Overclock.net banner
1 - 20 of 76 Posts

·
Old, Senile, and Happy
Joined
·
4,977 Posts
Discussion Starter · #1 ·
Hey gang,

I was hoping for a section for DIY electronics, but couldn't find anything like that.

So since this project is for aimed at being able to control PWM pumps like the 35X or new PWM version of the D5 from a normal variable voltage fan controller I figured the Water Cooling section was the best place for this thread.

As many of you know, a lot of my builds feature some custom electronics that make them unique and special.

All of my builds with PWM pumps have custom PWM controllers to set the pump speed manually, or be able to switch it to mobo control.

I got some PM's recently asking about being able to control a PWM pump from a fan controller, and one suggested a circuit that's been around for a while, that after giving it some thought, seemed adaptable as a base to work from.

Here's the link, the second one down is the one of interest:

http://www.overclockers.com/forums/showpost.php?p=6473318&postcount=25

I was already planning to use the new Lamptron CW611 controller in the "Mix 'n Match" NZXT 630's builds, so being able to use a pair of 35X's in a setup where I could switch control between manual or mobo speed control and control by the CW611 seemed like a cool idea.

One thing I've learned from building several iterations of DIY PWM controllers, is that you have to build them so that you have a fair amount of trim control for the minimum % and the max % points.

You typically don't want to run a pump down below 15 to 20% or so, and most are at max speed by 70% to 75%.

It's nice to have trim pots set up so that the main speed control runs the pump from the minimum desired speed to the max speed over all of its range, not just a small portion somewhere near the middle of its range. . . ..

And to make it more fun, not all pumps, D5 / 35X are the same, and PWM fans are different yet, as to how they react to the PWM %.

The key part of the circuit that makes it adaptable to voltage control, is the voltage divider network R3, P1, R4, which controls the trip point of the LM311 comparator.



With the circuit set up, it was easy enough to measure what voltage was needed to reach 100% duty cycle, and what voltage threshold was required to rise above 0%.

Knowing what the 0 to 100% duty cycle voltage span is, allowed me to set up a finely adjustable voltage divider network to convert the variable voltage from the fan controller down to a range equal to that span.

The small problem is that the fan controller output starts at 0V, while the 0% duty cycle voltage is about 2.2V.

To solve that problem, I made a finely adjustable voltage divider network and a non-inverting summing amplifier to add the first 2.2V, so that the swing of the fan controller output can swing the PWM from 0 to 100%

In the pic below, the smaller rear protoboard has the fan controller output voltage divider network on the far right, with the "base level" voltage divider network just to its left.

The IC to the left of the dividers is a dual unity gain buffer that feeds the summing amp in the IC to its left. Also in that IC is the unity gain buffer for the manual speed control circuit's voltage output.

The larger front protoboard, is essentially the circuit in the schematic pic, with a LM339 quad comparator IC in place of the single LM311, as the output of the right NE55 IC needs additional conditioning.

The trim pots at the far left are for the 0% and 100% adjustments.

The trim pot on the lower board is the frequency trimmer.



This pic shows the speed control and the switch for fan controller or manual speed control operation.



Here it is set up on my electronics bench: . . .It's in manual mode and set at about 12% as a minimum:



Here's ~95%, and you can see that the frequency is right at 25KHz:



Here's 50%, notice the frequency stays right at the 25KHz mark:



Close up of what I have set as the minimum %:



Here's some pics with the PWM being controlled by the Aerocool X-Vision (display position 4)

Here's 100% . . . full speed on that fan of 1100 rpm 11.9 volts out from the X-Vision:



~70% . . . 11.0V out from the X-Vision . . ..1000 rpm:



~40% . . . .5.6V out from the X-Vision 600rpm



~20% . . . 3.9V out from the X-Vision . . . . . 600rpm



Next phase is to solder it up on some perfboard.

Not sure yet whether I want to go to the trouble to design a pcb for it yet.

Hope you like my project,

Darlene
 

·
Registered
Joined
·
2,345 Posts
Well if you end up designing a PCB that I can just drop the parts onto and solder I would be interested in it. This is great and def something I would love to have.
 

·
WaterCooler
Joined
·
2,542 Posts
I really like this idea, at least what I understood
biggrin.gif
I want to control my second pump too (mcp35x) as it will only cool my graphics card. But currently I'm forced to hook it up to the mobo.
So does it matter that the mcp35x has a 4 pin control cable and the Lamptron CW611 controller only has 3 pin?
 

·
Old, Senile, and Happy
Joined
·
4,977 Posts
Discussion Starter · #5 ·
Quote:
Originally Posted by givmedew View Post

Well if you end up designing a PCB that I can just drop the parts onto and solder I would be interested in it. This is great and def something I would love to have.
I'll put it on perfboard first, and probably use a piggybacked dual board configuration. Makes for a more compact setup, and it's easier to work up one board at a time.

When I built the first tach signal translation board for FC9 controlled fans, I did it on perfboard, never figuring I'd need more than that one . . . then I needed another, and then when I needed a third one I facepalmed and realized I shoulda done a pcb from the get - go.

Hopefully, I've learned from that experience and I'll just do a pcb for this once the first perfboard one is working.

Quote:
Originally Posted by sadeter View Post

I think I'm lost somewhere between "Wow . . ." and "What . . ?"
Understanding how it works isn't all that important as just seeing that it does and recognizing its usefulness.

PWM controlled fans and pumps are becoming more mainstream as people come to see the advantages of PWM.

There's a wider range of control where the slower rpm range can be extended downwards for quieter operation without loosing the ability of the fan or pump to start from stopped/off while at that slow setting.

Thanks for looking,

Darlene
 

·
Premium Member
Joined
·
15,004 Posts
Nice little device, but may I ask why did you go with such a brutal amount of ICs? Two 555s can do the same work as well...
smile.gif


Edit: I see, you wanted duty cycle control with voltage. Said nothing!
 

·
Old, Senile, and Happy
Joined
·
4,977 Posts
Discussion Starter · #7 ·
Quote:
Originally Posted by SortOfGrim View Post

I really like this idea, at least what I understood
biggrin.gif
I want to control my second pump too (mcp35x) as it will only cool my graphics card. But currently I'm forced to hook it up to the mobo.
So does it matter that the mcp35x has a 4 pin control cable and the Lamptron CW611 controller only has 3 pin?
You'd need to make up some adapter cables.

The 35X's connector needs 12V from a molex to its power pins.

A wire from the ground and tach signal pins needs to go to the CW611, as it has to have a tach signal to output some controlled voltage level.

You'd then need wires from the +V and ground pins of the CW611 output to the PWM converter and wires from the converter output to the ground and PWM pin on the 35X's connector.

I have two CW611s coming this week, as well as a pair of 35X's and other stuff.

The 611 is so new that there's not much info on it yet, so I'll have to figure out all the details once it's here to work with.

Will be updating frequently with how it's all progressing,

Darlene
 

·
Registered
Joined
·
2,583 Posts
Quote:
Originally Posted by IT Diva View Post

You'd need to make up some adapter cables.

The 35X's connector needs 12V from a molex to its power pins.

A wire from the ground and tach signal pins needs to go to the CW611, as it has to have a tach signal to output some controlled voltage level.

You'd then need wires from the +V and ground pins of the CW611 output to the PWM converter and wires from the converter output to the ground and PWM pin on the 35X's connector.

I have two CW611s coming this week, as well as a pair of 35X's and other stuff.

The 611 is so new that there's not much info on it yet, so I'll have to figure out all the details once it's here to work with.

Will be updating frequently with how it's all progressing,

Darlene
Ha ha Like your brains "and your electronic skills" Subscribed
 

·
Old, Senile, and Happy
Joined
·
4,977 Posts
Discussion Starter · #11 ·
Quote:
Originally Posted by 47 Knucklehead View Post

If you go to production on the boards (from a house like PCB123), count me in. I think this would be slick.
I get email adds from them regularly, lol.

I've been thinking on this and how best to make it useful.

I'm going to perfboard it first, with some additions to what's now the smaller board, so that everything can interface efficiently, with pigtails long enough to reach back to the cable management area.

I believe I can size it so that 2 of them can fit in a drive bay with the switches and speed control pot able to mount to a front panel.

There will be a molex pigtail for power in, both 5V and 12V.

There will be two 4 pin pigtails out that supply 12V power, PWM signal, and tach lines for a pair of pumps. (and likely 2 additional molex pigtails with 12V only to supply pumps with molex connectors)

There will be a pigtail out to the CW611 to supply it with a tach signal, and bring its variable voltage output to the PWM controller.

There will be a pigtail out that supplies a tach signal to the mobo's CPU fan header and picks up the mobo's PWM signal so that can be used, optionally, to control the pump speed.

The CW611's are coming express mail, so should easily be here for the weekend.

Not so sure on the 35X's as they are coming priority mail, and because of the holiday, won't even ship until today.

Once I get the perfboard prototype built, I can start testing it on one of the dual D5 / dual res assemblies that are ready to go in the stretch build until I get the dual 35X / dual res assembly built for the 630 that's build with the GT-X rads. (oooops . . .let the cat out of the bag there a bit, didn't I)

Here's an old pic of one of the dual D5 / dual res assemblies, and since it has a flow meter installed, it'll be a good opportunity to see how the CW611 works with flow meters.



Stay tuned for more,

Darlene
 

·
Old, Senile, and Happy
Joined
·
4,977 Posts
Discussion Starter · #12 ·
Quote:
Originally Posted by Jetskyer View Post

I wonder...
Would it be possible to hook up a temperature sensor instead of a fancontroller to control the PWM setting?

Genius design and thanks for sharing, +rep!
Most higher end fan controllers have integrated temp sensors and control the fan speed based on a user selected setpoint if you want them to.

That's one of Lamptron's claims for the new CW611, that it's suited particularly to temp based pump control.

Darlene
 

·
Registered
Joined
·
235 Posts
Can I just say... ummm wow!
 

·
Old, Senile, and Happy
Joined
·
4,977 Posts
Discussion Starter · #15 ·
Quote:
Originally Posted by Electrocutor View Post

If you're wanting to go all out on your circuit, you should probably include a 12v burst on startup when using it for lower voltage to make sure that whatever is plugged in will start moving before being tuned down.
I believe that you're missing the concept of this project.

This controller outputs only the PWM signal for control of those fans and pumps designed to have their speed controlled by a PWM signal.

PWM signal, for lack of a better word, is a variable duty cycle square wave form with a high level of 5V and a low level of 0V at a frequency of 25KHz.

One of the key advantages of using PWM devices is their ability to start at low speed settings.

Perhaps think of this controller more as a converter, in that it can convert a variable analog voltage level input from a common fan controller into a proportional duty cycle fixed voltage square wave output.

Analog in translated to PWM out.

If the variable voltage controller this was being used with had a 12V startup burst, that would be translated as a ~100% duty cycle output burst.

Darlene
 

·
8.2 liters
Joined
·
5,574 Posts
Quote:
Originally Posted by Electrocutor View Post

If you're wanting to go all out on your circuit, you should probably include a 12v burst on startup when using it for lower voltage to make sure that whatever is plugged in will start moving before being tuned down.
Quote:
Originally Posted by IT Diva View Post

This controller outputs only the PWM signal for control of those fans and pumps designed to have their speed controlled by a PWM signal.
And therefore the fan or pump gets a direct 12V from elsewhere
smile.gif
 

·
Registered
Joined
·
2,345 Posts
Quote:
Originally Posted by Electrocutor View Post

Ah, I'm sorry. I misinterpreted. I thought you meant to allow a 4-pin header that could accommodate both pwm and voltage-controlled fans. I tried a number of ways to modulate a normal 3-pin fan, but the results were always too noisy (buzzing).
Yeh I experienced that buzzing/rattling with the bigNG fan controller in PWM mode.

What she is doing is awesome though... I only run PWM fans and pumps now and would love a little more control since right now I am limited by the single CPU tied PWM signal that my board outputs the other PWM is off the motherboard temp which is useless. Which means I have to run the same exact PWM signal to all my fans and my pump.
 

·
8.2 liters
Joined
·
5,574 Posts
Quote:
Originally Posted by givmedew View Post

Yeh I experienced that buzzing/rattling with the bigNG fan controller in PWM mode.

What she is doing is awesome though... I only run PWM fans and pumps now and would love a little more control since right now I am limited by the single CPU tied PWM signal that my board outputs the other PWM is off the motherboard temp which is useless. Which means I have to run the same exact PWM signal to all my fans and my pump.
As far as I recall, the bigNG modulates the 12V signal and doesn't even have a connection for the fourth PWM wire. There is really a dearth of true PWM fan controllers. Zalman's ZM-MFC controllers offer one PWM connection, and the rest are voltage controlled. The Aquaero 5 also has a single PWM header. Apart from that I don't know of any more, so you're often indeed stuck with what your motherboard provides.

One trick if you are watercooling, is that with a fullcover block, you have a spare PWM connector on the GPU. Unfortunately it's got a kind of mini connector, but Gelid makes an adapter for this. Thanks to this I'm using my GPU to control my pump, and the two PWM headers on the motherboard to control push and pull fans independently.
 

·
WaterCooler
Joined
·
2,542 Posts
1 - 20 of 76 Posts
Top