Hi all,

I have a 12vdc proportional coil which is controlled using a logic level N-channel enhancement MOSFET (FQP30N06L). Please see attached spec sheet for both and my circuit diagram.

In order to obtain proportionality I vary the current between 100mA and 1200mA using PWM control of the Mosfet Gate. At 100% PWM duty cycle the coil draws about 4A. I would like to limit the maximum current the coil is able to draw @ 1.2amps. I know I can do this via PWM, but then I wouldn't be driving the MOSFET gate as hard as it requires and as such heat dissipation is a problem and reliability is shocking.

So, I am thinking of using something like an LM317, LM338, or LM2930T configured as a current controller to achieve this. Given my voltage and current requirements can anyone recommend a suitable IC with accompanying circuitry / configuration? I want to keep heat to a minimum is possible.

Thanks in advance.
Jim

 

Just to add, would the LM350 be suitable? Even though it specifies max output current as 3amp, would I be able to limit this at 1.2amps?

 

You can just put a resistor in series with your inductor.

Why is it that if you reduce current by adjusting the PWM that you would not be driving the transistor as hard? What is that 10k pot doing in the circuit!?

 

10

You can just put a resistor in series with your inductor.

Why is it that if you reduce current by adjusting the PWM that you would not be driving the transistor as hard? What is that 10k pot doing in the circuit!?

The 10k pot is to adjust the PWM duty cycle. So using ohms law with a voltage of 14vdc (alternator) and a current of 1.2amps I would require a 12 ohm resistor?

 

Pulse-width modulating the gate drive reduces the average current through the inductor, but not the peak current. And how is anything surviving without a catch diode across the inductor? Also, if the FET is getting decent gate drive, like 10 V no matter what the PWM period is, the heat in the FET should be relatively constant across the range of PWM duty cycle percentages. For the heating issue, verify that the gate drive meets the FETs requirements for full enhancement.

ak

 

Ok thanks AnalogKid, I understand.

I'm going to put a 12 ohm resistor inline anyway. But I know what your saying regarding the PWM. I suppose during one frequency cycle as the PWM is either fully ON @ 5v and OFF at 0v the FET is getting the full 5v to drive into saturation. The FQP30N06L MOSFET I'm using is logic level so has no problem being driven straight from the board with 5v PWM rds(on) @ 0.035 ohms.

I do have a MOSFET driver chip UCC27424P which I could use, but is there any need for this as I'm already using a logic level FET? Or would it just help the circuit run more efficiently? I am using a schottky 1N5822 free-wheeling diode across the coil now, but that circuit diagram was a little dated.

My original problem was that the FET was getting ridiculously hot and eventually failed resulting in a short open between source and drain, regardless of gate voltage. That was before I was using the catch diode.

Thanks,
Jim

 

My original problem was that the FET was getting ridiculously hot and eventually failed resulting in a short open between source and drain, regardless of gate voltage. That was before I was using the catch diode.

Well, that's your problem, not some need for a resistor.

 

Surely not having a diode would cause excessive heating of the FET though? The only job of the diode would be to stop voltage spikes back to the FET when turning the coil off? Correct me if I'm wrong.

Would there be any advantage driving my logic level FET using the UCC27424P chip?

 

Just a few relevant facts from the coil datasheet:

Maximum Current 1150 mA
Nominal Coil Resistance at 122ºF (50ºC) Stabilized 9.4 ±8% ohms
Nominal Coil Resistance at 68ºF (20ºC) Cold 6.4 ±8% ohms

 

At 100% PWM duty cycle the coil draws about 4A.

How are you getting 4A through that coil? It's cold resistance is 6.4Ω according to its spec, so with the rated 12V supply the current shouldn't exceed 2A.

 

That's the thing, I have no idea. With full 100% duty cycle the coil draws an awful lot more than 2A. That's why I contemplating using a current limiter.

I haven't actually tried the updated circuitry with the schottky diode on the machine yet so I cant be 100% sure the diode will fix the issue of the FET popping.

When the manufactures specify a maximum current, I'm guessing more current could be applied on a slightly higher voltage and so the current would need to be limited @ 1.15A? The alternator on the tractor kicks out 14v when running. Therefore the coil should draw 2.18A until it warms up. This is why I need to limit the current.

 

Surely not having a diode would cause excessive heating of the FET though? The only job of the diode would be to stop voltage spikes back to the FET when turning the coil off? Correct me if I'm wrong.
...............

Surely it will.
The job of the diode here is more than just suppressing spikes.
It's to allow the inductive current to flow between PWM pulses and give a more or less continuous current through the inductor.
Without that, all the inductive energy is being dissipated in the MOSFET which either will cause it to overheat, or get zapped, take your pick.

 

A question I don't see yet is what are you trying to achieve with the solenoid?
Max.

 

That makes sense Crutschow. Essentially the diode acts as a buffer by allowing the current to freewheel and therefore maintain the electromagnetic field between PWM pulses.

The coil is used to control a hydraulic proportional valve. With increased current the solenoid armature adjusts the pressure in the system.

I have heard of an alternative configuration or is some instances the two are combined. this is the use of a Zener diode between the drain and source of the FET. What advantages would this have over the freewheel diode?

 

................
I have heard of an alternative configuration or is some instances the two are combined. this is the use of a Zener diode between the drain and source of the FET. What advantages would this have over the freewheel diode?

The Zener will absorb the inductive energy and quickly stop the current, which is not what you want.
You want a plain diode to maintain the current.
For a slight efficiency improvement you could use a Schottky diode, which has a lower forward voltage drop than a standard junction diode.

 

Hi Jim,

There are some details to consider.

The coil has a transient suppressor built it to which limits reverse pulses to approximately 68V peak. Add to this the 14V power source, and the max the MOSFET will see on the drain is 82 Volts peak. Your MOSFET is only a 60V Vds rating so it will take a beating until it blows.

Adding a catch diode across the coil will fix this problem, but would cause another problem. Since this is an hydraulic proportional valve it needs to be very responsive to changes in average current so as to control oil flow quickly. The catch diode would keep current flowing and therefore keep the valve activated longer than intended. That is why the coil manufacturer built in a transient suppressor instead which allows the coil to de-energize properly.

Sorry Crutschow I have to disagree with some of your post.

Don't use a diode across the coil, instead use a MOSFET with a max Vds of at least 100 Volts. The coils responsiveness will not be affected because the current sustained after turn-off will be greatly reduced.

If the hydraulics responsiveness is not an issue then forget what I said.

Ifixit

 

Thanks guys, much appreciated and very thorough explanations.

I'll see if I can find a MOSFET with a higher Vds rating, thanks Ifixit. For the time being, the hydraulic response time is not a major issue, this is something I can try, see how it responds and adjust if needs be. My concern at the minute is preventing the FET from going west as I need the reliability.

Going back to an issue I mentioned earlier, would wiring in the UCC27424P FET driver give me any advantage even though the FET is a logic level unit?

Thanks,
Jim

 

................
Going back to an issue I mentioned earlier, would wiring in the UCC27424P FET driver give me any advantage even though the FET is a logic level unit?..........

Depends upon your PWM operating frequency and the MOSFET gate capacitance.

 

MOSFET gate capacitance needs to be charged and discharged during switching cycle.

 

Thanks for the help chaps!

Fitting the Schottky diode across the coil made a massive difference. FET is now so cool I can hold it. Tested the circuitry on the scope and the difference is immense. After each pulse of the PWM the voltage was spiking up to 50vdc, thus causing the heating and eventually failure.

Before diode:


With diode:


Problem solved.
Thanks again