[amsat-bb] Re: inquiry about homebrew az-el systems

Gus 8p6sm at anjo.com
Wed Feb 27 23:18:17 PST 2013

I particularly like the idea of using bicycle gears/chain for gearing.  
Whatever we come up with, I'd prefer if the bulk of it was available 
locally, wherever you live.  There are no Home Depots in my country, but 
there are lots of bicycles.  Perhaps we can use the frame of the bicycle 
to provide materials for the main structural component too!  So, one old 
bicycle could provide gears, bearings, structural components, and more 
perhaps? :-)

I'm fine with the Arduino as a platform for the software.

The multi-turn pot would obviously work, but I'm a little dubious about 
it.  They aren't that easy to source, and would have to be connected to 
the rotating shafts somehow that there would be very little backlash and 
no creep/slippage.  I'd thought of the sensors from a couple old optical 
mice, mounted right over the shaft, and detecting rotation directly, but 
after a bit of googling, I'm not sure we can easily pull off movement 
info.  Then I thought about a hall-effect switch mounted near one of the 
(bicycle?) gear wheels, detecting the passage of each tooth.  An optical 
sensor, looking through your bicycle chain, detecting the passage of 
each link?  Or a simple micro-switch, being bumped to produce a pulse 
for each link in the chain?  But how would you determine direction of 

Perhaps a design that can accommodate either a voltage variable position 
sensor, OR a pulse train, making it easier to build the machine with 
whatever parts are available?

I'm chucking out a lot of silly ideas in the hope that someone can use 
them as the starting point, and come up with something practical.

On 02/28/2013 12:49 AM, Lizeth Norman wrote:
> Gus,
> I like your thinking. Simply put we need to make feedback position
> sensor with associated motor/motor control stuff.
> The motor rpm's and gear ratio are functions of the number of degrees
> per unit of time, as you say.
> One simple solution that I was going to try was  a reversible dc gear
> motor being polarity switched by a dpdt relay, and a three turn pot
> for position sensing. The pot is 16$US. Per copy. Need two as well as
> the associated hardware. Scaling the dac on the uP is a nit.
> Then again, since we can fiddle with the values in software, so our
> inputs can be anything (from a hardware standpoint) of reasonable
> accuracy.
> Just thinking, why use gears and gear motors?? Why not a dc motor,
> bicycle chain and sprockets? The HD bearings for the shaft are $6.95
> as I recall, a piece.
> Still set on using stock sizes/materials if I can. This would give me
> a start modelling this stuff with CAD and make it easier to source.
> A note regarding your comment on the  Arduino. It really does not
> matter what we use  in terms of antenna rotators at this point. I
> think that if we coherently model the design process, someone in cost
> reduced circumstances could reproduce whatever we decide with far more
> basic tools than available to us. The key is that there is a
> microcontroller with tons of IO attached to a serial port...
> Or then again, CO7WT is bit banging the parallel port. Now that's the way to go.
> blah, blah, blah,
> Norm n3ykf

73, de Gus 8P6SM
Barbados, the easternmost isle.

More information about the AMSAT-BB mailing list