4 pin CPU fan into 3 pin header
- Thread starter Aroldo
- Start date
- Oct 30, 2004
- 32,255
- 14
- 99,960
It's done, but certainly not recommended... at least not by fan manufacturers such as Nidec.
http://www.swiftech.com/pwmcontrollers.aspx
Original is here:
http://www.nidec.net/apps.htm#
You have to click "more" 3 times to see it
Actually, the problems cited in the articles linked above are for the reverse situation - that of trying to connect a 3-pin fan to a voltage supply that has already been modulated by a PWM signal. In that situation the fan receives what the motor of a true 4-pin fan gets: a voltage that is not constant, but really is changing from 0 to 12 VDC and back again at a frequency of 20 to 25 KHz. The article says this could wear out a true 3-pin fan fast.
BUT no computer can do this. 4-pin fan ports operating in PWM mode do NOT sent this type of voltage to any fan. They send full constant 12 VDC to the fan on Pin #2, plus they send a separate PWM signal on Pin #4. Inside the fan casing there is a small chip that uses that PWM signal to modify the voltage supplied into the pulsed voltage, and that goes to the motor. Of course, only motors designed for such use are sold as 4-pin fans suitable for computers.
What OP asked about is the opposite: what happens when a 4-pin fan is plugged into a 3-pin port operating in Voltage Control Mode? It is supposed to work just fine. In fact, this is one of the backwards compatibility features designed into the 4-pin fan system - you CAN use it on a 3-pin port. Why? Well, the way that PWM signal is used by the chip inside a true 4-pin fan is as an inhibitor. That is, when the PWM signal is "on", the flow of current from the +12VDC supply line to the motor is interrupted; when the PWM signal is "off", the current flows fully. If you plug a 4-pin fan prepared for that scenario into a 3-pin port instead, what it "sees" is there is never an "on" state of the PWM signal because there is NO PWM signal! Hence, whatever voltage comes in from Pin #2 goes to the fan motor unimpeded. Now, the 3-pin port operating in Voltage Control Mode is NOT sending a fixed +12VDC signal on Pin #2. It reduces the voltage to slow the fan down. Thus, the 4-pin fan actually behaves exactly like a 3-pin fan under Voltage Control.
There are two small disadvantages to this way of doing things. 4-pin fans truly under PWM control can start at lower speeds, and can be slowed down to lower speeds without stalling, than can 3-pin fans under Voltage Control. When you make the 4-pin fan behave as a 3-pin one under Voltage Control, you lose those two advantages. Now, even these items are not a big problem. All mobo designers know that fans under Voltage Control cannot be relied on to start at low voltages, and could stall if their voltage is too low. So all such fans always are started up at full voltage at boot time, and it takes a few seconds before the BIOS gets around to determining just how much the fan voltage can be reduced. By that time the fan is already running full speed. After that the automatic control systems have a lower limit on the voltage they send to the fan, so it should not stall (unless the fan is worn or defective). The advantage that IS "lost" is just that the 3-pin Voltage Control Mode port will not try to run that fan as slow as a true PWM-controlled fan MIGHT be able to run.
BUT no computer can do this. 4-pin fan ports operating in PWM mode do NOT sent this type of voltage to any fan. They send full constant 12 VDC to the fan on Pin #2, plus they send a separate PWM signal on Pin #4. Inside the fan casing there is a small chip that uses that PWM signal to modify the voltage supplied into the pulsed voltage, and that goes to the motor. Of course, only motors designed for such use are sold as 4-pin fans suitable for computers.
What OP asked about is the opposite: what happens when a 4-pin fan is plugged into a 3-pin port operating in Voltage Control Mode? It is supposed to work just fine. In fact, this is one of the backwards compatibility features designed into the 4-pin fan system - you CAN use it on a 3-pin port. Why? Well, the way that PWM signal is used by the chip inside a true 4-pin fan is as an inhibitor. That is, when the PWM signal is "on", the flow of current from the +12VDC supply line to the motor is interrupted; when the PWM signal is "off", the current flows fully. If you plug a 4-pin fan prepared for that scenario into a 3-pin port instead, what it "sees" is there is never an "on" state of the PWM signal because there is NO PWM signal! Hence, whatever voltage comes in from Pin #2 goes to the fan motor unimpeded. Now, the 3-pin port operating in Voltage Control Mode is NOT sending a fixed +12VDC signal on Pin #2. It reduces the voltage to slow the fan down. Thus, the 4-pin fan actually behaves exactly like a 3-pin fan under Voltage Control.
There are two small disadvantages to this way of doing things. 4-pin fans truly under PWM control can start at lower speeds, and can be slowed down to lower speeds without stalling, than can 3-pin fans under Voltage Control. When you make the 4-pin fan behave as a 3-pin one under Voltage Control, you lose those two advantages. Now, even these items are not a big problem. All mobo designers know that fans under Voltage Control cannot be relied on to start at low voltages, and could stall if their voltage is too low. So all such fans always are started up at full voltage at boot time, and it takes a few seconds before the BIOS gets around to determining just how much the fan voltage can be reduced. By that time the fan is already running full speed. After that the automatic control systems have a lower limit on the voltage they send to the fan, so it should not stall (unless the fan is worn or defective). The advantage that IS "lost" is just that the 3-pin Voltage Control Mode port will not try to run that fan as slow as a true PWM-controlled fan MIGHT be able to run.
- Oct 30, 2004
- 32,255
- 14
- 99,960
Swiftech quotes the article because peeps are running THEIR PWM PUMPS off variable voltage signals... not the other way around. They don't have DC voltage controlled pumps so if what you are saying was correct, there's be no reason for Swiftech to issue this warning.
What they are talking about is the "so called" PWM fan controller that are not modulating speed PWM but by variable voltage.
"Pulse-width modulation of DC operating voltage" - What they are saying is that the fan controller is converting a PWM signal from the MoBO and modulating DC voltage accordingly ....
This is the same thing you can do with a 4 pin header. I will note that a 4 pin header is not necessarily PWM. Most Z87 boards have 4 pin chassis headers which do not provide PWM control, but rather DCV control. In Z97 , for the most part, they may operate either way. So a 4 pin header can control via DCV... this is a bad thing for PWM devices. As the OP clearly states, "the 3-pin header provides a variable voltage." That's exactly what Swiftech is warning about .... PWM devices want 12 volts...not 7. not 5, not 3..
"[using] DC operating voltage to modify fan speed [edit: in PWM devices] is not recommended" - Again, clearly, the speed of the item being speed controlled is a PWM device.... not a DVC device. And using variable voltage to control PWM is not recommended.
If that is not clear enough
The OP clearly stated that "the 3-pin header provides a variable voltage".... exactly what Swiftech warns against. PWM fans and pumps, are designed to receive 12 volts so as not to experience the problems listed.
And again:
You do not want to use voltage modulation to control PWM which is exactly what Swiftech and Nidec are saying.
As for the advantages of PWM, the information is somewhat outdated and that's the very reason Swiftech has this warning. Well not outdated so much as it leaves out the "hybrid" category. The PWM controllers they are talking about all advertise PWM control because that is generally considered superior. And they do use the MoBo's PWM signal to control speeds.
They, like the Phanteks fan PCB use a PWM signal to output a DC voltage. This provides a "best of both worlds" approach. You get all the fine and low speed control w/o the low speed problem. These devices get fan speeds down to 20-25% an area not usually possible with DC direct voltage control. You also eliminate the low speed clicking or hum of PWM fans. In the end now we get PWM control using 3 pin fans that cost 1/3 to 1/2 as much, w/o the hum / clickety noise and we still get low speed operation, even ramp them up and down as well as turn them off and use passive cooling at low loads.
But as Swiftech and Nidec state this quite clearly. PWM is designed for 12volt operation and not for variable voltage operation. If it was the "reverse situation" as you suggest, Swiftech wouldn't have a full page dedicated to eliminating the misuse of their PWM devices devices.
What they are talking about is the "so called" PWM fan controller that are not modulating speed PWM but by variable voltage.
"Pulse-width modulation of DC operating voltage" - What they are saying is that the fan controller is converting a PWM signal from the MoBO and modulating DC voltage accordingly ....
This is the same thing you can do with a 4 pin header. I will note that a 4 pin header is not necessarily PWM. Most Z87 boards have 4 pin chassis headers which do not provide PWM control, but rather DCV control. In Z97 , for the most part, they may operate either way. So a 4 pin header can control via DCV... this is a bad thing for PWM devices. As the OP clearly states, "the 3-pin header provides a variable voltage." That's exactly what Swiftech is warning about .... PWM devices want 12 volts...not 7. not 5, not 3..
"[using] DC operating voltage to modify fan speed [edit: in PWM devices] is not recommended" - Again, clearly, the speed of the item being speed controlled is a PWM device.... not a DVC device. And using variable voltage to control PWM is not recommended.
If that is not clear enough
The OP clearly stated that "the 3-pin header provides a variable voltage".... exactly what Swiftech warns against. PWM fans and pumps, are designed to receive 12 volts so as not to experience the problems listed.
And again:
You do not want to use voltage modulation to control PWM which is exactly what Swiftech and Nidec are saying.
As for the advantages of PWM, the information is somewhat outdated and that's the very reason Swiftech has this warning. Well not outdated so much as it leaves out the "hybrid" category. The PWM controllers they are talking about all advertise PWM control because that is generally considered superior. And they do use the MoBo's PWM signal to control speeds.
They, like the Phanteks fan PCB use a PWM signal to output a DC voltage. This provides a "best of both worlds" approach. You get all the fine and low speed control w/o the low speed problem. These devices get fan speeds down to 20-25% an area not usually possible with DC direct voltage control. You also eliminate the low speed clicking or hum of PWM fans. In the end now we get PWM control using 3 pin fans that cost 1/3 to 1/2 as much, w/o the hum / clickety noise and we still get low speed operation, even ramp them up and down as well as turn them off and use passive cooling at low loads.
But as Swiftech and Nidec state this quite clearly. PWM is designed for 12volt operation and not for variable voltage operation. If it was the "reverse situation" as you suggest, Swiftech wouldn't have a full page dedicated to eliminating the misuse of their PWM devices devices.
Let's remember that the Swiftech article specifically addresses power supplied to their PUMP systems designed for PWM control by a mobo fan header. We also should distinguish between lower-voltage input power from a slowly-changing DC voltage (such as produced by a Voltage Control Mode mobo fan port) and "low voltage" which is really from an external PWM Controller that outputs a pulsing DC at higher frequency whose AVERAGE appears to be a reduced DC.
On the other hand, OP was asking specifically about using a common case ventilation FAN on the "wrong" port type.
As Swiftech points out, the term "PWM Control" has developed confusion. In wider circles there have been many "PWM Speed Controllers" marketed for DC motors that work by modulating a DC supply into a pulsed voltage (usually over 100 Hz, often over 1 kHz) of varying Duty cycle. They work well when paired with appropriate motors. More recently in the computer ventilation and cooling field we see what are called "PWM fans" and "PWM Fan Controllers" and "PWM Speed Control" from mobo fan ports. This latter field has a very important difference. The fans involved and their controller systems do NOT send a pulsed DC supply to the fan. Instead, they send a constant DC supply (normally 12 VDC) plus the PWM signal on a separate line to the motor. Inside the motor there are electronic circuits for several functions, and they use on the DC voltage supplied. Almost all such small motors are brushless DC motors which use sensors and electronic circuits to replace the function of a commutator and brushes in older designs. They have Hall Effect sensors on the motor shaft to generate a signal consisting of pulses (2 per revolution) to be sent back on Pin #3 to the mobo as a "Tach Signal" to be counted, generating a fan motor speed readout. They have a controller chip which uses the PWM signal from Pin #4 to switch the current through the motor from the DC voltage supplied on Pin #2 on and off, thus effecting speed control. There may be other functions.
Now, what happens if the supply to the motor is not as intended? There are two cases here to consider.
One, which concerns Swiftech especially, is what happens when someone provides the DC supply to their equipment NOT from a fixed 12VDC supply, but from one of those pulsed-output "PWM Controllers" designed for non-computer applications (It is probable in this situation that there might be no 25 kHz PWM signal supplied separately.) . Certainly such a situation could produce unpredictable performance from the pump's internal electronics, and I'm sure they are right to warn of likely damage to their equipment. This should never be done, they advise. Now, no computer mobo fan port operates this way. BUT perhaps there are third-part fan controllers that do. I really don't know if there are. Thus one should be very careful to investigate the controller details if one were planning to use them.
The other situation is connecting the PWM-style (4-pin) device to a 3-pin port that provides no PWM signal and provides a DC supply which is NOT pulsed, but is also varying slowly from 12 VDC down to some lower limit (maybe 5 VDC, maybe less). For Swiftech's concerns surrounding its pump systems, they recommend also that this should not be done, and I'm sure they have good reason for that.
But OP's question was about common 4-pin ventilation fans. Now, 3-pin fans have many of the same electronics built into them, including brushless electronic simulation of the commutator / brush system and Hall Effect Tach Signal generation. They just don't have any circuits to apply the PWM signal to current flow through the motor. Those fans have been designed to work properly with a DC supply voltage over a reasonable range suitable for fan speed control. With the introduction of 4-pin (PWM) ventilation fans we have been assured that the new designs are "backwards compatible" with 3-pin systems so that such fans CAN be plugged into 3-pin mobo ports and WILL work reliably, achieving varied fan speeds under the Voltage Control Mode of operation. I have to assume that the designers of 4-pin fans have used the same techniques as those of 3-pin fans to ensure that those electronic systems common to both designs work in both systems. I understand that it is quite possible that a liquid cooling system using a PWM-controlled pump may not be designed to work well on a Voltage Control Mode port, but that does not mean that ALL 4-pin devices cannot work under that control mode.
On the other hand, OP was asking specifically about using a common case ventilation FAN on the "wrong" port type.
As Swiftech points out, the term "PWM Control" has developed confusion. In wider circles there have been many "PWM Speed Controllers" marketed for DC motors that work by modulating a DC supply into a pulsed voltage (usually over 100 Hz, often over 1 kHz) of varying Duty cycle. They work well when paired with appropriate motors. More recently in the computer ventilation and cooling field we see what are called "PWM fans" and "PWM Fan Controllers" and "PWM Speed Control" from mobo fan ports. This latter field has a very important difference. The fans involved and their controller systems do NOT send a pulsed DC supply to the fan. Instead, they send a constant DC supply (normally 12 VDC) plus the PWM signal on a separate line to the motor. Inside the motor there are electronic circuits for several functions, and they use on the DC voltage supplied. Almost all such small motors are brushless DC motors which use sensors and electronic circuits to replace the function of a commutator and brushes in older designs. They have Hall Effect sensors on the motor shaft to generate a signal consisting of pulses (2 per revolution) to be sent back on Pin #3 to the mobo as a "Tach Signal" to be counted, generating a fan motor speed readout. They have a controller chip which uses the PWM signal from Pin #4 to switch the current through the motor from the DC voltage supplied on Pin #2 on and off, thus effecting speed control. There may be other functions.
Now, what happens if the supply to the motor is not as intended? There are two cases here to consider.
One, which concerns Swiftech especially, is what happens when someone provides the DC supply to their equipment NOT from a fixed 12VDC supply, but from one of those pulsed-output "PWM Controllers" designed for non-computer applications (It is probable in this situation that there might be no 25 kHz PWM signal supplied separately.) . Certainly such a situation could produce unpredictable performance from the pump's internal electronics, and I'm sure they are right to warn of likely damage to their equipment. This should never be done, they advise. Now, no computer mobo fan port operates this way. BUT perhaps there are third-part fan controllers that do. I really don't know if there are. Thus one should be very careful to investigate the controller details if one were planning to use them.
The other situation is connecting the PWM-style (4-pin) device to a 3-pin port that provides no PWM signal and provides a DC supply which is NOT pulsed, but is also varying slowly from 12 VDC down to some lower limit (maybe 5 VDC, maybe less). For Swiftech's concerns surrounding its pump systems, they recommend also that this should not be done, and I'm sure they have good reason for that.
But OP's question was about common 4-pin ventilation fans. Now, 3-pin fans have many of the same electronics built into them, including brushless electronic simulation of the commutator / brush system and Hall Effect Tach Signal generation. They just don't have any circuits to apply the PWM signal to current flow through the motor. Those fans have been designed to work properly with a DC supply voltage over a reasonable range suitable for fan speed control. With the introduction of 4-pin (PWM) ventilation fans we have been assured that the new designs are "backwards compatible" with 3-pin systems so that such fans CAN be plugged into 3-pin mobo ports and WILL work reliably, achieving varied fan speeds under the Voltage Control Mode of operation. I have to assume that the designers of 4-pin fans have used the same techniques as those of 3-pin fans to ensure that those electronic systems common to both designs work in both systems. I understand that it is quite possible that a liquid cooling system using a PWM-controlled pump may not be designed to work well on a Voltage Control Mode port, but that does not mean that ALL 4-pin devices cannot work under that control mode.
- Oct 30, 2004
- 32,255
- 14
- 99,960
You make a convincing argument, very different from the 1st post but reads well nonetheless
It sounds great but, as you stated is based upon your assumption of what designers did with no documented support for such assumption.
However what we have left is Nidec makes PWM fans and it's Nidec that is saying do not power a PWM fan via variable voltage.
Here we have a manufacturer of PWM fans stating very clearly not to do this. Yes, Swiftech is concerned about the psuedo PWM thing but they go on to state that PWM devices want to see 12 volts and only 12 volts. They are not saying that only psuedo pulse width modulation of DCV speed control is bad. They are saying use only 12 volts for PWM and that just because they are using a psuedo PWM pulse width modulation to control the voltage, that still doesn't make anything but a 12 volt signal acceptable.
again, Swiftech states:
It's a short, concise sentence and stands on its own. What the OP inquired about is using variable voltage and PWM fans, and all else PWM, are designed to receives a fixed, not variable, 12 volt signal.
http://www.overclockers.com/pwm-fan-controller/
Now in Z87, 4 pin chassis headers that claimed to be PWM but were not "true PWM" but psuedo PWM. This is a 4 pin header and exactly the situation which Swiftech and Nidec describe.
http://www.overclock.net/t/1462770/4-pin-fan-headers-on-motherboard-that-arent-pwm
http://linustechtips.com/main/topic/32579-4-pin-fan-header-pwm/
http://www.silentpcreview.com/forums/viewtopic.php?t=66283p=575840
I can see that we're not going to agree on all this. What I read from the Swiftech and Nidec website articles is that, when using fans designed to be "PWM-Controlled" (aka 4-pin) computer cooling fans, one should NOT use a power supply to them on Pins 1 and 2 that has been converted from a fixed DC voltage to a pulse-width-modulated 12VDC line that produces an AVERAGE voltage less than 12 V. Such a voltage supply would contain fast transients and possible overvoltage spikes that, they warn, could damage the components of the fan motor. More specifically, the damage is not likely done to the motor windings or field coils. Damage from transients is likely to affect the electronic components included inside the motor case that perform functions such as simulated commutator / brush systems and the internal PWM control chip. The Nidec article does show people how to make their own circuit for experimentation purposes that can feed the computer PWM Fan with proper signals - a fixed 12VDC supply plus a proper 5V p-p PWM signal at about 25 kHz - but with more range of variability in the PWM signal than a computer mobo would use. The Nidec cautionary notes, I believe, are especially directed towards technical people (and not so knowledgeable people) who have worked in the past with PWM motor speed controllers and DC motors and who misunderstand the new computer-style "PWM Fans" to be designs suitable for the same control systems.
I do not believe that either firm's articles mean to tell us that a constant DC voltage supply (not a 25 kHz PWM-modulated 12 VDC supply) at a voltage less than 12 VDC will damage their equipment. Certainly I can believe that at some very much lower voltage a 4-pin computer fan may not work properly, but I don't know exactly what that lower limit would be. In the case of Swittech's pump systems the tell us never to use less than 12 VDC of their system will malfunction, and I'm sure they say that for good reason. The advice should be followed for Swiftech pump systems. However, common 4-pin FANS are represented by their makers to be able to work properly when connected to 3-pin mobo fan ports that can only operate in Voltage Control Mode and reduce the voltage on Pin #2. We also know that all BIOS designs on mobos impose on themselves a lower limit of voltage output from a 3-pin fan port of somewhere between 5 and 7 volts. I put those two items together and assume that these limits in the BIOS satisfy the minimum voltage requirements of computer 4-pin fans connected to a 3-pin port.
I do not believe that either firm's articles mean to tell us that a constant DC voltage supply (not a 25 kHz PWM-modulated 12 VDC supply) at a voltage less than 12 VDC will damage their equipment. Certainly I can believe that at some very much lower voltage a 4-pin computer fan may not work properly, but I don't know exactly what that lower limit would be. In the case of Swittech's pump systems the tell us never to use less than 12 VDC of their system will malfunction, and I'm sure they say that for good reason. The advice should be followed for Swiftech pump systems. However, common 4-pin FANS are represented by their makers to be able to work properly when connected to 3-pin mobo fan ports that can only operate in Voltage Control Mode and reduce the voltage on Pin #2. We also know that all BIOS designs on mobos impose on themselves a lower limit of voltage output from a 3-pin fan port of somewhere between 5 and 7 volts. I put those two items together and assume that these limits in the BIOS satisfy the minimum voltage requirements of computer 4-pin fans connected to a 3-pin port.
- Oct 30, 2004
- 32,255
- 14
- 99,960
It says that minor variations are acceptable... just as a PSU can deliver 11.4v to 12.6v on a 12 volt rail and this will still be in compliance with the ATX spec. This is a far cry from 3, 5, 7 volt signals. Nidec is not a pump manufacturer. And Swiftech is a fan supplier and the same rule applies to their fans.
I might be willing to "assume" that will will be OK and try it out for myself, if I hadn't read the above documentation stating quite unequivocally to the contrary. That "constant 12v" statement is pretty clear and a drop from 12 to 5 volts (58%) can in no way be considered a "minor variation". I have discussed this subject with Swiftech as well as Phanteks support personnel and their stand on the matter was in no way unclear.
I might be willing to risk it myself ... I wouldn't care if I toasted a fan. But I would not risk making a recommendation to someone else based upon my assumption that a 58% or more drop in voltage represents a "minor variation".
Well Nidec, for one, has clearly taken a position to the contrary . Do you have any links to manufacturer sites that state this ?
I don't doubt that over time, this can kill the motor.... the question is, is that going to happen within a time period that we care about ? If my fan has a 1 year warranty, by all means go ahead....5 years ? maybe I don't wanna say that.
I might be willing to "assume" that will will be OK and try it out for myself, if I hadn't read the above documentation stating quite unequivocally to the contrary. That "constant 12v" statement is pretty clear and a drop from 12 to 5 volts (58%) can in no way be considered a "minor variation". I have discussed this subject with Swiftech as well as Phanteks support personnel and their stand on the matter was in no way unclear.
I might be willing to risk it myself ... I wouldn't care if I toasted a fan. But I would not risk making a recommendation to someone else based upon my assumption that a 58% or more drop in voltage represents a "minor variation".
Well Nidec, for one, has clearly taken a position to the contrary . Do you have any links to manufacturer sites that state this ?
I don't doubt that over time, this can kill the motor.... the question is, is that going to happen within a time period that we care about ? If my fan has a 1 year warranty, by all means go ahead....5 years ? maybe I don't wanna say that.
TRENDING THREADS
-
Question User RX 580 8gb red devil powercolor, asking original BIOS- Started by PizzaInGame
- Replies: 5
-
-
-
News Microsoft updates Windows 11 24H2 requirements, CPU must support SSE4.2 or the OS will not boot- Started by Admin
- Replies: 18
Facebook
Twitter
Instagram