Multi-channel fan controller for several PWM/Non-PWM fans Vs a decent MB?

[jalyst]

If one or some could offer their thoughts on the best direction to go here, I'd be most appreciative!

It seems my MB has the following:
1x CPU fan-header (PWM)
1x chasis fan-header_1 (PWM)
1x chasis fan-header_2 (3-pin)
1x Power fan-header (3-pin)
I was hoping there'd be more fan headers in-total, but at least there's my requisite minimum of 2 PWM headers.

I'm planning on 2->4 AP-15's for my Rad, & 4->6 AP-14's for the chassis.
Plus I'm ordering 1x >3000RPM Nidec PWM 120mm to experiment with (exact model YTBD, prolly 5400).
If testing goes well, I may order a second Nidec....

[proposed layout]
2x Nidec's, each sitting on their own PWM header, with their +/- going direct to the PSU?
Then I could put the up-to-4x AP-15's on one of the 3-pin headers, & the up-to-6x AP-14's on the remaining 3-pin header?
[/proposed layout]

Are 3-pin headers essentially linear voltage controllers (i.e. rheostat/bus) that can run 3-pin fans at 12v or less?
Or can they only run non-PWM fans at their rated top (12v) speed?
Would the 3-pin headers take the load, or would some of the GT's on them have to be rigged direct to the PSU?

Finally, is my MB's PWM likely to be high-frequency PWM?
How do I determine whether decent quality PWM has been implemented?
For instance this user's MB isn't able to regulate his fan down to 1000RPM via it's PWM header.
The lowest it can go is 2400, despite the fan being capable of 1000RPM.

Does this all sound too clumsy, is a dedicated controller looking to be imperative IYO?
I'd rather not have something that uses a bay, but would still like RPM/voltage feedback & at least auto-control if possible.

There does seem to be a few multi-channel controllers which do 3-pin, but also have at-least one channel for PWM fans.
Nothing I've seen that avoids using a bay, YTD if any offer 2x quality high-freq PWM channels*, w/high enough wattages for 2x 5400RPM PWM fans.
So hopefully no need for the self-build route, but I do get that some guys like doing that sort of thing for fun!

For the 2x 5400RPM 120x25mm fans I'm considering**, the options -as I see them- are:

(1)
Buy 2x D1225C12BBAP-31 & pair them with a controller that has at-least 2x LVC channels.

Granted I'd be using linear voltage control instead of PWM, but I'd still retain all the functionality of option (2) right?
e.g. RPM/voltage feedback, and speed control, just not the same power efficiency...

(2)
Buy 2x D1225C12BBZP-00 & feed +/- wires of 2x directly to the PSU (not sure how best to manage that), & then connect their PWM/RPM wires to the 2x PWM fan-headers on my mobo?

I haven't done the math but I suspect option (2) will be pricier than option (1), despite the latter not needing a discrete controller.
This is because the Nidec 5400RPM 25mm PWM fans, have to be imported at much greater expense than their Scythe equivalents.
This makes option (1) more attractive, because not only will it be cheaper, but I'd also own a muilti-channel PWM/Non-PWM fan controller!
Admittedly though I haven't done the maths, so I may be totally wrong...

I would prefer to start off with MB initially if do-able, hopefully saving $ + giving myself plenty of time to research the best controller.

*YTD how to ascertain if a quality PWM channel/circuit is being used.
**assuming that's the highest RPM 25mm fan from Nidec now?

[jalyst]

Thanks for confirming the AP-14/15's aren't PWM, and do under-volt well!

AP 15/14's etc are NOT PWM. 3 Pin.
So you need the standard Mosfet etc on the mobo or controller to handle the startup current of each. Yes, they are directly controlled by the raw voltage applied to the fan, they have no PWM hardware on the fan. And they do undervolt well. So for voltage control they have to be connected to the voltage control device, branching the power wires to the PSU will ensure 100% RPM all the time, no matter what you want to do with the RPM sensing wire.

I know the GT15's are about .3 amps on startup. Adjust accordingly for the 14's. Mobos have a higher amp CPU connector, some mobos are 3/4 pin and allow either type to be used. Most are PWM 4 pin I 'think' tho. 1 Amp is the common accepted CPU header max allowed, the others? .5 Amp.

According to Scythe's global site, peak draw for AP-15 is 0.083, & AP-14 is 0.049.
YTD the max. draw for each of my headers, but assuming you're right & the CPU_Fan PWM header is 1amp & the other 3 are 0.5...

Then that means one of the 3-pin headers will comfortably support up to 4x AP-15's at their top speed (0.332).
And the 2nd 3-pin will comfortably support up to 6x AP-14 at their top rated speed (0.294).

Think of this.... Using the mobo and all the availible fan headers on the mobo makes a absolute mess of wiring.
Get a solid quality fan controller, make do with one less slot.
1-6 GT14s spread between two mobo headers? Nope....
Sorry I didn't read every bit of the post you made, but maybe this will help.

Actually it's 2-4 AP-15's for one header & 2-6 AP-14's for the other, so as many as 10x AP15/14's between the 2x 3-pin headers.
Then we've got up to 2x PWM fans >3000RPM (prolly the 5400's) sitting on the 2x PWM headers. So yeah, prolly quite a mess!

I probably will get a dedicated controller in the coming months.
I just want to make sure it's do-able with my MB only in the meantime.

It certainly did help, and was most appreciated!

[jalyst]

I think Conumdrum answered most of your questions.

I've since identified at least one flaw with the multi-channel PWM/LVC controllers I've seen so far:
They don't interface with our MB/BIOS/EFI & adjust LV/PWM of their channels independently, based on temp readings passed from the MB.
Plus RPM data of each fan can't be passed back to the MB....

Such a set-up would be close to 'Fan-control Nirvana' IMO....
This 'kinda' comes close, cept that it has manual control & hence requires a bay, which I don't need.
And the input 4-pin molex connection is only suited for 60-80W, so even running 2-ch at full capacity could get dangerous.

Perhaps you can answer a question for us - why do you want control and RPM feedback of so many fans, motherboard or otherwise?

If it's do-able, why not?!

I'm building my controller because I want my system to tick all the air cooling boxes - silence, efficiency, and raw performance. To claim all three with the same fans simply isn't possible without a controller. Silent cooling may be efficient but cannot provide raw performance; efficient cooling may be silent and may perform well but excels at neither; and performance cooling is highly efficient but never silent.

This is precisely what I'm aiming for, but in a pre-built controller device, if not possible via MB.

As it stands, my silent rig is composed entirely of low speed fans; for efficient cooling (when stress testing etc) I have to swap out the CPU fan for something more powerful; and for raw performance during benching I have to set up several fans to cool the chipset, VRM, memory etc. in addition to swapping out the CPU fan again.

I don't envision needing to swap-out the chassis fans etc. very often...
But I may decide to get more heavily into OC, & hence need some stronger fans to cool chipset/vrm/mem etc.
And I may eventually replace the AP-15's on my Rad with 2x 5400RPM GT, if testing of 1x goes well.

Primarily the build isn't a 'hard-core' OC/gaming machine....
It's primary utility is HTPC/PVR (BE+FE), & to some extent initially, a NAS.
But I will play with mild OC's, & occasionally may want to go more 'hard-core'. (not to win benchmarking championships etc though)
The majority of the time though it'll run at stock, or even slightly UC.

[Philwong]

You can try using Artic cooling F12 with PST function which allows up to 5 fans to share a single PWM header. I've one of my rigs hooked up to 4 of them, but rewired the +12V and ground rails directly to the PSU to play safe.

In my gaming rig using Asus Rampage II Gene, I've 2 F12s hooked up directly to the CPU's PWM header and another pair connected to my 5870 graphics card's PWM header. CPU fan speed is controlled via the bios, while GPU uses a custom temp and RPM response curve charted out via MSI Afterburner.

All remaining 3 pin case fans are also thermally throttled by bios for quieter fans during regular usages such as web surfing and enhanced cooling for gaming. I've a Matrix Orbital Typhoon GX LCD, LCD studio and Everest for real time feedback of the entire cooling system and vital system stats.

More info on Artic Cooling PST here -> http://arctic.ac/fileadmin/arctic/we...CTIC_F_PWM.pdf

Phil

[jalyst]

Thanks Phil,

I think I'm happy with what my MB can provide for now, but LT I'll definitely want a more sophisticated discrete controller.
One that can do everything I want independently, or in tandem with my MB+EFI/OS.

Just a quick Qn; what are your thoughts on the current t-balancer offerings?

I think the BigNG may get closest to what I want... Just wish it 10-12 channels;
Two with enough juice to power the 5400's at peak load on start-up (50w?), they'd remain PWM whilst the others are LVC.
Or even better, every channel capable of 'independently' switching between PWM/LVC.

Granted what I've outlined prolly isn't catered for by any product around, so there'll have to be compromises

[asura]

I love my BigNG (though it's not currently installed... again...) it's a very versatile piece of kit.

I'm not sure exactly what it is you're doing, but two of them is probably overkill, and (if memory serves) can give hiccups if running them simultaneously on the system (rather than as independent units.) However, it seems likely that one BigNG and one miniNG would do you... the two 5400's on the miniNG and the rest on the BigNG

One thing to be aware of is that mCubed offer very little (no) support, if you aren't confident you can get it going then perhaps hesitate and get one of the pending new aquacomputers.

[Waterlogged]

Granted what I've outlined prolly isn't catered for by any product around, so there'll have to be compromises

. . .or you could learn about electronics (like I'm doing for a current project because I refuse to compromise) and build your perfect controller yourself.

[NaeKuh]

. . .or you could learn about electronics (like I'm doing for a current project because I refuse to compromise) and build your perfect controller yourself.

then we would start looking like your avatar and be grumpy from all the soldering..

[Waterlogged]

then we would start looking like your avatar and be grumpy from all the soldering..

My avatar is me squinting as a result of all the "WTF? moments" I come across here.

I actually enjoy soldering

[Alexandr0s]

. . .or you could learn about electronics (like I'm doing for a current project because I refuse to compromise) and build your perfect controller yourself.

Been there, done that. Actually today:



I'll post a small log in the Xtreme Hardware Mods subforum in a few minutes

Edit: log here

[jalyst]

. . .or you could learn about electronics (like I'm doing for a current project because I refuse to compromise) and build your perfect controller yourself.

Well with enough time/perseverance I'm sure could slap together 1-4 PWM channel controller.
But fabricating something with the functionality requirements I've outlined would be another matter.

Even a brilliant EE (which I'm not) with years of experience with PCB's/soldering would struggle to implement what I've described.
Which is why I have to go with what's out there, & perhaps "Franken-mod" it if possible!?

Any ideas apart from suggesting I build it all myself? Thanks for your time/thoughts!

[jalyst]

I'm not sure exactly what it is you're doing, but two of them is probably overkill, and (if memory serves) can give hiccups if running them simultaneously on the system (rather than as independent units.) However, it seems likely that one BigNG and one miniNG would do you... the two 5400's on the miniNG and the rest on the BigNG

I hadn't yet considered "daisy chaining" a bunch of t-balancers together.
I'm not sure what the limitations are there, if any...

Assuming I can chain some together, I think 2x BigNG, and 1x miniNG would be almost perfect!
Two 5400RPM's can sit on the MiniNG, & the remaining (up-to-10) AP14/15's can sit on their own channels (two will have to share one).

This (if possible to override any limitations) comes closest to meeting my entire check-list of wants/hopes.
Although I'm yet to entirely cross-reference that....

One thing to be aware of is that mCubed offer very little (no) support,

That's unfortunate, not a fan of Co's that don't like to stand by their products in a pro-active sense.

if you aren't confident you can get it going then perhaps hesitate and get one of the pending new aquacomputers.

Is this another type of fan control system or a complete build?
If the latter I'm not really interested, but thanks for the ideas!

[DumpALump]

So I've been messing around with electronics for a bit, just learning a few things. I had a zalman fanmate 2 sitting around that I had from an old VF900 VGA cooler. Opened it up, cut a larger heatsink for the voltage regulator. Attached 2 molex connectors to the ends. (all stuff just sitting in the electronics drawer). The voltage regulator heats up as you turn down the current (since the schematic sends more current to the mosfet), so the larger heatsink should help with removing the heat. There are better IC's out there that can handle the higher loads (cost like $1 each), but figured I might as well use whats around. I believe you can also use multiple to spread out the heat so you could still utilize smaller heatsinks.

Then I spotted a multifan power adapter. http://www.svc.com/mfpp.html Fry's had one for $10. I could have built one, but I was lazy. Attached my AP-15's to my power adapter and control 6 AP-15's from the potentiometer on the zalman fanmate 2. Simple stuff. I really didn't want one of those bay controllers since 1) Expensive 2) I have all internal WC and pretty much zero room in my P183 so the controller sits in the bottom bay near the radiator.

I am interested in building something that you are looking for as I would like the fans on my radiators to vary with cpu temps. I know there are IC's out there that can change output based on input from a third source. Its just a matter of finding a schematic for using one.

I would look around but I'm currently figuring out how to make a volcano vaporizer for my gf.

Considering that the zalman fanmate 2 can control up to 6watts already technically you could just put the 6 AP-15s or 6 AP-14s on it. I'd still put a larger heatsink on it though. Then again I put heatsinks on everything lol. This setup actually takes up very little room. If I find out some info about controlling an IC's output from the motherboard header, I'll let you know.

[Waterlogged]

WL/Asura/Anyone?
Thanks again very much everyone, night.

Don't look at me, I'm still trying to figure out my own headache. I'm just here to cause you more confusion.

Seriously though, a quick Google image search for "pwm circuit" revealed many possibilities. If I didn't know better, I'd say what your asking for is easier than what I'm trying to accomplish.

[jalyst]

Don't look at me, I'm still trying to figure out my own headache. I'm just here to cause you more confusion.
Seriously though, a quick Google image search for "pwm circuit" revealed many possibilities"

Yep & I've seen them, they're all for "one trick pony"* controllers, none come near the multi-functional device/s I'm seeking.

If I didn't know better, I'd say what your asking for is easier than what I'm trying to accomplish.

LOL, not sure how you work that one out...
But as you stated, your main goal is just to confuse the situation

Anyway good luck with your project!
*not that there's anything wrong with that, so long as it does that "one trick" well

*edit*
I confused you with Alexandr, yes what you're doing does seem pretty tricky.
Whether or not it's trickier than everything I've outlined (which I suspect is nigh on impossible) I don't know.

<SNIP>
I am interested in building something that you are looking for as I would like the fans on my radiators to vary with cpu temps. I know there are IC's out there that can change output based on input from a third source. Its just a matter of finding a schematic for using one.
<SNIP>

Thanks for your thoughts,
Alas nothing that offers any concrete solution/path for what I require.
But hopefully as I learn more this thread will be of some use to you.

I would look around but I'm currently figuring out how to make a volcano vaporizer for my gf.
<SNIP>

[Waterlogged]

Yep & I've seen them, they're all for "one trick pony"* controllers, none come near the multi-functional device/s I'm seeking.



LOL, not sure how you work that one out...
But as you stated, your main goal is just to confuse the situation

Anyway good luck with you project!
*not that there's anything wrong with that, so long as it does that "one trick" well

I have to admit, your OP had a "TLDR" vibe early so I more or less summarized it by the thread title and post header. I may have misunderstood the "dual mode" part. Are you saying you want to be able to do PWM or linear control from the same circuit? Maybe if you could summarize the need into some firm outline of requirements only (minus all the questions), it'd be a little easier to say how easy or impossible it'd be to do.

[jalyst]

Hmm, I thought I'd made it pretty clear what I'm wanting from a discrete controller.
Perhaps not in the 1st 3 posts, as my main focus there was suitability of my MB for what I'm wanting.

But since I've ascertained (mainly another forum) that my MB will be 'acceptable' for the short-term...
The focus has shifted to discussing whether there's a discrete controller that meets 'extra' requirements.
And I've subsequently outlined those requirements....
And if not, whether one can be "hacked together" using off-the-shelf controllers already out there.

Perhaps the main mistake I made was using one thread for two separate discussion points.
My apologies, if it's all a bit too verbose for you, I'll type out a thorough & concise summary ASAP.

[Dr.Joe]

The t-balancer does not support 4pin PWM fans. They do however work, if they can be controlled by voltage regulation. Some PWM Fans dont start if the 12V input is below 11V.

The t-balancer uses a technology that is identical to what i.e. lamptron uses, hence the very high wattage rating for the t-ban in PWM mode and the lamptron units, This technologz is also called pwm, but in that case it isnt used as a signal like in 4 pin fans but to reduce the source voltage for the fan.

[fgw]

They don't interface with our MB/BIOS/EFI & adjust LV/PWM of their channels independently, based on temp readings passed from the MB.

some (may be each) channel of the fan controller can be connected to the MB and use the PWM signal the MB sends to control the fan.

Plus RPM data of each fan can't be passed back to the MB

did not think of routing back the rpm signal to the MB, as rpm can be read through the fan controller anyway, but this might be useful when the MB/bios is used to monitor fan rpm and probably shutdown or alarm the system on failure. anyway, this can be implemented easy.

it requires a bay, which I don't need

no bay required here. the fan controller is a box to be mounted anywhere. but keep in mind, as the fan controller will use linear voltage control it will need some airflow to keep temps under control when high load fans are used. this should not be an issue when using standard fans which draw less than 250mA (most of the fans out there will fall into this category). anyway, max possible current per channel will be 3A. not decided yet how many channels will provide this.

And the input 4-pin molex connection is only suited for 60-80W, so even running 2-ch at full capacity could get dangerous.

have to check the specs to see what a single 4pin molex is able to supply. if one is not capable of doing the job, i might consider using a second one but i doubt it will be required.

And I may eventually replace the AP-15's on my Rad with 2x 5400RPM GT, if testing of 1x goes well.

powerwise this should not be a problem.