Okay, some updates from the weekend...  I am running hardcoded peak=20
times, and hardcoded PWM duty-cycles.  I'm using a 1N5822 schottky=20
directly across the solenoid here.  Both the MOSFET and schottky are=20
running very cool with this arrangement.  With some code tweaking, I=20
have a nice clean 4.1A peak and 1.1A hold...
http://tinyurl.com/j7p75x6
http://tinyurl.com/hd2hjfz

I added in inline 0.1-ohm resistor right at this solenoid to be able to=20
see the ripple during the hold phase, but the drive circuit does not use=20
this for feedback this time.  I am considering adding a current-sense=20
amp at that same point, and using that for feedback, so I can=20
dynamically control the current.  All because I expect that in a=20
practical circuit, variations in the +12V supply will vary the peak and=20
hold currents.

There is other code and other tests I want to try, but first, there are=20
some issues.  Specifically, I'm seeing what looks like ripple, but=20
there's a pattern to it.  I moved the scope probe wires, and that ripple=20
changed.  So I thought it must be noise being picked up in the probe=20
wires.  I did add a 5.6-ohm resistor to the gate, but that did not=20
change the "ringing".  See here (notice the sine-wave pattern of the=20
spikes)...
http://tinyurl.com/gmd9s2b

Here is Vgs at the top, and Is (voltage across 0.05-ohm resistor between=20
source and ground).  Again, the sine-wave pattern of the ripple is very=20
evident, and it looks like it's around 75 Hz.  I'm not aware of anything=20
else running at 75Hz on this board, and it's being powered off a +12V=20
car battery.
http://tinyurl.com/jatk5oq

However, I did fry 2 TC4427A's.  The first I thought was during a period=20
when I did not have the +12V connection to the flyback diodes, but when=20
I fried the second one, everything was hooked up and running fine.  I=20
can't help but think that this is related to the possible ringing.  So=20
that brings up the snubbers.  From what I know of them, snubbers should=20
be calculated based on the ringing frequency... is this so?  I'll go=20
research.

Cheers.


--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.