On 23 May 2006 08:31:29 -0700, "
wrote:

This was after I tweaked the PID to get rid of the oscillations but
still had some average position errors...

I guess with a digital system it would be easier to optimize. Mine was
all analog... and only one op amp, so the settings interacted :-)

Hi Dan,

If anything, going digital has all the prospects of thrashing the
antenna into shreds. The oscillations you speak of are eminently
suited for an analog solution - you simply have to understand the
nature of resonance, coupling (over, critical, and under), and attack.
If you don't, translating them into binary code will be impossible.

For instance, your average position errors were probably due to
under/overshoot and a too heavily dampened system with too much gain.
No computer program is going to improve that.

73's
Richard Clark, KB7QHC

Richard,
Agreed on the suitability of an analog solution... I certainly made
some steps toward critically damped operation. My circuit design was
bad, though.

I didn't have a very analyzable problem; the rotator rotates one of
those fiberglass surplus military masts with respect to another, so I
get a lot of stick-slip, especially if there's a breeze (there are 12
feet of mast above this thing).

This is not to say a suitable set of P, I and D couldn't be found, and
it would have been pretty slick if I'd set up right for empirical loop
tuning, but I hadn't done that. I got pretty close by soldering
components in and out.

I would have had a better time with a real multiple op amp circuit with
a knob for each of P, I and D.

I think you hit the nail on the head regarding too heavily dampened
with too much gain in the antenna rotator case. I just couldn't really
tune the thing.


In my case, I just decided to throw up my hands and go to the simplest
method so I could get on the air with a rotating antenna. I have a
quite successful analog servo going every day with my remote antenna
tuner (www.n3ox.net/projects/servo). That one is ever-so-slightly
underdamped, proportional only, and works nicely. There's a little bit
of hysteresis and it certainly could be improved, but I stopped at the
good-enough-for-daily-operation level.

I guess what I meant by "going digital" was that depending on whether
you're more comfortable with programming than a soldering iron (not
me!) you might think about having a simple motor drive circuit and
implement the admittedly complex PID loop in software/firmware.

You're absolutely right that you can't do this if you don't know how to
do the analog...

73,
Dan
N3OX

On 23 May 2006 09:38:07 -0700, "
wrote:

You're absolutely right that you can't do this if you don't know how to
do the analog...

Hi Dan, All,

The single loudest voice on these matters is Robert Pease at National
Semiconductor. For streaming courses/lectures on this very topic,
visit:
http://www.national.com/nationaltv/index.html

I've been following Bob's no-nonsense contributions for more than 25
years now.

73's
Richard Clark, KB7QHC

wrote:

Richard,
Agreed on the suitability of an analog solution... I certainly made
some steps toward critically damped operation. My circuit design was
bad, though.

I didn't have a very analyzable problem; the rotator rotates one of
those fiberglass surplus military masts with respect to another, so I
get a lot of stick-slip, especially if there's a breeze (there are 12
feet of mast above this thing).

This is not to say a suitable set of P, I and D couldn't be found, and
it would have been pretty slick if I'd set up right for empirical loop
tuning, but I hadn't done that. I got pretty close by soldering
components in and out.

I would have had a better time with a real multiple op amp circuit with
a knob for each of P, I and D.

I think you hit the nail on the head regarding too heavily dampened
with too much gain in the antenna rotator case. I just couldn't really
tune the thing.


In my case, I just decided to throw up my hands and go to the simplest
method so I could get on the air with a rotating antenna. I have a
quite successful analog servo going every day with my remote antenna
tuner (www.n3ox.net/projects/servo). That one is ever-so-slightly
underdamped, proportional only, and works nicely. There's a little bit
of hysteresis and it certainly could be improved, but I stopped at the
good-enough-for-daily-operation level.

I guess what I meant by "going digital" was that depending on whether
you're more comfortable with programming than a soldering iron (not
me!) you might think about having a simple motor drive circuit and
implement the admittedly complex PID loop in software/firmware.

You're absolutely right that you can't do this if you don't know how to
do the analog...

73,
Dan
N3OX

Written with the implicit assumption that you'd want to do it in
software, but quite applicable to analog solutions:
http://www.wescottdesign.com/article.../friction.html

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html