I'm looking for a servo with an accuracy of <1 degree I.E. can move in from say 1 to 2-3-4 etc degrees etc. Anyone know if a servo can do this ? I don't really want to use a stepper motor because I need to know an exact 'zero' start point. Maybe a 'digital' servo has this type of accuracy ?
I'm not 100% sure but it sounds doable with a digital servo. Only one way to find out, tho. :
However, if you are going to go down the digital servo road, why not an OpenServo you can control over I2C rather than a regular digital servo controlled over the standard pulse width system?
Do you need that precision over the full course of the servo, or can you gear down to a smaller range with higher accuracy?
Speaking of gearing, beware metal gear servos as they often have a little backlash in their gear train, which makes accurate positioning hard to attain.
I would be interested in hearing how you get on. All the digital hobby servos I have come across use conventional pots for position feedback and they are unlikely to have a linearity of anywhere near the 0.25% you would need to get accuracy to one degree or better.
I would think a stepper may be your best bet if you can figure out a way locate the shaft position when you start up
Many thanks for the reply, What I'm trying to do is make a pan and tilt head to connect to 'Orbitron' ( http://www.stoff.pl/ ) to trace/track space objects with a camera,laser,etc. I thought 1 degree accuracy was a good starting point (no pun :D)
This type of thing http://moro.fbrtech.com/ift/project.html - I know half it uses a stepper motor ! ;D
I would think you can get better than 1 degree with a stepper. In your application, can the position be calibrated from time to time. For example, swing the head until an optical sensor indicates that the head is at some reference position so you can check and adjust your step count if necessary.
pretty well everyone I know would use a stepper for this, with out any gearing my stepper moves in 0.225 degree steps. Not bad! I'm gearing this at 3:1 for 0.075 degree, im sure you could crank that up if you want even more though.
There are steppers that have a ir sensor for setting a 0 point, so it's all collaborated. I'll find the link and post it, there commonly used in printers.
Nick D
Yes you will need three known reference points in order to calibrate the transformation matrix needed to map between the co-ornate system of your platform with the co-ornate system of the sky. You will also need a real time clock as the sky co-ornate constantly change. It's quite heavy computationally and I am not sure if the Arduino has the capacity. From what I can remember you will also need SIN functions.
I made something similar in the mid 90s, in effect a digital setting circle that would read out the celestial co-ornate of where my scope was pointing. I used optical shaft encoders to feedback the position of the scope but I made no attempt to actually move the scope.
There is a book by "Trueblood" on designing computer tracking systems for telescopes but I am not sure if it is still in print.
Bits of my work is still on the net at:-
http://www.doc.mmu.ac.uk/STAFF/A.Wiseman/Acorn/BodyBuild/BB93.html
I would have thought that you would be better off with stepping motors as direct drive does have a lot of weight.
The transformation equations were described in a Sky & Telescope article but I can't put my hand on that at the moment.
please keep us informed of you progress on this one.
Hello Mike, most, if not all of the leg work would be done buy the PC running Orbition - as everyone seems to think stepper motors maybe the best way to go. I also have a nice purpose built pan and tilt head but alas no feedback - pots etc or room inside to fit any
oh yes, any chance your a 'ham' Mike ?
Definitely steppers for this, and for one reason having nothing to do with accuracy: jitter.
You won't be able to see anything thru a scope of any power that is being held on target by a servo, particularly if controlled by the pulse duration method. The servo will jitter around and the vibration will make the viewing experience rather unpleasant.
I agree with all the other reasons, and I also doubt you'd get enough torque unless the scope was exquisitely well balanced and there was no wind pushing on it.
Besides, steppers are really not that bad to play around with.
Instead of/ along with the encoders or contact switches to recalibrate position, I'd look into an accelerometer and if your budget will stretch, a compensated compass, for absolute orientation and while you're at it, a GPS receiver for both your position and for a good timebase.
Lastly, building the hardware so that you can compensate for the earth's motion with a single stepper rather than two (by pointing one of the axes parallel to the planet's axis of rotation) will make your life a LOT simpler in terms of calculation and accuracy, to Grumpy Mike's point.
any chance your a 'ham'
G8HBR (but lapsed)
Lastly, building the hardware so that you can compensate for the earth's motion with a single stepper rather than two (by pointing one of the axes parallel to the planet's axis of rotation) will make your life a LOT simpler in terms of calculation and accuracy, to Grumpy Mike's point.
Yes, but the price you pay is having to align the system with the axis of rotation of the Earth. The other axis has to be at the same angle as you are in latitude. This is not always too easy as I found out driving my scope.
By the way have you checked out the "Star Scout" but I am not sure how accurate that is.
Guessed as much with the CW mentioned on your pages - I'm G6ZKC BTW You should renew no renew costs now licence for life !
I've not seen Star Scout, Orbitron seems good for satellites or Stellarium.org for stars is really nice. I'm playing with the 'Wisp' plug-in for Orbitron that outputs Azi and el data to a com port.
Yes, but the price you pay is having to align the system with the axis of rotation of the Earth. The other axis has to be at the same angle as you are in latitude. This is not always too easy as I found out driving my scope.
What would be really handy, I suspect, would be a system that allowed manual input for correction, and deduced the error in alignment from those, reducing it over time until it got dialed in.
Having said that, with a GPS to give latitude, an accelerometer to give inclination and a compass to give magnetic north... ach, you'd need true north compensation tables wouldn't you?
Hum, add a GPRS module to download them from a webserver somewhere? :
Hi Syvwich,
What you just described is a Star Scout.
http://www.celestron.com/skyscout/
But what's the fun of buying one ready made when you can spend more money on parts and waste hundreds of hours putting together hardware and software and testing it...
oh.
Never mind.
I quite agree :
for my panorama robot i use 1:250 gearboxes that gives a resolution of 400.000 microsteps per 360 degree
more than enough to follow the stars with high precision
but have a look to the orion altazimuth teletrack. this device has an interface, so you can connect an arduino which calculates the move for both axis.
cheers
Paul
"pretty well everyone I know would use a stepper for this, with out any gearing my stepper moves in 0.225 degree steps. Not bad! I'm gearing this at 3:1 for 0.075 degree, im sure you could crank that up if you want even more though.
There are steppers that have a ir sensor for setting a 0 point, so it's all collaborated. I'll find the link and post it, there commonly used in printers.
Nick D "
Any chance of the link yet ?