Re: [CAD_CAM_EDM_DRO] Re: mounting encoder with pulley

Alan Marconett wrote:

>Hi Caudlet,
>
>That's what I thought! It just multiplies, and offers a choice of
>common +5 or common 0V.
>
>So I still need a way to divide the encoder quadrature signals!
>
>Alan KM6VV
>
>

It's funny - I just found a reason to use an encoder divider rather than
a pulse multiplier. I hadn't been able to think of any real advantage
until now.

The situation that prompted this is that I have a 1000 CPR encoder on a
5000 RPM motor. To run it at "ideal" speed, it should be at 4000 RPM.
That's 16M ppm or 266.66k pps, 20M ppm or 333.33k pps for full speed.
Since that's too fast for a Gecko, I can't get into the optimum speed
range for the servo. The application is for a rotary table. The motor
is only 50 oz-in, and I intend to run a large-ish rotary table with it.
I'll need to do some reduction to get the servo into the right speed
range. I also definitely don't want a 75-pound rotary platform spinning
at 45 RPM! I can get an 8:1 reducer pretty cheap, so that would be 4000
* 8 = 32000 ppr at the table crank. The table is a 90:1 ratio (1 crank
turn = 4 degrees table turn), so I end up with 2.88M ppr on the table
(and an effective torque of 36000 oz-in!). Of course, machining
something on a long radius will be slow, but requires torque. Machining
on a smaller radius will need more rotational speed (to maintain tool
speed in the part). Now, I don't need 2.22 steps per arc-second (!)
resolution, and the drive system won't be that accurate anyway. So, I
can easily sacrifice resolution to get around the speed limitation of
the Geckodrive. Driving the motor at 250kpps, I can only get it to 3750
RPM. That's not bad, but it is a little below the 80% speed that's
recommended for servos (it's 75%, not too far off). That would also be
a maximum speed, not the "normal" speed. Ideally, I'd like to be able
to get the motor to full speed for rapids, but expect to use it at
20-80% most of the time.

It looks like I need at least a 2:1 divider for this. So maybe I'll
design one after all :)

On the original subject (using a belt reduction for an encoder), there
is one possible problem. To eliminate backlash in a toothed belt
system, you have to put tension on the belts. This puts a radial load
on the encoder shaft, which may be a bad thing - they're generally not
deisgned to take much sideways pushing. Industrial encoders with ball
bearings should handle it OK, but any of the plastic case inexpensive
ones (Renco, HP, and US Digital low-end models) may not.

- Steve