Pwm silliness

honegod

[H]F Junkie
Joined
Aug 31, 2000
Messages
8,327
Since PWM gives the motor bursts of variable duration 12v from its input,
If that input was 17v would not the fan run MUCH higher rpm @ 100% pwm?
That seems obvious
Also obvious is the motor would start heating up dramatically.

But could this be helped by finding exactly which part of the motor was generating the heat and beefing up the cooling of that part ?

Could the first overvolt overheating part be fixed so it does not fail ?

My 200mm 800rpm case fan is on its way, and I like the idea of being able to overclock it to 1500rpm.

Possibly a cpu HS& fan to cool it down ?

OR A Waterblock, so I could have a watercooled HS fan on my air cooled CPU.
 
Last edited:
Just add another fan instead of "OC'ing" the fan. Adding more cooling to cool a cooling device is....well I think you get the point.
 
Just add another fan instead of "OC'ing" the fan. Adding more cooling to cool a cooling device is....well I think you get the point.

On the other hand 'heat kills' if only by promoting wear, so improving the cooling could well improve lifespan.

If one could cut a slot in the motor case and solder a copper 'fin' onto the hot bit and so have the fan NOT fail when it is spinning its hardest,and is needed most, sounds good to me.

Having a couple hundred units of raw power (rpm) added with no loss of reliability is a wonderful thing, and hardly silly atall.
 
Yup, the PWM modulates the pulse width of whatever voltage is fed to the + pin/wire, commonly 12v.
But the pwm has no idea what that voltage is, right ?
The only voltage that concerns the pwm circuit is whatever the signal is @.

So if a 5v + was plugged in to the power pin when the PWM signal was 100% the fan motor would see 5v.

I am thinking that a power overvolt would only really be 'seen' by the motor when the PWM nears 100%.

I am sure there is a curve that describes the relationship between Voltage seen by the motor and PWM %.

As in @ 18v @ 66.6 pwm = 12v
(not 12v literally, but effectivly)



Why not just a bigger/more fans ?
Already maxed out on available fan area.

Does anyone MAKE a 1500rpm 200mm fan ?
 
4-Wire Pulse Width Modulation
(PWM) Controlled Fans
Specification
September 2005
Revision 1.3

Yes, power input is specified @12v.

But... That is for compatibility, so you can plug your new fan into the same socket your old fan used and it will work properly.

Since it is MY computer I have no such constraints.

Plan A has all fans plugged into a header, with tach and PWM inputs going to whatever controlls that fan.
Cpu,case, on motherboard.
Gpu, a thermal probe regulated pwm controller.

Power & ground directly from the PSU.

It would be easy to add a voltage increaser circuit to pump whatever voltage to whatever fan could use it, I am thinking the case input 200mm.
The gpu & cpu with 120mm 2000rpm each should be fine.
 
I am not a brushless fan designer so I cannot answer your questions. There are a few different ways the circuitry inside the fan could be designed and each of those ways would mean different answers to your questions.

You seem to be focused on the rpm of the fan rather than its CFM and ability to move air.
Take a step back and think about what you actually NEED in terms of airflow given the case you are using.
Do you actually have fans installed right now that are not keeping your system cool or is this all hypthetical back-of-the-napkin stuff?
Also, have you considered noise as a factor? A faster spinning fan will be louder.
 
you can overvolt a fan only up to about 13.8V before it will just burn.. as inductors, resistors, ICMicrochips (those that handle PWM, fan tachometer) and even those who use little capacitors, are rated for very very specific voltage usage. 13.8V IIRC it's the most you will get of a fan before something smoke in front of your eyes.

if you really want to go silly (as referring to the PWM Silliness) go with delta =) 5800RPM 800+CFM 68mmh20 a really quiet 73dba operation at just 230W

5a72bb4edd2c5_Delta200mm.png.af51b937b86183f22bf1ce988d2f857e.png
 
Last edited:
Guys, why has no one suggested an experiment to the OP; just take a 12v wire and a 5v wire from your PSU, splice them together so you have 17v and hook that up to your fan then take an inferred temp probe and point it at the fan motor to see what heats up so he can cool the overheating part? For being the [H]est forum on the internet, I'm surprised no one suggested this yet.
 
Guys, why has no one suggested an experiment to the OP; just take a 12v wire and a 5v wire from your PSU, splice them together so you have 17v and hook that up to your fan then take an inferred temp probe and point it at the fan motor to see what heats up so he can cool the overheating part? For being the [H]est forum on the internet, I'm surprised no one suggested this yet.
Uhmm doesn't work like that
 
If a stock 200mm fan is not enough to cool your system, you need to rethink your enclosure, radiators, or lack there of.
 
You seem to be focused on the rpm of the fan rather than its CFM and ability to move air.

Yes, RPM is the focus of this thread.
I believe a well designed fan blade should be able to pump more air with an increase in RPM above the stock speed.

In that the flow curve should continue to rise with additional rpm, up to the point where the blade starts losing its 'grip' on the air. Faster than I want to go.


Do you actually have fans installed right now that are not keeping your system cool or is this all hypthetical back-of-the-napkin stuff?

In between :)
I got my processor back from Silicon Lottery, the 200mm intake fan has shipped.
the magnesium and titanium and copper sheets for the harddrive array mounting and ducting are on the slowest boat imaginable, from China.

So... Not blue sky, I use a spiral bound notebook.
I need the plans ready so when the [H]ardware finally arrives I can get right on to building "it". As properly as possible.


Also, have you considered noise as a factor? A faster spinning fan will be louder.

Absolutely, PWM means I get to choose how loud I allow it to get.
 
Lots of identity ten tea thoughts here. 18 volt on PWM at 66% is in no way 12v to motor. Each power pulse to motor is still 18 volt.

You can't take a 12 volt power source and add a 5 volt power source and get 18 volts. If it even works without burning something up you will most likely end up with 7 volts with 5 volts from the 12 volt source back-feeding into the 5 volt source. This is exactly what happens when someone uses the 5v and 12v leads from a PSU and end up with 7v.
 
Lots of identity ten tea thoughts here.

Jeepers, I feel all overweened by your self confidence.

So if a1000rpm 12voltsdc. Fan has the pwm set for the fan to run @ the same rpm as it would @100% with 7v input , what is the difference that the motor sees, other than the different taste of the different waveform ?

Would a 6v fan explode if it recieved 12v at 50% pulse duration ( whatever = 100% rpm )

Is there like a circuit in there that cuts off the peaks and stuffs that energy into the valleys so show the motor a constant varying voltage
That would be the bit that explodes because of overload at excess Volt ?

I want to join the non Id10t team, but my ignorance prevents me from asking the right questions.

The last powersupply i used HAD a [ - ] 5v rail .

I did use the 7v trick, +12v to +5v, pre pwm.

I used several 6v car gage lightbulbs on the 5v to load the rail and light my engraved window design.
 
Jeepers, I feel all overweened by your self confidence.

So if a1000rpm 12voltsdc. Fan has the pwm set for the fan to run @ the same rpm as it would @100% with 7v input , what is the difference that the motor sees, other than the different taste of the different waveform ?

Would a 6v fan explode if it recieved 12v at 50% pulse duration ( whatever = 100% rpm )

Is there like a circuit in there that cuts off the peaks and stuffs that energy into the valleys so show the motor a constant varying voltage
That would be the bit that explodes because of overload at excess Volt ?

I want to join the non Id10t team, but my ignorance prevents me from asking the right questions.

The last powersupply i used HAD a [ - ] 5v rail .

I did use the 7v trick, +12v to +5v, pre pwm.

I used several 6v car gage lightbulbs on the 5v to load the rail and light my engraved window design.
A 1000rpm 4-pin PWM fan plugged into a 3-pin variable voltage fan header will run the same speed on 7v variable voltage as a 1000rpm 3-pin variable voltage fan with same motor will run. Obviously the waveform of pulsed power to motor using PMW control is different than constant power to motor from variable voltage, but I can't say exactly what it will look like. It depends on many things including how the pulse is dampened to keep motor from clicking or even possibly vibrating. PWM power control is a combinaton of number of pulses per minute combined wiht length of each pulse to regulate motor speed. On a VOM this can sometimes appear to be lower voltage.

You can use an inline resistor like Noctua supply with their fans to drop the 12v power on a PWM fan to lower voltage and the resulting speed curve will be lower.

No there is nothing cutting peaks and stuffing that energy in valley. Simply explained it's pulses of 12v power to the fan motor.

Yes, people do sometimes combine the 5v and 12v rails to get 7v, but it is not a good idea and can damage things. Basically what is happening is the 12v lead overpowers the 5v lead and feeds the 5v back into PSU so whatever is running on the 12v lead is only able to use 7 volts .. because the other 5 is going back into power supply. I think you can see how that could result is the 5v smoke leaking out. :D

If I understand correctly, your using the 5v rail to power 6v lightbulbs is not the same as using +12v and +5v to get 7 volt.
 
One other thing to mention is that 200mm fans often have slower speeds because the impeller doesn't have the strength to spin at high speed without centripetal force tearing it apart. An 800rpm fan isn't just the motor's rating; it's also the fan blades' rating.

A 200mm fan rotating at 1500 rpm... by my rough calculation you're looking at a ~35mph tip velocity. That's quite a lot for a piece of thin plastic rotating in a ~7" circle.
 
If I understand correctly, your using the 5v rail to power 6v lightbulbs is not the same as using +12v and +5v to get 7 volt.

Not a bit in any way, the notion was to use up the extra potential feeding backwards into the +5v .
 
air cooling is awful why do you want that, loud and obnoxious.


EVERYTHING is air cooling.

radiators are air cooled
Heatpipes are air cooled

The heat that WAS in the liquid nitrogen is in the air now, put there by whatever chills the nitrogen gas.

So air cooling is a given.

The QUESTION is how to best move the heat from its source TO the air.

THIS post is regarding improving the airs ability to absorb heat, by conduction, through increased flow rate .

Wind chill.

If old PSUs could have a MINUS five volt header, then it is POSSIBLE to rig up something similar with modern circuitry.

I SEE voltage boosters on ebay, cheap.

I was mostly looking with this post to see if anybody had explored this notion and establish a dialog about methods.

I want to be ABLE to crank a fan up to 11 or even 13 should I desire to.
 


Yes, that kinda thing.

But why so much hostility ?

So many answers that boil down to " I don't know why but your idea is idiotic".

I SAID RIGHT ABOVE that voltage converters were small cheap and readily available so posting one example and telling me "to blow up all the fans you want" does NOTHING to further my understanding of this topic.

'tis indeed right and proper to sneer, jeer, and mock, but sharing hard won information is better.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
because as someone else pointed out the electronics in the fan will only take so much juice befor literally melting. you don't seem to get that so this will do what you want.
 
Puzzled look.

1) Melt from overheating.

2) Improve cooling to regulate temperature so it doesn't melt

3) Figure out what needs cooling.

4) Figure out how to cool it/them.

5)

6)Profit.

Other than 5) what is it that is so obvious, that i am totally not seeing ?
 
Puzzled look.

1) Melt from overheating.

2) Improve cooling to regulate temperature so it doesn't melt

3) Figure out what needs cooling.

4) Figure out how to cool it/them.

5)

6)Profit.

Other than 5) what is it that is so obvious, that i am totally not seeing ?

That you are trying to solve a problem in the worst possible, and least efficient way. If a standard speed 200mm fan blowing straight onto the motherboard is not enough, you need to rethink your enclosure, or cooling method. A smaller and noisy fan is readily available that can accomplish the same results, that is designed to operate 24/7 under those conditions. But water cooling is far more practical if you need to remove that amount of heat wattage continually. Lastly an open air bench case with a industrial fan blowing across the system would make more sense.

https://www.amazon.com/Computer-Bra...d=1519726191&sr=8-3&keywords=open+air+pc+case

https://www.amazon.com/Stanley-6557...qid=1519726325&sr=8-7&keywords=industrial+fan
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
Guys, why has no one suggested an experiment to the OP; just take a 12v wire and a 5v wire from your PSU, splice them together so you have 17v and hook that up to your fan then take an inferred temp probe and point it at the fan motor to see what heats up so he can cool the overheating part? For being the [H]est forum on the internet, I'm surprised no one suggested this yet.

12v adn 5v used together would give the unit 7v
You need -5v and 12v to get 17v



and to the OP. It is totally doable.
I have done that on as an experiment before heck with a multi switch i could give it up to 24 volt with the -12v and the 12v
I've have also done 150v fans and connected it to 230volt with a pot in the midle (meant for light dimming) and it worked. however at full speed the fan did make a eletrical crackle noise.

how is going to react with pwm i dont know this was before pwm time


if you need any intsruction on how to oc a fan let me know
 
how is going to react with pwm i dont know this was before pwm time.

There, that is the idea, try, and find out !
Sacrificial test unit + experiment to accumulate DATA so as to KNOW.


It seems that NOBODY knows, that actually MODIFYING stuff is no longer seen as a good thing.

Inconceivable, silly.

I will recieve my first ever pwm fan today.

A 200mm Noctua, which I presume to be overengineered enough to withstand a bit of loving rudeness.

next up is a Sacrificial Unit to experiment on.

Thank you SvenBent for reboosting my faith in the [H]
 
Last edited:
Puzzled look.

1) Melt from overheating.

2) Improve cooling to regulate temperature so it doesn't melt

3) Figure out what needs cooling.

4) Figure out how to cool it/them.

5)

6)Profit.

Other than 5) what is it that is so obvious, that i am totally not seeing ?
Your impression that all overvolting can be solved with cooling is mistaken.

You're talking about a 41% voltage increase over nominal. What do you suppose will happen if, for example, you drop 1.69v into an Intel chip? Cooling won't save you from that.
 
Over-volting puts a tremendous amount of stress on the junctions and gates inside chips. If you go too high, electrons will punch through the very thin materials that would normally block them and destroy the component. Then it just becomes a cascade of failures until you see smoke.
 
Your impression that all overvolting can be solved with cooling is mistaken.

You're talking about a 41% voltage increase over nominal. What do you suppose will happen if, for example, you drop 1.69v into an Intel chip? Cooling won't save you from that.

I have been giving way more than 1.69 volt into an intel chip aka the intel pentium ran at 3.3olt.
 
There, that is the idea, try, and find out !
Sacrificial test unit + experiment to accumulate DATA so as to KNOW.


It seems that NOBODY knows, that actually MODIFYING stuff is no longer seen as a good thing.

Inconceivable, silly.

I will recieve my first ever pwm fan today.

A 200mm Noctua, which I presume to be overengineered enough to withstand a bit of loving rudeness.

next up is a Sacrificial Unit to experiment on.

Thank you SvenBent for reboosting my faith in the [H]


Looking forward to your report. Sometimes you just gotta try stuff out for the heck of it.
 
I have been giving way more than 1.69 volt into an intel chip aka the intel pentium ran at 3.3olt.

The original Pentium was also built on a completely different design process than today's chips. 3.3v on an i7 would fry it.

The 8088 through the 80486 all ran on 5v. Earlier chips had to have several voltages to work (e.g. i8080 needed +12v,+5v,and -5v)

A little understanding of how electronics work and their limitations will get you a lot further than juicing shit up and seeing what happens.
 
The original Pentium was also built on a completely different design process than today's chips. 3.3v on an i7 would fry it.

The 8088 through the 80486 all ran on 5v. Earlier chips had to have several voltages to work (e.g. i8080 needed +12v,+5v,and -5v)

A little understanding of how electronics work and their limitations will get you a lot further than juicing shit up and seeing what happens.

Kinda like how a fan was different than a CPU chip?
I think you completely missed the point here

VanGoghComplex used a chip behavior to determine the behavior of a fan.
Even though another chip would already be fine with that.
it was merely to show how idiotic that comparison was because as you state: difference in design.
 
You do realize that there are electronics (i.e. chips) inside a PWM fan, right?

Yes, but are they "powered" by the signal pins or by the 12v power input ?

Some sort of voltage divider thing that drops 12v to whatever the circuits use ?

I do not like decisions based on "i don't know" whan SOMEBODY does know.

'A little knowledge can be a dangerous thing, by stopping one from continuing on to find out more.'
 
Over-volting puts a tremendous amount of stress on the junctions and gates inside chips. If you go too high, electrons will punch through the very thin materials that would normally block them and destroy the component. Then it just becomes a cascade of failures until you see smoke.

Yes.

Yet fans are exposed to harsh conditions, must be built to a price point, and must be dependable.

I see nothing that would require super delicate finicky electronics, rather the opposite.

I would expect the minimum componant count of as wide tolerance bits that meet the price point that are as beefy ( thickest insulation ) as possible
 
Yes, but are they "powered" by the signal pins or by the 12v power input ?

Some sort of voltage divider thing that drops 12v to whatever the circuits use ?

I do not like decisions based on "i don't know" whan SOMEBODY does know.

'A little knowledge can be a dangerous thing, by stopping one from continuing on to find out more.'

signal pins by definition do not power anything.


Yes.

Yet fans are exposed to harsh conditions, must be built to a price point, and must be dependable.

I see nothing that would require super delicate finicky electronics, rather the opposite.

I would expect the minimum componant count of as wide tolerance bits that meet the price point that are as beefy ( thickest insulation ) as possible

There's nothing "finicky" about a 12v part not being able to withstand a 50% increase in voltage. Normal voltage tolerances in modern electronics are +/- 10%. That means you can probably safely go up to +13.2v without harming the device. Beyond that is anyone's guess as to which component gives out first. You could pop a resistor, burn the varnish insulation off of one of the coil windings, or it might be that nothing happens at all. Give it a shot and see.
 
VanGoghComplex used a chip behavior to determine the behavior of a fan.
Even though another chip would already be fine with.
it was merely to show how idiotic that comparison was because as you state: difference in design.

I did not. I used chip behavior as an example to make a point: electrical devices have a nominal operating voltage, and bad things tend to happen to them when you exceed that voltage by a significant amount. They are designed to withstand a specific potential, and usually not much more.
 
Back
Top