Home     Forum     Projects     Experiments     Circuits     Theory     BLOG     PIC Tutorials     Time for Science

### Author Topic: 12V current amplifier from standard PC Fan controller  (Read 7154 times)

0 Members and 1 Guest are viewing this topic.

#### sgking2

• Guest
##### 12V current amplifier from standard PC Fan controller
« on: January 23, 2012, 08:27:01 AM »
Hello. Can you please help with what should be a quite simple circuit?

I have a really nice fan controller, touchscreen, the NZXT Sentry LXE http://www.nzxt.com/new/products/fan_control/sentry_lxe but it is underpowered.
It has 5 channels that are 8W max each. I need each chanel to control (3) fans that are .4A each. So, each channel would be controlling (3) 12V x 0.4A or 1.2A / 14.4Watts total. Obviously this is more than their 8W max. I would like to use this controller as a signal source instead and drive a current amplifier that can provide the unobstructed power directly from the PSU, on a 4-pin molex. Can you please provide the best way possible to use this controller as a voltage source to drive an independent amplifier with enough current to handle 12V 1.2A total with protection from kickback from the fans starting/stopping?

If it is going to dissipate heat, then perhaps you can tell me how much and if it requires a heat sink to remain functional/stable.

Thanks!

#### kam

• Hero Member
• Posts: 1849
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #1 on: January 23, 2012, 16:22:09 PM »
how are the fans controlled? if with PWM, then it dissipates no heat, and a simple mosfet (eg IRf540 - my favorite) can handle huge loads. A flyback 1N4007 resistor is enough .

If the control is by changing the voltage, then you need a big transistor such as a tip series. there is no need for flyback diode there. To calculate how much power it will dissipate, suppose that minimum voltage across the fans is 5 volts, this means 7 volts voltage drop, multiply this by the current. Eg:

Pdis = 7 x 1.2 = 8.4

sounds small, but without a heatsink it will toast within a minute.

#### sgking2

• Guest
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #2 on: January 23, 2012, 17:04:43 PM »
I should have specified, yes voltage controlled. Ultimately I need to drive 9 of them and would like to have the flexibility to do it from 1,2, or 3 channels. Is there a limitation on using the same channel output just spliced into several transistor inputs? Even though the channel can sink up to 8 or 10 watts should i bother with a high impedance buffer of some sort?

What if the control voltage was turned down to 2v or even 0 ( off) , to stop the fans when they were not needed, what cuurent would the transistor see then because the fan stalled or stopped? Is the circuit you are prposing called a voltage follower or emitter follower and how does this differ (for better or worse) than a regulator with variable input?

To plan the circuit i have to look at the worst case scenario in voltage drop/ max current terms. That will tell me how many fans (1-9) i can control per transistor. I want it robust so it has long life. How do you limit the input minimum to 5v and stop the fans? Does the transister see no load at that point with fans stopped or max current and how to determine current if the 1.2A was based on 3 fans at .4A each but no fans are running (stalled to off)?

Thanks for you fast reply!

#### kam

• Hero Member
• Posts: 1849
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #3 on: January 23, 2012, 18:16:44 PM »
controlling such high loads with voltage drop (power dissipation) is not a good idea. the circuit is indeed an emitter follower and you need only one transistor and some resistors. The problem is that the transistor will have to dissipate a lot of heat. 8Watts of heat are too much already. T rippled (for 3 fans) means 24 watts. This is totally inefficient, and you will need certainly an extra fan to cool down the heatsink of the transistors.

I refer to 5 volts minimum, because almost all fans do not operate bellow 6 volts and they usually stall. The math to calculate the power dissipation for any case is simple. First you need to determine how many volts you want to "block" across your emitter follower. Suppose that the power supply is 12 volts, and you want the fan to operate at 7 volts. This means that you want to drop 12-7 = 5 volts. Therefore, across the fan will be 7 volts and across your transistor (collector - emitter junction) will be 5 volts.

Now regarding the power that the transistor will dissipate. The formula for the power is P = V x I. To be accurate, you need to measure how much current your fans draw at 7 volts. Let's assume that each fan draws 100mA @ 7 volts (it will be different from the nominal current shown on the fan box). So, the transistor will have to dissipate:

P = V x I = 5 x 0.1 = 0.5 Watts = 500 mWatt

This is for one fan. So if you want to connect multiple fans in parallel, then you add the power for each fan. For 3 fans, it will be 500+500+500 = 1500mW = 1.5 Watts.

It seems that you want to make something BIG, therefore i  totally suggest that you make a voltage to PWM circuit and drive the fans with PWM instead, controlled by the voltage output of your gadget. Here is an example:

Voltage Controlled PWM Generator

This circuit operates from 0 to 5 volts, but that is not a big problem for you since you can convert the 0-12 output of your gadget to 0-5 volts, and drive the 5volts PWM of the circuit to a mosfet to control many fans with no dissipation at all. I've tried 6 fans in parallel with the IRF520, and the mosfet did not even got warm (no heatsink needed).

#### sgking2

• Guest
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #4 on: January 23, 2012, 19:26:49 PM »
I understand this is not ideal but I need to make it work. Since each channel can power up to 10 Watts, what if I use 2 channels to power two fans directly (4.8W) leaving the remaining 5.2W on each channel to act as the transistor drive (should be plenty). Now instead of 9 fans, I only need to drive 7 - not a big decrease but it is something. Then I I branch each channel into a group of 4 and 3 transistor circuits. For example the output of one channel from this controller drives one 4.8W fan directly and also acts as the input control for 4 individual transistor emitter followers - each transistor driving only (1) 12V .4A 4.8W fan. This is powering the triple radiator fan array of a PC liquid cooling box (there are three 360mm radiators, each with three 120mm fans). I don't think heat is a problem because I can use some sort of thermal paste or pad to mount the transistors directly to my copper radiator bodies to soak it up or I can arrange them in a nice array and just add a water block heat sink to the whole lot and dissipate it in my water loop.

The fans I am using I picked out very specifically because of their extremely high reliability, and low noise. This fan does not come as PWM and the unit is already built. I was originally looking at http://www.sidewindercomputers.com/tbalancerbigng.html as a fan amplifier but figured I could build something similar w/o all the bells and whistles that was more geared to this project.

There has to be a way.
PS. What happens in a voltage controlled fan circuit when the voltage drops to 4V and the fan stalls and is essentially off? Is the transistor seeing the 8V drop @ 400ma or is it seeing the full 12V drop because the fan is not consuming any? I don't quite understand what happens at this end of the line, from stall (5V) to zero volts on the input. If the fan is not moving how can 400ma be used anymore as the load parameter.

Thanks for your help on this project! Your insight is invaluable.

#### kam

• Hero Member
• Posts: 1849
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #5 on: January 24, 2012, 15:13:55 PM »
Better you can make one driver for each fan. A TIP series transistor for each fan connected as emitter follower is a good solution. You can connect as many drivers like this as you want on each channel, since the power required for each one is relatively small. This is my proposal.
You may also connect two transistors or more on each driver, since the TIP can dissipate a lot of power, but then you will need to take care all the heat before the transistor is toasted. For me, max 2 fans is ok per transistor.

Quote
There has to be a way.
Certainly, the one describe above is a way...

Quote
What happens in a voltage controlled fan circuit when the voltage drops to 4V and the fan stalls and is essentially off? Is the transistor seeing the 8V drop @ 400ma or is it seeing the full 12V drop because the fan is not consuming any? I don't quite understand what happens at this end of the line, from stall (5V) to zero volts on the input. If the fan is not moving how can 400ma be used anymore as the load parameter.

There is no way you can foresee what will happen, since the fans have different drivers on them (BLDC fans have built-in drivers). Best case scenario (and most common) is that the controller detects that the fan stalls and does not power the coils
Worst case scenario is the opposite. If this happens, then the current will be many times more than the nominal current, and this is what causes the fan to burn - this is also why the previous discussed precaution is taken from the controller manufacturers

But a stalled fan is something that you don't wanna happen. That is why you should not power the fan with less than 7 volts. If you want to stop the fan, simply turn the power off.

#### sgking2

• Guest
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #6 on: January 24, 2012, 16:24:32 PM »
Awesome, thanks. I can do one Q per fan. Do you have a sketch of what one circuit would be with whatever resistors or diodes might be needed and i can just duplicate 9x?

I think i am stuck driving these with only 2channels, the other 3 are used elsewhere. Any problem splitting 1 channel to 4 or 5 of the transistor inputs? How much current does the gate or base use of a tip?

Now that we are settled on 1 trans per fan, is there an ideal transistor type, maybe low dropout or darlington. Cost is not a factor.

And finally, is there a clever way to force the output based on drive voltage? For example, when the drive voltage on base is <= 6v then kill or disconnect power from psu to emitter/collector so there is NO heat, and when it is above >= 11v to bypass the trans and just apply full psu 12v rail without losing diode junction voltage drop? Maybe a relay or scr or other solid state way to go full on or full off on fan side and only use the tip for the area in between?

Thanks for all your help, i am eager to get your thoughts on these final questions so i can buy the parts today and get busy

#### kam

• Hero Member
• Posts: 1849
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #7 on: January 25, 2012, 08:12:39 AM »
Do you have a sketch of what one circuit would be with whatever resistors or diodes might be needed and i can just duplicate 9x?
LED Fade-In Fade-Out Dimmer at the bottom of the page, check the "Need more current?". T1 is your power transistor. The line that comes from Q1 is the output of your gadget. You will need to add a resistor in between. If you use a TIP142 series transistor (darlington pair), use a 1KOhm resistor for each transistor base.

Any problem splitting 1 channel to 4 or 5 of the transistor inputs? How much current does the gate or base use of a tip?
Each transistor with 2.2 KOhm base resistor will draw max 12/2200 = 5.4 mA so no there is no problem driving the transistor. Each one, if it is a darlington pair, will be able to provide 5.4 mA x 1000 = 5.4 Amperes.........

is there an ideal transistor type, maybe low dropout or darlington. Cost is not a factor.
TIP140, TIP145,TIP141, TIP146,TIP142, TIP147 choose one of those. There is no kind of "low dropout" in transistors. A typical 0.7 volts drop is considered for all silicon transistors connected as emitter followers. In other words, if max output of your gadget is 12V, your fan will get max 12-0.7 = 11.3.

And finally, is there a clever way to force the output based on drive voltage? For example, when the drive voltage on base is <= 6v then kill or disconnect power from psu to emitter/collector so there is NO heat, and when it is above >= 11v to bypass the trans and just apply full psu 12v rail without losing diode junction voltage drop? Maybe a relay or scr or other solid state way to go full on or full off on fan side and only use the tip for the area in between?
You can use 2 voltage comparators, or you can make a schmitt trigger, but that requires that you do some more research.

#### sgking2

• Guest
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #8 on: January 25, 2012, 09:06:02 AM »
You are awesome kam!

Just to clarify.... for the Darlington TIP142, use 1K or 2.2K base resistor - you mentioned both.
and the TIPs you mentioned, they are half NPN and half PNP, do I want the NPN (TIP142) ?

The "Add more power" circuit you referred to shows current limiting resistors, do I need any resistors in series with my fans or just straight to fan and GND?

« Last Edit: January 25, 2012, 09:20:58 AM by sgking2 »

#### kam

• Hero Member
• Posts: 1849
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #9 on: January 25, 2012, 15:29:20 PM »
1k and 2.2k are ok both, better use 2.2k.

The TIP140, TIP141 and TIP142 are NPN so you can use for example TIP 142.

You do not need limiting resistors because the fan itself has enough resistance to regulate the current. We use limiting resistors for leds, because if you exceed the forward voltage, the current climbs rapidly to toasting levels.

#### sgking2

• Guest
##### Re: 12V current amplifier from standard PC Fan controller
« Reply #10 on: February 07, 2012, 16:52:23 PM »
Better you can make one driver for each fan. A TIP series transistor for each fan connected as emitter follower is a good solution. You can connect as many drivers like this as you want on each channel, since the power required for each one is relatively small. This is my proposal.
You may also connect two transistors or more on each driver, since the TIP can dissipate a lot of power, but then you will need to take care all the heat before the transistor is toasted. For me, max 2 fans is ok per transistor.

^^^^ With regard to the above, Pd for each fan to use 1 TIP142 for each fan.
While experimenting I hooked up 1, then 2, then 3 and finally 4 fans to just one TIP142G (ON semiconductor), and it finally started getting warm. I had all fans down to their lowest voltage before stalling, about 4.2 volts. That 8 volt drop x 4 fans at 400mA each,  should have been about 12.8 watts total. I was monitoring the temp on the body, collector and the heatsink. The body and collector were cooler at 34C and the heatsink was the warmest at 39C. Is this possible or typical? Could it have been that efficient at dissipating 13W of heat? The room ambient was 27C with no active air circulation (no fans to cool the heatsink). The heatsink was a beefy RA-T2X-25E from Ohmite. It is a 25 gram black anodized heatsink with a thermal resistance of 4.8C/W (natural convection). It had almost 9,000 mm2 surface area in a package about 1" tall, 1" deep and about 1.7" long.

Does it seem possible that the TIP with that heatsink was able to effectively dissipate 12.8W through natural convection with no additional climb in temp above 39C ? If so, is it safe to use 4 fans or should I drop to maybe 3 per? At 3 fans it was barely 3C above ambient. I do not know if I am reading this right, but the fact that the heat sink was hotter than the transistor means it is working well to draw it away?