op amp controlling a mosfet

I doubt this simple circuit will do. The concept is to drive the gate negative (down to ground) with respect to the source any time the battery is more negative than the generator. So it gets real important what powers the opamp, and that it (both the supply pins and inputs) is protected from voltage spikes. Are you going to power the opamp at all times from the battery?

I expect to power the op-amp from the generator. I didn't indicate power pins, bypass caps, input resistors and such on the circuit diagram. I was more interested in the concept, the utility of the LF412 and and the like so I didn't flesh out the circuit like I guess I should have. So are you saying that conceptually it's ok, but that actual implementation is impractical? I get the impresson you don't like the idea of powering the op-amp from the generator.

Then you have a problem to solve. How do you keep the gate voltage at the battery positive (to keep the gate to source at zero) when the generator is off and the only voltage the opamp can output is zero?

That may be practical if you add an open collector or open drain output to the opamp so that it can pull down or let go, only. That way, when the opamp is unpowered, the gate to source resistor can pull the gate voltage up to the source voltage. you may want to lower the gate to source pull up resistor.

When the generator voltage falls below the battery voltage the op-amp should exhibit an active high output equal or near the generator voltage. As the generator voltage continues to fall, the op-amp's ouput should continue in an active high, tracking the generator voltage, until the generator voltage falls too low to drive the op-amp. Without the generator to power it, the op-amp's totem pole output will most likely be left open. Whereupon the gate-source resistor will take over in keeping the mosfet turned off. So I don't see a problem. And there's a relatively simple workaround if the op-amp output does go wack.

(Snipping some stuff)

As a side note, this blocking rectifier will be part of a voltage regulator. I built a previous version that used a regular diode, but I would like to make an improved version that doesn't require as much heat sinking. That's another reason I want to use the mosfet. The mosfet can do double duty -- not only as a blocking recitifier, but I can exploit the Rds as a sense resistor for the current limiting feature I'll be implementing. So in addition to eliminating much of the heat of a rectifier, I get to eliminate a separate sense resistor and its associated heat. My idea is to minimize the size of this voltage regulator and its heat sink. Then instead of a big chunky thing that needs its own separate mount, it will be a slender thing I can hide under the battery cover. I can use the second op-amp in the LF412 for the current limiting, with this circuit that I received in response to a question I posted to this newsgroup:

. + -----+--- Rsense ----+----> To Batt (+) . | | . Rc | . +------------, | . | __ | | . c\\| / +|--' | . NPN |---< | | . e/| \\_-|-----' . | . Rload . | . - -----+--------------------> To Batt (-)

with the Rsense being the drain-source resistance of the mosfet.

I think the + input has to connect directly to Rsense. Rc is just the source to gate resistor, so it is effectively on the battery side of Rsense (the channel resistance). You're getting there.

If you power the op-amp from the generator when the generator is stopped the non-inverting input is 12V above the positive rail.

and how will you get the gate high enough to turn the mosfet off ?

It may be easiest to just use a Shottky power diode.

Bye. Jasen

Why the opamp, wouldn't this be a *lot* simpler cheaper and have zero power drain when off?

o-+----+^+-----+-o | ||| | | === | | | ___ | .-. +-|___|-' | | | | | | '-' | | |/ +--| NPN |>

| o------+---------o (created by AACircuit v1.28 beta 10/06/04

I think the opamp is needed to detect the small drop across the channel resistance when the generator falls slightly below the battery voltage, so that the generator does not "motor".

I have seen only alternators with diodes to the output, after all you do want a DC-current to flow. So the whole idea is a bit say "strange"

I thought about it and the circuit I posted could have a problem when transitioning from the charging condition when the generator turns off. As the generator voltage drops, the op-amp output drops, dragging the mosfet gate down, turning it on and allowing the battery to discharge through the generator windings. I'd solve this pretty much as you suggested, by reversing the op-amp inputs and using a transistor on the op-amp output. Then the op-amp can turn the mosfet on ONLY when the generator voltage is higher than the battery voltage; under any other conditions, the mosfet's gate-source resistor will keep it turned off.

generator | ,------------------------+ | | | _ _|D '--|+\\ | STP80PF55 LF412| >--, ,--+-|| p-channel ,--|-/ | c/ | |_ mosfet | '--R--| | |S | e\\ 10K | | | | | | gnd | | '-------------------+----+ | | battery

I like the simplicity of your circuit, Ban.

o-+----+^+-----+-o | ||| | | === | | | ___ | .-. +-|___|-' | | | | | | '-' | | |/ +--| NPN |>

| o------+---------o

But What happens when you turn the motor off and the generator voltage starts to fall? The npn transistor will keep the mosfet turned on, the battery will discharge through the mosfet into the generator windings, which will keep the voltage up at the transistor base. From the looks of it, the transistor will never turn off.

There could be a glitch when the generator's turning off, and you would end up with battery current flowing through the windings of the generator. That would surely keep the voltage at the base of the npn high enough to keep it and the mosfet turned on. Maybe this would work, though:

o-+------+^+-----+-o | ||| | | === | |/ | | /_ 12v | | | zener| | | | ___ | .-. +-|___|-' | | | | | | '-' | | |/ +----| NPN |>

| o--------+---------o For the transistor to turn on, the voltage at the generator output has to be at least 12.7 volts. If there's a glitch while turning off and for some reason the voltage at the transistor's base doesn't fall far enough to turn it off, the 12 volt zener gives you a little insurance. The heavy drain on the battery as it discharges through the low impedance armature windings will probably drag the voltage below 12.7 volts, turning off the npn transistor and the mosfet. Adding a common diode in series with the 12 volt zener will bump the insurance to over 13 volts. I like it.

If the LF412 really can stand at least 14 volts positive (and I think it can) on its inputs when it is unpowered and you can protect both the supply and input pins from larger damaging transients, it just might work.

I assumed he was talking about a shunt wound DC generator, not an alternator.

That will work better, in fact I had already drawn a similar circuit, but you were faster. I have used another divider 1k/10k in the gate drive to increase the voltage drop with a weak input. But I think a Schottky diode will do the job as well with only 0.4V drop. Put a transil into the input line in case the switch shuts off very fast to avoid the spikes. A DC-generator will not be as reliable as a normal alternator and there will be the problem with overcharging/regulating the battery voltage as well.

will that turn off?

old vehicles had generators with a commutator and brushes.

Bye. Jasen