MOSFETs unexpectantly hot!!!

buz11 overheating

I have three MOSFETs (BUZ11) turning on and then off in a row and then repeating the sequence. The MOSFETs are driving the coils of a HDD motor. After some time, I smelt that familiar hot smell and the MOSFETs were literally on fire. They were unbearable to touch and I'm only using 9v from a PP3 sized battery. The motor used to work at 5 400 RPM, but is bearly spinning more than 200 RPM (I don't have any means of measuring it). What is causing them to heat up so much, is it the 330Ω resistor connected to the gate, have I wired them incorrectly or what? This is the first time I'm actually using MOSFETs.

Sputnik :idea:

have you checked the circuit connections?

check the biasing of the MOSFETs

also maybe the MOSFETs are damaged

With the inductive load of the motor windings you could be either getting inductive voltage spikes that exceed the breakdown of the transistors or there is a dV/dt that is turning on the transistor. You need to examine the circuit with an oscilloscope.

Given my financial position, I can't afford an oscilloscope (I really want one), although it may be possible for me to make use of my school's oscilloscope. So, perhaps if a put a diode in series with each of the windings, that will help. I don't think any of them are damaged, yet. I haven't used them excessively (< a minute total).

Sputnik :idea:

The Rds(on) is inverse proportional with the applied gate voltage, therefore a low gate voltage will cause higher power disipated around the device. Also, make sure that you drive the mosfet with a push-pull output, not with an open-collector one. Driving it with an open-collector and a resistor to Vcc will cause a slow rising edge for the gate signal, depending on the resistor value, thus increasing the power disipation in the transition period. A schematic will be usefull....

/pisoiu

...The Rds(on) is inverse proportional with the applied gate voltage...

Click to expand...

...are U sure? Cause the gate voltage is definitely over Vg (9V)...
...in my oppinion the big problem it's "the inexistent" turn-off circuit...

maybe it's very useful to read this article: h**p://focus.ti.com/lit/ml/slup169/slup169.pdf

hope this helps,
//a

you should be driving the gates of the mosfet with a few inverters in parallel. this is the cheapest way to get strong drive in both the turn-on and turn-off direction.

9v seems ok for the value of the gate drive, i think it's probably that your circuit switches slow, and you are dissipating a lot of power in the rise/fall times..

aNdreiBuuu said:

...The Rds(on) is inverse proportional with the applied gate voltage...

Click to expand...

...are U sure? Cause the gate voltage is definitely over Vg (9V)...

hope this helps,
//a

Click to expand...

Very. A higher gate voltage, over the treshold voltage means a lower rds(on).
https://www.fairchildsemi.com/ds/BU/BUZ11.pdf , figure 7, even if a 0,02ohm span between Vgs=5 and Vgs=9 is not a good reason for such overheating.
From the initial description, I understand that the coils are powered from 9V, not necesarily the command circuit.
I have the sensation that we are taking wrong assumptions, a schematic will be usefull.

/pisoiu

HDD motors that I'm familiar with are synchronous motors that do not readily self start with a high frequency drive. They must be started at a low frequency, and then have the drive frequency slowly increase. If the motor is running at a speed that is below the synchronous frequency, the back emf from the motor coils will be low, and the coil currents will approach V/R, which is much higher than they would be at synchronous speed.

Okay, okay! I haven't drawn up a schematic yet, but I will describe it as best as I can. The microcontroller (PIC16F84) is programmed to turn each MOSFET on in sequence. The 5v from the controller goes through a 330Ω resistor to the Gate, another 10kΩ resistor from the Gate to ground. The motor does not start up by itself (I need to spin it by hand at the moment). The 9v is connected to Drain. The output from the Source goes through 1N5822 diodes (I hoped to stop back EMF going through MOSFETs), then to each of the motors coils. From the motor common connection to ground.

I think MOSFETs are overkill for such an application, I'll find some transistors instead.

Sputnik :idea:

It looks like you are using the source follower configuration. This works for BJT emitter followers because of the 0.7 Vbe drop leaves much of the remainder of the voltage for the motor windings. With MOSFET there is a several volt Vgs required for turn on. This leaves much less of the voltage for the motor windings. The rest of the voltage drops across the transistor, that is the difference between the output (few volts) and the 9 V.

You will have to switch to the pull down mode. Put the motor winding common point to 9 V and the transistors from the other end of the winding to ground. That way they act as switches on and off with very little drop across them.

This may require you to change the pulsing sequency to the transistors to get the required rotation direction.

What is this "push-pull output" that I read in an earlier reply? How does it work and how would I use it?

Thanking you,
Sputnik :idea:

Something like this.
(Disregard the 555 - I picked this picture from an old post of mine.)

Sputnik said:

What is this "push-pull output" that I read in an earlier reply? How does it work and how would I use it?

Thanking you,
Sputnik :idea:

Click to expand...

I did not know what you use for driving the mosfet. The push pull configuration is better than the output collector configuration, because it reduces the power disipated by the mosfet. In your case, this is not important, because the Rds(on) is quite low, since it is a low voltage mosfet. The problem is however important when working with mosfets which have a higher Rds(on), especially those working at higher voltages. When you turn off the mosfet in an open collector driving configuration, the transistor used for driving quickly discharge the gate of the mosfet and turn it off. But when you want to turn on the mosfet, the only element which is charging the gate is the resistor from the collector of the driver. This resistor will not charge the gate instanteniously, it depends on the gate capacitance. The mosfet will turn on when the voltage is above the Vgs(th), but in this point the Rds(on) is at higher values than at a higher Vgs, causing the device to disipate power. The goal is to reduce the time spent in this area, and to charge the gate as quick as possible. You can do this by reducing the value of the resistor in the output collector configuration ( the driving element will dissipate more power in on state), or you can use push-pull configuration.

/pisoiu

I do wish to use MOSFETs in a DC-AC converter, or won't it be possible cause I'll be using a sine wave and therefore the MOSFET will be inbetween the off and the on state.

Sputnik :idea:

I think you can drive the transformer with mosfets and square waves and place a filter at the output to obtain sine wave. Check this UPS reference design, it will give you some answers.
**broken link removed**

/pisoiu

With what voltage must I drive the Gate with? I have used the above mentioned push-pull configuration and they are still getting hot! I tried using 2N3904s (before trying the BUZ11s) that passed their maximum current (100mA) and didn't get nearly as hot. Have I connected it wrong, I have the Drain connected to +V and the Source to each motor windings with 1N5822 diode in series. I'm getting really frustrated with this now!

Sputnik :idea:

Sputnik,
the source must be common for working in switching mode, and in the drain put the load (motor)... (your source load act as negative current reaction... and it's not good for switching mode operation!)

...for what's your series dioda?

...put the schematic on the forum; we need that for helping you,
//a

PS: for sine wave you need to implement a PWM controller (the PWM will be generate based on sine table)

why you don't use logic level gate Mosfet (ie 2sk2232) ?

I believe that the MOSFET is not turned on completely. The 330 Ω resistor u use, seems to keep the mosfet in a high on resistance than in a fully on.
Use a 10Ω resistor for the gate and check the results.
If no improvement is noticed, then u need a driving cct as indicated in a previous post. A logic level mosfet is also recomended sicne it is turned on with lower gate voltage levels.

Another idea is to use a transistor with low on saturation voltage.