Re: My stepper getting hotter

--- In CAD_CAM_EDM_DRO@y..., "mariss92705" <mariss92705@y...> wrote:

> John,
>
> Everything marked "Amps" was measured, everthing marked "Watts"
> was calculated. I use ACAD for nearly everything; the data was
> keyed into ACAD for the graphs but if you like I will enter it
> into Excel and post that as well.

No need - you've already done far more than I expected. Typing
lots of numbers into a spreadsheet is a pain - been there, done
that. This is gonna sit on the back burner until I get finish
another project that I have to do, but then I'm gonna make a dyne.

>
> You pretty much have it right in your assumptions except for one
> point; data indicates motor losses are independent of load, not
> speed. You may have misread that.

Right - you did all your testing at a constant speed.

>
> Motor power output is the difference between the max load and
> no-load currents multiplied by the power supply voltage. The
> result meshes closely (+/- 2%) with dyno measurements.

So the "power out" graph is calculated at each data point, and
checked against measured dyno results at a few points?

>
> Speaking of dynos for testing step motors; a DC brush
> servomotor makes an excellent load. <snip>

I was thinging more along the lines of direct torque measurement
by putting the load motor (or the test motor) on gimbals that
would let it spin around the shaft centerline. On a dual shaft
stepper this is as easy as another pair of bearings, one on each
shaft. The new bearings support the shaft, and the motor frame
and stator are supported on the shaft by the stator bearings and
free to spin - until the lead wires wind up. ;} Prevent it from
spinning with a lever arm of known length, and measure the force
at the end of the arm to determine torque. I've used dynes of
this basic construction for larger motor tests, and they work
well. Linearity of Kt vs. armature current, etc., is unimportant,
in fact the loading device can be extremely non-linear.
Some of the dynes I have used have DC motors (running as
generators) for loads, but the big ones use eddy current brakes
that have non-linear torque/excitation curves which also
vary with speed (no eddy current torque at zero speed).

Your suggestions filled in the other half of the problem - control.
I don't have a servo motor or amp, but your description started
wheels turning in my brain. The key is that "percent load" is
proportional to position error between the stepper's commanded
and actual position. (Actually sine of the error.) I'm used to
thinking of load as proportional to speed error, not position
error.

How about this - use a stepper as the load motor. Feed the same
step signal to both stepper drivers. Both motors turn at the
same speed, and develop no torque. By moving the lever arm,
I can rotate the two motor stators with respect to each other.
As I do, the motors will develop torque, which I can measure with
the lever arm. So I can plot torque as a function of relative
angle at any speed. Then I fix the angle at whatever gives max
torque (should be 90 electrical degrees), and I can measure
torque vs. speed.

>
> There are other refinements necessary but you get the idea.
>

Yeah - the devil is always in the details....

Thanks for everything,

John