Fan splitter problem

[Mtt]

I have more fans than the number of plugs available on my motherboard and I have been using Y splitters for a couple of my fans. I only recently realized that using Y splitter and have both fans plugged in actually decreases the rotation speed of both fans. That means if i plug both fans to the splitter, both fans run slower than when i plug the fan directly onto the motherboard. can anyone explain why this is happening? I thought fan splitter is like power strip. Each slot gets the same voltage.

 

[Billb2]

Are these 3 or 4 wire fans plugged into a fan header that controls their speed?

 

[guskline]

I'm confused. I thought if the splitter still provided enough power to each fan to run at max so long as the rated amperage of the fan controller channel was not exceeded.

 

[SimsReaper]

Have you ensured that your motherboard headers have enough of an amp/ wattage rating to run more than one fan at full power? In a 3 pin fan connector, using a splitter is like putting the fans in a parallel circuit, which does mean that the fans all receive the same voltage, which should be the full 12V. However, you still need to be able to supply the fans with the required current, and even though your fans may be receiving a full 12VDC, without enough current to them, they would still be "starved" for power. Full solution would be to run the fans off of a molex connector straight from your power supply. Current will no longer be an issue and you can run as many fans as you want. I run 10 off of one molex connector.

 

[Mantrid-Drone]

It could be a header power issue but a pair of fans uses so little I'd be surprised if that was the case.

I'm thinking maybe something to do with PWM (Pulse Width Modulation) where the whole voltage is continually applied but it is pulsed on and off at different rates to control the voltage the fan receives and thus its speed.

If you're using the Akasa two way splitter you might have noticed one cable is only 3 pin ie. is not PWM controlled. What that means is that the second fan, PWM or not, is being controlled by the speed of the fan attached to the 4 pin PWM one and, of course, that fan does have to be PWM capable for it to controlled by the PWM.

I was just wondering myself what happens with a splitter because you have two fans drawing power. I'm new to this sort of thing so I'm just guessing but the PWM is, surely, not recognising it as two separate fans. It supplies the same power as if it was just one unit and so, together they'd both run slower.

This is a temperature control issue so does that matter? If the temperature rises the PWM will increase the RPM of the controller fan and thus the second fan too. The pair must be acting more efficiently than just a single fan so, I would have thought, that the advantage is neither needs to spin up as fast as a single fan would to do the same cooling job ie. equal heat reduction for less fan noise.

If this is the case then a solution isn't actually necessary is it? The two fans are in fact doing the cooling job more efficiently and that's why they're both running slower.

However apart from that SlimReaper suggested, maybe an adjustment could be made to the Slope PWM settings in the MB BIOS for the header used with the fan splitter.

Just an idea.

 

[coffeejunkee]

I'm confused. I thought if the splitter still provided enough power to each fan to run at max so long as the rated amperage of the fan controller channel was not exceeded.

Me too. Assuming a standard 3-pin y-splitter my experience is both fans run at same rpm as without a splitter. If this is some kind of pwm splitter it might explain things.

Have you ensured that your motherboard headers have enough of an amp/ wattage rating to run more than one fan at full power? In a 3 pin fan connector, using a splitter is like putting the fans in a parallel circuit, which does mean that the fans all receive the same voltage, which should be the full 12V. However, you still need to be able to supply the fans with the required current, and even though your fans may be receiving a full 12VDC, without enough current to them, they would still be "starved" for power. Full solution would be to run the fans off of a molex connector straight from your power supply. Current will no longer be an issue and you can run as many fans as you want. I run 10 off of one molex connector.

Usually headers supply 1A * 12V = 12W which should be plenty for 2 normal fans. Need some Delta 6000rpm fans to go past the limit. Obviously with molex you give up the ability to adjust rpm.

I would first install Speedfan to check if your fanspeeds are in fact abnormal.

 

[Mtt]

I will provide more details of my experiment. Maybe the problem is the black and short cable. The purpose of that cable is to reduce the speed of both fans. When i unplug one fan, the other fan spins faster!

 

[aigomorla]

looks like there maybe a speed reducer (resistor) in that black cable...

 

[Mtt]

looks like there maybe a speed reducer (resistor) in that black cable...

Yes. But when i unplug one fan, why does the other one spin faster?

 

[aigomorla]

Yes. But when i unplug one fan, why does the other one spin faster?

i guess that resistor scales to amperage draw..

So the more amps it pulls though the resistor, the larger the effect.
(speculation)

So by putting 2 fans though the resistor, ur pulling more amps, hence the resistor is cutting more power.

 

[guskline]

looks like there maybe a speed reducer (resistor) in that black cable...

That's it! Don't use a resistor.

 

[BonzaiDuck]

Every time this sort of topic pops up, I'm inclined to suggest using a device like the Swiftech 8W-PWM-SPL-ST:

http://www.sidewindercomputers.com/sw8waypwmsp8.html

But there's a slight drawback that you need to use PWM fans exclusively, although there are now plenty of such choices available. The Akasa splitter may work the same way.

In past years I had wired fans in parallel while monitoring the speeds of each tach wire on a separate motherboard plug. It always seemed that the balance between the fans would vary, or was otherwise not quite even. And the average RPM per fan was slightly depressed.

With PWM fans and the Swiftech, there's no such problem, except that you can only monitor a single fan connected to the device, while controlling all connected devices no matter the 12V amperage spec, size or RPM.

 

[coffeejunkee]

That's it! Don't use a resistor.

Well, if you want you can use a resistor but you need 2 and put them after the splitter. But it would be better to just use the motherboards fancontrol or Speedfan to regulate speeds.

 

[BonzaiDuck]

Well, if you want you can use a resistor but you need 2 and put them after the splitter. But it would be better to just use the motherboards fancontrol or Speedfan to regulate speeds.

It's nice to look at these problems, and here or there the discussion provides guidance for someone with a similar problem.

I think at one time, I had three 140mm side-panel or bottom-panel fans wired in parallel, with the tach-wire from one connected to a motherboard header. Truth be told, I don't want to do such things anymore.

If you need a lot of fans inside the case, there are ways to do it which don't require any parallel wiring. You can buy any of a number of fan-controllers for three-pin fans, or use the Swiftech/Akasa devices/splitters. I just look carefully at what the motherboard provides to drive my "fan strategy." If more are needed, or you want to reduce the draw on the motherboard, there are ways to do that . . . too . . .

 

[Mantrid-Drone]

Sorry to open this up again but does anyone know, for sure, if when using the Akasa two-way fan splitter cable which has one 4-pin and one 3-pin connector does the fan on the 3-pin connector actually need to be PWM compliant?

I can't find any authorative information on this.

If yes, please explain why and what would happen if you used a non-PWM fan.

 

[BonzaiDuck]

Sorry to open this up again but does anyone know, for sure, if when using the Akasa two-way fan splitter cable which has one 4-pin and one 3-pin connector does the fan on the 3-pin connector actually need to be PWM compliant?

I can't find any authorative information on this.

If yes, please explain why and what would happen if you used a non-PWM fan.

We'll either need to do some quick searches, or someone else may chime in. I cannot be absolutely certain of what I remember about this, having immersed myself in it just long enough to get PWM fans to work on such a splitter within the last four months.

The splitter I use cannot be too different from the Akasa splitter, although you can review the two columns of specs to be sure. I use the Swiftech 8W-PWM-SPL-ST model. "ST" stands for "SATA power-plug," and there was an earlier model that used Molex.

I THOUGHT I read somewhere that 3-pin fans -- non-PWM fans -- would run uncontrolled at their top-end as auxiliary devices on the splitter. [On my splitter, only one designated PWM device can be monitored; the rest are controlled according to duty-cycle percentage applied regardless of the device's amperage and top-end RPM.]

I'm wondering if I didn't get this impression about the 3-pin devices from some comment in a user forum and its "take-with-a-grain-of-salt" credibility.

ON the other hand, it now occurs to me that the way you achieve PWM fan-control requires something in a PWM fan that responds to a PWM signal. And 3-pin fans have no such component. But I am also a bit confused or befuddled that one of the plugs on the Akasa splitter has only three pins. I think the splitters are just wiring-- there aren't any transistors, diodes, resistors as far as I can tell. So if the motherboard PWM pin sends a signal to the splitter, it is either replicated or distributed to all the four-pin plugs to provide PWM control with PWM fan or pump devices. So the Akasa splitter would likely only power a 3-pin device on the 3-pin plug at its top-end speed.

the other thing of which I'm not sure and too lazy at the moment to find at resellers offering the Akasa splitter: I'm not sure whether the Akasa splitter provides actual POWER to all the devices from the motherboard PWM pin-set, or powers them directly from the PSU as does the Swiftech splitter. Again, I only remember an impression that the two splitters do essentially the same thing and likely the same way.

 

[Aikouka]

Sorry to open this up again but does anyone know, for sure, if when using the Akasa two-way fan splitter cable which has one 4-pin and one 3-pin connector does the fan on the 3-pin connector actually need to be PWM compliant?

Properly-wired fan splitters simply do one thing: they remove the tach signal wire for all but one of the fans. This is true for 3-pin splitters and 4-pin splitters. So, yes, it should work fine; however, you cannot take advantage of PWM control simply because there's one PWM fan on the line. PWM control requires that fourth wire as that's the signal from the controller telling the end-device (doesn't have to be a fan as some water-cooling pumps have PWM support too) how fast to operate. 4-pin splitters do include the fourth wire for all fans.

EDIT:

I'm pretty sure I have an Akasa PWM fan splitter lying around somewhere, but I'm not sure where. Also, in regard to fan splitters, I've found that Akasa and BitFenix make some nice sleeved units, but I'd probably steer clear of Modright. The cables are great... when they work. They seem to use a cheap way to split the single cable into multiple (there's actually a section of electrical tape on it) and this 4-way splitter won't even work unless I compress the area where the cables split. I like them because they're black wires without requiring thick sleeving, but the last thing you want are your fans just randomly shutting off because of poor cabling.

 

[Mantrid-Drone]

I think I'm more confused now.

The Akasa fan splitter definitely has one 4-pin and one 3-pin connector:-

http://www.akasa.co.uk/update.php?tpl=product/product.detail.tpl&no=181&type=Cables&type_sub=Fan%20Cable%20Adapters&model=AK-CBFA04-15

Aikouka seems to be saying that even with a PWM compliant fan on the 3-pin connector it won't respond as a PWM fan because there is no 4th pin. It seemed likely to me that you couldn't have two fans on one header sending or receiving separate PWM signals. But then he mentioned there are all 4-pin splitters available......that is confusing.

The advice I've just had on another forum was that if a non-PWM fan is used on the 3-pin only plug it would run at 100% all the time even if a PWM fan is connected to the 4-pin plug.

However the advice given suggested if instead a PWM compliant fan is attached via the 3-pin plug as long as there is another PWM compliant fan on the 4-pin plug it will sync to the speed of that fan despite there being no data connection.

The theory stated being that when the MB senses a PWM fan is attached via the 4th pin it supplies the PWM controlled voltage via the header to both fans. Both fans therefore run at the same speed despite one of them not having the 4th pin data connection.

But I'm still not convinced about these explanations, particularly as regards what happens with the fan on the 3-pin plug.

I guess testing is the only way to answer the question conclusively unless someone here actually has a definitive answer.

 

[Aikouka]

I think you're misunderstanding the PWM signaling. The PWM pin is from the PWM controller (motherboard) to the end device. The controller is supposed to use the signal from the tachometer to understand what exactly its dealing with. That's why fan controller software (such as ASUS's Fan Xpert) will test the fans first to understand their RPM curve.

So, if we look at the pin-out:

1. 12V
2. Ground
3. RPM
4. PWM

The idea is that we only want one fan reporting its speed. If all the fans do, it becomes a jumbled mess... trust me on that. One time, I used a splitter that didn't remove the third pin from the other fans, and my fan controller's RPM kept jumping around. Anyway, so what will happen is that one fan will report its speed, and then the controller will tell all fans how fast to operate. To my knowledge, it doesn't actually adjust the voltage or amperage. Since the power isn't adjusted, any fan that ignores or simply isn't designed for PWM operation (i.e. a 3-pin fan) will operate at normal speeds.

The advice I've just had on another forum was that if a non-PWM fan is used on the 3-pin only plug it would run at 100% all the time even if a PWM fan is connected to the 4-pin plug.

That's exactly what I'm saying will happen.

 

[coffeejunkee]

That Akasa splitter doesn't really have a 3-pin connector. Both are 4-pin, except one header has the rpm pin missing, for the reason Aikouka already stated: it would mess up the rpm readout if the pin was there.

To properly use that splitter you need to use 2 4-pin pwm fans. You can connect a 3-pin fan too but it will run at 100% speed all the time.

 

[Mantrid-Drone]

The above explanations are what I thought and understood was the case and I was really just looking for confirmation. But I was confused when the fact there are all 4-pin splitter out there was mentioned.

BTW I wasn't saying the voltage was adjusted but as I understand PWM it is either full on or full off. The PWM system pulses these very short off phases at different cyclical rates thus controlling the fan speed with reference to temperature sensors and/or BIOS settings. If this is wrong blame the sources I've read for any mistake or incorrect terminolgy.

 

[Ayah]

i guess that resistor scales to amperage draw..

So the more amps it pulls though the resistor, the larger the effect.
(speculation)

So by putting 2 fans though the resistor, ur pulling more amps, hence the resistor is cutting more power.

technically, temperature. power = resistance*(current^2).
indirectly, current causes the resistor to heat up.
crap resistors can have TCRs (temp coeff. of resistance) of %s/Kelvin.

 

[BonzaiDuck]

http://www.sidewindercomputers.com/sw8waypwmsps.html

Now that I see the Akasa splitters, I can see we're talking about two "slightly different" items. The Akasa splitter apparently powers the PWM units from the same motherboard fan-plug.

The Swiftech unit powers all units from the PSU, but passes on the PWM signal from the motherboard fan header to control up to eight PWM devices proportionally -- by duty cycle -- regardless of the power draw from each individual device.