Re: [CAD_CAM_EDM_DRO] Re: Replacing Anilam Controller with EMC? / Another Choice

jim_clamp_it wrote:

>Hi,
>
>I'm also converting a Bridgeport Mill that had an Anilam Crusader II
>control. I've tried LinUx Redhat 5.2 with EMC but, I'm a bit
>challenged on the PC side of a CNC system so. I went with Mach2 since
>I've used it with a stepper driven router table. The machine has
>Westamp style servo amps (+/-10v). I bought three Rutex 997's (they
>have a newer and better version out now) step/direction to +/-10v to
>interface the amps to the PC. The motors are SEM's 30Amp on X & Y
>with a 20Amp motor on the Z. The linear encoders on X & Y pulse
>every .0005" so, with a little backlash from the motors to the table
>the motors will never be quiet and nasty things happen. I had to
>mount rotary encoders on the motors. There's a 2:3 timing belt
>reduction from Ballscrew to Motor. I've tried the encoders on the
>screw but, I've found the motors quiet down with just a little
>pressure pressed on the belt with that configuration. I think the
>more ridgid the coupling of the encoder to the motor, the more the
>motor will settle.
>

Well, yes, if quiet motors are your purpose, that's right. I have a similar
system, with encoders on the end of the leadscrew. The belt is EXTREMELY
tight, tensioned by a bearing with a wheel pressed on it that is adjusted by
an eccentric. I believe the belt tension runs something like 100 Lbs.
I would not have done this except the original setup was done that tight.

> The linear encoders are most accurate but there's
>really is no deadband. The D/A conversion has to have at least a
>little error bouncing back and forth to hold position. I don't want
>to replace ballscrews so I'll live with .0005" or so of backlash.
>
>

If you'd used EMC, there would be deadband, adjustable from the software.
Also, you could graph out the actual position vs commanded during a move
to verify the accuracy of the servo following. With the Rutex 997s you
only know the error is less than the following error limit, which I was not
able to determine from their documents. So, your actual position is
+ or - how much? Do you know?

>So far, I have the X axis motor strapped to the table (not belted to
>the screw) so I can tune the servo. That's another reason I like the
>encoder on the motor rather than the screw. When the screw turns my
>table moves and I need to have a fast hand for the kill switch.
>
>

But, tuning the servo WITHOUT the machine load is pointless. It will
change when loaded.

>The motor is nice and quiet and responds well to step and direction
>commands. I mounted a 200 CPR, $15 ebayed encoder to the motor. 800
>steps rotates the shaft one complete turn. Rutex has free software
>for download to help tune your servos to their products. The software
>interfaces thru the LPT port and will plot nice graphs produced from
>data between commanded position to actual position. (It's easy)
>
>

Hey, I'd like to compare that with what I've done with EMC.

Jon