How do I drive these Delta fans with varying speed?

[iamdarkyoshi]

Hello everyone, I have recently obtained a box of like 30 or so delta fans from a source that shall remain unnamed for now.

The red and black are positive and negative, but the yellow and blue wires have me stumped.

 

 

I thought they might be the standard yellow tach and blue PWM wires of a standard PC fan, but they aren't, I terminated them in a 4 pin fan header and the fan ran at full speed. So I thought maybe my dell test motherboard was being... dell, so I just connected the fan to a 12v power supply and touched the blue wire (normally a PWM input) to ground, expecting it to shut down the fan (0% PWM) Nothing changed, still ran at full speed. So I tried the yellow wire (normally a tach wire on most fans) and oddly, it shut the fan down. So I figured maybe delta just switched the tach and PWM colors around, so I did the same, and the fan didn't move at all when connected to my dell testbench. Then I tried voltage controlling it, it was still pretty fast at 7v and wouldn't even move at 5v, so voltage control seems unlikely.

Any ideas? I can't seem to find any datasheets for the 4 pin flavor of this fan, only 2 and 3 pin versions.

 

 

Thanks!

[W-L]

4 minutes ago, iamdarkyoshi said:

-SNIP-

Those look to be a 4 pin PWM variant see if it plays well with an external power source and a PWM control signal, the PWM based signal is not just a simple voltage but a frequency waveform that controls the speed,

 

Looks like this unit here:

http://www.pchub.com/uph/laptop/656-67218-19058/Delta-Electronics-AFB1212VHE-Server-Square-Fan.html

[iamdarkyoshi]

2 minutes ago, W-L said:

Those look to be a 4 pin PWM variant see if it plays well with an external power source and a PWM control signal, the PWM based signal is not just a simple voltage but a frequency waveform that controls the speed,

 

Looks like this unit here:

http://www.pchub.com/uph/laptop/656-67218-19058/Delta-Electronics-AFB1212VHE-Server-Square-Fan.html

I know how normal fan PWM signals work, and thats why I used a motherboard fan header to generate the supply. Anyway...

 

Update: Broke a bit of the plastic off the back of one of the fans and found the PCB labels for the wires:
Black - GND
Red - VIN
Yellow - PWM
Blue - O/F

Thinking maybe they use 12V PWM instead of 5V PWM, I put 5V on the PWM pin to simulate my motherboard outputting 100% PWM, and unlike the motherboard, this simulated setup ran the fan at full speed.

Now I'm really confused.

[iamdarkyoshi]

2 minutes ago, CUDA_Cores said:

I suppose if you were really despite to get these things working, you could bypass the brushless motor controller and install something like an RC ESC. Then who cares about the controller and what it does. You have your own now.

I was hoping to fit standard headers on these fans and sell them, the box I have is like 500-700$ worth of fans, they are brand new

[iamdarkyoshi]

1 minute ago, CUDA_Cores said:

If that's you plan, first I dentify the pins for power. Then keep randomly switching around pins until you get something to work if that is your goal. If that doesn't work for your purposes, you are ultimately SOL.

Read above, I found the PCB labels for the wires

[W-L]

5 minutes ago, iamdarkyoshi said:

I know how normal fan PWM signals work, and thats why I used a motherboard fan header to generate the supply. Anyway...

 

Update: Broke a bit of the plastic off the back of one of the fans and found the PCB labels for the wires:
Black - GND
Red - VIN
Yellow - PWM
Blue - O/F

Thinking maybe they use 12V PWM instead of 5V PWM, I put 5V on the PWM pin to simulate my motherboard outputting 100% PWM, and unlike the motherboard, this simulated setup ran the fan at full speed.

Now I'm really confused.

If you just shorted them to ground or giving them 5V isn't the best way to simulate a PWM signal, was there nothing when you tried varying the speed when connected to a motherboard header?

[iamdarkyoshi]

3 minutes ago, W-L said:

If you just shorted them to ground or giving them 5V isn't the best way to simulate a PWM signal, was there nothing when you tried varying the speed when connected to a motherboard header?

Its a dell studio motherboard, which starts at full speed then goes down. I don't have a display plugged into the machine, but I confirmed it works with the stock cooler.

[W-L]

3 minutes ago, iamdarkyoshi said:

Its a dell studio motherboard, which starts at full speed then goes down. I don't have a display plugged into the machine, but I confirmed it works with the stock cooler.

Can you manually vary the speed to see if it's affected, but it sounds like it should work to me. Only thing like you said is if they's proprietary for sever hardware but the PWM control should be standardized. 

[iamdarkyoshi]

1 minute ago, W-L said:

Can you manually vary the speed to see if it's affected, but it sounds like it should work to me. Only thing like you said is if they's proprietary for sever hardware but the PWM control should be standardized. 

I'll spill the beans on where these came from, tesla. They were being thrown out...

[iamdarkyoshi]

A user on the EEVBlog forums might be on to something:

The document also mentions that the PWM input voltage ranges are 0.0V to 0.4V for a LOW, and 2.8V to 20V for a HIGH. Thus, connecting the yellow wire to +12V should have it spin at its maximum RPM. The document also mentions that the PWM input will draw 40µA to 20mA. Finally, keep in mind that the PWM signal's frequency must be somewhere between 30Hz and 30kHz.


Now, if you look at the Intel fan specification (http://formfactors.org/developer%5Cspecs%5CREV1_2_Public.pdf) document provided by mariush, section 2.4 states that a fan must not sink more than 5 mA. 

Your motherboard probably cannot source enough current on its PWM control output pin for this fan.

[W-L]

19 minutes ago, iamdarkyoshi said:

I'll spill the beans on where these came from, tesla. They were being thrown out...

Hmm seems like a waste just to toss a brand new box of fans, hey their loss your gain  

 

So in that case the issue is with the board probably not able to provide a strong enough PWM signal.

[iamdarkyoshi]

1 minute ago, W-L said:

Hmm seems like a waste just to toss a brand new box of fans, hey their loss your gain  

 

So in that case the issue is with the board probably not able to provide a strong enough PWM signal.

And then I thought harder about this theory...

 

Now that I think about it, maybe not. If the motherboard can't sink enough current, then it should theoretically go full speed rather than not doing anything, which is what it does unless I disconnect the PWM pin.
 

[W-L]

4 minutes ago, iamdarkyoshi said:

And then I thought harder about this theory...

Now that I think about it, maybe not. If the motherboard can't sink enough current, then it should theoretically go full speed rather than not doing anything, which is what it does unless I disconnect the PWM pin.

Yes that is normally what happens, it should just run full speed, I remember hearing this with the SP series of Corsair fans where they required a very strong PWM signal if more than 6 or so where wired to a power splitter. If it just flat out stops then I'm not too sure why that would be the case. 

[cameroncoats]

I found the datasheet here via digikey - not sure if you've found it yet; it's a slightly different model but as far as I can tell the only difference is airflow.

 

The PWM input has a minimum frequency of 30Hz and is pulled to 0.9Vcc when disconnected which should drive the fan at full speed.

 

The datasheet gives the yellow wire as an open collector tach output - do not short this to a power rail as it doesn't look to have any protection.

 

Are you able to test with hardware other than your dell mobo? I wouldn't put it past dell to use a non standard fan control scheme.

 

No formal spec is given for input impedance but looking at the schematics listed it's a FET input with high value resistors for pull up - I wouldn't expect it to draw more than 100µA @ 5v so I'm not convinced that it's drawing too much current from the PWM pin.

[iamdarkyoshi]

56 minutes ago, cameroncoats said:

I found the datasheet here via digikey - not sure if you've found it yet; it's a slightly different model but as far as I can tell the only difference is airflow.

 

The PWM input has a minimum frequency of 30Hz and is pulled to 0.9Vcc when disconnected which should drive the fan at full speed.

 

The datasheet gives the yellow wire as an open collector tach output - do not short this to a power rail as it doesn't look to have any protection.

 

Are you able to test with hardware other than your dell mobo? I wouldn't put it past dell to use a non standard fan control scheme.

 

No formal spec is given for input impedance but looking at the schematics listed it's a FET input with high value resistors for pull up - I wouldn't expect it to draw more than 100µA @ 5v so I'm not convinced that it's drawing too much current from the PWM pin.

The yellow wire is PWM, I removed some of the plastic and saw the label on the pcb

[Unimportant]

25 minutes ago, iamdarkyoshi said:

The yellow wire is PWM, I removed some of the plastic and saw the label on the pcb

The datasheet he found seems close enough tough. The yellow wire is indeed PWM, it takes a 30Hz to 300Khz PWM signal, with the preferred frequency at 25Khz. The PWM signal voltage is MAX. 15V with logic 0 beeing below 0.8V and logic 1 above 2.8V. (Seems like a schmitt trigger input).

With the PWM signal wire left unconnected (Hi-Z) the fan defaults to 100% speed. Connecting it to ground is the same as a 0% duty cycle PWM signal which is why the fan does not move under that condition.

 

The blue wire is connected to a frequency generator that allows you to read actual fan speed. (Like a digital tacho). The signal is described under the header "Frequency generator waveform" in the datasheet.