I'm looking for some suggestions

[Roddy Wayne]

Hi All,
I'm using PicBasic Pro, a PIC16F877, a STE250NS10 Power MOSFET and a AD536A RMS to DC converter in this project.
The programs I've written work well almost all of the time but, there are some exceptions.

I need to run a Hydrogen Generator at between 17 and 20 Amps in order for it to be most effective and efficient. Because of greatly varying alternator output voltages on some vehicles, this is difficult to realize. Newer Chevrolet Pickup's have voltage swings greater than 3 volts, from 12.0 to 15.1 volts. This 3 volt range can cause the Generator's current to vary way to much! The AD536A's output (mentioned above) is stable and is being fed into an onboard A2D (8 bit). I'm using the HPWM command and some Potassium Hydroxide to control the Generator's current. While monitoring the generator's current draw, if it's lower than 17 amps, I need to turn the HPWM on harder. If the generator's current draw is greater than 20 amps, I need to turn the HPWM on less.

I'm looking for some suggestions on how to accomplish this because I certainly don't know how!

Any and all suggestions are very welcome and will be greatly appreciated!

[Archangel]

Hello Roddy Wayne,
It is a matter of measuring the voltage drop across a known resistance, the more voltage lost, the higher the current being used by the load. I would use the ADC to measure the voltage on both sides and subtract the lower from the higher. Build a lookup table with PWM duty cycle calibrations.

[Jerson]

Since you have a RMS-DC converter, I suspect you already know the output current via this dc reading. If this is true, all it takes is a little proportional control to correct the output in relation to the input.

The PWM will be proportional to the error in the actual output.

[Charles Linquis]

If you need to measure current, use an Allegro Hall-effect sensor. Isolated, accurate, high-output.

[Roddy Wayne]

Since you have a RMS-DC converter, I suspect you already know the output current via this dc reading. If this is true, all it takes is a little proportional control to correct the output in relation to the input.

The PWM will be proportional to the error in the actual output.

Yes, you are right in assuming that I have detected this current draw correctly. So, I guess I'll need to build a lookup table with the PWM percentages in it.

Thanks again Jerson!