Re: SLO-SYN 12.7 Amp motors

Jon Elson wrote:

> > Ahha (and Flashcut also) uses external hardware to avoid
> > this problem, but I am reluctant to adopt that approach until the "software
> > only" approach has been thoroughly developed. Other "software only" solutions
> > avoid the problem by going into a tight loop in their program with only
> > periodic polling of the keyboard buffer to check user input. This option is
> > unavailable to us as we want to multiprocess and keep Xwindows alive, etc.
> > The problem comes in when you try to control a large machine (like a
> > Bridgeport) at high speeds (60+ inches/minute or thereabouts). Since the
> > torque output of stepper motors drops with speed, any timing variations in
> > the step rate (at high speed) can lead into resonance problems or stalling.
> > Fred has made some good progress lately and will have an improved version of
> > the stepper software available in the near future.
>
> Well, that's just another argument for servos!
>
> Jon

I have finished the mechanics on my big gantry mill and am ready to hang motors. I am faced with big bucks to
buy 4 axes of servos (400+ in oz peak torque) and amplifiers.

What got me interested in this thread is an acquisition I made today. Due to a facility move at work I got 8
slo-syn MH112-FJ4201 1200 in oz steppers with 3180-pi mosfet indexers. These monsters give 900 in oz at 4000
half steps per second and 100 in oz at 20000 steps. Max speed appears to be 2500 rpm. Even with the typical midrange
stepper torque droop these things have more oomph that my motor sizing software calculated is needed. Inertia match is good too.
The indexers are rs232 or parallel programmable in EIA L, H, N, G,X, and F codes. They use 20kHz switching supplies and plug in
the wall. (120v 5A)

I had no intention of using an open loop system on my machine. I have seen the unfortunate effects of
slipped steps before. That would be fairly violent when one is moving an 800 lb gantry.

I sure would like to save a couple of grand though. Although new or used for experimental stuff only,
the ten year old indexers run at 9600 baud only. I calculated that my mill has 5.5X10^18 addressable
locations using 1024X4 renco encoders- sounds like a real bad combination! I don't think the indexers
can do precise coordinated moves either. Too slow.

I will consider bypassing the indexer and just using the amplifier portion under PC control with
encoder feedback for at least slipped step detection and correction. Will EMC do this?

I also got some .5 and .25 hp pm dc 90v motors. With encoders these are a fair match. But no amps
other than a bunch of SCR speed controls (no good for cnc).

If I don't use the steppers I will need:

1) 4 10 amp Copley brush servo amps ($1080) or
2) 4 Apex PA60 10A hybrid PWM servo driver chips ($520 + a month of weekends fiddling with heat sinks
and power supplies) or
3) Jons amp design

and spare motors and Renco encoders.

I would appreciate any comments anyone could offer.

>
>

Leslie M.Watts
Illinois Tool Works
Corporate Advanced Technology group
(847) 657-4559
http://www.netcom.com/~leswatts/leswp.html