My home-brew focuser.

In ,
Steve Taylor typed:
Jo wrote:

The greatest benefit of electric
focusing is that the focus may be adjusted without mechanical
disturbance to the OTA and the consequent settling time. For
amateurs there is no requirement to achieve focus in a given
number of milliseconds.

Surely the greatest benefit here is that you can do it remotely ?

I didn't write that, it was Graham.



Take a look at http://www.epanorama.net for other design ideas, but
what you want is a comparator with a little bit of hysteresis, to
look at the signal from your pot and decide whether the pot is
"forwards" or "backwards" - then use that and a bit of logic to set
the legs of the bridge in forward or reverse. Its odds on there are H
bridge drives on the VCR. The comparator can be an LM311 or just a
bog-standard opamp.

OK, thanks for the advice. I also have a box full of 2N3055's so could
easily fabricate an H-bridge if I need to.

Incidentally, PWM is usually very good at eliminating the low speed
"stiction" you are experiencing - what is the chopping frequency you
are using ? If it is too high, the driver will have a struggle to
fight the motor inductance.


It is in the low audio range, I can hear the motor "singing". Good point...I
will do some tests. Any suggestions on optimum chopping frequencies?


A peek with an oscilloscope would see
whether you were actually getting the current in the motor when you
think you are.

Yes. One of the few items of test gear I do have is a scope.

Thanks,

Jo

Jo wrote:
In ,
Graham W typed:

Then it isn't geared right and/or the motor is running at too high
a speed for the application.

I assume you are talking about Steve's hypothetical focuser. In my
real focuser the speed is fine (now).

Yes, correct.

On my controller, which is a PWM back EMF f/b type, I have
arranged a 'speed-up' button which I hold in at the start of a
session to rev the motor up and down its mechanical range
to ensure smooth control at its 'normal' speed.

I would really appreciate some more info on your controller,
especially the back EMF feedback loop. One awkwardness with my
current system is that the current required to start the motor is
much higher than the running current. I have to turn up the speed
control, then wind it back quickly. Fixing that one problem might be
enough to make it work fine for my needs.

OK. I'll look out my notes for this (if I can remember where they are
since I last saw them in 1998!). This 'poor' starting was one reason
that I put the 'fast' switch in my controller. Once the motor has had a
little burst of higher energy to warm up its brushes and agitate the
lubricat film in the gearbox etc., it is good for the rest of the session.

Here is another feature I would like to add:

It would be great if my rotary speed control pot could give
bidirectional control...at mid point it should be zero speed. At
either side of mid point the motor should speed up in either
direction. I know about H-bridges..it is the rest that I could use
some help with. This would give more intuitive feel to the controller
than my current up/down switching method.

The thing about H-bridge's outputs is that you don't want the active
side to be other than fully 'on' or fully 'off' since anything in between
is going to generate a fair amount of heat and demand much more
thermal consideration and potentially larger devices! This does
rather limit what can be done for speed control but nevertheless
I would imagine that PWM could be employed in a SMPS to supply
the output stages.

I might have some brainfade here but what are you currently using for
a speed controller that has this undesirable characteristic?

--
Graham W http://www.gcw.org.uk/ PGM-FI page updated, Graphics Tutorial
WIMBORNE http://www.wessex-astro-society.freeserve.co.uk/ Wessex
Dorset UK Astro Society's Web pages, Info, Meeting Dates, Sites & Maps
Change 'news' to 'sewn' in my Reply address to avoid my spam filter.

In ,
Graham W typed:
OK. I'll look out my notes for this (if I can remember where they are
since I last saw them in 1998!). This 'poor' starting was one reason
that I put the 'fast' switch in my controller. Once the motor has had
a little burst of higher energy to warm up its brushes and agitate the
lubricat film in the gearbox etc., it is good for the rest of the
session.


Thanks.

The effect I'm seeing is that (motor current to start rotating)( slowest
running speed) hence I start off with a speed surge every time..not good for
fine focussing. Steve made some suggestions regarding this which I shall
check out.


The thing about H-bridge's outputs is that you don't want the active
side to be other than fully 'on' or fully 'off' since anything in
between is going to generate a fair amount of heat and demand much
more
thermal consideration and potentially larger devices! This does
rather limit what can be done for speed control but nevertheless
I would imagine that PWM could be employed in a SMPS to supply
the output stages.


Understood.

The http://www.epanorama.net/ site apparently offers a bidirectional
controller with H-bridge and speed control circuit but when I tried to go
there I got a "not found".


I might have some brainfade here but what are you currently using for
a speed controller that has this undesirable characteristic?


The PWM circuit I'm using is on the website at :
http://www.nu-ware.com/Focuser/

Jo

Jo wrote:


OK, thanks for the advice. I also have a box full of 2N3055's so could
easily fabricate an H-bridge if I need to.


I was thinking more BC108- ZTX300 ! Any idea of the running current you
are seeing ?

It is in the low audio range, I can hear the motor "singing". Good
point...I
will do some tests. Any suggestions on optimum chopping frequencies?


That depends on the waveform, but if its "singing" audibly, I have my
doubts that the frequency is too low. Anyhow, sniff the CURRENT waveform
with a low value shunt in the ground lead - the waveform should
definitely have a nice flat bit at the top.

Did you fit any suppression, like a 100 Ohm/100n is series across the
brushes ? You'll probably see some horrible RF scruff on the current
waveform, when the transistor is off and the circuit is ringing through
the diode.

You could try using a higher supply voltage too: if you watch the
current waveform with a scope, just watch for the current going off
scale as the motor armature saturates. Then you have too much supply...

Yes. One of the few items of test gear I do have is a scope.


You have THE piece of test gear.
I moderate an embedded microcontroller forum, and you wouldn't believe
the number of folks over the years there who have been struggling for
MONTHS to fix "software" problems that are actually down to hardware
they could sniff with a scope in 30 seconds...."What use is a scope"
they say...

Steve

Jo wrote:

The http://www.epanorama.net/ site apparently offers a bidirectional
controller with H-bridge and speed control circuit but when I tried to go
there I got a "not found".
Try this:

http://www.reed-electronics.com/ednm...es/51602di.pdf

Its an H bridge with PWM. You'll need to add a comparator to drive the
directions signals, if you want to use your pot idea.

Steve

"Jo" wrote in message
...
In ,
Graham W typed:
OK. I'll look out my notes for this (if I can remember where they are
since I last saw them in 1998!). This 'poor' starting was one reason
that I put the 'fast' switch in my controller. Once the motor has had
a little burst of higher energy to warm up its brushes and agitate the
lubricat film in the gearbox etc., it is good for the rest of the
session.


Thanks.

The effect I'm seeing is that (motor current to start rotating)(
slowest
running speed) hence I start off with a speed surge every time..not good
for
fine focussing. Steve made some suggestions regarding this which I shall
check out.



This is probably a really stupid thing to say, but why can't you just shove
a capacitor into the circuit, so that the first time you power the motor,
it's gets an extra power boost? That way, the capacitor will discharge
after giving the boost, then the current will drop back to normal.

Just a passing thought.

Regards

Colin Dawson.
www.cjdawson.com

In message , Graham
W writes
Jo wrote:

After all, it is going to be the camera/MK1 eyeball which decides
if it is in focus, n'est ce pas?

Yes, with the help of a focusing aid, yet to be built.

Now, that would be interesting! How about an FFT function looking at
the image sharpness on the screen - all done in software! Anyway ...

It is easier and faster to find a bright point like object near the
centre of the field and then measure its height and curvature. Sharpness
and curvature is maximised when the image is in focus so a pretty good
in focus indicator is:.

x[-1,0] + x[1,0] + x[0,-1] + x[0,1] - 4x[0,0]

Where x[0,0] is a local maximum. In a real implementation you may have
to keep track of the position since it could drift or move as you focus.

Regards,
--
Martin Brown

In ,
Colin Dawson typed:

This is probably a really stupid thing to say, but why can't you just
shove a capacitor into the circuit, so that the first time you power
the motor, it's gets an extra power boost? That way, the capacitor
will discharge after giving the boost, then the current will drop
back to normal.

Just a passing thought.


The C would have to go somewhere that ensured large pulse widths momentarily
on switch on. Hmm...it is something that would be very easy to try out with
my existing controller circuit.

Jo

In ,
Steve Taylor typed:
Try this:

http://www.reed-electronics.com/ednm...es/51602di.pdf

Its an H bridge with PWM. You'll need to add a comparator to drive the
directions signals, if you want to use your pot idea.


Thanks. That looks like a good starting point for my Mk 2 controller.

Jo

"Jo" wrote in message
...
In ,
Graham W typed:
Now, that would be interesting! How about an FFT function looking at
the image sharpness on the screen - all done in software! Anyway ...
I wasn't going to be that ambitious :-) I was simply going to construct a
Hartmann mask. Actually, I do know a bit about software development and
FFT,
but that would still be a rather daunting project :-) Does anybody know
of
existing software that does anything like this?

Not having read the previous posts, I may not have grasped your
requirements, but for what they are worth ... I find Mathcad - check out
www.adeptscience.com - extraordinarily good for playing around with all
sorts of analyses as well as throwing in some extra, cut-down, programs that
interface with Mathcad beautifully.

It's very easy to use, and you can get as complex as you like without any
programming skills (if you have some, there are additional functions you'll
find a doddle to use though anyone can get to grips with them). A 2D FFT
transform is included in the basic package. For those situations where the
analyses are more complex than you'd care to tackle, there are additional
modules you can get, for instance the Image Processing Pack.

One of the 'freebies' is a cut-down version of VisSim supplied with Mathcad.
If you want to simulate the dynamics of the control feedback system of your
focuser, it will let you wire up your system on the computer and check its
response to all sorts of inputs. It's easier to change the design in VisSim
than in the garage.

Another 'freebie', a cut-down version of Smartsketch (it may have changed
its name since I got it), helps with mechanical design (bit like a CAD
program but you can dynamically change lengths and/or angles of components -
even to the extent of driving your mechanisms).

The other one is a cut-down version of Axum, a graphing program far superior
to anything Excel can handle.

Excuse my enthusiasm for this little suite of programs, but I've been using
them for some time and never fail to be impressed by what you can do easily
do with them. I use Mathcad to calculate the dimensions and weights of
variously shaped arrow points which saves me a lot of headaches with
integral calculus (can you apply a chain-rule several times in succession
without error?) and currently use Axum with a large database of stars to
view them in stereo, from anywhere. I'm hoping to use Mathcad with Axum to
include proper motions in the displays (one day).
Grim