[Nova-developer] RE: About microstep ( and other solutions ... )

[Liam G. <l_g...@bi...>]

Hi everbody,

I'm about out of my depth here, regarding electric motors, but I'm interested in the HP motion controller chip.

Would it be possible to transfer this thread to the nova mailing list. Registration can be found  at https://sourceforge.net/mail/?group_id=4413 . The mail site is sometimes slow, so it may need several attempts at loading the registration page. The list is private to us and archived by sourceforge so we can browse it at our leisure.

Thanks 

Liam

-----Original Message-----
From:	fma...@es... [SMTP:fma...@es...]
Sent:	31 May 2000 09:42
To:	l_g...@bi...; jc...@ca...; E97...@st...
Subject:	About microstep ( and other solutions ... )

Juan wrote:
>Connecting a small microcontroler (or PC) its possible to command two
stepper
>motors , with a precision in microstep mode up to 1/32 of a step, what
would give
>for a 1.8º motor a total of 6400 microsteps per tour.

>Logically the motor can be operated in full step, halfstep and microstep
modes, and can be connected directly to
>the parallel port of a PC.

I have speak with Massimo about this, and this is our opinion :
-------------------

Mainly I think you are in right...

Just 1-2 years ago I've thinked the same...
BUT...
there are some practical problems related to the phisical limitations of
stepper motors.

I speak about the high NON linearity of the microsteps.
If you run a stepper motor at 1/32 1/64 of step, you will note that in the
"full step " condition (when one phase is at maximun of the current and the
other to the minimum) each microstep is much smaller than when you are
between two full steps.
The progress is about a sine wave and is due to the carachteristics of the
magnetic field between two steps.

Some american factories have well thinked to sample the error and to
correct
it varying the current of the phase for each microstep by an eprom stored
table.
BUT
1) the total microstep resolution is reduced
2) surely the non linearity of the microstep is different from one step to
the other, this because any little difference in the coils can change the
magnetic field distribution.........

Another limitation is due to the elastic properties of the magnetic field.
If you put a stepper motor in full torque but stopped state and try to move
the shaft you will note that you are able to bend the rotor.
This is because the magnetical attraction of the poles is elastic...
The amount of this error is not big in normal condition ... 2-3' of degree
but for a motor with 180 steps it means about 1-2 microsteps if a 1/64
microstep drive mode is used !!!
So if you have an unbalanced telescope, the result is a different tracking
on the telescope position, because the load applied to the motor shaft is
different from one telescope position to the other...

Our Cavezzo telescope have a great tracking capacity because the stepper
motor coupling is greatly reduced by a gear, and the motor run at 100
step/second (about...), in this manner every non linearity of the steps are
unsignificant and the different telescope position don't change the
tracking
precision even with unbalanced telescope.
By this way we have reached "no moved stars" in a ccd field even with 20
minutes exposures on celestial equator !!!!!!!!
The counterpart is the impossibility to achieve the high moving speed
needed
by an automated telescope system.
In fact our telescope is frictioned so we can move it by hand and actually
all the positioning are made manually.


For this reason I have excluded exotic microstep drivers from the Cavezzo
automation...
Actually we are evaluating brushless motors...
the cost is much higher, but the torque at very low speed is much higher
then any other motor, the resolution achieved nowdays is very high thanks
to
sinusoidal drivers and is practically limited to the encoder resolution.
Japanese Fanuc for example achieves 1.000.000 pulse/revolution !!!!!
for a cost of about 1000 Dollars...

The secret is the encoder.
Even a brush or brushless motor have the non linearity problems of
steppers,
but the position feedback gived by the encoder eliminates every error !!!

Brush motors can still have some imprecision near the switching point of
the
brushes, brushless motors are made in a three phase configuration to
limitate this problem.

Dc motors have brushes, wich limitates the life of the motor itself
compared
with brushless...
but this is true if it runs many hours (for example in an industrial
machine
tool... 8 hours/day !!!)
But if the cost is a problem DC brush motors are still nowdays quite good
with an uncomparable price.
This because the DC drive technologies are much stable and tested.
Moreover dc motors since to 200W can usually be found at scrap prices !
Simply by adding an encoder you can achieve a precision/high speed drive.
Hewlett Packard sales complete motion controller chip at a quite good
price...

Just connect the encoder outputs to the chip, the two drive output to an H
bridge witch drive the motor, and the 8bit bus to your system and send
position command strings !
moreover you can buy it even in small quantity directly in internet from HP
!!!!!

If you want more information... write me !
 (fu...@2a...)

Aboyut the external intelligent chip philosophy...
I agree completely !!!
In our Cavezzo project we use the same philosophy !!!
We have a Z80 system which take care of motor positioning and encoder
reading...


Bye !
Fausto