Servo PWM signal to drive a FET in Linear?

[gkamysz]

Anyone have a link for a circuit to drive a FET in linear mode from an RC servo signal?

Greg

[alexcmag]

Linear mode????

FETs are not designed to be linear. How about making a ESC with a linear PWM curve?

http://www.designsoft.com.au/ahome/rc/PIC-ESC/

[JohnMuchow]

Linear mode????

FETs are not designed to be linear. How about making a ESC with a linear PWM curve?

http://www.designsoft.com.au/ahome/rc/PIC-ESC/Well, linear FETs are. :)
Their response isn't linear (to the gate voltage) but they're designed to be biased in their linear region. But, MOSFETs like that are typically optimized for audio amp applications and are often rated for lower current than we would need. And very expensive too.

Many switching MOSFETs are rated for linear operation though and a large number of electronic loads, including the West Mountain Radio CBA and my CC-400, use linearly-biased MOSFETs very successfully.

But, Greg, you won't be able to drive a MOSFET directly with a servo signal. You'll need a microprocessor in-between to take the 1mS-2mS ontime of the servo signal and turn it into a 0%-100% ontime PWM signal. I haven't seen any circuits for this but the hardware is very simple. It's the software you'll need to hunt down. :)

At a quick glance, alexcmag's circuit looks like it would work for you though. Just take the output of the PIC and filter it to create the adjustable DC voltage you'd need to drive the MOSFET's gate. You might need an op-amp in between in order to buffer the PIC though.

Greg, what are you building? :)

John

[gkamysz]

Well, I want to have a variable electromagnet for something I'm doing. Linear mode will eliminate switching noise. As would a higher switching frequency. I looked at AstroFlight's little coreless motor control with 95kHz switching. It's current capacity is OK, and it's only 20USD. But, will it work OK with high inductance coil?

I have a little knowledge of electronics but not enough to design a circuit for this or program a processor.

I tried an old analog ESC to test my idea and that worked ok other than the switching noise. I'm thinking about trying AVR programming. I might just have take the time to learn it.

Greg

[JohnMuchow]

I don't know if Astro's ESC will work or not for your application....sorry.

Does this device have to respond to a servo signal? I only ask because the circuit to bias the FET linearly and have it hold the desired current level is very simple and can be adjusted with a potentiometer or anything with a voltage output (in a certain range, of course). I have a couple of schematics I could send and here's an example that's easy to build: http://www.webx.dk/oz2cpu/radios/dc-load.htm

John

[gkamysz]

I've had that DC load on my project list for years. Yes, I know it's a very simple thing without servo PWM, but I need it. I'll try it with a regular ESC first. The noise might not be much of an issue.

Greg

[shanghai_fool]

But, will it work OK with high inductance coil?


Well, a motor is about as inductive as you can get. Your main concern may not be when you turn it on, but when you turn it off. ESC's have reverse diode protection but they have their limits. With a large electromagnet, there could be considerable energy "stored" in the magnetic field and when it collapses, it will produce a large reverse voltage spike. It may be enough to damage the ESC.

Be careful

[gkamysz]

Would a Schottky diode across the ESC outputs protect it?

[alexcmag]

Yes, but only in brushed DC without reverse, ok?

Brushless and reverse can reverse voltage in motor, so a Schottky diode there will short-circuit the output.

[gkamysz]

I've used Schottky diodes on brush ESC that didn't have brake FETs to recirculate current when the drive FET was off.

My electromagnet won't need reverse.

[Acetronics]

Hi, Gkamysz

The solution looks simple, as the ~ 50 Hz PWM runs at a constant freq.

1) use a RC Lo pass filter or Sallen and Key filter to convert PWM to a DC voltage ( High impedance output ...)
2) use an OPA ( say a Rail to rail one ...) as a voltage to current converter ... that will drive your FET .

As I think ( ? )The magnetic field is a linear function of the current applied...

Note here ... your Mosfet gate will need more than your esc supply voltage , if N mos used ... lol. Sooooo ... also think to a Logic Level one. Or simply a PNP transistor ! :rolleyes:

Alain

[vintage1]

Hi, Gkamysz

The solution looks simple, as the ~ 50 Hz PWM runs at a constant freq.


Don't bet on it.

Some transmitters do, others don't.

[JohnMuchow]

Hi, Gkamysz

The solution looks simple, as the ~ 50 Hz PWM runs at a constant freq.

1) use a RC Lo pass filter or Sallen and Key filter to convert PWM to a DC voltage ( High impedance output ...)
2) use an OPA ( say a Rail to rail one ...) as a voltage to current converter ... that will drive your FET .Won't that result in just a very tiny change in the voltage level as the signal changes from 1mS on to 2mS on (out of 20mS)? Not nearly enough to swing a MOSFET from off to fully on once filtered IMO. And you'd probably have to offset the voltage swing anyway for various MOSFETs and their varying threshold voltages too.

I think, to make it work, that the servo PPM signal still needs to be converted to a 0% to 100% duty-cycle PWM signal, probably via a micro, to drive the MOSFET's gate from 0V to v+ (full off to full on), once filtered.

John

[vintage1]

Classic way is to use a 'pulse stretcher'

You charge capacitor via a diode and a resistor, and let it discharge via a much larger resistor. Run that through a low pass filter and bang it into a voltage amplifier with a FET output stage..or a damn big power transistor. No need for a FET if doing it all linear and not switching ..