Hi All,
I have been designing step motor drivers and multi-axis
controllers as a consultant to several companies for over 15 years.
Almost all of the drives use microstepping, but there are a few
half-step designs mixed in.
There seems to be a lot of confusion about lost steps and current
reduction. I hope the following helps to clear the mist a little:
1. Stepmotors cannot loose one step, but the driver can. A 1.8 degree
200 fullstep/rev motor can only slip in increments of four fullsteps to
one of the 50 indent positions per revolution.
2. Drives can loose or gain steps for several reasons, but one common
cause is noise on the input lines. In the old days (before 9/11) the
input opto-coupler acted as an effective low pass filter with about a
200KHz cut-off frequency. New opto-couplers are good up to 15MHz, and
unless the driver design includes a simple low-pass filter, there can be
lots of false steps.
3. I never used a simple RC for current reduction. It was either a
timer IC or a timer function in the microcontroller. The response to
full current was typically less than one millisecond, and this did not
give a problem unless the motor was supporting a load that could produce
a torque overload at the reduced current. If that was the case we did
not use the current reduction feature.
Regards
__________________________________________________________
John Harris
E-Mail jdhhar...@customstage.net
On 9/12/2011 8:52 AM, gene heskett wrote:
> On Monday, September 12, 2011 10:10:17 AM BRIAN GLACKIN did opine:
>
>> Gene,
>>
>> You mentioned that you are having/had issues with the idle current setup
>> on your new drives. The problem (imho) is that the hardware ICR
>> solutions require that steps be lost in order to wake the chip. So no
>> matter what, at some point your likely to have missed steps. I played
>> ad nauseum with the velocities and accelerations and kept getting
>> frustrated by the creep (lost steps) on the Z axis of my gantry router.
>> Since I mainly cut 2.5 D, the Z is idle 95% of the time with very
>> short moves deeper into or out of the cut. I do not know how many
>> times my router drove across my 2' X 4' surface at full depth on G0
>> back to the origin. Its amazing how powerfull 200 oz-in stepper motors
>> are.
>>
>> The Hobbycnc Pro board also has ICR. Take a look at the Wiki regarding
>> this, Kim Mortensen has a nice writeup on the issues with ICR that
>> probably go beyon just that board.
>>
>> ICR on the HCNC board uses an RC pair to trigger the reduction. I
>> suspect that it typical of most setups. In my case, I simply
>> eliminated the RC pairs and direct wired the axis chips to the
>> parrallel port. Amp enable from EMC triggers the chip to wake before
>> any step is issued. This eliminated all issues I had with ICR.
>>
> I found that these amps have an "enable" input, but that it was in fact a
> disable if driven. And since emc enables the amps when out of e-stop, that
> isn't a lot of help anyway.
>
> But thanks for the heads up about lost steps being rather endemic to ICR
> (nice abbreviation) equipt drivers. I have a 1" stroke dial indicator, so
> I will setup some lost steps detection moves before I actually make any
> more swarf with it.
>
> The ICR recovery can be many times faster if a re-triggerable
> multivibrator/timer is used, and in modern integrated circuitry that is far
> cheaper then an rc circuit. I would think, as an electronics type, that it
> would be more of a function with the motor inductance impeding its being
> ramped back up to full current. This could take a small, but important
> amount of time. Something in the millisecond range that would effectively
> make the first step a weak one. Since my z motor is a triple stack 425oz,
> it is higher inductance and would likely show that effect first. But there
> is no data included with them on how long it must be paused to do the ICR.
> They only have 2 leds, a green one to indicate power ok, and a red one to
> indicate a fault. And my cabling does not allow an amprobe to be used.
>
>> Just a datapoint.
> I will post what I find. Possibly I can just clip onto the wire insulation
> near the motor and see tha ICR by stray pickup on the scope. That of
> course won't be a quantitative measurement, but a relative one that will
> give me the timing info. In the case of my Z motor, it is an 8 wire motor,
> I could pull a wire nut off and insert my home-made spindle current ammeter
> for a visual indication. At full song, it should be about full scale. IF,
> note caps, the rectifiers I used to make that AC ammeter out of a DC meter
> are fast enough. SI power diodes have notariously slow reverse recovery's
> and likely not very efficient at 200khz.
>
> Cheers, gene
------------------------------------------------------------------------------
Doing More with Less: The Next Generation Virtual Desktop
What are the key obstacles that have prevented many mid-market businesses
from deploying virtual desktops? How do next-generation virtual desktops
provide companies an easier-to-deploy, easier-to-manage and more affordable
virtual desktop model.http://www.accelacomm.com/jaw/sfnl/114/51426474/
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
