# Troubleshooting a fan speed controller UPDATE

#### MisterBill2

Joined Jan 23, 2018
8,799
The duration of whatever event occupies 2.6 divisions is 2.6mS. If that event is the full pulse period then the pulse frequency is 1000/2.6 = 385Hz.
And 385 Hz is not fast by any transistor standards. So the challenge now is creating an interface that allows the PWM device to correctly drive the IGBT that you have. And given the current draw of your motor you will need to be driving that transistor into saturation quite rapidly, and also you will need a heat sink. And you will need at least one resistor, which the value will take a bit of math to figure.
I am thinking that your IGBT needs to have the gate voltage positive relative to the emitter terminal in order to start conducting. So at this point it is useful to look at the data sheet for the device, which I have not done yet. Do you have the data sheet, or at least a link to it? At this point it is needed to be able to know how to switch it on and get it into saturation, because it will need to be switched fully on for driving the motor.

#### Tonyr1084

Joined Sep 24, 2015
5,993
IGBT Data Sheet says (below)
td(on) Turn-On Delay Time = 45 nS (typ)
tr Rise Time = 38 nS (typ) [conditions] TJ = 25°C
td(off) Turn-Off Delay Time = 220nS (typ) / 340nS (max) [conditions] IC = 42A, VCC = 960V
tf Fall Time = 160 nS (typ) 250 nS (max) [conditions] VGE = 15V, RG = 5.0W

How does this relate to the frequency of my PWM Dimmer (385Hz)?

#### Tonyr1084

Joined Sep 24, 2015
5,993
Do you have the data sheet
Yes. Post #1 has the link. What should I be looking for?

#### shortbus

Joined Sep 30, 2009
8,593
When connected the way drawn and with the PWM set to zero - I get a positive signal, voltage at 10 VDC according to the scale readings. As I turn up the PWM I get downward dropping square waves (not "Square" just square like). As I increase to approaching full ON, more and more of the waveform holds down to near zero (on the scale). At full ON the scope line falls to zero volts. I may be doing that all wrong, but I'm seeing that I get a square wave type change as I adjust the PWM. Since the PWM dimmer "Sink" IS to the negative side (ground as I might refer to it) I decided to use the IGBT in the same manor. Perhaps that's wrong. Maybe it should be on the high side of the motor.
Your dimmer seems to be working like it should. Think about it, when the PWM is set to "0" it is what when attached to a light be full on. As you turn the knob and lower the PWM the light would start to dim, because the time "on" is less. The dimmer when doing it's real job is putting the plus voltage from the lamp to ground or negative, it's sinking the voltage. A car or most that I'm familiar with the dash lights come on full bright and when you turn the dimmer it is dimming, making the light less.

Now to do it like you want for a motor, you would put the positive volts on one side of the motor, and the other side to the sink on the dimmer. To use an outside IGBT or mosfet you need some kind of pull up on the gate and then parallel the gate with the sink. But can't really see it working, since it(the dimmer) isn't made to do this. I don't know how old of an IGBT that is but the older ones suffered from something called "tail voltage", or something similar to that, don't remember the correct term. That meant though that to shut them off fast you needed to put a short negative voltage pulse on the gate. The newer ones have gotten around that somewhat but it is the reason they don't like fast switching like a mosfet.

Drain Peak Current of the 2SK972 is 100 Amps *PW ≤ 10 µS Duty Cycle ≤ 1 %. Takes longer than that for the motor to spin up to speed

You forget the mosfet inside is turning on and off, or doing dimming. Personally I wouldn't be afraid of using it. But then again I take things from a "practical" view, not a "theoretical" one.

#### Tonyr1084

Joined Sep 24, 2015
5,993
Personally I wouldn't be afraid of using it.
I lack the courage you have. And in 'theory', it looks like it should work. After all, you described the circuit I drew initially in post #1. So a pull up resistor IS in circuit, a 1KΩ resistor, but the iggy didn't turn on. First thought was that something was malfunctioning. But disassembling the circuit and testing the dimmer showed that it IS functioning. Also tested the iggy by tapping the gate to V+. It lit. When disconnected from the V+ the iggy shut off. That surprised me because I thought the gate had an internal capacitance, and without pulling that capacitance to zero volts the iggy would have stayed on. Guess I was wrong about that too.

You mentioned "Tail Voltage". Saw that in the data sheet. Acquired this device (and others like it) in 2005, not knowing how old the device was then. The company was getting rid of some stuff they weren't using. Possibly because they had better control devices. Still, older devices DO work. The dimmer is out of an 89 Toyota Celica. It works too. So taking a working "This" and connecting it to a working "That", one would assume, SHOULD produce a working "Thing". No?

Last edited:

#### shortbus

Joined Sep 30, 2009
8,593
Of course I made a mistake in editing it to say it was a source when it actually IS a "Sink". So a pull up resistor IS in circuit, a 1KΩ but the iggy didn't turn on.
That pull up is way, way to high. IGBT and mosfet gates are basically a capacitor, but the speed they turn on is based on how fast that cap gets charged. The problem That I see in using the dimmer with the outside IGBT is that when it is in the "off" position, the motor will be running full speed. This is due to the pullup turning the gate on and not getting PWMed. That's why I said in my last post I didn't think it would work like you want or think.

I lack the courage you have. And in 'theory', it looks like it should work.
Your thinking like I did when first working with mosfets. That the D-S amps was the amps through the thing you're switching. But that isn't what "amps D-S" means. It means the voltage times the RDSon. There is a withstand voltage, that's the voltage limit the mosfet will pass reliably with out getting destroyed, but the amp limit is that voltage times the RDSon. So I can't see that mosfet in the dimmer failing with driving the motor directly.

If wrong, what's the worse case? You go to a junk yard and get another one, or build a motor PWM circuit or even a less expensive solution is to get one from Ebay already built for ~\$5.

#### Tonyr1084

Joined Sep 24, 2015
5,993
OK. Something I don't quite understand: With the 1KΩ resistor @ 12 VDC, current is 12 mA (through the resistor). That's 144 watts (oops, MilliWatts). If replaced with a 470Ω resistor, current goes up to 26 mA, wattage = 306 W (oops, that's MilliWatts). Would need a half watt resistor. I MIGHT have one. Odds are not good though. But the puzzlement came when I bypassed the pull up resistor. The iggy still didn't turn on. Didn't try that for any extended period of time. Full voltage to ground - not good.

Last edited:

#### Tonyr1084

Joined Sep 24, 2015
5,993
NEW TEST:

Applied 12 volts through the iggy CE. Powered the G with 5V and the test lamp came on about half brightness. Does that tell us anything?

Last edited:

#### Chris65536

Joined Nov 11, 2019
270
I don't see why your PWM circuit won't work. Maybe you'd be better off replacing the pull up resistor with a second transistor? Are you still using that old power supply, with a significant voltage drop? How many watts is your 12v test lamp?

#### Tonyr1084

Joined Sep 24, 2015
5,993
@Chris65536 Using an ATX Computer power supply. The old school supply just isn't all that reliable, and it's easy to have an accident with it. Can easily over volt something. I like the isolation aspect of the Ferro-Resonant transformer and the adjustability of the Auto Transformer. Aside from that it doesn't get a whole lot of use. For quick and dirty tests it does enough at low currents.

The bulb is a standard auto tail light / brake light bulb. Using the tail light part, not the brake light part. I'm going to GUESS at the wattage, probably around 25 watts. Headlights (incandescent) are commonly 55 watts and this is no where that brightness.

Been thinking of a transistor. Probably a 2N2222 or a 2N3904.

 seems to draw 870 mA at 12 V. Suggests 10.4 watts.

Last edited:

#### MisterBill2

Joined Jan 23, 2018
8,799
OK. Something I don't quite understand: With the 1KΩ resistor @ 12 VDC, current is 12 mA (through the resistor). That's 144 watts. If replaced with a 470Ω resistor, current goes up to 26 mA, wattage = 306 W. Would need a half watt resistor. I MIGHT have one. Odds are not good though. But the puzzlement came when I bypassed the pull up resistor. The iggy still didn't turn on. Didn't try that for any extended period of time. Full voltage to ground - not good.
The error in your math is that milliamps and ohms gives you MILLIwatts, a thousand times less.

#### Tonyr1084

Joined Sep 24, 2015
5,993
OOPS! Milliwatts. 144 mW and 306 mW.

That happens to me a lot! Not good for a professional electronics inspector who should pay more attention to details like that.

#### MisterBill2

Joined Jan 23, 2018
8,799
I suggest an experiment to determine if your controller is a sinking device or a sourcing device. You will need to have that PWM controller and a 12 volt light bulb. First connect the light bulb between the +12 volt supply that is connected to the controller +12 volt terminal, and the PWM output terminal of the controller. Then connect the 12 volt negative terminal to the power supply negative. At this point I am guessing that the light can be controlled just like it is supposed to be controlled, torning the knob clockwise to make it brighter. Next, on to the second part of the experiment, remove the wire ti the light bulb from the positive terminal and connect it to the negative terminal, and see if it can be controlled brighter and dimmer.
The purpose of this experiment is to determine if the controller is a sinking device or a sourcing device. It is not likely to be both. When we know that it will be simple to create the correct circuit for you.
And the information that I wanted from the data sheet was the gate to emitter voltages for cutoff and for saturation.

#### BobaMosfet

Joined Jul 1, 2009
1,777
@Tonyr1084
Stop. Slow down. You need to understand some things, not just stab around in the dark. You're dealing with 12V at serious Current levels. The voltage won't hurt you, but the current can blow your hand off.

Rule #1- Put a fuse in this thing. Save your circuit, and maybe your life or limb.
Rule #2- Never EVER short positive to ground via any path- make sure when you remove a resistor, that it isn't there to prevent a short from the positive pole through any components to ground.
Rule #3- Study & learn how to read a datasheet- NO COMPONENT operates at the maximums. The TEST conditions stated in the datasheet are the ideal (and vendor recommended) state the component is meant to operate in. Any variation away from that could have risks you need to understand and mitigate.
Rule #4- If something does not work as expected- RE-EVALUATE YOUR EXPECTATION, as your expectation is likely faulty.

Now--- please restate in 1 or 2 simple sentences, what it is you are trying to do with the blower motor. Nothing else is important. Are you simply trying to power the blower motor effeciently? Are you aware of how the blower motor is designed- it may not be designed to handle PWM well.

#### Tonyr1084

Joined Sep 24, 2015
5,993
@MisterBill2 check out post #30. That's how I tested and proved out the "Sink" part of it.

@BobaMosfet "The goal": FIRST is to control the speed of the blower. It's a 12 volt DC blower out of a car. It originally was for blowing fresh air into the exhaust system. I plan on using it as a blower to add oxygen to a fire. By oxygen I mean ambient air, not O2. That's the ultimate goal. To have the blower blowing at full power is likely NOT going to be desired, so I plan on controlling the speed of the motor via PWM. Because the blower will draw tons of current at startup I'm taking it quite slow. When I tapped the sink to hot it was through a protected power supply, the ATX Computer supply. A short automatically cuts power. SECOND; I'm using a test lamp to prove out whether the circuit is working or not. So far it's not. I know about current and explosive forces. I've dinked around with the likes before. I've deliberately blown up a capacitor just to see what would happen. The capacitor was placed inside a copper pipe and the end cap had two small holes drilled. The leads of a 16 VDC cap protruded. For sake of not giving anyone else any ideas I'll forego the rest of the explanation and just say that I launched the body through the trees in my neighbors back yard at simi-ballistic speeds.

In my life I've managed to dodge my own stupidity. I've learned to take precautions. That's why I'm not yet connecting a blower to a car battery, which it is designed for. Baby steps. First to get the circuit working. Then to install a reasonable fuse (yet to be determined), then to finally wire it up to the battery and blower. Till then this is all small time stuff on my workbench. If I burn anything it's likely to be one of the small jumper wires I have, the kind with alligator clips on the end. Gauge is probably 20 or 22 gauge. And they DO get warm while running the lamp from the 12 V supply. Worse comes to worst I'll burn out one of those jumpers. Until the circuit works the way I want it to - slow and safe is the #1 factor.

Last edited:

#### shortbus

Joined Sep 30, 2009
8,593
At this point I am guessing that the light can be controlled just like it is supposed to be controlled, torning the knob clockwise to make it brighter.
Just the opposite happens in my cars and trucks. The dimmer brings the lights dimmer as it is turned clockwise.
And earlier he said the scope showed his output to get smaller when going clockwise.

#### Tonyr1084

Joined Sep 24, 2015
5,993
@shortbus I don't recall stating clockwise or counterclockwise. Maybe I did. Nevertheless, for the sake of clarification; CW = brightening CCW = dimming. However, because of the way the circuit (post #1) is set up - I would expect that (eventually) the motor will spin faster with CCW. I think.

 OK, I finished reviewing all my posts. I didn't see anywhere where I said clockwise or counterclockwise. Maybe I missed something. But I don't think I ever said that.

Last edited:

#### MisterBill2

Joined Jan 23, 2018
8,799
OK, Yes, I neglected to realize that it had been shown that the PWM dimmer control is a current sinking device. What that means is that something else will need to do the sourcing of the drive for the IGBT because it takes a positive voltage relative to the emitter to switch on. At least that is what would be needed if it were a regular device.
Now I realize that I have no clue as to if the IGBT is PNP or NPN, and that makes a great deal of difference. If it is a PNP device then the emitter can connect to the +12 volts and the collector can connect to the + terminal of the motor, with the motor negative connected to the supply negative.The gate would be connected to the controller sinking output.
If it is an NPN type of device things get a lot more complicated.

#### Tonyr1084

Joined Sep 24, 2015
5,993
Did you look at the data sheet? Says it's an N channel.

IRG4PSH71K

Looking like I'll have to add a transistor to drive the gate of the IGBT.