On 12/10/2015 3:05 AM, Richard Prosser wrote:
> ...
>
> Something else to watch -
> The inductance of the wires feeding the coil can still produce
> significant spikes if the diode is placed directly across the coil -
> especially if there is any length involved. It may pay to put diodes
> at both ends, or at least an RC snubber at the MOSFET end.
I've wondered about this, as it's all being done on a bench, with short=20
wires.  I may have to try with some long wires to see how things change.

> The application appears to have 2 conflicting requirements - a slow
> current decay to take advantage of a reduced hold-in current following
> the higher pull-in pulse, and a fast current decay to turn the thing
> off quickly when that's what's required. There may be a happy medium
> position, or an additional control line and load dump circuit may need
> to be added to turn the coil off quickly.
Not really... there's no requirement to turn off quickly, as the on=20
period can be shorter to compensate for a longer decay time.
>
> In addition, there is a desire to monitor the coil current whenever it
> is supposed to be on - to ensure it energises as quickly as possible,
> and then is turned down to a level at which it will stay energised,
> but at reduced power. So the current sense element is ideally in the
> PWM flywheel loop. Maybe a magnetic sensor or a high-side resistor and
> differential amplifier?
I've been considering a high-side sense resistor, and drew up a circuit=20
with it, but can't easily test yet, as I don't have a PCB laid out for=20
it, so it will take some time to cobble together.  But I want to try the=20
other options first, of using a pre-determined PWM duty-cycle and see=20
how it holds up to changing supply voltages, etc.

> It would be useful to know: The unenergised and energised inductance
> of the coil, the DC resistance of the coil, The pull-in current, the
> drop-out current. Then we could set up something in spice or something
> to get into the ballpark. These could be derived from your origonal
> scope plots if that's all thats available but I haven't tried (yet).
I've been using/assuming ~5mH (from research) unenergized, 2 ohms=20
resistance (measured).  Can't remember the pull-in current, but I can=20
measure from the oscilloscope plots I have.  FWIW I know that just=20
switching it on with 1.3A (using a 6.8-ohm resistor in series and 12V=20
power) will energize the solenoid... but after ~4ms, with no pressure.

Cheers,
-Neil.

--=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
.