Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

DC motor selection for REALLY low speeds

  1. May 18, 2007 #1
    Hi everyone.

    I'm in the process of designing a telescope mount, and am now trying to figure out how to drive the thing. Ideally I would like to go with an off the shelf dc motor and gearhead setup, as this appears to be the cheapest option. My problem is, that I don't know how slowly you can run a dc motor before you start to cause issues - like cogging, or maybe burn-out, or who knows what. So I was hoping that you guys could give me some pointers....

    Here's some specs that I think are required:

    -The system must turn at around 1 revolution per day.
    -Start-up torque, with safety factor included, is about 20Ncm.
    -Normal running torque will be significantly less than above.
    -Budget is around AU$200 per axis (this is the main problem!)

    Stepper motors are not an option, unless someone can help me prove that the cogging of a micro-stepped stepper is comparable or less than that of a standard dc motor.

    Becuase I've come in on this project three quarters of the way through, I'm a bit limited by the design so far - so a great big worm gear arrangement probably will be a little too tricky to implement (and would play havoc with the budget too I think).

    I guess what I would primarily like to know (at this stage:biggrin: ) is how to figure out how slow a dc motor can safely go given the data provided by manufacturers such as mclennan.co.uk and pittmannet.com.

    I'm sure there are a million other questions too, but I just don't know them yet! I've trawled the net and various textbooks too, so if anyone can suggest some good texts etc, that would be great.
  2. jcsd
  3. May 18, 2007 #2


    User Avatar
    Gold Member

    Can't help you with most of your questions, unfortunately. I would, however, recommend looking into epicyclic gear trains. One alone can give you something on the order of 1,000:1 reduction, and you can stack them.
    Russ might be your best source here; he has a bit of experience with telescopes.
  4. May 19, 2007 #3
    Thanks for the tip Danger - I'll have a look into epicyclic gear trains.

    I've managed to find some good info on the Pittman website, and it seems that slowest you can run any of their servo-motors is about 500rpm, which means I still need a reduction of around 700 000:1. If I were to use a worm drive of say, 60:1, on there, I could get away with a gear train of about 11 000:1, which is in the realm of possibility (I've seen a 15000:1 gearbox made by Mclennan). It's just the cost that appears to be the killer! If Russ is hanging around, I'd love to hear your thoughts on driving telescopes....
  5. May 19, 2007 #4


    User Avatar
    Gold Member

    Russ is always lurking somewhere; he's one of the Mentors.
    Strangely enough, when I Google epicyclic trains all that turns up are various planetary set-ups such as in an automatic transmission. I have not been able to find anything at all about the one I'm thinking of. It consists of 2 sun gears back to back, connected by a single planetary. One of the suns has one more tooth than the other. That's for instance 100 teeth on one and 99 on the other. It's not an ideal situation because obviously there isn't a perfect pitch match between the 3 gears, but you can get some awesome ratios.
  6. May 19, 2007 #5


    User Avatar
    Science Advisor
    Homework Helper

    I don't know anything specifically about telescopes but I'm curious why you don't want to use stepper motors. I would have thought they were very suitable for this, since they are easy to control, high torque, and run efficiently at any speed from hundreds of steps/second right down to zero.

    I would have thought the inertia of the scope and drive train would smooth out the "stepping" effect pretty well. You might want to consider the position errors caused by tolerances in the reduction gears (imperfect tooth profiles, gear wheels not perfectly centered, etc) and compare those with errors from a stepper motor drive.

    Speculating, you might even be able to compensate for the errors in a cheap gearbox with clever software that drives the stepper at slightly variable speed to compensate. That sounds pretty simple, compared with trying to control the speed of an "analog" motor accurately.

    One final thought (hope it's not so obvious as to be patronizing): cheap and simple mechanical devices that rotate accurately at 2 rev/day have been around for a long time. They are called clocks. Maybe high-rratio epicyclic gears etc are overkill for this application.
    Last edited: May 19, 2007
  7. May 19, 2007 #6


    User Avatar
    Gold Member

    It's not patronizing at all; sometimes we overlook the obvious. My only concern in this case would be getting enough torque out of a clock to move the telescope. Good post.
  8. Jun 5, 2007 #7
    I can't use stepper motors because the guy I'm building this for has said "NO STEPPERS". Which is unfortunate....

    Clocks though, eh? Good suggestion, although as Danger said, I may need a little more torque than that required by the hour hand of a clock :) Will look into it though.

    I have found *a* solution - a motor with a 800:1 gearbox on the front, plugged into the 15000:1 gearbox I mentioned earlier. I'm not really happy with it, as you mentioned, Alephzero, all those gears are sure to introduce a heap of errors. Some sort of feedback loop may help to sort this out though (although I feel like I'm just handballing the problem onto the next person).

    Slightly off topic, I've read on a few places the term "base speed" referring to DC motors - does anyone know what this means? Rated speed perhaps?
  9. Jun 5, 2007 #8


    User Avatar
    Science Advisor
    Homework Helper

    Well, clockwork used to be the answer...

    http://www.aas.org/~pboyce/mma/cooke-tel.htm [Broken]
    Last edited by a moderator: May 2, 2017
  10. Jun 5, 2007 #9
    Ahh, I thought you were talking about modifying an old wall clock!
  11. Jun 13, 2007 #10


    User Avatar

    If you're really stuck, ask me and I'll send you plenty of more
    specific links and references that I've accumulated over the

    My best suggestion for inspiring ideas / dialog is to take a
    look at the various "ATM" sites (amateur telescope
    makers / making) out there. There's a LOT of
    good information on all kinds of homebrew motorized
    drive mechanisms for various kinds of mounts / platforms
    out there. There are all kinds of options from the simple
    to the byzantine and the elegant to the rube goldberg
    "wow I can't believe that works".

    There's nothing inherently wrong with steppers; being
    a fairly old time amateur astronomer myself, I realize
    that the two objections are (a) complexity -- needing
    a less common motor, and digital/analog/power drive
    controller board electronics to drive it at the proper rates,
    and (b) smoothness -- if you're tracking something at high
    magnification, or doing long exposure astrophotography,
    then the jitter due to the jerky steps / pulses can cause
    irritating vibration / dancing / blurring of the object
    even though the jitter motion may be quite small.
    That being said, almost all the modern scope designs
    for modestly big instruments including ones that
    are excellent at astrophotography do use steppers and
    feedback control system technology to control the
    motors. In fact many scopes use ALTITUDE-AZIMUTH
    mounts and it's ONLY the smart positioning/tracking
    systems that keep them tracking anything at all. The
    cost / size / mechanical inconvenience of an massive
    equatorial mount is FAR worse a burden than something
    effectively as cheap/simple as an alt/az 'dobsonian' like
    mount added with the cost of the relatively more
    complex stepper / computerized tracking & motion control
    system and position sensing systems etc.
    Also you can do things like micro-step steppers even
    to the point of using them as somewhat linear thrust
    motors in the space between one step and the next
    so in effect it' acting more like a DC motor (which it
    is, really, just one with several individual phases of
    stators to control) than a digital 'step' motor.

    You can get things called slow / synchronous motors that
    are AC sine wave motors that are powered by household
    power, and they run synchronously to the AC power
    frequency which is only 60 or 50 Hz. Given that and very
    modest up/down gearing, you can end up with very slow
    speeds indeed, but you're limited to places where you have
    sources of sine wave AC power to run them.

    If you pick the appropriate DC motor you can run it
    quite slowly, you just need to find one that's got plenty
    of torque and can run with up to 12 hours * many days
    per year of service duty in a near stalled condition with
    very little energizing current if it's being used at much less
    than it's usual rated speed. Generally that'll be a very
    over-rated motor used at a fraction of its capacity.

    There are also lots of ways to 'cheat' and get an excellently
    smooth tracking with relatively cheap parts. One being
    using an 'equatorial platform' which is basically just
    a board / 'ramp' that the telescope sits on that happens
    to point north and be tilted at the local angle of latitude.
    Think a couple of doors screwed together by one set of
    hinges, one laying flat, the other tilting open at whatever
    angle is needed, and a brace set between them to keep
    them open at the right angle. Just a crude description;
    there are lots of designs for building them out of
    anything from plywood to concrete etc. etc. check the
    ATM articles.

    Given such an adjustable equatorial platform you're
    already north aligned and equatorially tilted, then you
    set an alt-az mounted scope on it e.g. like a dobsonian.

    Then use a manual ALTITUDE adjustment on the scope
    since ALTITUDE (now == declination given the
    e.q. platform) never needs to change once you're pointed
    at the object of interest. So then the only axis that
    needs mechanical tracking is the AZIMUTH
    (== right ascention now given the eq. mount). This
    can be done ever so smoothly, cheaply with a
    tangent arm drive. Get a long threaded rod with modestly
    fine threads per inch of its length.
    Fix one end of the threaded rod to your favorite motor,
    DC, AC, synchronous, whatever. The motor and rod are
    fixed to the platform and do not move with the scope.
    Put a "nut" aka "follower" on the fixed position rotating
    threaded rod, and hold the nut / follower from rotating
    and have it 'push' tangentially against a rotation circle
    attached to the azimuth (aka R.A.) axis of the scope.
    As the motor turns, the nut/follower 'unscrews' along
    the length of the threaded rod and pushes the scope
    along. Disadvante being it can only track for a limited
    period of time since the follower is moving in a line and
    the scope is rotating, so the line is basically a chord of
    the rotation circle of the scope and once the nut is at
    maximum extension you have to mechanically 'reset'
    the scope by rotating the tube mount relative to the base
    so that the 'retracted' position now again points at the
    object of interest and you're ready to track for another
    length of time until you reset again. With even a modestly
    slow motor and modestly resolute threaded rod you can
    devise these to track for hours at a time which is usually
    quite sufficient even for photography, though it's unsuited
    for something that should function from dusk to dawn
    entirely automatically. Of course you can add
    altitude / declination motor drive if you want too, but
    that's quite optional.

    Another variant is to make the motor and threaded rod
    so they move linearly on a slide as the rod screws against
    a fixed nut/follower; in that case you use the end of
    the screw rod itself to push tangentially against a
    fin/driving point on the rotation axis of the scope so that
    the tip of the screw pushes along the rotation of the
    scope. Again, smooth, simple, and the drawback is
    that it'll work over a certain angle of arc driving duration
    before needing to be reset to bring the rod back to
    the retracted position, bear the fin against the rod, and
    rotate the scope relative to the fin for another period
    (minutes, hours, depends on your mechanical design)
    of driving.

    Another simple / cheap option is again basically like
    a dobsonian idea; instead of using a very very fine
    toothed right ascention gear and a fine gear to rotate
    against it, just abandon the gears and use a drive
    circle attached to the rotation axis of the scope that's
    not a gear at all but is just a very large diameter CIRCLE
    that's FRICTION DRIVEN by the rotation of a small
    motor mounted next to it. If you use something like a
    small circular table top, circular piece of plastic
    countertop, circle-cut plywood, etc. it's hardly difficult or
    expensive to get a drive circle that's perhaps two, three,
    four feet in diameter. Bicycle wheel, same idea.

    It's not particularly in the way, heavy, or inconvenient
    if your scope tube is even longer (e.g. 12" F/5 = 5' long
    tube or 8" F/8 tube etc.), and the drive circle can even
    be disassembled into two or four pieces if need be for
    transport / storage compactness.

    So given something like a 3' e.g. diameter drive circle,
    that's 113" of circumference over 360 degrees of arc.
    Drive against it with something like a 1" diameter
    friction wheel and that's a 1:36 rotation reduction right
    there from the gear ratios. So for the big wheel to
    turn at 1/1440 RPM (1 turn/day) the small wheel
    would turn at a rate of 1/40th RPM which is a lot more
    easily achievable with common gear box reduced motors
    due to the large drive circle to small drive rotor ratio,
    and it'll help minimize the effect any irregularities in drive.
    The friction drive against the drive circle is of course
    easy to mechanically or electrically (solenoid/spring)
    disengage for manual repositioning of the scope, and it's
    advantageous since there is no precise alignment or
    gearing of the scope to the drive motor, so no worries
    about stripped gears or precise assembly.
  12. Jun 13, 2007 #11
    Wow - heaps of info there, thanks. I'm just going to pick on one or two points at this stage though...

    First, I've looked at AC motors, but the mount needs to run from a portable power source. The addition of inverters etc makes things a little expensive.

    So onto the DC motor comment. When you specify "little energizing current", is that to limit the rate at which the batteries powering the motor are drained, or is there another reason?

    Thanks for all the other design info - I've read of some very weird designs....
    I have it pretty much nutted out now though, I've found some reasonably priced, AGMA 10 quality wormsets, and will use a couple other spur gears to couple the motor to the worm. The only thing remaining is the motor selection. I'm considering a Como Drills 919D series geared motor, but I'll have to check out a few things with it....
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook