4-pin or 3-pin? Simple question.

[redsquirrel0249]

I know what 4-pin and 3-pin fans are. Basically. 4-pin fans are controlled via PWM and 3-pins can only be controlled by voltage adjustment.

 

But wait... if you can control them both, why do you need (or why would you want) 4-pin over 3-pin if 3-pin is cheaper? Is it easier? Is it more stable?

 

Just curious. I'm trying to get the greatest value in air cooling for a Meshify C.

[minibois]

4 pin has a wider range (mostly on the 'lower end' spectrum) of speed adjustment, while with voltage control the fan basically just stops spinning at a certain point.

Plus, unless you have a good controller, speed control on 3 pin fans in motherboard is a bit less granular.

[redsquirrel0249]

@Minibois Thanks, good to know. Lower range is important for my use case, and more specific controls are always useful. Doesn't tightness of PWM control vary by motherboard, though? Maybe I'm not well versed enough on my Gigabyte B350, but I didn't have a million knobs and buttons for fan curve settings.

[mariushm]

Controlling via PWM is easier... the fan controller simply sends a series of pulses through that 4th wire to the fan. A chip in the fan counts those pulses and depending on the number of pulses counted each second (or something like that) it figures out a percentage between let's say 15% and 100% and then adjusts the rpm accordingly. The smart chip is configured and knows all the time how long it has to send power to the hub to start spinning and what's the minimum rpm to prevent fan from stalling.

 

The voltage remains fixed at 12v all the time.

 

Three pin fans have no "smart chip" in them, they only send the rotation speed to the pc through the 3rd wire.

So the only way for the motherboard to control the fan is by adjusting the voltage going to the fan... and there's somewhat a linear relationship between the fan rpm and the voltage but it's not a straight line. Also, the fan may not turn on if the voltage is too low because of inertia and friction in the hub, but once it spins you may be able to adjust voltage down and get a minimum rpm.

For example, you may need 6.5v to start the fan and get it spinning and afterwards you may be able to lower the voltage down to 4.5v but if anything touches the blades or adds resistance or something like that the fan may stall.

Adjusting the voltage in a chip on the motherboard causes losses and the chip on the motherboard can be warm and such a chip would use more space on the motherboard.

[minibois]

3 minutes ago, redsquirrel0249 said:

@Minibois Thanks, good to know. Lower range is important for my use case, and more specific controls are always useful. Doesn't tightness of PWM control vary by motherboard, though? Maybe I'm not well versed enough on my Gigabyte B350, but I didn't have a million knobs and buttons for fan curve settings.

What I mainly meant by easier control, is that the motherboard can more easily make a real curve, instead of 'steps' which it has to do with voltage control.

[minervx]

may as well do 4 since its not that much more expensive anyway

 

whichever number you pick, pick for all the fans

[redsquirrel0249]

@Minibois That makes sense in comparison to analog/digital audio. Analog is clearly the original sound source, but digital does a decent approximation in certain cases using bits.

 

@mariushm I didn't know PWMs had chips in them to read the signal, though that makes sense.

 

@minervx I disagree; many cases come with 3-pin fans, and I have two with mine that are valued at $20. Deciding whether or not to value such products is important for a lot of people. At the same time, I'm planning on getting a 5 pack of 4-pin Arctic F-12s for $25 or so haha. Why do you say to use the same pin number for all of the fans? I'm getting all the same fans just for continuity's sake.

 

On the note of F-12s, if anyone knows if you can run them outside of their PST syncing mode, I'd be interested to hear about it. I bought some recently and had to return them because they didn't work at all in my testing of multiple configurations and trying to run them at max speed in BIOS.

[mariushm]

6 minutes ago, redsquirrel0249 said:

 

@mariushm I didn't know PWMs had chips in them to read the signal, though that makes sense.

 

 

See the Intel's fan specification / standard, I'm attaching it below.  Check "Fan Speed Control" chapter (3) in particular, it explains how the fans must account for various conditions, more things than what a "dumb" 3 pin fan know.

 

The document also explains rules on what 3 pin fans must follow.

4_Wire_PWM_Spec.pdf

[redsquirrel0249]

@mariushm So Intel fan controllers (and surely others) well have better minimum speeds than 3-pins, but will readjust input below a certain threshold to the minimum spec.

 

I didn't actually know Intel made fans other than their stock coolers, so it's interesting this document comes from them. Thanks. Don't know what exactly a duty cycle is, though, but that's probably more about the electronics side of the product.

[mariushm]

It's not Intel fan controllers and it's not fans made by Intel, Intel just wrote the specifications for various standards, like the ATX standard (where the screws are on motherboards, cases, how power supply voltages should be, where to mount power supplies to cases etc)

 

The manufacturers of fans were supposed to follow the specification written by Intel, if they wanted their fans to be "approved" for use in servers and computers built by various manufacturers with Intel motherboards and processors. But as usual, makes no sense to make 2 or 3 different fan standards so once fan manufacturers accepted this specification, everyone else followed and accepted these specifications.

 

So fan controllers , IO chips on motherboards, all expect fans to follow those specifications, and all fans designed to be sold for use in computers are designed with the assumption that the fan controllers and chips on motherboards will send the signal according to the specifications written by Intel.

 

Through PWM, the motherboard can only say "set the speed to this percentage, between 15% and 100%" but it can't set a particular rpm speed. The software can determine the rpm by simply reading the rpm through the sensor wire (the 3rd wire) a few seconds after the software tells the fan the desired percentage... so if it wants to, the software can determine a "fan profile" ... for example set percentage to 20%, wait, read rpm, set percentage to 30%, wait, read rpm

So if it's patient enough, the software can use 10% steps, or 5% steps, or could be as slow as 1% steps...  then user could simply say "I want 1000 rpm" and then the software could look up at what percentage it would get closest to that rpm and set that percentage.

 

With 3 pin fans the motherboard can only adjust the voltage, but how fine the output voltage adjustment is depends on the controller chip. Some motherboards could adjust voltage of the fan between 5v and 12v in 0.5v steps, others could be more fine and use let's say 0.1v steps

This one also isn't capable of setting an exact rpm value, but like with the other scheme, the software or controller chip can "learn" by reading the rpm at various voltages.

 

[redsquirrel0249]

Didn't know that. Makes sense when you put it in context of motherboard approval. And you can use simple fractions to determine the output RPM from a percentage.

 

I think my motherboard does 5% steps.

 

Good to know background. Definitely feel more confident about the type differences. I think I still would've gone PWM without knowing all that, but I need to look into the specific model I'm going to go for.

[LogicalDrm]

One more thing I haven't seen in here are those who use RGB or LED. Most older style fans (or cheap fans) have LED power on same cable with fan itself. When you have 3pin and control is done via voltage, LED dim when voltage is lowered. With PWM fan receives full power all the time, but as control is done with pulses LEDs will remain at full brightness.

[redsquirrel0249]

@LoGiCalDrm Good to note. I personally don't care for RGB fans, but it's good to know if I ever decide to.

[WoodenMarker]

On 4/25/2018 at 12:51 PM, redsquirrel0249 said:

Why do you say to use the same pin number for all of the fans? I'm getting all the same fans just for continuity's sake.

On the note of F-12s, if anyone knows if you can run them outside of their PST syncing mode

It makes it easier to control the fans. If you run a 12v signal to a 3pin fan, it'll run at 100% so you'll end up needing to run different groups of wires if you want to use both pwm and dc fans. This isn't a hard rule though since a motherboard can have a certain number or pwm or dc headers and it might be more convenient to let that decide which fans you go with.

'PST syncing' is just a daisy chain and you don't need to use the extra wire.

[redsquirrel0249]

Why would you need multiple groups of wires for two different fan types if you're solely relying on mobo fan headers @WoodenMarker? I understand you can't daisy chain them together, but you could put the two 3-pins together, no problem.

 

And like I asked in the other post, just trying to see if I can use F-12s individually w/o PST syncing.

[WoodenMarker]

52 minutes ago, redsquirrel0249 said:

Why would you need multiple groups of wires for two different fan types if you're solely relying on mobo fan headers @WoodenMarker? I understand you can't daisy chain them together, but you could put the two 3-pins together, no problem.

And like I asked in the other post, just trying to see if I can use F-12s individually w/o PST syncing.

 If you run a 12v signal to a 3pin fan, it'll run at 100%. Running both 3 and 4pin fans using a pwm signal is fine if you're fine with the 3pin fans running at 100% or controlling the 4pin fans with dc.

Yes, daisy chaining the fans is optional.

[redsquirrel0249]

Oh, you're talking about chaining the two types of fans together.

 

And that's the problem with PST... I want to daisy chain without syncing. I only have so many fan headers, but I want to control the fans individually. Is that not possible? I at least want front and back fans separate speeds, and I'm sure you can do that, right?

 

 

[WoodenMarker]

3 hours ago, redsquirrel0249 said:

And that's the problem with PST... I want to daisy chain without syncing. I only have so many fan headers, but I want to control the fans individually. Is that not possible? I at least want front and back fans separate speeds, and I'm sure you can do that, right?

No, you can't do that with daisy chaining and this isn't the fault of the fan. Each fan header has one rpm pin. This rpm signal can be duplicated but you can't have more than one signal on the pin. If you want to control separate groups of fans with different settings, use more headers on your motherboard or get a fan controller that has controls for several channels. 

[redsquirrel0249]

So you can definitely control back and front separately. Ok, that's really all I need.

[WoodenMarker]

14 minutes ago, redsquirrel0249 said:

So you can definitely control back and front separately. Ok, that's really all I need.

Yeah, the position of the fan doesn't make a difference.

[redsquirrel0249]

@WoodenMarker So why did you say that GP-12s are better than F-12s? I just sold my GP-12s today lol

[WoodenMarker]

5 minutes ago, redsquirrel0249 said:

@WoodenMarker So why did you say that GP-12s are better than F-12s? I just sold my GP-12s today lol

GP-12's are quieter while performing similarly. That's usually what it means to be better.

[redsquirrel0249]

Yeah, I suppose you're right. I mean, 30 dB is still pretty quiet in my book. Obviously significantly louder, but still not very loud. I may just get some better PWMs. What do you recommend? I feel like spending $20 each on noctuas is stupid. $80 on air cooling sounds stupid. Maybe it's not, but it seems like it...

 

Enermax Twister Cluster Advance (wow that's a mouthful) fans don't look bad.

 

[WoodenMarker]

1 hour ago, redsquirrel0249 said:

Yeah, I suppose you're right. I mean, 30 dB is still pretty quiet in my book. Obviously significantly louder, but still not very loud. I may just get some better PWMs. What do you recommend? I feel like spending $20 each on noctuas is stupid. $80 on air cooling sounds stupid. Maybe it's not, but it seems like it...

Enermax Twister Cluster Advance (wow that's a mouthful) fans don't look bad.

30dBA is quiet but you shouldn't take the fan spec at face value. The spec is likely measured in an unrealistically quiet environment and a real world reading will be much noisier. There's also little to no regulation on how fan specs are measured so they're not a great point for comparison either. Look at in-depth reviews and benchmarks instead.

 

Why did you want pwm in particular and what are you looking for in fans? The included GP-12's are pretty good already. What did you find unsatisfactory about them?

Twisters are still worse than GP-12's. The biggest drawback is the plastic used which is more resonant and pwm which isn't as acoustically smooth as the otherwise similar Magma Advance.

[redsquirrel0249]

How does a quiet environment make the reading of the fan noise quieter? Wouldn't it make it quieter in the perspective of louder noises? I understand testing isn't going to be perfect, but neither is the 19 dB reading on a fractal fan. It's not like I'm looking at poop fans either, I'm only looking at fans well backed by hardware.info, popular among users, or with good reviews otherwise.

 

I just figured extending fan range would be a good idea. What do you mean by acoustically smooth? The transitions between speeds, or the electrical noise of the fan's operation? And I'm planning on using anti-vibration mounts anyway, so plastic shouldn't make much of a difference, right?

 

Besides, plastic is a subjective point that I'd never be able to measure, anyway. If I try to research the plastic or get a review that talks about the plastic for every fan I consider, I'll be pulling my hair out before I find fans to buy. The one thing I'm totally tired of is trying to buy something with very simple criteria and needing put everyone's two cents under a microscope. I'm tired of looking at Abe Lincoln's face at such a high magnification.

 

I need something reasonable in loudness, with effective RPM range, efficient CFM through mesh (which I hear called static pressure, despite the inability for individuals to directly measure this statistic), a worthwhile price, and a decent aesthetic if I can get it.