WC specific fan controller, any interest? Thoughts?

I was contemplating fan controllers and such and had a sudden thought: Are there any fan PWM controllers out there specifically aimed at water cooling?

This is what my brief daydreaming came up with, an automatic radiator fan controller.


It'd have a USB plug for programming your values in via software. This would not need to be plugged in for daily use.

It would not have an always-on/required-to-run piece of software. This has upsides and downsides. Upside: No system resources consumed, nothing to load on boot, I don't have to try to write a driver. Downside: No CPU temp based control. I don't think this is a major negative, especially if there are GPUs in the loop.

It would have 3-9 powered 4pin PWM fan headers. Probably either three or six. These would have a strong enough signal that they could drive multiple fans. A tremendous number of fans if the 12V and GND for fan drive power were coming from somewhere else.

It would have at least two temp sensors.
One to be attached to the radiator end tank (yay hot glue, or something) or stuffed into a gap in the fins with the airflow to that area blocked by foam or something (to avoid airflow cooling the sensor). An in-loop water temp sensor could be connected and calibrated for.
The other temp sensor would be in the radiator intake air flow, be it inside the case or external air that is being used.

Fan speed would be calculated based on a couple variables, and an arbitrarily large number of static values.
My current thoughts:

Fan speed primarily set by the deltaT of radiator to intake air. The larger the difference between the radiator temp and the air temp, the faster the fans run. The fan curve of dTWM% would be programmable via software (with the value being saved through power cycles), at least ten points on the curve. Potentially a lot more (probably not more than 255 points, though).

Additionally there would be a given loop temperature (adjustable via software) at which the fan turn on regardless of dT, in case of very strange airflow due to a window being open, cats eating the sensor, kids poking the sensor into the fans which eat the sensor, etc.

Optional manual control via physical knob or software is certainly possible as well.


Now Water Cooling Forum, I ask you: What are your thoughts? Would you buy something like this? What is it missing? How much would you pay if it existed? What does it have that it does not need?

The Aquaero products do most of what you describe already. Their software is only required for setting up your fan curves or if you want to base a fan curve off some system temperature such as hard drive or CPU temp. Once it's set up, you don't need the software running at all if you're not using system temps for anything.

The only thing that you describe which the Aquaero can't do is the part where you describe a sort of "fan override" in case of a sensor malfunction.

Not sure I quite understand ?

What would be the point as would you not want the controller to use CPU and GPU temps as a barometer to controll the potency of the the fans ?

Controlling the fans using water temps would it not be too sensitive to ambient temp swings ?

I do not see any way around not haveing the software enabeled so it can ajust on the fly, hence the usfulness of a PMW controller

With out software running you could argue that simply placing temp probes at the varibels that you want the fans to run under and measuring it to make adjustment manually with a rheostat.

What it would be great for is measuring Delta T on the fly and how load effects it, just not convinced I would want my fans taking orders from these readings

Still do not think I have my head wrapped around this, these are questions that just poped out as I read it

Control would primarily be based on the wster:air dT, so ambient temp swings wouldn't be an issue. The thought behind going off the dT is that it's the one and only variable that the fans have direct control over, as well as the only one that takes
both cpu and gpu load into consideration.
Example: 130w cpu and 300w gpu. A cpu-only load is going to cause much less heat in the cooling system and hence require less fan speed. Going off core temps you get a big spike in fan speed that isn't going to do anything very useful if the loop is sized for the 430w total combined load, odds are that it can eat the 130w without raising fan speed much at all.
It also avoids the issue of a gpu load not increasing the fan speed due to the fan speed being based off cpu temps.

Mostly it would be for allowing noisier fans to be used with a minimum of noise in general conditions.
An automatic version of you turning up the fan speed knob when you play games and turning it down when you're browsing.

All the above said, this is exactly why I'm asking questions of the WC section before I spend time and money developing such a thing!


EDIT: This also stems from how annoying I find the always-on software from the software fan control AIO systems. I like the functionality, I hate having software eating resources.
Also how annoying the massive jump in fan speed in load is when using cpu temp based control on the motherboard. Fine for normal boring fans, really annoying for serious business fans.

Hmmm.. that is interesting take.

I will have to put some thought into this.

I am trying to get a grasp on how it would work on the fly with out software controll. The unit it self would simply have to measure temp at chosen locations as you explained. you would have to get a switch that could work simply off of temp, in the same essence as your thermostat works in your home. A thremostaic switch would be one thing, but to get one that is not on or off but rather throttels the voltage would be daunting assuming no software support ?

If I understand this correctly, have temp probes in certain spots and having the fans adjust to the set curve based off of the DT of the system? Hmm, I'd possibly be interested if I could do that and do among other flexible things like base it on CPU, GPU, water temp or Delta if I understood this or set fan curves etc. Also, being able to control pumps and read out temps etc. Basically what a Aquaero 6 does because it has PWM now at a much LESS $220/$230 price tag. The only other product in the states that has tried and failed miserably is the $150+ paper weight corsair link I am still using. If you could say do all that for $100 price tag around I think I most likely would be interested. Capable of controlling 9+ fans + pumps + temp probes. I really don't mind having the software monitoring going on on one screen.

Edit.....ideas popped to my head that i forgot......edited whole post

A microcontroller would be running it. Said MCU would be reading the values and comparing them to the programmed fan map and generating the PWM signal based on that.
So there is software, but it's only running on the MCU not on your PC.

The moment it starts displaying anything, the price goes way up. Even just a four line alphanumeric LCD display adds $30 to it in hardware alone.
Getting temps of the CPU/GPU requires using a USB port and active software on the PC. Using a USB port directly (rather than a USB-serial dongle) means buying an ID number from the group that controls USB, and that starts at $5k or so.

Fans or pumps wouldn't make a difference, as long as they accept PWM input. Now running a pump based on water:air delta probably wouldn't work well. You could slap a sensor on the waterblock though.

As far as i know/remember either the mCubed/T-Balancer or the Alphacool HeatMaster work that way... you just need the software to "program" the settings/curves and once happy with it all , save to chip.
As of then the soft is no longer needed unless you want to monitor what's going on and display it on screen. I think the Aquaero works the same way. There a forum somewhere dedicated to the Aquaero

Of course the Heatmaster & T-balancer are A LOT cheaper as the Aquaero

PS: doesn't the Bitfenix Recon has similar (but a lot less) features?


PPS: doesn't the NZXT Kraken work that way already? Software only needed to "program". (not that it'll help unless you take the whole thing apart & re-purpose it)


PPPS: or you could just bolt an arduino inside your case somewhere

.

Bobnova said:

A microcontroller would be running it. Said MCU would be reading the values and comparing them to the programmed fan map and generating the PWM signal based on that.
So there is software, but it's only running on the MCU not on your PC.

The moment it starts displaying anything, the price goes way up. Even just a four line alphanumeric LCD display adds $30 to it in hardware alone.
Getting temps of the CPU/GPU requires using a USB port and active software on the PC. Using a USB port directly (rather than a USB-serial dongle) means buying an ID number from the group that controls USB, and that starts at $5k or so.

Fans or pumps wouldn't make a difference, as long as they accept PWM input. Now running a pump based on water:air delta probably wouldn't work well. You could slap a sensor on the waterblock though.

Click to expand...

You are basically describing a AQ6 Pro/XT. It can run without software with it's micro processor and lcd screen, or with software for keyboard/monitor display.

The AQ5LT can do this with software, no lcd display on the controller, at one third the cost of the AQ6.