Question DC/PWM mismatch effects

[Deleted member 2969713]

What happens if you set a DC fan to PWM in the BIOS, or a PWM fan to DC? Anything terrible? I ask because earlier when I was messing around with the fan curves I tried switching back and forth between the two because I couldn't remember which ones my fans were, and with the CPU fan at least (which it turns out is PWM) it just kinda works, albeit with different (slower, IIRC) RPMs reported.

I have two case fans as well. Only one system fan header, so I use a splitter. I don't actually know for sure whether they are PWM or DC voltage control. I have them set to DC.

Is it that a PWM-capable (4-pin) fan can be controlled via either PWM or voltage control, and that a DC-only (3-pin) fan will simply run at full speed since it's not actually receiving the PWM signal?

 

[Lutfij]

Make and model of your motherboard and the fans? You'll see that the PWM fans on a DC header will be running at full speed.

 

[Deleted member 2969713]

Make and model of your motherboard and the fans? You'll see that the PWM fans on a DC header will be running at full speed.

Not sure of the make/model of the case fan I put in the front, I can't seem to find the email receipt for it. The back fan came with my Cougar MG120 case. The CPU fan is part of the Thermalright Assassin X120 Refined SE cooler package. Motherboard is MSI PRO-M2 Max B450. Both headers are 4-pin, but in the BIOS I can switch between DC (voltage control) and PWM for either.

I believe either my case fans or the splitter I'm using are 3-pin, since if I remember right they vary their speed properly on DC but spin up to 100% on PWM. I don't really want to go back into the BIOS and experiment with it again in case the mismatch could cause damage to the fans.

Edit: I got curious, so I went to the effort of opening up my case (unplugging everything, etc). The second case fan is a 4-pin Thermaltake TT-1225 (A1225s12s), and the fan splitter it's plugged into, which is plugged into the sole system fan header, is also 4-pin. But the case fan is 3-pin. So if I set the system fan header to PWM I'd get 100% full speed at all times on the case fan, which is very noisy.

Okay, that's really weird. Not only can I not find an email receipt for my fan purchase, but I can't find any official documentation of my particular model even existing online.

Found it. The official model name is totally different from the A1225s12s printed in the center of it.

 

[CountMike]

Not sure of the make/model of the case fan I put in the front, I can't seem to find the email receipt for it. The back fan came with my Cougar MG120 case. The CPU fan is part of the Thermalright Assassin X120 Refined SE cooler package. Motherboard is MSI PRO-M2 Max B450. Both headers are 4-pin, but in the BIOS I can switch between DC (voltage control) and PWM for either.

I believe either my case fans or the splitter I'm using are 3-pin, since if I remember right they vary their speed properly on DC but spin up to 100% on PWM. I don't really want to go back into the BIOS and experiment with it again in case the mismatch could cause damage to the fans.

Edit: I got curious, so I went to the effort of opening up my case (unplugging everything, etc). The second case fan is a 4-pin Thermaltake TT-1225 (A1225s12s), and the fan splitter it's plugged into, which is plugged into the sole system fan header, is also 4-pin. But the case fan is 3-pin. So if I set the system fan header to PWM I'd get 100% full speed at all times on the case fan, which is very noisy.

Okay, that's really weird. Not only can I not find an email receipt for my fan purchase, but I can't find any official documentation of my particular model even existing online.

Found it. The official model name is totally different from the A1225s12s printed in the center of it.

4pin(PWM) fans can be controlled in DC mode , just with less accurate and less finer peed adjustments. 3pin fans only in DC mode.
Why ? Because in PWM mode voltage is always maximum at 12vDC and speed is regulated by practically speaking turning power off and on in certain intervals and length of those intervals. That keeps power of motor constant.
3pin fans are controlled by varying voltage to motor so it has variable power to turn, lower that voltage is gives motor less power so air and bearing resistance slows it down.
Never mix 3 and 4 pin fans on same header in PWM mode.

 

[Deleted member 2969713]

Never mix 3 and 4 pin fans on same header in PWM mode.

Why, just because the 3-pin will simply run at full speed, or is there another reason? I'm not planning on doing so, I'm just curious. I'm pretty sure I had in the past briefly done so when messing around with the system fan curve, to make sure it was supposed to be on DC and not PWM, and I put it back to DC when the case fan got really loud.

 

[CountMike]

Why, just because the 3-pin will simply run at full speed, or is there another reason? I'm not planning on doing so, I'm just curious. I'm pretty sure I had in the past briefly done so when messing around with the system fan curve, to make sure it was supposed to be on DC and not PWM, and I put it back to DC when the case fan got really loud

Yes, that's what happens. Pulse width modulator is circuit inside the fan motor that regulates pulse width and frequency on command from MB.
Those are no ordinary motors but poly-phase with permanent magnet (which id in the rotor/fan blade part).
If you are interested this is how that type of motor works.
https://en.wikipedia.org/wiki/Brushless_DC_electric_motor

 

[Paperdoc]

You have it mostly right. For the core concern: no, you will not cause damage by feeding the "wrong" signals to a fan. The major issue from "mis-match" is speed control.

In the newer design - 4-pin fans - the header supplies power from Pin #2 at constant 12 VDC, plus sending the new PWM signal to the fan on Pin #4. Inside the fan there is a chip that uses that PWM signal to modify the flow of current from that power line through the windings to achieve speed control.

In the older 3-pin fan system, the header supplies power by Pin #2 that VARIES, from 12 VDC for full speed to about 5 VDC for minimum speed without stalling. A mobo header with only 3 pins can send out ONLY this signal set. However, if it has 4 pins you normally can set (in BIOS Setup options for each header) whether the header acts like an older 3-pin one and does NOT send out a PWM signal on Pin #4, or acts as a full 4-pin header in PWM Mode.

IF you connect a new 4-pin fan to a header using the older Voltage Control Mode (aka DC Mode) (whether the header has 3 or 4 pins) it gets no PWM signal so that chip cannot modify current flow. BUT it does receive from Pin #2 the VARIABLE Voltage supply so its speed IS controlled. This is not the ideal way, but t works. The main advantage lost in this is that a 4-pin fan fed 4-pin PWM signals can be run at a slower minimum speed without risking stalling.

If you plug a 3-pin fan into a header set to use the newer 4-pin PWM Mode, it gets constant 12 VDC from Pin #2 and no PWM signal from pin #4 it does not connect to. Further, it has no special chip so it could not use that signal anyway. So that fan runs full speed all the time. You get good (maybe excessive) cooling but no ability to reduce speed.

You can identify which fan type you have easily by eye. The 3-pin type has a cable with THREE wires (always Black-Red-Yellow) ending in a female connector with 3 holes. For a 4-pin fan the wire colours can vary, but there are FOUR wires, and hence four holes in the connector.

Mechanically and electrically these fan systems are as similar as possible, so you CAN plug any 3-pin or 4-pin fan into any 3-pin or 4-pin header and they will work, subject to the small performance limits outlined above. For that reason almost all Splitters now are 4-pin, and they work with any fan and header. Be aware, however, that Splitters and Hubs are two DIFFERENT devices for connecting more than one fan to a single header, but makers mis-use the names. They treat them as equal, and they are NOT the same device. The clear difference is in the connections the device makes. A HUB has a connection directly to a power source from the PSU, and a Splitter does not. Because of the differences, when using 3-pin fans you can use only a SPLITTER if you expect speed control, and the mobo header must be set to Voltage Control Mode. A HUB can be used only with a header in PWM Mode because it requires the PWM signal from Pin #4, AND it can control the speed of ONLY 4-pin fans that can use that PWM signal with their special chip.

 

[Deleted member 2969713]

All interesting information, thanks. Especially this tidbit:

Inside the fan there is a chip that uses that PWM signal to modify the flow of current from that power line through the windings to achieve speed control.

I had assumed the PWM signal was simply fed straight to the motor, turning it on and off rapidly and thus having the effect of modulating the speed. But that makes more sense, I think.

The main advantage lost in this is that a 4-pin fan fed 4-pin PWM signals can be run at a slower minimum speed without risking stalling.

Yeah, I think I witnessed this when I was messing around with the CPU fan curve and changing between PWM and DC voltage control.

 

[Paperdoc]

Your understanding of PWM fans is pretty close. More details.

In classic use of PWM technique for speed control of larger motors in industry, they ARE DC motors, and the power fed to them by the motor control circuit is not just DC. It is "switched" on and off rapidly, so that the flow of current (and hence magnetic fields and torque developed) through the motor windings is working only a part of the time. As you say, this means the motor SPEED is reduced since it is driving a load with that torque force. The system used in computers used for small fans does the same thing but with an important difference. SOME of this is because the designers wanted to make the new design as similar as possible to the older motors controlled simply by varying the VOLTAGE supplied to the fan. This was a marketing strategy to ease the introduction of the new design into a mature market. In this system the header supplies to the motor on Pin #2 a constant 12 VDC, not an interrupted supply. But it also sends on Pin #4 the PWM signal. This is the Pulse Wave Modulation signal. In some ways it looks like a classic square wave which is either on or off, no in-between. But whereas a square wave is on exactly 50% of time, a PWM signal has a variable % On. So it can be On 35% and Off 65%, or On 55% and Off 45%, etc. The little chip inside the motor on its circuit board simply uses that to switch the fixed 12 VDC supply line on and off to the motor winding. The result is exactly the same as the how large industrial motors are controlled this way. The difference is simply in WHERE the modulation of the power flow is done. In the computer fan system, the PWM signal typically operates around 20 to 22 KHz frequency, and has a 5 VDC amplitude. The chip that uses the signal draws VERY little current from the source, so it is no problem to have many fans tied into that one signal from a header to do their individual jobs. That's what a Splitter or Hub does for this fan type. Between those two sharing devices, the main difference is in where the 12 VDC power comes from, since that is a significant current draw. A Splitter gets that motor power from the header which has a max current limit of 1.0 A. A Hub does NOT connect to that power source, and instead gets it direct from the PSU. If the power connection from the PSU is a SATA power output from there, the SATA connector imposes a limit of 4.5 A because of contact current capacity. Lots more! If the connection is a 4-pin Molex connector with heavier contacts, the limit is even larger - about 10 A. But most Hubs themselves do not contain components that could handle 10 A! Why would it when it is powering a group of fans that can use no more than about 0.20 A per fan?

A small note. Because of this difference between industrial and computer case fan versions of "PWM control", makers of the computer fans specifically caution users NOT to try to supply such fans with the common form of PWM power supply from an industrial-style control system. The system in those small fans cannot tolerate a power supply that is rapidly pulsing on and off because there are several solid state components in their circuit board other than the chip that modulates current flow. For example, an essential job of that board is to provide switching of power flow through windings synchronized with the position of the rotor to replace the effect of the older design of a commutator and brushes. This is the heart of a brushless DC motor design.

 

[Gururu]

I found that Auto works fine for 3 pin fans and Pwm for 4 pin. Maybe your mobo has an Auto setting. All can be speed controlled. Note all my mobo headers are 4 pin.

 

[Deleted member 2969713]

Your understanding of PWM fans is pretty close. More details.

In classic use of PWM technique for speed control of larger motors in industry, they ARE DC motors, and the power fed to them by the motor control circuit is not just DC. It is "switched" on and off rapidly, so that the flow of current (and hence magnetic fields and torque developed) through the motor windings is working only a part of the time. As you say, this means the motor SPEED is reduced since it is driving a load with that torque force. The system used in computers used for small fans does the same thing but with an important difference. SOME of this is because the designers wanted to make the new design as similar as possible to the older motors controlled simply by varying the VOLTAGE supplied to the fan. This was a marketing strategy to ease the introduction of the new design into a mature market. In this system the header supplies to the motor on Pin #2 a constant 12 VDC, not an interrupted supply. But it also sends on Pin #4 the PWM signal. This is the Pulse Wave Modulation signal. In some ways it looks like a classic square wave which is either on or off, no in-between. But whereas a square wave is on exactly 50% of time, a PWM signal has a variable % On. So it can be On 35% and Off 65%, or On 55% and Off 45%, etc. The little chip inside the motor on its circuit board simply uses that to switch the fixed 12 VDC supply line on and off to the motor winding. The result is exactly the same as the how large industrial motors are controlled this way. The difference is simply in WHERE the modulation of the power flow is done. In the computer fan system, the PWM signal typically operates around 20 to 22 KHz frequency, and has a 5 VDC amplitude. The chip that uses the signal draws VERY little current from the source, so it is no problem to have many fans tied into that one signal from a header to do their individual jobs. That's what a Splitter or Hub does for this fan type. Between those two sharing devices, the main difference is in where the 12 VDC power comes from, since that is a significant current draw. A Splitter gets that motor power from the header which has a max current limit of 1.0 A. A Hub does NOT connect to that power source, and instead gets it direct from the PSU. If the power connection from the PSU is a SATA power output from there, the SATA connector imposes a limit of 4.5 A because of contact current capacity. Lots more! If the connection is a 4-pin Molex connector with heavier contacts, the limit is even larger - about 10 A. But most Hubs themselves do not contain components that could handle 10 A! Why would it when it is powering a group of fans that can use no more than about 0.20 A per fan?

A small note. Because of this difference between industrial and computer case fan versions of "PWM control", makers of the computer fans specifically caution users NOT to try to supply such fans with the common form of PWM power supply from an industrial-style control system. The system in those small fans cannot tolerate a power supply that is rapidly pulsing on and off because there are several solid state components in their circuit board other than the chip that modulates current flow. For example, an essential job of that board is to provide switching of power flow through windings synchronized with the position of the rotor to replace the effect of the older design of a commutator and brushes. This is the heart of a brushless DC motor design.

You know your fans! I appreciate the in-depth explanations.

I found that Auto works fine for 3 pin fans and Pwm for 4 pin. Maybe your mobo has an Auto setting. All can be speed controlled. Note all my mobo headers are 4 pin.

My motherboard doesn't have an auto setting, but that's all right, since I know I need to use voltage control to get variable speed from both a 3 pin and a 4 pin fan sharing a splitter.