I need a fan controller!

[Biggu]

I see the pin your talking about in the picture and Id say it is fine since that's not the contact point. that's just excess from the board.

 

[BonzaiDuck]

I see the pin your talking about in the picture and Id say it is fine since that's not the contact point. that's just excess from the board.

I still can't "see" it, so I was puzzled. What I see is a long row of pins at solder joints. And if it isn't the pin that fits a proprietary plug, I wouldn't mess with it unless it's touching another pin with prospects for an electrical short. In such a case, there's also the possibility you could cause a solder-joint to crack by fiddling with it. So -- yeah -- leave it alone . . Hook it up and see if it works like it's supposed to. If it doesn't work -- send it back; if it does work -- well -- experience will temper your imagination in the future.

EDIT--ADDENDUM: OK . . . I didn't go back through the posts, but it looks as though Bubbles got an NZXT "Sentry." These have been around for awhile. I haven't anything bad to say about them. Looking at the "stuff" in the package at the Egg, I don't see any thermal sensors -- or if I do -- it looks like a single one. This would have worked like a unit I bought some years back -- maybe a CM product. With the latter, no "touch-screen" but fancy-looking buttons and a blue temperature and fan-speed LCD. It had a single thermal sensor, if I recollect.

If the NZXT has a single analog sensor, you'll want to make it count. In those days, with a 533-FSB Pentium Northwood CPU, I didn't bother trying to install the sensor between the CPU and the mobo -- which I'd described here, or on a similar thread that's still active.

Here's the deal. I was ready to blow this NZXT off as techno-bling-candy -- a front-panel controller that's still manually controlled. Well -- I could've got some mileage out of my old CM front-panel, if I'd just taken the time (and risk) of sticking that single sensor under the CPU.

Think about it. Used to be that HDDs would get warm or hot; other devices or hotspots in the case . . also. You could stick a sensor on one or more HDDs, the CPU, the chipset. But between cooler-running HDDs and fewer, ascendancy of SSDs, better motherboards with heatsinks as good as the modder would buy to glue permanently to certain mobo components with Arctic Epoxy -- the only essential temperature needed for SIMPLE-but-excellent fan-control is the CPU temperature.

Likely -- the NZXT has no communication with the mobo. You'd be best to get that single sensor (if there is one) installed to the CPU so it measures the greater variability of temperature without attenuation by a heatpipe cooler base.

So -- Yeah! I won't guarantee it, but maybe these cheaper units (as opposed to the Aquaero 5 LT circuit-board for $80) -- could be a bang-for-the-buck choice. Frankly, I never thought about this before I started hankering for USB-connected, onboard-processor equipped fan control.

 

[Bubbleawsome]

Would I be correct in thinking these use less than 10w?
http://www.newegg.com/Product/Produc...82E16835345070

 

[birthdaymonkey]

Would I be correct in thinking these use less than 10w?
http://www.newegg.com/Product/Produc...82E16835345070

Those are 140mm PWM fans. I thought your case took 120mm fans? If you want to hook them up to a controller, you don't want PWM fans anyway. If you're planning to use these with your motherboard CPU PWM headers, they would only fit on a radiator that accepts 2x140mm.

According to the specs, they use 0.4 amps. So 12v * 0.4a = 4.8 watts. You can use this mathematical wizardry to calculate the power draw of any fan.

I just read through this thread with some puzzlement. Why spend all this money on new fans and fan control when you could have spent a bit more on a case with decent airflow? You really don't need as many fans as you think you do, especially with a single-GPU system. Cheaping out on a $40 case, then spending an extra $60-70 on fans and fan controllers is not a wise plan.

You should have just bought something like a Fractal R4, which comes with a built in fan controller.

http://www.newegg.com/Product/Produc...82E16811352024

 

[Bubbleawsome]

Those are 140mm PWM fans. I thought your case took 120mm fans? If you want to hook them up to a controller, you don't want PWM fans anyway. If you're planning to use these with your motherboard CPU PWM headers, they would only fit on a radiator that accepts 2x140mm.

According to the specs, they use 0.4 amps. So 12v * 0.4a = 4.8 watts. You can use this mathematical wizardry to calculate the power draw of any fan.

I just read through this thread with some puzzlement. Why spend all this money on new fans and fan control when you could have spent a bit more on a case with decent airflow? You really don't need as many fans as you think you do, especially with a single-GPU system. Cheaping out on a $40 case, then spending an extra $60-70 on fans and fan controllers is not a wise plan.

You should have just bought something like a Fractal R4, which comes with a built in fan controller.

http://www.newegg.com/Product/Produc...82E16811352024

Ok, going to reply in order.

My case has 3 120/140mm fan slots. I didn't know about the 3rd one until today. I currently have 1 PWM fan attached to my fan controller, is that bad?

Ok, wasn't sure of the equation.

This case has rather nice airflow, but only came with meh fans. 2 more fans should keep it silent. (1 for double fan CPU cooler)

I didn't know cases came with fan controllers, hm. I like my case though, and $99 would have overshot my budget by $60 at the time. I have since got delayed Christmas money so I have enough for a few fans.

 

[taq8ojh]

I was thinking about some sort of fan control before, but there is some confusion I could use help with.
Could someone very briefly explain how PWM works? I know those four pin CPU fans can change their speeds, but I don't quite understand how, based on what?

My board has several of those connectors, and while the idea of variable fan speed is nice, I don't understand how to make it work. The settings in BIOS are pretty useless too (blabla according to system temperature - what the heck is that?).
Do I simply need some software to benefit from that?

So with fan control, do I need 4 or 3-pin fans?

 

[Bubbleawsome]

So with fan control, do I need 4 or 3-pin fans?

I am running 3pin on a 2 pin van controller. YMMV

 

[BonzaiDuck]

I'd read the white-paper on PWM, and forget the details. Three-pin fan control is a matter of variable voltage, and supposedly PWM is more efficient.

EDIT-INSERT-AFTERTHOUGHT: My own motherboard allows for use of either PWM or 3-pin fans on the CPU fan-ports. It will control either type.

Because of a minor glitch I'd been troubleshooting with my rig [see "CPUs and OC'ing"] -- and because I'm just "innerested" in this issue of fans and fan control -- I am focusing a lot of attention on my rig right now.

So I tried to find the amperage spec limits on my P8Z68 mobo, which apparently weren't included in the user manual (this time -- ASUS had done it for earlier gens and models). The wisdom seems to be "1A per channel" but there is some sort of dynamic allocation of power to the fan plugs in a priority order -- beginning with CPU_FAN and CHA_FAN1.

I had noticed this for putting low amperage fans (first -- a 200mm @ 700 RPM and then a 140x25 @ 1,700 RPM max) -- on the CHA-FAN2 and OPT-FAN1 headers. These both would only spin up to a percentage of their top-end -- even if set in BIOS for 100% and no thermal control.

Again -- my mantra of wisdom: Focus high-output fans on CPU-exhaust, CPU and if desired -- ducted motherboard components. If your highest output fans show full range of control on the motherboard, then -- it's adequate. There's no need to control input fans if they are 140mm or larger -- they're going to be much more quiet. The only thing one should attempt to do after that: Balance the exhaust CFMs against intake CFMs and attempt to make the latter a tad greater.

IF you WANT thermal control of the intake fans, it would seem (as a lonely bystander here) that even the $30 fan controller with a single sensor wire and four 10W fan plugs would be quite adequate -- even elegant. When is exhaust going to be at maximum? When the CPU temperature rises. When do you need more intake? When the exhaust is higher.

As I also said before -- with these cheaper thermal fan controller -- try sticking the analog sensor under the IHS if there's a little quarter-inch gap somewhere where it meets the CPU-PCB. OR install the folded analog sensor between CPU and mobo and run the thin sensor wires between the CPU pins -- care--- fully.

Here's something interestin' . . . I created a separate thread for it:

http://forums.anandtech.com/showthread.php?p=35944254#post35944254

 

[taq8ojh]

So let's say I don't need thermal control/monitoring and only would control like 3 fans with it. I wouldn't mind something a bit on the useless visual gimmick side, but also not something that would resemble disco. So maybe some coloured leds, and/or a display with rpm etc.
Any recommendations?

 

[Bubbleawsome]

So let's say I don't need thermal control/monitoring and only would control like 3 fans with it. I wouldn't mind something a bit on the useless visual gimmick side, but also not something that would resemble disco. So maybe some coloured leds, and/or a display with rpm etc.
Any recommendations?

The sentry 2 doesn't detect rpm. It's display can turn off though.

 

[taq8ojh]

I think I found something. Zalman ZM-MFC1 Combo.

 

[BonzaiDuck]

I think I found something. Zalman ZM-MFC1 Combo.

And I also looked closer at the NZXT Sentry. People complain that it resets its automatic settings upon reboot. There were some other drawbacks.

This is why -- if I'm going to add equipment and therefore complexity -- I'd want a USB-connected controller with its own processor and memory (such as the ones we've seen.)

There may be others out there that cost less than half the Aquaero 5 LT, have at least one thermal sensor wire (all you need, really) and provide reliability for your custom settings.

The Zalman? That's still your typical manual fan-controller: no thermal control. But I'd also said it before and I'll repeat:

If you can thermally control your CPU fan and the main chassis exhaust fan from the motherboard, you shouldn't need much more. You can either hook up your input/intake fans directly to the PSU and run them full-bore, or you can get a simple fan-controller like this one, set them for the speeds you want at all levels of idle and load -- then forget them.

If the intake fans are noisy at full-bore, you have two choices again. If the airflow CFM at lower speed is adequate even for computer load conditions and variable exhaust speed determined by load temperature, then -- again -- set 'em and forget 'em. If they aren't noisy at full bore, then you shouldn't really need a fan controller of any type.

 

[Bubbleawsome]

Mine keeps settings.

 

[BonzaiDuck]

Mine keeps settings.

Lucky that, then. I'll also file that away per the Sentry, in case I ever need such a controller. Did you try to install the analog sensor to the CPU yet?

 

[Bubbleawsome]

Lucky that, then. I'll also file that away per the Sentry, in case I ever need such a controller. Did you try to install the analog sensor to the CPU yet?

I did, but I wasn't getting good contact by putting it between the IHS and cooler, so left it out. I have sensor 1 sitting above the 5.25 drives, sensor two is wedged in the CPU heatsink, 3 is taped to some wires in the middle of my case for "case ambient", 4 is taped to the HDD and 5 is wedged into the gpu shroud. All report accurate tems AFAIK.

One thing though is that I am not sure that it is spinning up to 100%. :/ It stays fairly quiet and temps are in the low 60s after running F@H for over 24h though. 5.25 bays are at 27, HSF is 32, HDD is 32, case ambient is 24 and gpu shroud is 37. True ambient is ~22c.

EDIT; 700th post

 

[BonzaiDuck]

I did, but I wasn't getting good contact by putting it between the IHS and cooler, so left it out. I have sensor 1 sitting above the 5.25 drives, sensor two is wedged in the CPU heatsink, 3 is taped to some wires in the middle of my case for "case ambient", 4 is taped to the HDD and 5 is wedged into the gpu shroud. All report accurate tems AFAIK.

One thing though is that I am not sure that it is spinning up to 100%. :/ It stays fairly quiet and temps are in the low 60s after running F@H for over 24h though. 5.25 bays are at 27, HSF is 32, HDD is 32, case ambient is 24 and gpu shroud is 37. True ambient is ~22c.

So the NZXT has several analog sensor pinouts and wires. That's good to know.

I've used more different Intel CPUs over the years than I can count. Well -- I could count them, but there were enough. I'd always noticed -- especially when lapping the CPU IHS, that the IHS is not completely sealed to the PCB. I have often noted a little 1/4"-wide, 1/16" or 1/32" gap in the middle of one side of the IHS. Since I now realize that there's nothing "in there" but the processor die, I am still somewhat sure that you could take the analog sensor (which is usually very flat) and stick it into that little slot (if there is one) to literally touch the processor die. You could then secure the sensor wire to the IHS or PCB with thermal tape or even just a dab of "Pit-Crew" automotive adhesive (no volatile chemicals).

An equally tedious effort would be routing the sensor and wire to avoid fouling the processor's seating in the motherboard, but I don't think this is a terrible insurmountable problem.

Even so, I can see you would have to remove your heatsink-fan assembly to get at the processor, and then re-install it. But I'm thinking -- if I'm correct about this -- you'd get good readings of processor temperature that might be closer to the core sensor values than anything else.

 

[taq8ojh]

The Zalman? That's still your typical manual fan-controller: no thermal control.

That's exactly what I want. Well, maybe thermal control over the exhaust fan(s) might be nice, but by the time I have everything up and running, I typically have a good idea about temperatures under different conditions, so manual control is perfectly fine.

I did, but I wasn't getting good contact by putting it between the IHS and cooler

Jesus Christ dude

 

[Bubbleawsome]

Jesus Christ dude

Yeah, it didn't go well. :$ I was going to put it in the IHS kinda like bonzai said, but I ran out of time. :awe:

 

[BonzaiDuck]

Yeah, it didn't go well. :$ I was going to put it in the IHS kinda like bonzai said, but I ran out of time. :awe:

Before I realized how retarded I am for not realizing long ago that the flat analog sensor might slip under the IHS, it was pretty obvious that you can't just put the analog sensor between HSF-base and the IHS without some "modification." Those two surfaces -- perfectly flat ideally -- need to mate up without some obstruction between them.

So the "modification." It would require either a Dremel tool and a very steady hand to cut a slot in the HSF-base just deep enough to wedge the sensor, but not deep enough to go all the way through or come close to the heatpipes. You'd only want the slot to reach from the edge of the HSF base to somewhere close to the middle of it.

The other possibility: find a machinist shop with the tools to cut a 1/32"-deep "dado" from edge to center of the base -- maybe 1/8" to 3/16" wide. In this method, you'd lay the analog sensor flat into the channel -- glue it in there with a dab of thermal epoxy just to secure it, put your diamond paste or Liquid Ultra on the HSF-base and IHS, and re-install the HSF as usual. Of course, this would reduce the effective surface area of the HSF that would mate with the IHS.

The temperature you likely would see coming from either of these two methods would be close to the fabled "TCase" temperature, or about 10C below the average of the core T-Junction sensors.

But that's a heck of a lot more trouble than slipping the sensor through the space between the IHS and PCB. You'd get near-TJunction-sensor temperatures this way. Your HSF would not be deprived of full-area contact. You wouldn't have to pay a machinist, and you wouldn't incur the risk of dremeling your HSF-base.

 

[taq8ojh]

...or just use any software that reads the temperature directly from the CPU

 

[BonzaiDuck]

...or just use any software that reads the temperature directly from the CPU

Sure -- if the software can communicate with the controller, but you need a USB connection for that. Otherwise -- you seem more comfortable with the idea -- you'd find yourself reaching for the knob on a manual controller.

 

[taq8ojh]

I was just kidding.

Wouldn't it be awesome if all boards had standardized interface for fan control hardware? Like, one universal port on the board, so you wouldn't have to drag 10 cables around the whole case? The thing would internally connect to all of the PWM fan headers and let you manage everything from a panel.

 

[BonzaiDuck]

I was just kidding.

Wouldn't it be awesome if all boards had standardized interface for fan control hardware? Like, one universal port on the board, so you wouldn't have to drag 10 cables around the whole case? The thing would internally connect to all of the PWM fan headers and let you manage everything from a panel.

Yeah . . . I remember I had a Northwood Pentium 4 running on an ASUS Springdale motherboard. Seemed to me you could at least give thermal control to the CPU fan, and the port was more than adequate for a 0.8A Delta Tri-Blade that could punch out as much as 140CFM. Now I remember the mobo: It was the P4P800.

Then I had an ASUS 680i Striker board. It had fan control for several ports and on-board pins for thermal sensor wires which came in the package.

After that, NVidia came out with their "ESA" standard and Silverstone produced the "Commander" which was supposed to work with NVidia chipsets. But they never perfected the software, so the Commander was left to be a "manual" controller without "manual knobs." You could set the speed through software, but it didn't provide thermal control.

I've accommodated myself to the limitations. My 2007 Q6600 build with the Striker had seven (7) fans: two 140mm intake in a cage on the case-bottom with the case raised up on 3" double caster wheels; two front-intake 120's; a 140mm "pusher" fan in front of the TRUE cooler but hung from the drive cage; and two 120x38 San-Ace exhaust fans. These latter were controlled by the CPU temperature off the mobo. The rest were "set it and fah-get it" with a Silverstone Commander.

The 2011 Sandy build has two 200mm and one 140mm intake fans. I'm pretty sure I can dump the 140 and block the vent at case bottom -- less dust. [I also had fitted the HAF case with 2" double-casters.] The exhaust is a 120x38 Panalfo pulling less than 0.50A and thermally controlled from the mobo. That leaves the two "limp-with-quality" Noctua fans on the NH-D14 cooler. I'm initially dumping the latter for replacement with a single Akasa Viper 140 "R" PWM -- to see if I can thermally control this latter single fan.

Frankly, I think I could just build a foam-art-board or Lexan duct for the D14 and let the Panaflo NMB-MAT do all the work. What's that? It's just three . . . freakin' . . . fanz!!

 

[Bubbleawsome]

The sentry can set alarm temperatures and then adjust fans per that.

 

[BonzaiDuck]

The sentry can set alarm temperatures and then adjust fans per that.

. . . But would not the alarm be tripped from thermal level(s)? So far, what you're telling me is that the Sentry is a darn good deal if it works as you say.

Does it allow the fans to spin up at continuously-increasing RPM with the temperature? Or do they just spin up from a preset level to full-bore after a thermal threshold? If the latter, that's a minor drawback from the more expensive, USB-connected "on-board processor" models I'd mentioned.

But it's only minor . . . .