Overclock.net › Forums › Cooling › Water Cooling › DIY PWM Pump / Fan Control . . . from a most unlikely source
New Posts  All Forums:Forum Nav:

DIY PWM Pump / Fan Control . . . from a most unlikely source

post #1 of 8
Thread Starter 
As many of you all know, I'm kind of an electronics geek, and have a bit of a romance with using PWM pumps in my builds, both D5's and 35X's.

I have, over the years, built a number of PWM controllers, both manual and controlled by an external analog variable voltage input.


I recently ran across something rather unique as I was about to do a little review of the Reeven 6 Eyes and the Phobya Maxguide 6, as both feature 6 displays, and I like to have that level of information available at a glance on my builds.


I made a series of posts in the watercooling club thread as I was investigating what I discovered.

Now that I have all the information that my experimenting has yielded, I'm putting it all out in a thread of its own here, so that it's easy to find, and in hopes that it promotes some discussion.


The easiest way to do this, I think, is to just copy/paste the original 4 posts in sequence. They are easy to follow.


Darlene


* * Geek Alert * *


As most of you who've followed my build logs know, I've shelled out not an insignificant pile of $$ looking for just the right controller / display for the current builds.

I've got a pair of FC 9's, a pair of CW611's, three Phobya Max Guide 6's, and most recently the little Reeven 6 Eyes.


I was going to post a little review of the MaxGuide 6 and the 6 Eyes, but as I started taking some pics and seeing how the new 6 eyes works, . . . .

It struck me that the Max Guide while rated at 36W per channel, has no output transistors, heatsinks, coils, or capacitors that you'd normally expect.

I found that intriguing, and set out to see how it could do what it does without those parts.


I looked up the big chips on the PCB, and the ones with markings are just VFD display driver chips.

There are a pair with all markings removed, however, and that was even more mysterious, so off to the lab I went.

The only way this could work is if they use something akin to a PWM buck converter.


Putting a multi meter on the output to check voltage revealed that without any load, either a fan or resistor, the voltage stayed at ~12V regardless of the control setting.

That's a sure tip off that they're doing things differently . . .

As soon as a load was attached, the multi meter showed ~0.5V in the OFF position, and at the lowest speed attainable, about 5V, scaling nicely to ~12V at max setting.

I also determined that the output is referenced to the system's gnd, as opposed to being referenced to the +12V line which is common with the Lamptron controllers.

Then I put the frequency counter and scope on the output to dig a little deeper . . . .

And here it got totally unique.

With the control at zero, where the display reads "OFF", there's almost zero DC output, maybe a couple tenths of a volt or so.

As the control is increased just to its minimum point above OFF, the output jumps up to 4V DC level . . . With an 8V PWM rider.

The PWM frequency measures at 45.4 to 45.5 Hz.


You can see in this pic, that the output voltage on the multimeter, (yellow one) is showing 4.62V . . . . The duty cycle of the PWM pulse is 11.9 % (gray multimeter), . . . . The frequency is 45.4 Hz, . . . The scope trace has it's zero at the first horizontal line up from the bottom, with a 2V per division vertical setting.

You can see the PWM pulse adds to a 4V base and extends 4 divisions (8V) up to the 12V mark.




The output range from about 4V to 12V is directly proportional to the pulse width, which seems to control the 7.4V over a range of 88% that adds to the base 4.6V that comes at the minimum of ~12%.


What the h3ll does this mean to me, and how is it related to water cooling, you're asking yourself . . . . . . .

And here. I believe is the answer . . .


Most all of us would like to be able to control pump speed easily and on the fly.

Most pumps do not respond very well to speed control from analog voltage output fan controllers, at least not below about 8V, which is often a lot higher RPM than we'd like.

I don't have a plain jane fixed speed pump to test with, but I'll go out on a limb here and predict that using this controller to control pump speed will work really well with any non-PWM pump.

Not as well as a real PWM pump, but damned decently.

Hope some of you find this interesting;

Darlene


Here's a few pics as I raise the output voltage . . . you can see the fan RPM increase and the pulse width increase.

Here's a better pic of the baseline minimum output:




About 25% output:




About 50% output:




About 75%:




Almost at max, ~99%:




And at max, 100%:







Thanks,

And it just sort of occurred to me that I may be able to pick off just 5V of the PWM part of the output, which would then be able to control any 4 wire PWM fan or pump. . . . .

That is if they'll run OK with a 45Hz PWM frequency instead of the 25,000Hz that's standard.

That would be really effin' cool . . . . biggrin.gif


Let you know if it works in a couple hours or so . . .


Darlene




I'm encouraged thumb.gif

I was able to pick off just the PWM part of the output, and limit it to 5V so that it matches the voltage for a standard PWM signal.

Better yet, it only took a few passive parts and can be simplified a bit more with a 7V zener diode.

Naturally, I'll have to order some since all I have close are 5.1V ones, necessitating a divider network to knock down the 6.9V to 5V.

To test so far, I used a PWM CPU fan to see what kind of results I got since the PWM frequency from the controller is only ~45Hz, and PWM devices are normally run with a 25KHz PWM frequency.


I got nice linear speed control with the fan powered from a 12V molex and controlled by the 45Hz PWM component of the controller's output.

The controller displays the RPM just fine as well.

I'll let it run a while tomorrow morning while I have my coffee and decide whether to try it on a pair of $100 pumps.


I'll post an update with pics in the morning,

Darlene





* * * Serious Geek Alert * * *


This post is an update to the really interesting discovery I ran across and detailed in this post:

http://www.overclock.net/t/584302/ocn-water-cooling-club-and-picture-gallery/48470#post_20676443


In short, the Phobya Maxguide6 dual bay controller has a really unique output, whereby once turned above the minimum or OFF position, (where there's only a few tenths of a volt output) there's a 4VDC base voltage, and an 8V PWM rider that starts at about 10% and increases to 100% as you turn the control to maximum position.

It's a very low PWM frequency, only 45.4Hz, but it is PWM.

I thought it might be interesting to see if I could pick off 5V of the PWM and use it as the control signal to control a regular PWM controlled fan or pump that takes its power from a PSU molex.

I posted here that I had pretty good luck with my initial attempt to do so, but some refinement was needed:

http://www.overclock.net/t/584302/ocn-water-cooling-club-and-picture-gallery/48500#post_20678661


The one issue was that the Maxguide6 needs a significant load before it produces a decent looking pulse shape, which prompted a RadioShack run this morning.

With a minimal load, the trailing edge of the pulse drags out instead of cutting off cleanly which effectively makes it behave as a wider pulse width percentage than it is.

That yields a higher effective output voltage and higher speeds on whatever it's running.

One of the prime reasons for using PWM devices is to gain lower operating speeds and better low speed control and start up.

With that concept in mind I worked up a little adapter board that takes the Maxguide 6 output, picks off 5V of PWM and conditions the pulse shape by adding a resistive load, which then can be used as the PWM control signal for normal 4 pin PWM fans or pumps.


To give you some more details, here's the pics:

These first few show the actual output of the Maxguide6, before splitting out the PWM component.


Basic test bench setup;

The scope is showing the 4V base with the 10% duty cycle 8V PWM rider. The effective voltage is 4.6V on the yellow meter and the fan rpm is 750. . . . With the 3W fan and a 7W resistor pack attached, the wave form is pretty fair.





Now here's the same setup with the load resistor plugged out:

Fan is at 1020RPM, voltage is showing at 5.7V, PW showing at 19.5%, The trailing edge of the pulse is badly drawn out, not nicely cut off.





Here's a close up to show the details:




And here's a close up with the load resistors plugged back in:

The effective voltage is back where it should be, along with the PW percentage and the fan RPM




These next few pics show the picked off PWM component of the output.


The little board loads the output to shape up the PWM pulse, and then a zener diode is used to drop out the first 5.1V, so all that's left is 7V of pulse, which a resistive divider then gives me the 5V needed.


Here's the little board:




Here's the 5V PWM signal; The shape is nice and clean, 11% duty cycle, and an effective voltage of 4.6V

The PWM cpu fan speed is 1260RPM, which is really close to the 1200 RPM minimum that it gives from an actual 0% pulse width.




Since it seemed like everything was working surprisingly well, controlling well at 45Hz where the norm would be 25KHz, I plugged up the Maxguide6 and the little board to one of my dual 35X setups, hoping like h3ll that I wasn't blowing $200 down the tubes . . . . .

Damn this was an awesome hack, works fine, with the lowest I can get at ~1600RPM . . . and of course it maxes out easily.




The really cool thing here is that with one controller, you could control all your rad fans (36W per channel) and with one little board, you could use another channel to control a PWM 35X or D5.


Hope you found this interesting,

Being able to PWM control regular 4 wire PWM pumps from a controller that isn't supposed to be a PWM controller at all, with only a few inexpensive passive parts from RadioShack is a pretty awesome hack IMHO.



Darlene
Edited by IT Diva - 8/30/13 at 10:08am
post #2 of 8
Wonder if your brilliance can shine doing the same thing for Laing DDC-1T pumps that have a PWM 4th wire in the mix. Basically, people just clip the wire off and run the pump a full max.

Nice work.

BMaverick
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
Athlon 64 x2 5000 Black 2.6Ghz OC 3.3Ghz GigaByte GA-MA78GM-S2H Full Radeon 780G HD 3200 4GB Corsair DDR2 6400C4 
Hard DriveOSPowerCase
Seagate Barracuda 750GB 7200 RPM 32MB Cache Windows 7 Pro x64 Corsair HX620W Meridian XClio High Tower I 
  hide details  
Reply
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
Athlon 64 x2 5000 Black 2.6Ghz OC 3.3Ghz GigaByte GA-MA78GM-S2H Full Radeon 780G HD 3200 4GB Corsair DDR2 6400C4 
Hard DriveOSPowerCase
Seagate Barracuda 750GB 7200 RPM 32MB Cache Windows 7 Pro x64 Corsair HX620W Meridian XClio High Tower I 
  hide details  
Reply
post #3 of 8
Quote:
Originally Posted by bmaverick View Post

Wonder if your brilliance can shine doing the same thing for Laing DDC-1T pumps that have a PWM 4th wire in the mix. Basically, people just clip the wire off and run the pump a full max.

Nice work.

BMaverick

Yeh I really want to know this. If bmav isn't willing to I will give you one of those pumps to play with. Although I never really believed that wire was actually PWM but bmav would know more than me about it.
Silence
(20 items)
 
Secondary Rig
(9 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5820k  ASRock x99 extreme 4 AMD R9 290X AMD R9 290X 
RAMHard DriveCoolingCooling
ballistix ddr4 dual channel Samsung 850 EVO 500GB DT 5noz CPU Waterblock XSPC Razor Full Cover GPU Block 
CoolingCoolingCoolingCooling
XSPC Razor Full Cover GPU Block 4x120 Black Ice Xtreme GTX480 Swiftech MCP655 Danger Den Monsoon Pump/Reservoir 
CoolingOSMonitorKeyboard
Corsair Link Fan Controller Windows 8.1 Nixeus 2560x1440 Logitech G15 
PowerMouseMouse PadAudio
Kingwin PF-850 Ttesports Level10M Razer Vespula Sennheiser HD590 
CPUMotherboardGraphicsRAM
Intel i5-2500k gigabyte z68xp-ud4 Gigabyte GTX 970 G1 Gaming DDR3 1600 4x2GB 
Hard DriveCoolingOSMonitor
Western Digital Velociraptor 300GB w/ Intel RST... Corsair H100 Windows 8.1 1920x1200 HP 
Power
PC P+C 1000w 
  hide details  
Reply
Silence
(20 items)
 
Secondary Rig
(9 items)
 
 
CPUMotherboardGraphicsGraphics
Intel i7-5820k  ASRock x99 extreme 4 AMD R9 290X AMD R9 290X 
RAMHard DriveCoolingCooling
ballistix ddr4 dual channel Samsung 850 EVO 500GB DT 5noz CPU Waterblock XSPC Razor Full Cover GPU Block 
CoolingCoolingCoolingCooling
XSPC Razor Full Cover GPU Block 4x120 Black Ice Xtreme GTX480 Swiftech MCP655 Danger Den Monsoon Pump/Reservoir 
CoolingOSMonitorKeyboard
Corsair Link Fan Controller Windows 8.1 Nixeus 2560x1440 Logitech G15 
PowerMouseMouse PadAudio
Kingwin PF-850 Ttesports Level10M Razer Vespula Sennheiser HD590 
CPUMotherboardGraphicsRAM
Intel i5-2500k gigabyte z68xp-ud4 Gigabyte GTX 970 G1 Gaming DDR3 1600 4x2GB 
Hard DriveCoolingOSMonitor
Western Digital Velociraptor 300GB w/ Intel RST... Corsair H100 Windows 8.1 1920x1200 HP 
Power
PC P+C 1000w 
  hide details  
Reply
post #4 of 8
Thread Starter 
Quote:
Originally Posted by bmaverick View Post

Wonder if your brilliance can shine doing the same thing for Laing DDC-1T pumps that have a PWM 4th wire in the mix. Basically, people just clip the wire off and run the pump a full max.

Nice work.

BMaverick


I was rather at a loss in my understanding what you were asking, until givmedew's post tuned on the light bulb.


If you were wondering if I'd have a go at figuring out the 4th wire to find a way to have it control speed, I could do that.


I would have thought though, that this would have already been documented in its industrial application origins.


Darlene
post #5 of 8
Very nice read, and makes sense to collate it here. thumb.gif

You will have to correct me here, i think, but can this be done with any controller that uses 12v pwm signals? Or is this exercise only repeatable on this device?

If not, then i will order a phobya controller, but if so, then i would love to try this with http://www.crystalfontz.com/product/CFA633TFHKU.html
post #6 of 8
Thread Starter 
Quote:
Originally Posted by NimbleJack View Post

Very nice read, and makes sense to collate it here. thumb.gif

You will have to correct me here, i think, but can this be done with any controller that uses 12v pwm signals? Or is this exercise only repeatable on this device?

If not, then i will order a phobya controller, but if so, then i would love to try this with http://www.crystalfontz.com/product/CFA633TFHKU.html


The way this particular device creates its output is rather unique, and probably only shared by other controllers produced by the OEM, Sunshine Tipway electronics.

http://www.stw.hk/english/nproducts.asp?page=1&newid=2&classid=

Notice that it has both a pure fixed level DC component of 4V and then the 8V variable pulse width is added to it.


The crystalfontz should be easier to set up.

According to its tech sheet, it has 18KHz PWM on its power pins on each fan connection.

Keep in mind that it does not natively support 4 wire PWM fans or pumps . . . . it actually pulses the output power to normal 3 wire fans/pumps.

What it doesn't tell you, or at least I couldn't find, is what the max load is rated at, but used in its native configuration, that would be important to know.

The up-side is, if it's correct that the output power is fully PWM, is that a simple zener diode circuit could be used to translate the output to a 5V level, whereby it could be used exactly as the Phobya output, as just a PWM control signal with the fans/pumps being powered from a PSU molex.

That would render it's max output level rather insignificant, as PWM signal current is at the milliamp level at most.

The down side is that when the crystal fontz reads the fan/pump RPM, it changes the pulse width momentarily while doing so.

I have no idea whether or not this would impact being able to read the RPM if you were translating the output to be just a 5V PWM control signal.

You'd have to try that and see.

You really need to look a the output on a scope to know some things about it for sure.


Darlene
post #7 of 8
Max recommended amperage is up 1.5A per set of pins, up to 4A with all four in use.

I originally put the idea of using this controller aside because i couldn't use a pwm fan, but if you were willing to make a list of parts i may need to create a zener diode circuit to drop the 12 v to 5 v i would order one of these right away and try and give it a shot.

Hell, i might be willing to have one shipped your way to see if you could come up with something. Martin uses one of these for his temp readings (sensors are digital temp probes), you might even rethink your entire pwm controller idea you created for him if you can come up a working daughter board that would give a crystal fontz a proper 5volt signal and rpm reading.

Also, i read through the cfa633 manual (i actually have done that a number of times redface.gif ), the controller only changes the pulse width when it has a problem reading the rpm signal. On startup it ramps all connected devices to 100% for a few seconds so it can get a good pulse reading off the rpm wire, then ramps down to desired. I dont know how it could lose the signal after that, but oh well...
post #8 of 8
Thread Starter 
Quote:
Originally Posted by NimbleJack View Post

Max recommended amperage is up 1.5A per set of pins, up to 4A with all four in use.

I originally put the idea of using this controller aside because i couldn't use a pwm fan, but if you were willing to make a list of parts i may need to create a zener diode circuit to drop the 12 v to 5 v i would order one of these right away and try and give it a shot.

Hell, i might be willing to have one shipped your way to see if you could come up with something. Martin uses one of these for his temp readings (sensors are digital temp probes), you might even rethink your entire pwm controller idea you created for him if you can come up a working daughter board that would give a crystal fontz a proper 5volt signal and rpm reading.

Also, i read through the cfa633 manual (i actually have done that a number of times redface.gif ), the controller only changes the pulse width when it has a problem reading the rpm signal. On startup it ramps all connected devices to 100% for a few seconds so it can get a good pulse reading off the rpm wire, then ramps down to desired. I dont know how it could lose the signal after that, but oh well...



The Dallas Semiconductor temp sensors, (DS1820series) are sweet.

Each one has its own ID string, so you can string lots of them along a 3 wire buss, or even a 2 wire buss with the ones that allow parasitic power sourcing from the data line if the microcontroller employed allows for that.


Zener diode circuits are super simple voltage regulator circuits with only 2 required parts, the zener and a resistor.

If it's for DC, you'd usually add a capacitor to smooth the output, but to chop off part of a PWM source voltage, just the zener diode and a resistor is all you need.

Here's a little idea of how simple it can be:

http://www.electronics-radio.com/articles/electronic_components/diode/zener-diode.php


You could get the parts at a local RadioShack. They have 5.1V zener diodes and use a 270 or 330 ohm, 1/2 watt resistor.

I'd have to have a crystalfontz and look at the output on a scope to be totally sure it's a normal PWM pulse, but being that it's all digital, it's probably a reasonable assumption that it is.


If you get the controller, get a proto board and some zeners and resistors at radioshack and use the pics in the link to put the circuit together.

If it works as expected, solder your parts on a perfboard and you could put several zener circuits on one piece of perfboard and have multiple PWM channels.


Darlene
Edited by IT Diva - 9/4/13 at 2:37am
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › DIY PWM Pump / Fan Control . . . from a most unlikely source