One way to motorize a dobsonian telescope - emphasis on simple materials and techniques.

NOTE: This page uses thumbnail images. Click on the thumbnails for more detailed images of 100KB size.

At the bottom of this page I'll briefly discuss what I learned, what I'd do differently the next time, and some other drive system approaches that appear promising.

What started as an experiment and prototype that I expected to quickly discard turned into a reasonably well working drive system for my sixteen inch f/6 dobsonian. It still performs well after two years. Here are photos that highlight some of the design and construction details of my JB Weld molded worm gears. A big thanks to Eric Greene for scanning my negatives and originally posting them to his website. Now I am adding explanatory text and the hindsight of experience.


Overall view of the rocker box, showing the altitude motor and worm (nylon threaded rod), with the azimuth worm gear also visible. The threaded rod is 3/8-16.


Different view, allowing one to see the azimuth motor, it's nylon threaded rod worm, and spring loaded worm retaining flap (which is merely 1/16 inch aluminum bar stock). This design requires the ground board to be circular so that you can mold the worm gear on it.


Close-up showing elevation motor, and home made, square flywheel. Flywheel is merely 1/4 inch scrap steel. Painted yellow marks on flywheel and motor housing were to check if stepper motor was missing steps over multiple slews. Below the worm is the spring loaded retaining flap. Above the worm is a cam follower (and there's one on the other side of the rocker box) to keep the mirror box centered without scraping inner surface of rocker box...which increases friction and can cause motor to stall when slewing. Above stepper motor is one of the undriven rollers. Because I made my rocker box sides about 1 1/2 inches thick, I was able to 'bury' the roller bearing within the plywood...a very stiff and low profile arrangement.


Another close-up of elevation motor from a different angle. Note that the cam follower is attached in a very simple manner...strapped in place with 1/16 inch aluminum bar stock and drywall screws. Simple but effective. The undriven roller is mounted on a 3/8 inch bolt that is cut to length and pressed into a slightly undersized hole. The recess of the roller bearing was 'fabricated' with a dremel tool. It looks ugly, but it works.


Another close-up of elevation motor and worm, this time focused on the worm and worm retainer. Note that the retainer is worn from rubbing against the nylon threaded rod. The worm retainer's 'tension' against the worm is adjusted by tightening its mounting screws. One nice feature of the layout of this elevation drive is that the mirror box merely drops in place over the worm during assembly in the field.


Extreme close-up of elevation worm, worm retainer, and cam follower. Wear on worm retainer is evident, however in use this has not been a problem. After an initial wear period, a reasonably large surface area of the retainer is in contact with the worm, and rate of wear slows considerably.


View looking straight down on elevation motor. Motor mounting brackets are 1/8 inch aluminum bar stock, with sorbothane (from shoe store insole) in between motor and bracket for vibration isolation.


Extreme close-up of undriven elevation roller on rocker box. This scope was originally a teflon/formica, hand-pushed dob. Sheet metal now covers all formica surfaces for smoother/lower friction motion with the roller bearings. The sheet metal is 24 gauge scrap from a local heating/air conditioning/ventilation contractor. They make sheet metal ventilation ducts and have lots of scrap laying about. I got mine for several dollars.


Close-up of azimuth drive train, showing JB Weld molded worm gear, worm (nylon threaded rod), worm retaining flap, flywheel, stepper and mounting brackets.


Close-up of azimuth worm, worm gear and retaining flap.


Close-up of azimuth stepper and flywheel. This is a NEMA frame 23 size motor. The square mounting face of the motor is about 2 1/2 inches on each side.


Close-up of azimuth motor's mounting brackets. Because this motor hangs close to the ground I had to make my ground board's feet about one inch taller, and when I set this scope up at a new observing site I have to trim the weeds around the ground board so that they do not foul the motor and flywheel.


Close-up of circular bubble level glued to bottom of rocker box. This helps me quickly level the ground board on uneven ground.


Close-up of the south leg of the ground board, showing the extension added to provide ground clearance for azimuth motor. Also, the leg is labeled for a reason. Mel Bartels' software can measure and store in a file the pointing errors of your mount. My scope's rocker box bottom is made of plywood. Guess's not perfectly flat. It's a little bit potato chip shaped. As the scope moves in azimuth, this imparts an elevation error of plus or minus several arcminutes. Fortunately this error is consistent and repeatable, so I measured it and captured it on disk. However, this pointing error is tied to the physical setup of the scope, so I need to place my ground board in the same orientation each time I use it. That is why one foot is labeled 'south.' I find that a small price to pay for improved pointing accuracy with Mel's software.


Overall view with mirror box now in place on rocker box. You can now see how the elevation worm meets with the elevation sector worm gear. Also visible is the sheet metal covering the elevation trunnions. Why is there a fan on the side of the mirror box? This is the second fan...the first one is behind/below the mirror. I use both fans to cool the mirror. The side blowing fan helps 'scrub' the warm layer of air off the face of the mirror, while the lower/rear fan draws air around the back of the mirror to cool its rear surface. I can get rapid cooling this way. I have the option of running both fans, one, or the other, or none of them. Almost all the time I find that running both fans helps improve image quality...especially in the early evening. In the rare case that the fans do not improve seeing I shut them down. It's good having such options. If you do not install fans on your scope for active cooling you do not have this option.


Close-up of side mounted fan. Note duct tape on fan hub. This holds a small washer underneath, and slightly off-center with respect to the fan's hub. Why? Through trial and error I found that placing a small washer in a certain location on the hub improves the dynamic balance of the fan and reduces vibration. Because many fans are not precision manufactured items, they often benefit from such tweaking. Some fans, through luck or good construction, will not benefit from this, but some will.


Close-up of elevation sector gear. Note the numerous imperfections in molding this worm gear. As I mentioned earlier I did not expect this method of fabricating a worm gear to work, and did not take the greatest care in molding. However despite the flaws, I have no complaints with its performance. A significant contributing factor to this success is that with such a large worm gear, many teeth are meshed with the worm at one that gaps/imperfections with parts of the worm gear are not a problem if enough of the rest of the worm gear can contact the worm. BOTTOM LINE: you don't need to mold a flawless worm to get decent performance.


Close-up of elevation worm meshing with elevation gear, but the retaining flap is 'disengaged' (i.e. the mounting screws are loose).


Close-up of elevation worm meshing with elevation gear, but the retaining flap is 'engaged' (i.e. the mounting screws are tight).


Overall rear view of mirror box resting on the rocker box, showing motors for both axes.



Lessons learned: (strengths, weaknesses, and what I'd do differently if I were to rebuild the system)



- Simple and low cost. With one motor, one worm, and one worm gear per's hard to reduce the parts count further. Molding a worm gear with JB Weld (in the big Industro size tubes) is far cheaper than any other source I know...although the price seems to have risen recently. You can often find this stuff at auto parts stores. (NOTE: there is no clutch system in my drive train in order to keep things simple, but as I explain below, use of simple spring loaded strips to retain the worm make this an acceptable work around. Is this work around acceptable for you? That depends on your design goals.)

- Using the stepper motor to handle the thrust loads of the worm further simplified the design and kept parts count even lower. (But that raises other problems.)

- Nylon threaded rod worms are very forgiving of misalignment. I did not need to worry about precision fabrication and installation of drive components...great for us basement hobbyists that lack a machine shop. The six inch long threaded rods handled small misalignment with ease. I could get away with 'fabricating' my motor brackets with a hammer, thanks to that flexible worm.

- The drive train is limp and viscous. In other words the combination of nylon threaded rod, rubber/metal shaft couplings, and compliant mounting of the motor to the telescope make for a decent amount of viscous damping. When the scope is disturbed by a breeze or clumsy hand the scope returns to position after only one oscillation. Many would consider that a pretty well damped system.

- The drive train is durable and can take abuse. The simple spring flap that holds the worm in place means that you can't damage the worm gear or worm. Adjust the spring tension so that the worm is held in place (and friction between worm and worm gear is not too high for the stepper)...and if a gust of wind comes up, or you slew the scope into a ladder...the worm rises out of the worm gear and skips teeth. No physical damage to the JB Weld gear, and the nylon threaded rod is pretty abrasion resistant. I've had this happen many times and have noticed no visible damage or loss of performance.

- You can fabricate just about any size worm gear you need, depending on the size of your gear blank (disk of wood or plastic), and the pitch of your threaded rod. This means you have great flexibility in choosing your overall system gear ratio. NOTE: I have not tried to make worm gears with fewer teeth...such as 200 or so. With fewer teeth you may get too much bending and distortion as you wrap the threaded rod around the disk when molding the worm gear...but that remains to be seen.

- No need for encoders, and no problems with friction roller slippage (a problem with some friction roller drive systems). As long as the stepper motors don't stall...Mel Bartels' software does a great job slewing the scope 'in the blind' to the next object. When the system is working well I do not have to remove my 220x eyepiece (approx. 8 arcminute field of view) and use a wide field eyepiece...just command a slew to the next small NGC galaxy and there it is. (Note, some nights I have zero stepper stalls...after hours of observing and hundreds of slews. Other nights I have a couple stalls, such as when it's windy or the scope is not well balanced.)



- This is not a stiff drive system. That six inch length of nylon threaded rod is flimsy. For heavier scopes, or scopes with large wind profiles this may prove to be an inadequate approach...but you can take steps to stiffen this drive design (see comments below).

- The nylon threaded rods are not perfectly straight. This imparts some periodic error to the system. (Because I use this scope for visual observing, I have no complaint with my periodic error. If I were using this scope for imaging the periodic error would be unacceptable.)

- Because I chose to make the stepper motor bear thrust loads, I need to have a tight hold on the motor with its mounting brackets. Those brackets are attached to a wooden scope, which amplifies the motor vibrations. This is noisy and can sometimes make vibrations visible at the eyepiece, especially if the stepper motors and software are not well tuned. (Note: After two years of use the motor bearings show no signs of being worse for this wear from bearing the thrust loads. Probably because my spring loaded retaining system lets the worm release from the worm gear if the force gets too high...typically at about five to ten pounds of force.)

- There is no clutch in my design. (However, spring loaded worm retention is a work around that I find acceptable as a sort of 'safety clutch.')

- You can't easily 'declutch' the drive and manually move the scope by hand. (However, the scope slews at over 3 degrees per second, so the wait is not long between objects, and the motor drive/wiring harness/computer/battery system fails so seldom that I do not mind having no 'manual backup' in this scope's drive system.)

- The spring loaded worm retaining flap, while simple, puts more friction in the drive train. This forces me to use more current to generate more torque in the motor while microstepping. More torque means greater vibration from the motor, which means more noise and greater chance for vibration to show up at the eyepiece.


What would I do differently:

- Stiffen the drive. (Explained in the next two paragraphs.)

- Isolate the motor so that it does not carry thrust loads. This will accomplish several things. First it will allow a shorter worm, which is stiffer. (The large body of the stepper motor meant I could not get the motor very close to the worm gear...which required a long, flimsy worm.) Second it will allow me to mount the motor less tightly to the telescope, which will mean less motor vibration is transmitted to the wood body...less noise, less chance of vibration showing up at the eyepiece.

- Experiment with metal threaded rods. That would stiffen the drive system, and reduce periodic error. Maybe this would make it good enough for some imaging applications if an autoguider were used. However, a metal worm may wear the JB Weld worm gear too much in the long run...this remains to be seen. (Also, a metal threaded rod may be less forgiving of alignment errors.)

- Use a lower friction method to retain the worm against the worm gear. This will allow the motor to use less current when microstepping, and keep vibration and noise to a minimum.

- Mold worm gears with fewer teeth. My gears have about 1200 teeth. Now that Mel's software can handle 40 microsteps per can probably get away with direct driving a worm gear with about 300-400 teeth (assuming you use a 200 fullstep/1.8 degree step motor). I may not go that coarse, but even with 600-800 teeth I could easily get 5-7 degrees per second of max slew speed, yet have a not-too-large microstep size in terms of acrseconds of sky to retain smooth motion at the eyepiece.

- Use smaller flywheels on the stepper motors. The purpose of flywheels is to change the resonance properties of stepper motors when half or fullstep slewing. Why? When steppers are running with no load, and ramped up in speed...rotor resonance problems crop up at about 2-3 revolutions per second...which is a speed regime where motor torque starts dropping off a bit. (As steppers spin faster, available torque drops lower and lower.) Put flywheels on the steppers and you lower the rotor/flywheel resonant a speed regime where steppers have more torque and can 'bull through' this problem/resonance speed zone. You should use the smallest flywheel you can get away with to fix this problem. Too big a flywheel means that the motors can't handle fast accelerations while microstepping (which has less torque than when full or half stepping) and the motors can miss steps.

- Use smaller stepper motors. Most of us use NEMA frame 23 (the motor is about 2 inches in diameter) steppers...more than enough torque to move your scope unless it's really large and heavy. If you use NEMA frame 17 size...smaller motors can accelerate much faster...better for handling microstep accelerations, yet still have plenty of torque for the typical sized amateur scope.

- Add a clutch? Perhaps, but I am happy with how the spring loaded worm retainers work, and overall system reliability is high enough that I don't lose much time due to computer/drive/battery problems.


Other approaches with promise: (more than one way to skin this cat)

- 'V' shaped worm gear. Create this by wrapping two threaded rods around a disk, meshing and nestled beside each other. One way to keep them close together would be by placing retaining flanges on either side of the threaded rods. This creates a 'V' shaped groove between the threaded rods that will mesh with a worm of the same pitch. Some may find this an easier and faster way to make a large worm gear, without the mess of molding.

- Industrial Velcro gears. Instead of a molded worm gear, wrap a strip of industrial velcro around a disk. Industrial velcro has its 'buttons' arranged in very orderly rows, and you end up with a very large spur gear of sorts. Sounds funny? Rick Singmaster has successfully used it on his Starmaster dobsonians. However, you then need to come up with a spur gear that meshes with the velcro's 'pitch' and tooth profile, and you'll need further gear reduction between the stepper and the final, small spur gear. (If you are nice to Rick Singmaster he may give you some info and ideas on the details of this drive.)

- Timing belts. Similar to industrial velcro, you can wrap toothed timing belts around disks and make whatever size 'gear' you need. Places like Small Parts Inc. carry various belts and gears. Again, you'll need further gear reduction between the stepper and the final gear.

- Chains. They can be quite stiff, and I've seen them successfully used on 36 inch f/5 monster dobs. Ask Andy Saulietis for advice in this regard. (his email is i-s-s-@-p-v-t-n-e-t-w-o-r-k-s-.-n-e-t (remove the dashes)


All feedback and comments are welcome, and will make this a better information resource for amateur telescope makers.


email: t-k-r-a-j-c-i-@-s-a-n-.-o-s-d-.-m-i-l (remove the dashes)

Last update: 25 Oct 2002