uC based glow driver ideas

[Ralph Weaver]

==Measuring current w/o inline resistor:

I'm attempting to do my own glow driver (running a glow plug from 12V). It's basically a switching regulator run by a uC. I don't really need to know exact current, but would like to be able to shut down in case of over current (shorted outputs).

Typically the voltage across a current sense resistor is amplified and read using an A/D - no big deal, but I'd like to be able to get rid of the resistor and amplifier.

If I know Vin, Vout and the PWM %duty cycle then I think I can detect over current (example Vin = 12, Vout = 0.5, PWM = 100%). I'm thinking that I can get a reasonably good idea what the actual current is.

==Eliminating the typical switching power supply inductor

The glow plug does not care if it's run on DC, AC etc. as long as the average voltage (really average power) is correct. Typical switch mode power supply's need the inductor to give a clean DC output. As long as I filter the feedback I shouldn't need the inductor.

I'm a software guy, so what would be the best way to filter the feedback? I can do averaging in the SW, but am pretty sure I'll need some sort of RC filter as well.


Thoughts???

Thanks.

[AndyKunz]

You can skip the RC filter if you make the measurements at the same point of the drive waveform. A simple avg would probably suffice. Use power-of-2 samples so the division is just right shifts.

You may me making yourself a radio station that you carry along on the airplane, especially if you don't have an inductor to round off those PWM edges.

Andy

[Ralph Weaver]

As far as I know the currently available glow drivers all use a simple PWM with no inductor. One that I have uses a uC with no A/D. If the input is 0 then it's <2.0V and assumed to be 1 cell else it's assumed to be 4 cell and a PWM with no feedback is used. I don't think the RF has been a major problem.

Most commercial SMPS's use a high frequency PWM to minimize inductor size. Without that concern, a lower PWM frequency can be used. At a lower frequency and at only 2-3A I think a slight slope on the FET gate would not cause too much switching loss and help with the rf. As long as the frequency is out of the audible range, I don't think it has to be very high.