From my experience the torque in µstepping would be equal to the weak full-steps.
The strong full-steps are 1.4x stronger but I think it will be fine.

Removing torque ripple i smy main goal for µstepping as it is the main source of vibration when I come to a halt at new position.  I have implemented accelleration/decelleration control and it reduces vibration well, but my test with a laserpointer shows I could use more smoothness. The positional accuracy is very good due to zero slop in gears and no backlash.  (I use single stage gearing with tight mesh and no worm-gear with slop.)

Here's a shot of a 30 second exposure from my cam on a seperate tripod and a laserpointer plotting on my silverscreen:
http://kreature.org/ee/avr/pt1/pt2_laserpointer_1.jpg
As you can see, it oscillates a bit when the y-axis decellerates vs the x but the vibration dies out before the shot is taken as the systenm comes to a smooth stop.
I would like to make this even better and it can be done with same cost of electronics.

Here is also some closeup shot of one of my axis stepper and gear:
http://kreature.org/ee/avr/pt1/parts_3.jpg
http://kreature.org/ee/avr/pt1/parts_5.jpg

The shaft is very robust at 15mm diameter and the bearings are under tension when the unit is assembled through a square alu beam. This was done to avoid any slop and rattle to give a rigit linkage able to hold large cameras and lenses.

Here's one of my toys and the reason the arm is so rugged:
http://kreature.org/images/zoomify/pt2/ … kustov.jpg
It's cheap, but effective. 1.5x factor on Minolta 7D camera, 1000mm f10 Makustov fixed focus tele and 2x teleconverter.
End result is 3000mm f20. Weight: 3.3 kg

Thanks for the explanation. Stepper motors are easy to use, because they don't need feedback. But as you said, as long as they do not slip. Adding an encoder to a stepper motor is not a good idea. And this is not a problem here: the Merlin/Orion works the same way: the encoders are just incremental ones, so, the (yaw=0, pitch=0) position is the one when you turn on the head. Don't worry, Papywizard use en internal offset, so you can set the reference position where you want.

About the vibration, I think you are experiencing a resonance problem; try to increase the minimum speed, or use a faster acceleration. At work, we often have such problem at low speed; the idea is to cross the resonance speed as fast as possible during the acceleration/deceleration phases, or, if you can, to start at a speed over that resonance speed (but you may start too fast, which can be a problem for the camera sensor). Another solution is too increase the gear box ratio, ta use the motor at a faster speed. But here, you may see the tork falls down (

Yes, but is an absolute position really usefull? The yaw=0 position does not really matter, unless you need precise coordinates¹. For the pitch=0, I use a bubble leveler, and set the reference when the camera is horizontal.

¹ This is the case for me these days, as I made a little tool to map the sun course over a full spherical pano, to see the natural masks from trees and mountains: I'm about to build my house, and want to determine the best orientation. So, I set the yaw=0 at north, using a compass.

Using a driver chip is not acceptable. That removes all the fun of driving the stepper.
Half the reason I do this project is for the fun.

In addition, such drivers are not always easy to come by and if they can be ordered they are often expensive.
My driver design uses easy to aquire parts and is not sensetive to replacing most of them with locally available alternatives.
In fact, most FET's you have in your parts bin will work.

Is the acuracy enough, on such low-cost gearboxes?

fma38

I originally looked at purchasing a Merlin/Orion. Once the import costs, exchange rate and modifications were added up; using the Merlin/Orion is too expensive. I started looking at alternatives.

KreAture

I do not doubt your electronics skills. Building equipment is therapeutic. If you can obtain free components, use them. 

In my region using dedicated chips is the only option. No-one in my region stocks suitable electronic parts; mail order can work out expensive. There is a generation of people who are just electronics card jockeys. The electronics hobbyists no longer exist to sustain the electronics shops.

I made a few enquiries in my local hobby shops and electronic component suppliers. The local hobby shops sell radio controlled toys; electronics shops sell kits published in Silicon Chip.

There is a professional photographer in the USA who used two cordless drills for a pan and tilt mechanism for a professional rig. The cordless drills I used were free. Purchasing an equivalent planetary gearbox is expensive. You have to dismantle the gearbox. Remove the hammer mechanism, grind off the hammer notches and re-assemble. The grease used in the gearbox is not of a high quality. I modified some cordless drill gearboxes, used quality grease and found the slop in the gearbox acceptable. If you can obtain a discarded cordless drill; try the modification. You may be surprised with the results.

Precision worm gears are expensive. If you have access to a lathe and the right tools you can machine a precision one from a blank. The blanks are available over the internet. An alternative supply of worm gears is your local motor vehicle wrecker. Electric car seats have three worm drive gearboxes. Damaged car seats are left to rot. The gear and worm are a reasonable quality. You need to machine the shaft and bearing mounts. Another source is windscreen wiper motors. The gearboxes are good for the worm and gear parts.

Unfortunate the harsh Western Australian summer climate I cannot control. It destroys LCD displays when used outside.

Imagine one third the size of Europe with a population of just under two million people. Imagine 80% of those people living in one city. Once you leave the city all you have is small country towns. The reason the city thrives is because of massive iron ore and gas deposits in the North West region of Western Australia. Commerce is just a support infrastructure for mining. Living in Perth has its benefits. Unemployment under 5%; crime rate is low; standard of living is high; 9 months of outside living weather. Once you leave the city the North West is a photographer’s oasis with beautiful rugged scenic country waiting to be photographed.

The only frustrating part of living in my region is obtaining suitable parts to build your own equipment.

Here's a new video with large movements between images to show how it accellerates and such.
http://www.youtube.com/watch?v=-Yu6nDVvPyA

Here's images of the rig:
from left side - http://kreature.org/ee/avr/pt1/pt2_1.jpg
from right side - http://kreature.org/ee/avr/pt1/pt2_2.jpg
tilt gear and axle - http://kreature.org/ee/avr/pt1/pt2_3.jpg
horizontal/vertical joint - http://kreature.org/ee/avr/pt1/pt2_4.jpg
pan gear - http://kreature.org/ee/avr/pt1/pt2_5.jpg

Except from my lathe, only a hacksaw, dremel and power-drill was needed to make it.

(Excuse my long setscrews but it was all I could find right now, hehe.)

Hi Greg,

Papywizard has allready a plug-in for Polulu servos.
Frederic (fma38) have also a "running" prototype controled by a "arduino"

See also:
http://www.gigapixel.nu/info_project.php

Last edited by claudevh (2009-10-08 00:04:56)

Arduino is the way to go; there is a ready-to-use mezzanine controller for stepper-DC-servo motors.

BTW, is your head already motorized? What kind of motors do you plan to use? For which usage?

I think that servos are very simple to use/drive (using a pololu controller), but are not very accurate (2°). Look for Phil's Panoduino head on google... As Claude said, Papywizard is able to drive such head using the Pololu servo controller. We never made tests with *this* head, but I build a little prototype with simple servos, and it works fine. It should have enough parameters to drive any kind of servo-based panohead.

I'm sure GuzZzt can help you on the Arduino side.

And I can help you on Papywizard side: I think we should write better protocol than the one used in the Merlin. Why not create some sort of universal protocol for all people who want to develop a motorized panohead, and want to drive it from Papywizard (even if it is easy to add a new plugin)? For example, we should try to support things the Merlin can't do, like goto funtion at different speeds, or shoot while moving, or so... I can start to write all low-level functions needed to interface with Papywizard in a good way.

I will be pleased to see how you will design your head with motors

Greg Nuspel wrote:

This sounds great, I'm looking at using 18:1 reduction from a 200 step motor to give a 0.1 degree step. With timing belt drives backlash will be near zero (nothings perfect). It would be great to design a more open interface allowing better control.

If there is interest we could even look at doing a group buy of the mechanical parts from my employer so that there would be a reasonable price per unit. I'm sure it would cost more than the Merlin but it would be a much better unit for handling larger equipment with great precision.

a version 3.0 with blootooth is interesting for me.
Canon EOS 400D (+ Battery Grip) x1.6 factor with Sigma 400mm + sigma 2.0 konverter + AT Bluetooth Modul + PW so yes I'm interested ;-) (for default merlin config I currently run with its not possible to center it correctly hori/vertically)
what price span do you expect?

Liebe Gruesse,
Georg

I would prefer to use an ascii-based protocol; it is much more simple to debug, and you can test the controller with a simple terminal.

About the commands, I think we should just execute them as soon as they arrive to the embedded controller:

dr 1 54.34 -> start moving axis 1 to position 54.34°
rd 1 -> return axis 1 position
st 1 -> get axis 1 status

What do you think? We can use encoder units for positions, rather than degres (it is easier to adapt on Papywizard side for different heads; if gearbox changes, no need to reflash the controller).

This will need a multi-threaded controller, but I think the Arduino is OK.