How much until PWM signal degrades? Fan question

kimura14

n00b
Joined
Dec 17, 2016
Messages
30
I'm sure this question has come up but I've spent about 2 hours looking for a real answer and haven't found one. I'm currently running 12 fans on 2x360 radiators and each rad (6 fans) has it's own hub with sata power adapter. Currently I'm using a single cable (2 wires) to supply 1 tach signal and retrieve 1 PWM signal from the mobo for all 12 fans. My question is, how much is too much? I would assume eventually the PWM signal has to degrade enough to be detrimental. Does anyone know the answer?
 
Not enough to notice in a home PC application.

The PWM signal isn't power delivery; there's no load on it at the end device, so you shouldn't expect any degradation at all. Plus, your hubs each have their own power source, so I'm betting they're actively repeating that PWM signal to each fan with their own power as opposed to a pass-through.
 
The PWM is a powered signal delivered to fan's PWM circuitry.
When there is no PWM signal to fan's PWM circuitry the fan motor runs on constant 12 volt power.
When there is contant PWM signal to fan's PWM circuitry the fan motor has no power to motor.
The fan's 12v power is on whenever the PWM signal power is off, so every time the PWM signal is powered the fan's motor power receives no power every time the PWM signal is off the fan motor receive power.
Hope that makes sense.

I believe this is by design so PWM fans can be used without PWM signal. So when a 4-pin PWM fan is plugged onto a 3-pin fan header a PWM fan functions as a variable speed fan.

In my experiences 10-12 fans is about the most that most PWM signals will support. It all depends on fan's PWM circuit vs signal PWM circuit. Not all are build to same design standards. For example some Corsair PWM fans will only work with maximum of 2 or 3 fans without problems. I don't know if there is a correlation, but PWM controlled splitter hubs all seem to have 8 fan headers on them
 
I'm sure this question has come up but I've spent about 2 hours looking for a real answer and haven't found one. I'm currently running 12 fans on 2x360 radiators and each rad (6 fans) has it's own hub with sata power adapter. Currently I'm using a single cable (2 wires) to supply 1 tach signal and retrieve 1 PWM signal from the mobo for all 12 fans. My question is, how much is too much? I would assume eventually the PWM signal has to degrade enough to be detrimental. Does anyone know the answer?

PWM = pulse width modulation. It doesn't carry high voltage instead the signal tells the controller how many cycles to turn on and off. Otherwise, the fan is on 100%. Think of it like the remote signal on a car stereo which can be spliced a ton of times.
 
While it is a remote signal, it is not wireless. The wired signal is of a specified strength, and each receiver of that wired signal takes some of the signal. Depending on how much signal each PWM fan circuit uses and how powerful the signal is determines how many fans can run on the signal. While 8 of most fans work fine, while with some other fans only 2 or 3 (early Corsir SP and AP series for example) will function properly. As I said before, generally 8 of most fans work fine,
 
There is a power draw in the presence of a PWM, but the amount of power drawn depends on the Max Power (i.e. fan @100%) and the duty cycle of the PWM signal. Roughly speaking, a 50% duty cycle would draw 50% of max power. So a 3W fan would draw 1.5W at 50% duty cycle.

The same concept can be applied to an LED. It's how one can get a dimming effect.

So, worst case scenario is all 12 fans are running at 100%. If they're all same type of fan running at the same speed, then you can do simple maths to figure out the effective power draw.

Edit: To answer the OP's question, "too much" would be when all fans (in concert) are drawing more power than the hub or whatever is supplying power is rated to handle. My guess is you'd need a larger quantity of fans or fans that pull more power to break things. Even then, worse case scenario is all said fans are running at 100%, which is rarely the case when fan speeds are being controlled and monitored by a system.
 
Last edited:
There is a power draw in the presence of a PWM, but the amount of power drawn depends on the Max Power (i.e. fan @100%) and the duty cycle of the PWM signal. Roughly speaking, a 50% duty cycle would draw 50% of max power. So a 3W fan would draw 1.5W at 50% duty cycle.

The same concept can be applied to an LED. It's how one can get a dimming effect.

So, worst case scenario is all 12 fans are running at 100%. If they're all same type of fan running at the same speed, then you can do simple maths to figure out the effective power draw.

Edit: To answer the OP's question, "too much" would be when all fans (in concert) are drawing more power than the hub or whatever is supplying power is rated to handle. My guess is you'd need a larger quantity of fans or fans that pull more power to break things. Even then, worse case scenario is all said fans are running at 100%, which is rarely the case when fan speeds are being controlled and monitored by a system.
The PWM signal is a completely separate system from the pulsed 12v power going to fan motor.

Too many PWM fans on a PWM signal have nothing to do with the 12v power the PWM fan motors run on.

The PWM signal from motherboard or other source (think of it as 'control master') is on pin for of fa header and goes to an PWM control system in the fan (think of it s 'control slave').
'Master' tells 'slave' what to do.
When 'master sends signal 'slave' stops the flow of power to fan motor.
The PWM system does not power the PWM fan's motor.
The PWM system controls pulses of constant 12v power coming 12v power on pin-2 of 4-pin PWM controlled fan header.

An overly simple comparison is a remote power relay that when the control sends3-5v power to the relay (PWM system in fan) it hold it open so there is no power goes to fan motor, and when control does not send power to relay (PWM system in fan) relay is closed and there is 12v power to fan motor. The more time the 3-5v PWM signal is on the less time there is not power tofan motor. ;)

No PWM control means power to fan motor is constant, so a PWM fan can be use on variable voltage fan header and have constant power of variable voltage from pin-2 to fan motor. ;)
 
The PWM signal is a completely separate system from the pulsed 12v power going to fan motor.

Too many PWM fans on a PWM signal have nothing to do with the 12v power the PWM fan motors run on.

The PWM signal from motherboard or other source (think of it as 'control master') is on pin for of fa header and goes to an PWM control system in the fan (think of it s 'control slave').
'Master' tells 'slave' what to do.
When 'master sends signal 'slave' stops the flow of power to fan motor.
The PWM system does not power the PWM fan's motor.
The PWM system controls pulses of constant 12v power coming 12v power on pin-2 of 4-pin PWM controlled fan header.

An overly simple comparison is a remote power relay that when the control sends3-5v power to the relay (PWM system in fan) it hold it open so there is no power goes to fan motor, and when control does not send power to relay (PWM system in fan) relay is closed and there is 12v power to fan motor. The more time the 3-5v PWM signal is on the less time there is not power tofan motor. ;)

No PWM control means power to fan motor is constant, so a PWM fan can be use on variable voltage fan header and have constant power of variable voltage from pin-2 to fan motor. ;)

We're basically talking about the same thing. You're correct in that the PWM signal, these days, is a separate channel to the fan header (i.e. 4 pin vs 2/3 pin). When the PWM channel pulses, the fan's motor input power pulses. When the PWM signal drops to zero, the power to the fan motor is (probably) gated. The end result is the same as if you were to pulse the +12V using an external PWM generator. It does make more sense (again, these days) to place the PWM in a separate channel as you can use any source to generate the PWM signal (presumably even a low powered device) while maintaining power distribution through a standard 12V power supply. My argument is that the end result is going to be the same, regardless of who is doing the pulsing. The power consumption is going to be the same. A fan's power consumption should be a function of the fan's duty cycle. The duty cycle is going to be controlled by the PWM signal. Now, depending on the fan circuitry, it may not be precisely 1:1, but I'd bet it's very very close.

An overly simple comparison is a remote power relay that when the control sends3-5v power to the relay (PWM system in fan) it hold it open so there is no power goes to fan motor, and when control does not send power to relay (PWM system in fan) relay is closed and there is 12v power to fan motor. The more time the 3-5v PWM signal is on the less time there is not power tofan motor.

That seems more like an inversion of the duty cycle. If your duty cycle were, say, 70%, then by your example's logic the fan's motor would only be receiving power 30% of the time. I don't think that's how most 4-pin fans work, but I get where you are going with that.

No PWM control means power to fan motor is constant, so a PWM fan can be use on variable voltage fan header and have constant power of variable voltage from pin-2 to fan motor.

This is a good observation as some fan headers, despite supporting 4-pins, do not support PWM on the 4th pin. So to compensate, they adjust the voltage on 2nd pin. The problem with this method is that some fans will refuse to spin when the voltage is set too low. With a PWM, you can get much lower fan speeds as the minimum voltage threshold is not a factor.
 
Last edited:
We're basically talking about the same thing. You're correct in that the PWM signal, these days, is a separate channel to the fan header (i.e. 4 pin vs 2/3 pin). When the PWM channel pulses, the fan's motor input power pulses. When the PWM signal drops to zero, the power to the fan motor is (probably) gated. The end result is the same as if you were to pulse the +12V using an external PWM generator. It does make more sense (again, these days) to place the PWM in a separate channel as you can use any source to generate the PWM signal (presumably even a low powered device) while maintaining power distribution through a standard 12V power supply. My argument is that the end result is going to be the same, regardless of who is doing the pulsing. The power consumption is going to be the same. A fan's power consumption should be a function of the fan's duty cycle. The duty cycle is going to be controlled by the PWM signal. Now, depending on the fan circuitry, it may not be precisely 1:1, but I'd bet it's very very close.



That seems more like an inversion of the duty cycle. If your duty cycle were, say, 70%, then by your example's logic the fan's motor would only be receiving power 30% of the time. I don't think that's how most 4-pin fans work, but I get where you are going with that.



This is a good observation as some fan headers, despite supporting 4-pins, do not support PWM on the 4th pin. So to compensate, they adjust the voltage on 2nd pin. The problem with this method is that some fans will refuse to spin when the voltage is set too low. With a PWM, you can get much lower fan speeds as the minimum voltage threshold is not a factor.
All computer PWM fans I have ever seen have 4 pins, being ground & 12v power, rpm signal from fan and PWM signal to fan.

Your explaination made it sound like the power was pulsed going to the PWM fan .. and that is not the way PWM control works on computer PWM fans.

The point I was trying to make is PWM computer fans have a constant 12v power lead to them, and use an external PWM signal to pulse that 12v power to the motor .. and if there is no PWM signal PWM fans have contant 12v to their motor.

There is no pulsed power lead on a PWM fan's plug.

Exactly! The PWM 'duty cycle' is the reverse of fan's power duty cycle. That is exactly how it works!! If there is 0% PWM signal the fan has 100% power.

That is why PWM fans change speed when plugged into 3-pin variable voltage fan headers the same way 3-pin variable voltage fans do.

This is also why people plug a PWM fan into a 4-pin fan header and assume it is functioning on PWM because the fan changes speed when it is just as likely the 4-pin fan header is only using 3 pins and is really a variable voltage fan header. ;)
 
Last edited:
https://en.wikipedia.org/wiki/Pulse-width_modulation#Power_delivery

VanGohComplex is right - PWM signal won't degrade at all. Your limit is strictly amps on your 12V rail or power line that your using to drive your fans.

A typical PC wire and molex connector at 18GA is usually rated for 4.5A - 5A per line, although Molex lists the design spec of the connector at 11A per pin. SATA has up to 3 12V pins, at up to 4.5A per pin, but given that most are fed from a single 12V rail on 18GA, your back down to 4.5-5A per connector.

So I guess to sum all that up - typically, you can drive about 50-60W worth of fans off a single power connector. Not sure what fans you are driving, but looking at a typical high static pressure fan for use on radiators (Corsair 120), it draws about 0.2A at full speed.... so you could, hypothetically, drive 300 fans from a single Molex or SATA connector at full speed, and possibly much more. Your fans may draw more (or less) power than that, but it should be easy to look up exactly what they do draw, and unless you have some seriously high powered fans, your 12 fans are probably no problem at all for your current setup.

Just for reference, a typical motherboard fan header is usually rated at about 2A.

*edit* incorrect link in the article, corrected
 
All computer PWM fans I have ever seen have 4 pins, being ground & 12v power, rpm signal from fan and PWM signal to fan.

Your explaination made it sound like the power was pulsed going to the PWM fan .. and that is not the way PWM control works on computer PWM fans.

The point I was trying to make is PWM computer fans have a constant 12v power lead to them, and use an external PWM signal to pulse that 12v power to the motor .. and if there is no PWM signal PWM fans have contant 12v to their motor.

There is no pulsed power lead on a PWM fan's plug.

Exactly! The PWM 'duty cycle' is the reverse of fan's power duty cycle. That is exactly how it works!! If there is 0% PWM signal the fan has 100% power.

That is why PWM fans change speed when plugged into 3-pin variable voltage fan headers the same way 3-pin variable voltage fans do.

This is also why people plug a PWM fan into a 4-pin fan header and assume it is functioning on PWM because the fan changes speed when it is just as likely the 4-pin fan header is only using 3 pins and is really a variable voltage fan header. ;)

Not quite. To settle things, a PWM fan header, at least a 4-pin standard one as you see on a motherboard, is generally going to be an open-collector type output with a pull-up resistor on the fan side. Very similar to a relay, but fundamentally different. That means the fan is providing the actual voltage for the PWM signal, while the job of the motherboard is to only open and close the circuit (via some type of FET or similar IC). So in a way, you're right, the power part of the pulse does not come from the motherboard side. However, the "pulse", i.e. the opening and closing of the circuit, is coming from the motherboard side. I went back and re-read what I've written and I don't see where I ever explicitly mentioned power was coming from the PWM pin itself. I was only ever implying that the act of PWM'ing the fan's motor draws power, and that power draw can be directly related to the PWM duty cycle. Not sure where things derailed with your interpretation there. *shrug*

And just to be perfectly clear, and I just looked up the spec, btw, the PWM signal is not inverted. It can't be. Because at 100% duty cycle, the circuit will close, implying the pull-up will go 5V, hence the "active" part of the pulse. If the duty cycle goes 0%, or the fan is disconnected, the circuit will become "open".

Source: http://www.formfactors.org/developer\specs\4_Wire_PWM_Spec.pdf
 
Last edited:
https://en.wikipedia.org/wiki/Pulse-width_modulation#Power_delivery

VanGohComplex is right - PWM signal won't degrade at all. Your limit is strictly amps on your 12V rail or power line that your using to drive your fans.

A typical PC wire and molex connector at 18GA is usually rated for 4.5A - 5A per line, although Molex lists the design spec of the connector at 11A per pin. SATA has up to 3 12V pins, at up to 4.5A per pin, but given that most are fed from a single 12V rail on 18GA, your back down to 4.5-5A per connector.

So I guess to sum all that up - typically, you can drive about 50-60W worth of fans off a single power connector. Not sure what fans you are driving, but looking at a typical high static pressure fan for use on radiators (Corsair 120), it draws about 0.2A at full speed.... so you could, hypothetically, drive 300 fans from a single Molex or SATA connector at full speed, and possibly much more. Your fans may draw more (or less) power than that, but it should be easy to look up exactly what they do draw, and unless you have some seriously high powered fans, your 12 fans are probably no problem at all for your current setup.

Just for reference, a typical motherboard fan header is usually rated at about 2A.

*edit* incorrect link in the article, corrected[/QUOTE


Normal PWM motor control and computer fan PMW control are not the same beast.

What you talk in theory does not work in reality. What is the maximum number of PWM fans you have controlled on+ 1 PWM fan header on a motherboard? I have my doubts you have every tried running as many as your fan headers will control.

I have and I have found in actual use 8-10 fans is generally the limit of what can be controlled with a single motherboard PMW header.

Indeed, typical PMW hubs can handle 4-5 amps. But it all depends on wire size, traces on board, etc. what it will actually handle.


Not quite. To settle things, a PWM fan header, at least a 4-pin standard one as you see on a motherboard, is generally going to be an open-collector type output with a pull-up resistor on the fan side. Very similar to a relay, but fundamentally different. That means the fan is providing the actual voltage for the PWM signal, while the job of the motherboard is to only open and close the circuit (via some type of FET or similar IC). So in a way, you're right, the power part of the pulse does not come from the motherboard side. However, the "pulse", i.e. the opening and closing of the circuit, is coming from the motherboard side. I went back and re-read what I've written and I don't see where I ever explicitly mentioned power was coming from the PWM pin itself. I was only ever implying that the act of PWM'ing the fan's motor draws power, and that power draw can be directly lated to the PWM duty cycle. Not sure where things derailed with your interpretation there. *shrug*

And just to be perfectly clear, and I just looked up the spec, btw, the PWM signal is not inverted. It can't be. Because at 100% duty cycle, the circuit will close, implying the pull-up will go 5V, hence the "active" part of the pulse. If the duty cycle goes 0%, or the fan is disconnected, the circuit will become "open".

Source: http://www.formfactors.org/developer\specs\4_Wire_PWM_Spec.pdf

The term "100% duty cycle" is missleading. If 100% duty cycle was constant PWM signal the fan would stop turning. 100% duty cycle is what the fan motor is doing, not the PWM signal. 100% duty cycle does not mean the PMW signal is sending pulses so close together that the PWM signal is on 100% of the time. At 100% duty cycle the fan will run at 100% of it's speed when the PWM signal is sending no signal/pulse at all.

Simple way for you to prove to yourself that our PWM fans have constant power with no PWM signal is to plug one onto a 3-pin variable voltage fan header and see if it runs. It will run just like any other 12v computer fan.

How can it do that if your claim of 100% duty cycle is full power PWM signal? Answer is it cannot.

100% PWM duty cycle is 100% fan speed.

Your 100% duty cycle is really 100% fan speed .. and that is the same as when the PWM signal not functioning at all .. constant 12v power to fan.

I went through all of this when first computer PWM fans came out and had a hell of a time until I figured out the PWM 'duty cycle' signal is the opposite of what it sounds like.
 
Last edited:
The term "100% duty cycle" does not mean the PMW signal is sending pulses so close together that the PWM signal is on 100% of the time. It means at 100% duty cycle the fan will fun at 100% of it's speed.

What part of "100% duty cycle" do you not get? What do you think the "duty" part of "duty cycle" actually is? If one were to hook up an oscilloscope and monitor a PWM signal, one would expect a 100% duty cycle to appear as a constant voltage (or very near). If there is anything that resembles a periodic signal, then, by definition, that's not a 100% duty cycle. The very definition of 100% duty cycle is that output remains at a constant quantifiable value (i.e. non-zero). The polar opposite would be 0%, assuming no voltage bias is present, which, in digital circuits, implies an open circuit or near-zero voltage.

Simple way for you to prove to yourself that our PWM fans have constant power with no PWM signal is to plug one onto a 3-pin variable voltage fan header and see if it runs. It will run just like any other 12v computer fan.

I never stated otherwise.

How can it do that if your claim of 100% duty cycle is full power PWM signal? Answer is it cannot.

Because the spec says that it has to. And funny enough, that's how it actually works.

100% PWM duty cycle is 100% fan speed.

On that we agree.

Your 100% duty cycle is really 100% fan speed .. and that is the same as when the PWM signal not functioning at all .. constant 12v power to fan.

Whoah, hold up there... There's a reason why the fan runs at 100% when the PWM signal is not present, and that maybe is where you are in error with your logic. First off, 0% (or 100%, or any other value) duty cycle is not the same as PWM signal not present. PWM signal not present is simply PWM signal not present. That is a completely different state than when the PWM signal is present.

It is clearly stated in the Intel spec I linked above (btw, damn you for forcing me to resort to looking it up ;)):

If no control signal is present the fan shall operate at maximum RPM.

And just in case you're still not crystal clear on things, this is also stated in the spec:

The maximum fan speed shall be specified for the fan model by the fan vendor and correspond to 100% duty cycle PWM signal input.

According to you, these two things would be a contradiction. But the truth is, they are not. They are, as I said, two completely different requirements unrelated to each other. What this really means is that while a PWM signal is present (and that's the important keyword), then the fan speed is a function of the duty cycle. However, if the PWM signal is not present (again, keywords), then the fan will operate at maximum speed. This implies that PWM and PWM sensing are, in fact, mutually exclusive.

However, what this truly implies is that each fan has a sensing circuit (in addition to the PWM part) for when a PWM signal is present or not present. Indeed, it is those very two states that determine what drives the fan speed if at all.

Basically, when the PWM pin is connected, the sensing circuit drives the fan speed based on the PWM signal. If the PWM pin is disconnected, then the sensing circuit drives the fan speed to 100% regardless.


Edit: Forget the whole 'sensing' notion, I already answered this in the previous post explaining how everyone does it. The pull-up resistor in the fan part of the PWM circuit solves this problem. When the PWM signal is disconnected, the pull-up will ensure logic high. This is the same thing as 100% duty cycle which equates to 100% fan speed per fan logic.

Visuals and simple explanation here.

I went through all of this when first computer PWM fans came out and had a hell of a time until I figured out the PWM 'duty cycle' signal is the opposite of what it sounds like.

As a side-note, I would just like to add that before we had "standards", and before we even had a 4th pin in general to drive an isolated PWM signal, we still had PWM controllers and could (as enthusiasts) drive fans (PC fans mind you) through PWM circuits that didn't require anything but the right know-how and tools. I can speak with some first-hand authority there. PWM isn't some new thing that just popped up overnight. It's been around forever. I can recall 20 years ago, as a student in my digital circuits lab courses, we'd create PWM circuits to drive all sorts of things: fans, motors, LEDs, etc.

And to be perfectly honest, I have no idea how a modern fan really takes a PWM signal and drives the motor internally. I can only speculate. I think it's going to depend on the implementation and implementor. I've never actually taken a modern fan apart and studied it. But that's why we have a specification today. It's why I can take the stance of "I don't really have to know how it works internally, I only need to understand how the interface works." And the interface part is clearly defined in the Intel's 4-pin spec. The spec is there to put everyone who makes a 4-pin fan on the level. The spec is a "good" thing.

And it's very possible what you've been describing is what is going on internally within the fan circuit. Again, I have no idea right now if that's the case. Without sounding too sarcastic, if you are a fan maker, you are free to design your fan however you want. But I'm pretty darned sure the way you describe things is NOT what's going on, at least externally, with the majority of PC 4-pin fans out there. And if you don't follow the spec (or standard), then you're probably not going to be making a lot of money off the PC crowd.

Sorry, if any of that came off as patronizing or condescending. I was holding back being snarky until the very end. ;)
 
Last edited:
What part of "100% duty cycle" do you not get? What do you think the "duty" part of "duty cycle" actually is? If one were to hook up an oscilloscope and monitor a PWM signal, one would expect a 100% duty cycle to appear as a constant voltage (or very near). If there is anything that resembles a periodic signal, then, by definition, that's not a 100% duty cycle. The very definition of 100% duty cycle is that output remains at a constant quantifiable value (i.e. non-zero). The polar opposite would be 0%, assuming no voltage bias is present, which, in digital circuits, implies an open circuit or near-zero voltage.

I'm not going to bother with the rest of your rant.


Have you hooked up an oscilloscope and monitored the PWM signal?

My understanding is 4th pin on PWM fan header is 3.3 - 5.25 volt, and when there is voltage on 4th pin (PWM pin) the fan motor receives no 12v power (and does not run).

My understanding is when 4th pin has no voltage (no PWM signal) the fan motor received constant 12v power.

Based on what above my understanding is there is no PWM signal voltage at 100% duty cycle.


At 100% duty cycle the fan runs at full speed because fan motor is receiving 100% 12volt power.

At 100% duty cycle there is no PWM control signal ..

"The fan RPM (as a percentage of
maximum RPM) should match the PWM duty cycle within ±10%. If no control signal is present
the fan shall operate at maximum RPM
. See Figure 3."



100% fan motor power = 100% duty cycle

100% duty cycle = no 3.3-5.25v signal on pin-4 (PWM pin on 4-pin PWM fan header).


From "4-Wire Pulse Width Modulation(PWM) Controlled Fans" in link Brian B so kindly posted above


"2.1.4 PWM Control Input Signal

The following requirements are measured at the PWM (control) pin of the fan cable connector see
Figure 7 and Table 1:
PWM Frequency: Target frequency 25 kHz, acceptable operational range 21 kHz to 28 kHz
Maximum voltage for logic low:
VIL = 0.8 V
Absolute maximum current sourced:
Imax = 5 mA (short circuit current)
Absolute maximum voltage level:
VMax = 5.25 V (open circuit voltage)
This signal must be pulled up to a maximum of 5.25V within the fan.
Note: New fan designs are strongly encouraged to implement a 3.3V pull up for compatibility with
buffer design limits on Hardware Monitor Devices e.g. Super IO devices."



To further clarify what I keep saying I have underlined the relevant part:


"3.3 Fan Speed Response to PWM Control Input Signal

The PWM input shall be delivered to the fan through the control signal on Pin 4 (see Section
2.1.4). Fan speed response to this signal shall be a continuous and monotonic function of the duty
cycle of the signal, from 100% to the minimum specified RPM. The fan RPM (as a percentage of
maximum RPM) should match the PWM duty cycle within ±10%. If no control signal is present

the fan shall operate at maximum RPM. See Figure 3."

 
You're just regurgitating everything already said. I don't see how you can interpret 100% duty cycle as 0 RPM on the fan. That's just backwards. You even contradict yourself a couple of times.

My understanding is 4th pin on PWM fan header is 3.3 - 5.25 volt, and when there is voltage on 4th pin (PWM pin) the fan motor receives no 12v power (and does not run).

And that's the part you're failing to get. If you can't understand what a Duty Cycle means and how it relates to driving the fan logic, then I can't help you. The Intel spec even demonstrates a real-world example PWM on an oscilloscope graph. It's pretty clear what active-high and active-low are. And, as I've said multiple times, the pull-up takes care of the active-high requirement when the 4th pin on the fan is not connected to anything.

It can't be more clearly stated in the spec - when the fan driver logic sees active-high (> 0.8V to 5.25V) from the PWM circuit it will power the motor. When the fan driver logic sees active-low (0V to ~0.8V), it will gate the power. Active-high = Power = fan rotation. Active-low = No power = no fan rotation. I can't explain it more simple than that.

Edit: Here's a scope of the PWM you wanted:

zLIs00M.png
 
Last edited:
You're just regurgitating everything already said. I don't see how you can interpret 1
00% duty cycle as 0 RPM on the fan. That's just backwards. You even contradict yourself a couple of times.



And that's the part you're failing to get. If you can't understand what a Duty Cycle means and how it relates to driving the fan logic, then I can't help you. The Intel spec even demonstrates a real-world example PWM on an oscilloscope graph. It's pretty clear what active-high and active-low are. And, as I've said multiple times, the pull-up takes care of the active-high requirement when the 4th pin on the fan is not connected to anything.

It can't be more clearly stated in the spec - when the fan driver logic sees active-high (> 0.8V to 5.25V) from the PWM circuit it will power the motor. When the fan driver logic sees active-low (0V to ~0.8V), it will gate the power. Active-high = Power = fan rotation. Active-low = No power = no fan rotation. I can't explain it more simple than that.


Where in in spec does it say "when the fan driver logic sees active-high (> 0.8V to 5.25V) from the PWM circuit it will power the motor"?

Maybe this quote from specs will clear it up for you:


"2.3.1 PWM Output Signal

The Hardware Monitor Devise is required to provide an open-drain or open-collector type output
for the PWM signal on pin 4
(see Table 1 and Figure 8) with the following properties:

Frequency: 25kHz nominal, 21-28kHz is acceptable

Current sink capability: 5mA required, 8mA recommended

Maximum voltage capability: 5.25V

Maximum VOL: 0.8V

Signal is not inverted, 100% PWM results in Max fan speed"


Isn't 'an open-drain or open-collector type output' and open circuit? In other words a circuit with no signal because it is open?


100% duty cycle is fan running 100% speed. I've never said that was not the case.

What I keep saying is PWM 100% duty cycle is not not constant PWM signal, but the absence of any PWM signal .. in our applications motherboard, GPU, aquaero fan controller, etc. Exactly how this is done by the electronics in master and slave PMW controller is well over the heads of most forum readers. But understanding no signal give fan constants power is what most users can understand .. except it seems you.


I think we are saying basically the same thing, but your explaining it in technical form and I'm explaining it in a way non-technical layman can understand.

0% PWM signal = 100% PWM open signal = 100% duty cycle = approx 100% fan speed
50% PWM signal = 50% PWM open signal = 50% duty cycle = approx 50% fan speed .. depends on what fan PWM signal to power pulse curve program is.
100% PWM signal = /0% PWM open signal = 0% duty cycle = approx 0% fan speed .. or whatever minimum rpm PWM signal to power pulse program is.
 
The fan driver logic will see active high because there's a pull-up resistor in the fan to bring the PWM signal high when the motherboard opens it's side of the interface. Without the pull up to 12v, the fan will not run.

So 100% PWM duty cycle as seen by the fan logic will mean 100% fan speed. 0% PWM duty cycle (again as seen at the fan's logic) will result in 0% fan speed.
 
I didn’t bother to read the spec, or to fully read everyone’s discussion here.

But wouldn’t an open collector on a line with high voltage read the same on a scope printout voltage as an active source with voltage on a passive line?

So, it seems your arguing semantics - does the controller send an active high to a passive circuit, or does it drain from a live circuit? Either way - the oscope waveform looks the same, the definition of “duty cycle” doesn’t change (it should be with reference to the load speed, i.e. fan speed), and the fan would work the same, just slight change in how the electronics drive the system.

Does it really matter if the PWM is an active drive or a passive sink? Technically, yes. It’s the difference in driving your car forward or reverse - your still getting from point A to B, it’s just a bit different. But realistically - your using the same device to get from A to B in more or less the same fashion.

But with regard to how many fans
you can drive - to the controller it’s still just current. Regardless of if you are actively pushing it, or passively sinking it. The limit is just how many amps that circuit can drive in total.
 
I think we are saying basically the same thing, but your explaining it in technical form and I'm explaining it in a way non-technical layman can understand.

I disagree with the technical part. I didn't start getting technical until you started refuting everything I originally stated. My first three posts were very high level, devoid of anything technical in nature. If anything, you were the one who initiated dragging it down to a technical level by refuting my statements or making wild assertions by contradicting them. But I'm willing to share the blame here for the sake of keeping this thread from derailing further into some tehnical essay on how a PWM fan works.

So let bygones be bygones, or something, and move on.
 
I disagree with the technical part. I didn't start getting technical until you started refuting everything I originally stated. My first three posts were very high level, devoid of anything technical in nature. If anything, you were the one who initiated dragging it down to a technical level by refuting my statements or making wild assertions by contradicting them. But I'm willing to share the blame here for the sake of keeping this thread from derailing further into some tehnical essay on how a PWM fan works.

So let bygones be bygones, or something, and move on.
 
I didn’t bother to read the spec, or to fully read everyone’s discussion here.

But wouldn’t an open collector on a line with high voltage read the same on a scope printout voltage as an active source with voltage on a passive line?

So, it seems your arguing semantics - does the controller send an active high to a passive circuit, or does it drain from a live circuit? Either way - the oscope waveform looks the same, the definition of “duty cycle” doesn’t change (it should be with reference to the load speed, i.e. fan speed), and the fan would work the same, just slight change in how the electronics drive the system.

Does it really matter if the PWM is an active drive or a passive sink? Technically, yes. It’s the difference in driving your car forward or reverse - your still getting from point A to B, it’s just a bit different. But realistically - your using the same device to get from A to B in more or less the same fashion.

But with regard to how many fans
you can drive - to the controller it’s still just current. Regardless of if you are actively pushing it, or passively sinking it. The limit is just how many amps that circuit can drive in total.

Ok, so if the fans do indeed provide the power, in effect in parallel thanks to the hub, this would be driving the amps up creating a potential limit to the maximum number of fans. Is that what you meant by the last sentence? If so, it's completely the opposite of what I had assumed would be happening, but does the same thing in the end, creates a limit.
 
So an open circuit on pin-4, the PWM control pin is fan receiving 12 volt power all the time .. and a closed circuit on pin-4 lowers fan speed to minimum rpm the fan's internal PWM circuity is programmed for.

Is that a simple way to interpret how PWM fan header controls PWM circuity in fan?
 
Ok, so if the fans do indeed provide the power, in effect in parallel thanks to the hub, this would be driving the amps up creating a potential limit to the maximum number of fans. Is that what you meant by the last sentence? If so, it's completely the opposite of what I had assumed would be happening, but does the same thing in the end, creates a limit.

The simple answer is there is definitely a limit to how many PWM fans can be controlled by a single PWM fan header. The reasons there is not finite number like '10 fans' is because not all fans have same circuitry, and the standard has a rather wide range of specifications for both master and slave sides of PWM controller.

I know this because I have added fans to PMW motherboard fan headers as well as on aquaero PWM fan controller and I could almost always get 8 or 9 fans to work, but how many more depended on what fans were used on what controller. The only really bad ones that would only allow 2-4 fans to be controlled were Corsair AF and SP fans. The last time I did this testing was about 3 years ago, but as far as I know the PWM standards have not changed and the same compatabliity problems still exist.
 
Makes sense. My main concern from the start was that I don't want to burn out a fan or a header prematurely. My rads are in a pedestal so I combined everything into 1 PWM & TACH cable / header for convenience. At this point in the thread I'm considering bringing 2 cables up for peace of mind. Not at all saying I haven't enjoyed reading the responses so far. (y)
 
Makes sense. My main concern from the start was that I don't want to burn out a fan or a header prematurely. My rads are in a pedestal so I combined everything into 1 PWM & TACH cable / header for convenience. At this point in the thread I'm considering bringing 2 cables up for peace of mind. Not at all saying I haven't enjoyed reading the responses so far. (y)
As long as you use a PWM splitter or hub with PSU power you will not damage the motherboard header or fans. How many and what fans are you using?
 
I'm using 12 Vardar F3 120s. I was running 1 360 rad on top of my case but got a second one and moved them both to a pedestal. FWIW I did learn they are a much quieter fan mounted upright vs flat as I had them before.
 
I'm using 12 Vardar F3 120s. I was running 1 360 rad on top of my case but got a second one and moved them both to a pedestal. FWIW I did learn they are a much quieter fan mounted upright vs flat as I had them before.
We sometimes rant on a bit. :D
My guess is you can run all 12 on one PWM source with no problems.
F3-120 are rated 1.4watt each, so even if they draw twice that at startup or stalled it is only 2.8w, for a total of 33.4 watts or 2.8 amps and that is only a guess, they may only draw 1.4 watt at startup. .. Either way not a problem on a decent PWM hub with PSU power. Silverstone PWM splitter hub has a capacitor to supply extra startup power.

As a comparison Nidec-Servo Gentle Typhoon 1850rpm fan model number D1225C12B5AZ-00 is rated 1 watt running and 4.3 watt startup.
 
I'm using two Swiftech hubs, still coming straight from my PSU. What's the point of the cap if your plugging into the PSU?
 
I'm using two Swiftech hubs, still coming straight from my PSU. What's the point of the cap if your plugging into the PSU?
The Swiftech hubs are good.
The cap is just to store power for that instantaneous extra load when fans start up and often draw 2 to 5 times as much power as when runnng. Many electric motors have starter capacitors on them. ;)
 
Back
Top