synchronous rectification/current limiting

I wanted a synchronous recitifier for a voltage regulator in an old-
style dc generator on an antique truck (to block the battery from
discharging into the generator).
This is what I came up with. I built it on a breadboard with junkbox
parts:
For the battery under charge used a junk motorcycle battery, and I
used a 12 volt battery in series with a 6 volt battery to serve as a
simulated generator.


generator
|
| 1N4148
+--->|--- ,--,
| \ / |
,---+--||-+ D V___/ |
| | ||-> | |
| 15K ||-+ S | |
| | | 1N4148 | |
| '-----+-->|-- | |
| | \ | |
| | >| | |
| ___ |--+ |
| _ /| | 27K
| ___ | | |
| _ '----+ |
| ___ | |
| _ 220K |
| | | |
| gnd gnd |
| |
\| |
|------+-------------------'
<| |
| 100K
| |
+--------'
|
gnd

The diodes are there only because the transistors would zener without
them.
Any pair of matched pnp transistors ought to work.
The circuit functions pretty well on the bench. The mosfet turns
fully on even at very low forward current, and there's no reverse
leakage when I remove the "generator" and put a load there, by way of
simulating the generator's armature windings when the truck engine is
not running.
Of course I will have to use a different mosfet(s), like several
STP80P in parallel.
I'm pretty confident this will work at high current, but if there's
something I should look out for I'd like to know.
My next question is about current sensing. I'd like to use the Rds of
the
mosfet bank as a sense resistor, but a mosfet channel's temperature
coefficient of resistance is even worse than copper, and the area
under the hood undergoes extreme temperature variations, deltaT 100C
or more maybe, from northern winter mornings to hot summer stop-n-go
driving, especially if the voltage regulator is bolted onto something
that gets hot -- like the generator.
So I'm thinking of connecting one end of a small mosfet like a BSS84
or ZVP3310 to the power mosfet drain and its other end to a constant
current sink, and compare the voltage of the small mosfet against the
voltage of the power mosfet. I'll use the same type of comparator
circuit for this as for the synchronous rectifier; the diode-connected
transistor can even do double duty in both comparator circuits.
diagram later... getting tired of typing.

 

(snip)

I think your source and drain are reversed.

 

Scratch that. And I definitely should have known better.

 

How about the "intrinsic diode?" As I understand it, power mosfets
have a reverse biased diode intrinsic to the device because of the
silicon construction.

To do what you want may require a pair of series Mosfets connected
source to source, or an additional silicon diode to prevent the
reverse voltage from being applied to the battery under charge.