Can a motherboard control the speed of 2 pin fan (using DC voltage control mode)?

[andydubson579]

Put shortly: Using DC voltage mode, will a motherboard be able to control the speed of a SYS_FAN through a 3-pin connection that is missing the yellow wire (basically making it a 2-pin connection)? If so, if two of these 2-pin fans were connected to one header, will the MOBO know to supply enough voltage for the two fans?

---------------------------

Put longly:
I'm pretty sure that...
--a MOBO can control a 4 pin fan by supplying a constant 12 Volts and using the 4th wire to control speed.
--some MOBOs can control a 3 pin fan by being set to "DC mode" and supplying the fan with a varying amount of Voltage. The yellow wire supplies the fan's RPM data, allowing the MOBO to know the speed of the fan.

Now, using DC mode, could a MOBO control a SYS_FAN that is missing that yellow wire and thus missing the RPM data? I have a few chassis fans that use a molex connection. I can get an adapter that converts the molex into a 3 pin connection without the yellow wire (basically a 2 pin connection), and then plug that into a MOBO header.

As for connecting multiple of these fans to one MOBO header, I have two options.
--attach a Y-splitter to one MOBO header, then attach an adapter to the two ends of the Y, then attach one molex fan to each adapter. (header > Y-splitter > two adapters > one molex fan on each adapter)
--or, I could simply attach an adapter to one MOBO header, then attach one molex fan to that adapter. THEN, I daisychain another molex fan to the first molex fan. (header > adapter > one molex fan > another molex fan daisychained to the first)

With either of these two options, will the MOBO (in DC mode) know to how much Voltage to supply since there is more than one fan connected to a single header (would the MOBO even know there is more than one fan on a single header?) I'd prefer the second method since it allows me to buy fewer adapters and I wouldn't need a Y-splitter.

---------------------------

The reason I ask is because instead of plugging the molex fans into the PSU or getting a manual fan speed control thing, I would like the MOBO to automatically control the fans.

Example of the adapter I would use: https://www.amazon.de/Caseking-GmbH-4043718029068-4-Pin-L%C3%BCfteradapterkabel/dp/B005H0FSCI/ref=sr_1_4?tag=georiot-de-new-21&ie=UTF8&qid=1483723721&sr=8-4&keywords=3+pin+auf+4+pin+molex&ascsubtag=tomshardware-3352933774575687473-21

Thanks for any help.

 

[lthampton]

No. That is simply power, so it is going to be full on or off. Maybe try investing in a fan controller, that would allow you to adjust the speed.

 

[andydubson579]

No. That is simply power, so it is going to be full on or off. Maybe try investing in a fan controller, that would allow you to adjust the speed.

Thanks for replying. I'm not sure I understand your reasoning for why it wouldn't work though. Wouldn't DC voltage control mode on the motherboard regulate the voltage sent to the fan, and NOT send a full 12 volts? Molex is usually just for power, true, but if adapted into a 3 pin connection (with just 2 wires, no yellow RPM wire), could it be regulated by a motherboard in DC mode?

I guess the crux of my question is whether a motherboard NEEDS the RPM data from the third wire for DC voltage control mode to work. If it doesn't need that, then adapting the molex fans into a 2-wire-3-pin connector should allow their speed to be controlled, yes?

Perhaps Paperdoc could provide an assist

fake quote so Paperdoc receives notification. Sorry Paperdoc

 

[lthampton]

No. That is simply power, so it is going to be full on or off. Maybe try investing in a fan controller, that would allow you to adjust the speed.

Thanks for replying. I'm not sure I understand your reasoning for why it wouldn't work though. Wouldn't DC voltage control mode on the motherboard regulate the voltage sent to the fan, and NOT send a full 12 volts? Molex is usually just for power, true, but if adapted into a 3 pin connection (with just 2 wires, no yellow RPM wire), could it be regulated by a motherboard in DC mode?

I guess the crux of my question is whether a motherboard NEEDS the RPM data from the third wire for DC voltage control mode to work. If it doesn't need that, then adapting the molex fans into a 2-wire-3-pin connector should allow their speed to be controlled, yes?

Perhaps Paperdoc could provide an assist

fake quote so Paperdoc receives notification. Sorry Paperdoc

Well, now having read it over, it should in theory work, as long as the fans are controlled by the voltage input and not some other factor. Definitely get paperdoc in here, i might've been slightly over my head with my previous response.

 

[Paperdoc]

Yes, what you suggest will work. I'll add a few cautions.

You have recognized the key question: does the mobo need the RPM signal in order to control fan speed? The answer is NO, interestingly. For a 3-pin fan, control of fan speed is done entirely by the voltage supplied on Pin #2 (Red wire) and it ranges from 12 VDC (full speed) to about 5 VDC. Any lower voltage and the fan motor may stall and fail to re-start until the voltage is raised. Pin #1 (Black wire) is Ground. As you have deduced, a fan designed for connection only to a Molex output from the PSU gets only those two connections, since the PSU has no way to do anything with a speed signal. Moreover, the PSU only supplies a fixed 12 VDC on the Red wire of the fan (interestingly, that's the YELLOW wire from the PSU) so there is no speed control that way.

The mobo's automatic speed control design is based entirely on the TEMPERATURE measured at a suitable sensor. For the CPU chip cooler that sensor is inside the CPU chip and its signal sent out on one chip pin. For SYS_FAN or CHA_FAN headers to be used for case ventilation the normal sensor is one built into the mobo by its maker at a spot they consider the best representation of general cooling needs. Normally multiple SYS_FAN headers use the same mobo temp sensor. The control strategy is a simple feedback loop. That is, the system compares the actual measured temp at the sensor to a target or Setpoint which is pre-programmed into the mobo. (Many headers also allow you to set custom settings if you wish.) If the current temp is too high, the system will speed up the fan. If it stays high, it will speed up again. If it is on-target, the fan speed will be kept unchanged. etc.

Note in that control strategy that the system never actually looks at the RPM value! It really does NOT care about the fan speed - it concentrates on temperature - so the RPM value does not matter at all for speed control.

So why have an RPM signal? Two reasons. One is just to provide info to YOU, the user, for whatever you want to do. But the other is FAILURE detection. The header's system checks to verify that the fan is turning. If it gets NO speed signal (or, in some cases, gets a speed which is below a minimum value setting), it will send out a warning that the fan has failed. For the CPU_FAN header, many mobos do more than that. If they detect failure of that CPU cooling system, they will send out an immediate warning and, in a very short time, shut down the entire system without even waiting for the temp sensor inside the chip to indicate a too-high temperature. Such systems often will refuse to allow you to boot up if the CPU fan fails to start up immediately. These extra precautions are to protect the very expensive CPU chip from permanent damage from lack of cooling.

It is those fan failure detection systems that prompt one caution. If you do as you plan, the fan that has no Yellow speed signal wire from it will be seen by the mobo as having failed and you will get constant alarms. For your situation look at the options in the SYS_FAN or CHA-FAN header configuration for an option to Ignore the fan speed.

The second caution is a general one for using Splitters etc. A normal mobo fan header can supply up to 1.0 A current max to all fans connected to that one header. Most fans these days consume 0.1 to 0.2 A max, but some (especially the ones with single-colour LED lights in them) can use 0.3 to 0.4 A. So when connecting more than one fan to a header using Splitters, look up the fan specs (may even be on the fan label) and verify the total load connected to a fan header does not exceed 1.0 A.

 

[andydubson579]

Yes, what you suggest will work. I'll add a few cautions.

You have recognized the key question: does the mobo need the RPM signal in order to control fan speed? The answer is NO, interestingly. For a 3-pin fan, control of fan speed is done entirely by the voltage supplied on Pin #2 (Red wire) and it ranges from 12 VDC (full speed) to about 5 VDC. Any lower voltage and the fan motor may stall and fail to re-start until the voltage is raised. Pin #1 (Black wire) is Ground. As you have deduced, a fan designed for connection only to a Molex output from the PSU gets only those two connections, since the PSU has no way to do anything with a speed signal. Moreover, the PSU only supplies a fixed 12 VDC on the Red wire of the fan (interestingly, that's the YELLOW wire from the PSU) so there is no speed control that way.

The mobo's automatic speed control design is based entirely on the TEMPERATURE measured at a suitable sensor. For the CPU chip cooler that sensor is inside the CPU chip and its signal sent out on one chip pin. For SYS_FAN or CHA_FAN headers to be used for case ventilation the normal sensor is one built into the mobo by its maker at a spot they consider the best representation of general cooling needs. Normally multiple SYS_FAN headers use the same mobo temp sensor. The control strategy is a simple feedback loop. That is, the system compares the actual measured temp at the sensor to a target or Setpoint which is pre-programmed into the mobo. (Many headers also allow you to set custom settings if you wish.) If the current temp is too high, the system will speed up the fan. If it stays high, it will speed up again. If it is on-target, the fan speed will be kept unchanged. etc.

Note in that control strategy that the system never actually looks at the RPM value! It really does NOT care about the fan speed - it concentrates on temperature - so the RPM value does not matter at all for speed control.

So why have an RPM signal? Two reasons. One is just to provide info to YOU, the user, for whatever you want to do. But the other is FAILURE detection. The header's system checks to verify that the fan is turning. If it gets NO speed signal (or, in some cases, gets a speed which is below a minimum value setting), it will send out a warning that the fan has failed. For the CPU_FAN header, many mobos do more than that. If they detect failure of that CPU cooling system, they will send out an immediate warning and, in a very short time, shut down the entire system without even waiting for the temp sensor inside the chip to indicate a too-high temperature. Such systems often will refuse to allow you to boot up if the CPU fan fails to start up immediately. These extra precautions are to protect the very expensive CPU chip from permanent damage from lack of cooling.

It is those fan failure detection systems that prompt one caution. If you do as you plan, the fan that has no Yellow speed signal wire from it will be seen by the mobo as having failed and you will get constant alarms. For your situation look at the options in the SYS_FAN or CHA-FAN header configuration for an option to Ignore the fan speed.

The second caution is a general one for using Splitters etc. A normal mobo fan header can supply up to 1.0 A current max to all fans connected to that one header. Most fans these days consume 0.1 to 0.2 A max, but some (especially the ones with single-colour LED lights in them) can use 0.3 to 0.4 A. So when connecting more than one fan to a header using Splitters, look up the fan specs (may even be on the fan label) and verify the total load connected to a fan header does not exceed 1.0 A.

Thanks for the thorough response. I guess I'll have to see within the MOBO's bios about disabling any SYS_FAN speed-detection safety feature. (I won't be disabling any CPU_FAN safety features of course.)

Also instead of using a true splitter on a single header, I'll try daisychaining the molex fans, and then using a single molex-to-3pin adapter on a header, but both fans should still receive the same voltage from the MOBO I think.

 

[Paperdoc]

Yes, that's right. I did not comment directly on your query about mobo adjusting for number of fans. When you use these devices to connect more than one fan to a single header, they all are connected in PARALLEL with each other. So they all receive exactly the same Voltage signal. If the fans are identical, they all will run at very nearly the same speeds. If they are of different designs the speeds may not match, but that does not matter. What the mobo automatic speed control system is doing is making sure that the AIR FLOW through the case is sufficient to keep the measured TEMPERATURE on target, and it does not care about actual fan speed.

From what you say, it sounds like your fans have on the end of their two wires a "stacking" type of Molex connector with male and female ends so several can be stacked together. So yes, if you can arrange to connect a single mobo male fan header to one of hose and then stack others onto it, it will all work. That MAY take some custom work by you, though. I know you say you have the required adapter, but I have never seen one which is a FEMALE fan connector (to plug into a mobo male header) and a 4-pin Molex (male or female) on the other end.

 

[andydubson579]

Yes, that's right. I did not comment directly on your query about mobo adjusting for number of fans. When you use these devices to connect more than one fan to a single header, they all are connected in PARALLEL with each other. So they all receive exactly the same Voltage signal. If the fans are identical, they all will run at very nearly the same speeds. If they are of different designs the speeds may not match, but that does not matter. What the mobo automatic speed control system is doing is making sure that the AIR FLOW through the case is sufficient to keep the measured TEMPERATURE on target, and it does not care about actual fan speed.

From what you say, it sounds like your fans have on the end of their two wires a "stacking" type of Molex connector with male and female ends so several can be stacked together. So yes, if you can arrange to connect a single mobo male fan header to one of hose and then stack others onto it, it will all work. That MAY take some custom work by you, though. I know you say you have the required adapter, but I have never seen one which is a FEMALE fan connector (to plug into a mobo male header) and a 4-pin Molex (male or female) on the other end.

Example of the adapter with female molex: https://www.amazon.de/Caseking-GmbH-4043718029068-4-Pin-L%C3%BCfteradapterkabel/dp/B005H0FSCI/ref=sr_1_4?tag=georiot-de-new-21&ie=UTF8&qid=1483723721&sr=8-4&keywords=3+pin+auf+4+pin+molex&ascsubtag=tomshardware-3352933774575687473-21

I'm not surprised that you've never seen one, only like 5 pages on the entire interweb had an adapter like this. Ones with male molex are much more common though. Just of curiosity, would you happen to know whether a MALE molex connector would work? Since the fans' molex connections are daisychainable, and thus double sided, could I plug a male molex at the back end of the chain to supply the fans with power (instead of plugging in a female molex at the front of the chain, as the adapter linked above would do)? Does power only run through one direction and not the other?

I have a feeling we're dealing with some pretty esoteric stuff here, the best way to know would be simply try it both ways....

 

[Paperdoc]

If you have fans with double-ended (male + female) Molex connectors and you do NOT plug them into any PSU Molex output (you certainly should not for what you're doing), then you certainly could use an adapter that ends in a male Molex and stack them all. All the connectors are simply making parallel connections, and there is no "direction" for power flow through them. HOWEVER, finding such an adapter with a FEMALE fan connector on the other end might also be hard. The common adapters allow you to connect from a PSU female output Molex (hence male Molex on one end) to the female fan connector (hence male fan pins on the other end).

 

[andydubson579]

Alright folks, that about answers everything. Thank you Paperdoc and lthampton! (I must say I'm simply amazed at Paperdoc's knowledge and helpfulness with this.) I'll pick Paperdoc's first response as the solution, although I'd advise anyone reading this thread to look at Paperdoc's later responses too. They were very helpful.

If anyone tries this adapter thing like me and it DOESN'T work for them, please reply with your situation. I'd be curious about it.

 

[Paperdoc]

Thanks for Best Solution.