you happen to have any info on undervolting options?
Have you heard of RM Clock Utility?
This tool undervolts the vcore and lowers the multy when idle and ramps them back when needed...it's quite an awesome tool...something like Cool and quiet on steriods 8)
At idle, my 2400mhz overcloked 3000+ venice stays at 4x300mhz with 1.1 v (that's the minimum RM clock alows on my venice. I've seen one opty go as low as 0.8v)
I just got a pair of Corsair TWINX2048-4000PT and a Ninja that i'll install today :D With the ninja and the new ram i will probably stay at 270mhz fsb, memory 1:1 and 9x.
Fingers crossed

EDIT : just a note about Asus A8N32-SLI on your list...
Have you read MikeC's comment about it? :roll:
MikeC wrote:I strongly recommend the Premium. You pay a big price in power consumption w/the A8N32 -- it basically doubles the chipsets. Naturally the power consumption (and heat) jumps -- some 35W in idle and ~17W at full load w/a X2-3800+, even when only one vidcard is used.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Thu Feb 09, 2006 9:23 pm

The problem with nforce4 or nhot4, is that it is all on one chip. See, if it was broken up to two actual north and south bridges, it could run passive.

Now I know people say that their fancy board has some heatpipe passive quality to it.

Well it sorta does. It is trying desperately not to melt. I duno if I like that. No, I know I dont like that.

Asrock dualsata II has a special perfection, it has a very fast performing board for gaming, easily undervoltable, it cost 70 dollars shipped in retail box, and well, it naturally has small passive heatsinks because it splits the north and south bridges up. Kinda simple solution, but it costs Uli and nvidia more money to do a double chip solution, hence, they want unified Bic Lighter single northbridges. after this 1695 chipset, all will go downhill permanently.

cyberman
Posts: 10
Joined: Wed Apr 20, 2005 8:34 pm
Location: Silicon Valley

Post by cyberman » Thu Feb 09, 2006 10:23 pm

~El~Jefe~ wrote:The problem with nforce4 or nhot4, is that it is all on one chip. See, if it was broken up to two actual north and south bridges, it could run passive.

Not to nit pick, (ok, so I'm nitpicking) but the nforce4 can in fact be run passive:

http://www.dfi-street.com/forum/showpos ... tcount=198

Although you do have to sacrafice a slot or 2 for that method. :/

There's some more examples of passive nforce4 heatsinking in that thread, I think.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Fri Feb 10, 2006 10:00 pm

passive with massive heat output is KINDA passive.

I could say I can fly out of a building without wings or an engine, or a parachute, and yes, yes I could, But I would make a nice red mist splat on the sidewalk soon after.

spcr techniques + nforce4 sli chipset passive = trainwreck

choo choo!
:shock:

Firetech
Friend of SPCR
Posts: 680
Joined: Fri Dec 16, 2005 4:50 pm
Location: Sydney, Australia
Contact:

Discontinued model

Post by Firetech » Sat Feb 11, 2006 2:53 am

The ABIT AN8 Ultra was looking like a good bet until they canned it... Passive NB heatpipe and multiple BIOS temp controlled fan headers. They weren't PWM proportionally controlled but you can't have everything.
I see Newegg have a reconditioned one but that's no use to me over here :wink:
The KN8 SLI now looks like the best bet from ABIT.

I'd like to be able to recommend my Foxconn but with the GPU card sat on top of the 40mm NB fan, I can't! It does have a nice proportional CPU fan control though (If you can work out how to set it up :lol: )

Le_Gritche
Posts: 140
Joined: Wed Jan 18, 2006 4:57 am
Location: France, Lyon

Post by Le_Gritche » Sat Feb 11, 2006 7:12 am

jaganath wrote: Calculating CMOS Power
(I made a link to the original article, not the google cache) Growing off-topic but very interesting read indeed.
The first thing to notice about the equation P = CLV2f + IqV is that power consumption varies linearly with frequency. If the leakage current Iq is relatively small[...]
And later :
As the semiconductor industry transitions to smaller and faster 0.13- and 0.09-micron (90 nanometer) CMOS, the required supply voltage goes down and the achievable frequency goes up, but leakage current grows. When we reach 65 nm and 45 nm CMOS technology in a few years, some experts believe that the leakage current will be so high that IqV will account for as much as 50% of total chip power consumption.
(article was written in June 2004)
Even though AMD uses SSOI to reduce current leakage, we have to expect that the rule of thumb about a voltage drop of x % resulting in a power drop of 2x % and a frequency drop impacting linearily the power consumption are going to be farther from the truth with die-size shrinking.

Regarding this thread, finally the article is not off-topic as it says :
One key for designing a low-power embedded system is component integration. Not only does a low chip count and a small printed circuit board lead to lower cost and physically smaller devices, it can also help to dramatically reduce power consumption. When signals propagate within a chip, they experience minimal signal loss because the wires are extremely short and the loads are very small. On the other hand, signals that must be driven off-chip must travel a much longer distance and must drive much larger loads. Thus, to minimize energy consumption, whenever high-bandwidth communication must take place between two components, it is best to integrate the two components into one chip. From a power consumption perspective, the most important components to integrate are typically the processor and memory due to their high-frequency interaction.
(emphasize is mine)
So integrating NB and SB reduces cost, complexity and power consumption, sadly the heat - even though reduced - is concentrated in one single place, making it harder to run passive.

Operandi
Posts: 243
Joined: Mon Nov 08, 2004 5:20 pm
Location: Wausau, WI
Contact:

Post by Operandi » Sat Feb 11, 2006 9:24 am

frostedflakes wrote:I would consider 0.1V to be significant, as it will increase power consumption / heat output of a processor with default 1.4V by ~12%. :shock:

I've never used this feature, but the board appears to have support for dynamic fan control via Q-fan. Doesn't look like SpeedFan supports it -- at least the ASUS A8V-E SE is not listed in the program. It may support the board's clock generator, but I don't know which one it uses, so I can't confirm this. I've also used CrystalCPUID, but it locked the system up. Not sure about RMClock, haven't tried it yet, although I've been meaning to.
Ok, but if undervolting is possible wouldn't you be able to compensate for the 0.1v increase?

Also I'll assume that Q-Fan controls two fan headers since that seems to be the norm for Asus, let me know if it's something other then that.
Tzeb wrote:EDIT : just a note about Asus A8N32-SLI on your list...
Have you read MikeC's comment about it? :roll:
MikeC wrote:I strongly recommend the Premium. You pay a big price in power consumption w/the A8N32 -- it basically doubles the chipsets. Naturally the power consumption (and heat) jumps -- some 35W in idle and ~17W at full load w/a X2-3800+, even when only one vidcard is used.
That may be true but it's hardly a deal breaker. Also remember the A8N32 uses an 8 phase power design that is far more efficient then lesser designs. While not enough to overcome the hotter chipset the benefits would be more pronounced when using a more powerful CPU then the 3800+.

Regardless, the goal of this thread is to find several boards that have features that lend themselves to use in quiet systems, not to find the single absolute best and exclude the rest.

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Post by nici » Sat Feb 11, 2006 1:23 pm

DFI nF4 boards offer 0-100% fan control for three fans headers, either a fixed speed or start-up voltage where the fan starts and desired speed at that temperature and a temperature where you can set it to run full speed. It also starts allfans on 100% so theres no risk of a fan not starting. They are also undervoltable. I would say DFIs have the best fan control of any mobo out there.

The chipset position is crap crap crap, right under the upper PCI-E 16x slot. The only DFI nF4 mobo with good chipset placement is the SLI-DR Expert

I got tired of the DFI though for various reasons, and am now running a ASUS A8N-SLI Deluxe wich has plenty of room to replace the chipset HS. Im running the Reserator again, and a waterblock fits perfectly on the chipset :) I chose the Deluxe over the Premium because they have identical features but the heatpipe thingie on the Premium couldn´t handle the heat without airflow, and because im watercooling i went with the Deluxe and saved 10€.. I dont need great fan control since the only fan in the whole system is a 120mm nexus at constant 500rpm.

Most Intel boards have passive cooling on he chipset(s). And if you want to replace the standard passive HS with a bigger HS its easier because the heatsinks are mounted with hooks instead of pushpins(i hate these f*ckers).

Operandi
Posts: 243
Joined: Mon Nov 08, 2004 5:20 pm
Location: Wausau, WI
Contact:

Post by Operandi » Sun Feb 12, 2006 9:03 am

nici, alright I added in the DFI board. Do you have any information about SpeedFan support, or CPU under-volting options?
nici wrote:Most Intel boards have passive cooling on he chipset(s). And if you want to replace the standard passive HS with a bigger HS its easier because the heatsinks are mounted with hooks instead of pushpins(i hate these f*ckers).
Yes, I know what you mean. Oddly enough my Aopen nForce2 board has a fairly large passive heatsink that is secured via the same hooks you see on many Intel boards. To me it seems like they went backwards when them moved to pushpins for nForce 3/4...

Also if you have any recomendations for Intel boards I'll take those too. :wink:

frostedflakes
Posts: 1608
Joined: Tue Jan 04, 2005 4:02 pm
Location: United States

Post by frostedflakes » Thu Feb 16, 2006 9:28 pm

operandi wrote:Ok, but if undervolting is possible wouldn't you be able to compensate for the 0.1v increase?
Yes, but unfortunately you can't compensate with CoolnQuiet enabled. :(

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Post by nici » Fri Feb 17, 2006 5:13 am

operandi wrote:nici, alright I added in the DFI board. Do you have any information about SpeedFan support, or CPU under-volting options?
I dont know about speedfan support because i don´t really think you need it on those mobos, three fans customizable as you wish from the BIOS or the included ITE SmartGuardian program. SmartGuardian sets the bios values AFAIK, so you can quit the program after you have configured it once.. I like the program though, it´s easy to read and set-up. Black backgound and large 7-segment display lookalike numbers in green for temps, along with the rpm of three fans. 0-100% control, and no alarm so it wont howl if you run your fans "too slow" as most of us do :lol: Maybe SpeedFan has an alarm funtion though..

I never used the automatic fan control thing though, i just set the cpu and exhaust fan at a constant speed wich worked good. The fans always start up at 100% so no risk of them not starting.. :)

IIRC the bios allowed undervolting to something like 0.8 or 0.85V, wich should be low enough for most people.. :)

I am pretty shure these applies to the SLI-DR Expert too, i cant see why they woould have less adjustment possibilities on the top-of-the-range model than on the "cheap" nF4 Ultra-D. The layout of the Expert is very similar to the A8N-SLI, just that the CPU and memory slots are "swapped around" Here is a pic. Looking at that pic i see theres an 8-pin connector beside the 24-pin atx-connector, wich i think is a little puzzling.. :?
operandi wrote:
nici wrote:Most Intel boards have passive cooling on he chipset(s). And if you want to replace the standard passive HS with a bigger HS its easier because the heatsinks are mounted with hooks instead of pushpins(i hate these f*ckers).
Yes, I know what you mean. Oddly enough my Aopen nForce2 board has a fairly large passive heatsink that is secured via the same hooks you see on many Intel boards. To me it seems like they went backwards when them moved to pushpins for nForce 3/4...

Also if you have any recomendations for Intel boards I'll take those too. :wink:
I have an Epox mATX mobo for socketA wich has a VIA chipset and uses these hooks, but it has a large passive HS stock so i havent changed that :)

The ASUS nF2 sktA mobo i had also had a passive HS, it had pushpins but its also glued on :shock: I removed the pushpins and noticed i still couldnt get the HS off.. Can´t remember why i wanted to take off the HS, but anyway i ended up leaving it without the pushpins and using them for a nF4 mobo :lol: No wait i think that was the mobo that got fried when my PSU blew up.. and i needed a couple of pushpins so i took them off. heh :)

I also think pushpins are a step in the wrong direction..

Never had an Intel machine of my own, but my parents bought a PII 400mHz with a whopping 6Gb HDD and Voodoo2 gpu in 1998 8) Thats the only Intel experience i have... :lol: It was still used in 2003 though, but with a Riva TNT2 gpu after the Voodoo2 crapped out.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Fri Feb 17, 2006 6:06 am

my parents bought a PII 400mHz with a whopping 6Gb HDD and Voodoo2 gpu in 1998
Damn that thing must be slooooooooowww.....

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Post by nici » Fri Feb 17, 2006 9:39 am

jaganath wrote:
my parents bought a PII 400mHz with a whopping 6Gb HDD and Voodoo2 gpu in 1998
Damn that thing must be slooooooooowww.....
Well it is by todays standards yes, but it was blazing fast back in 1998 8) When we got that computer is still the biggest "WOW!" i have experienced with computers, with 3d games looking unbeliavably good :) Price was something around 2000€ at the time, without a monitor. Add inflation to that.. It still got old pretty fast though, 8 years ago compueters were developoing a hell of a lot faster than they are currently. It would still be good as a surf-box or file-server though :)

stupid
Posts: 544
Joined: Tue Nov 29, 2005 2:21 pm
Location: NYC, NY

Post by stupid » Fri Feb 17, 2006 10:27 am

jaganath wrote:
my parents bought a PII 400mHz with a whopping 6Gb HDD and Voodoo2 gpu in 1998
Damn that thing must be slooooooooowww.....
Actually, I still have my PIII 450 (slot 1) that I've mothballed. I am currently fixing it up, and looking to donate it to a charitable organization (Salvation Army?). When I was using it way back when, I OCed it to a whopping 550MHz (whoa, stand back). It's an old clunker now, but good enough for someone who wants to surf the web, and do Excel. It also gives me a way to finally get rid of a 20GB and 40GB hard drives that have been mothballed as well.

Erssa
Posts: 1421
Joined: Sat Mar 12, 2005 9:26 pm
Location: Finland

Post by Erssa » Fri Feb 17, 2006 11:38 am

frostedflakes wrote:I would consider 0.1V to be significant, as it will increase power consumption / heat output of a processor with default 1.4V by ~12%. :shock:

I've never used this feature, but the board appears to have support for dynamic fan control via Q-fan. Doesn't look like SpeedFan supports it -- at least the ASUS A8V-E SE is not listed in the program. It may support the board's clock generator, but I don't know which one it uses, so I can't confirm this. I've also used CrystalCPUID, but it locked the system up. Not sure about RMClock, haven't tried it yet, although I've been meaning to.
I would be suprised if it didn't support Speedfan. The CrystalCPUID lock up could just be a result of bad settings, not necessarily incompability.

timosa
Posts: 11
Joined: Mon Dec 20, 2004 5:27 am
Location: Finland

Passive cooling DFI SLI-DR Expert chipset

Post by timosa » Mon Feb 20, 2006 11:25 am

Do you have any recommendations for cooling DFI SLI-DR Expert motherboard chipset passively? I have read it runs quite hot.

nici
Posts: 3011
Joined: Thu Dec 16, 2004 8:49 am
Location: Suomi Finland Perkele

Re: Passive cooling DFI SLI-DR Expert chipset

Post by nici » Thu Feb 23, 2006 8:02 pm

timosa wrote:Do you have any recommendations for cooling DFI SLI-DR Expert motherboard chipset passively? I have read it runs quite hot.
I doubt it runs any hotter than other nF4 SLI mobos, a Zalman NB47J should fit nicely on the chipset and cool it pretty good if there is some airflow over it :) If the chipset sensor is placed similarly to the Ultra-D its pretty accurate as explained above, so you will see how well the new heatsink performs compared to stock.

Depending on how quiet your system is, slowing the orginal chipset fan might be enough for now. The stock thermal pad is crap, replace that with silver goop and it runs about 22c above ambient temp when only spinning 1400rpm. 3000rpm was enough to cool it while gaming for hours or looping 3dMark.

defaultluser
Posts: 82
Joined: Mon Feb 20, 2006 9:39 pm

Post by defaultluser » Mon Feb 27, 2006 7:00 pm

jaganath wrote:
I would consider 0.1V to be significant (for a silencer, at least), as it will increase power consumption / heat output of a processor with default 1.4V by ~12%.
Yeah, it doesn't sound much but arguably the supply voltage is the most significant factor in determining how much heat a chip will produce. The relationship between clock speed (frequency) and power dissipation is linear, but the relationship between supply voltage and heat is quadratic:

Calculating CMOS Power
Calculating CMOS Power
A simplified equation for power consumption in a CMOS gate is P = CLV2f + IqV, where CL is the load capacitance, V is the supply voltage, Iq is the leakage current, and f is the switching frequency.

The first thing to notice about the equation P = CLV2f + IqV is that power consumption varies linearly with frequency. If the leakage current Iq is relatively small, reducing the operating frequency of a typical CMOS chip by a factor of two also reduces the power consumption by about a factor of two. More importantly, though, notice that the first voltage term of this equation is squared. This implies that if the supply voltage can be reduced, this also has a significant effect on power consumption. Unfortunately, driving a CMOS chip at lower voltages also means that its maximum operating frequency is reduced, which in turn results in lower performance. Thus, when attempting to optimize battery life, designers must carefully weigh the performance consequences of reduced voltage and chip frequency.

Figure 2 shows the combined effects of frequency and voltage on power consumption in the Analog Devices Blackfin DSP. Notice the quadratic shape of the curve due to the change in voltage.
Yes, but do you know WHY power usage is related to the square of the voltage? it's easier to understand this relationship when you look at the basics:

V = I * R must always be satisfied. Thus, if you increase the voltage, you get a proportional increase in current.

Given Power = V * I, you see how you end up with the V^2 relationship.

And, just to stay on-topic, my Asus A8V Deluxe is fully passive, and runs cool. Sure, it's not exactly new, and yes, it lacks PCIe, but you can still pick them up. Two Q-fan controlled headers, with voltages as low as 9/16 * 12v. Undervolting.

Post Reply