intro and questions

[scudzuki]

Hello all,
This is my first post so a little about me. Worked for 15 years in manufacturing... machinist/programmer/designer/supervisor for 10 years building rollforming equipment and 5 years moldmaking. Now I am an independent small business computer network consultant. For years I've longed for access to equipment like I had in the good old days. I bought a South Bend lathe on eBay recently which got the ball rolling. I am picking up a CNC retrofit RF-45 mill (retrofit done by Boca Trading) in a few weeks which the seller boasted is equipped with servos rather than steppers. Only after firming a deal with the seller did he disclose that he had problems with repeatability when rapiding over 50 ipm. This kind of conflicts with what I've learned about the superiority of servos over steppers, i.e. theit immunity to lost steps. Possibility that there are other forces at work? What other factors affect repeatability in a servo drive? Thanks in advance to all those with way more knowledge than me.

Joe

[Stepper Monkey]

Properly designed, neither type of system can just "lose" steps in normal operation. Ever. Improperly designed either type of system can, anytime.

There are many types of servo system, and only the most expensive and complex actually hold claim to some of the features servo salesmen often use as examples of servo greatness over steppers. Even then they have to be designed properly and engineered for the specific machine and purpose to work properly. It takes some knowledge and engineering. You can just bolt on the needed parts and have a system that works, but not one that works well or properly all the time, much less optimally. Unfortunately this is pretty common with both types of system. Some cheaper brushed servo systems are far worse than any stepper system ever could be in terms of repeatability.

Servo systems aren't all the same, find out what type is specifically mounted on this machine first!

[The Blight]

I bet it's a servo system that runs closed loop between the driver and motor. The software doesn't know what's going on. Some of these drives have a fault line which can communicate with the software to detect lost steps. I don't know if the driver can remember how many steps it's off by when it has been reset, so you might loose some steps there. A servo motor is after all just a motor with feedback. A servo system on the other hand should have feedback to the controller which in turn can take actions against lost steps and other variables.

Even a properly designed system has it's limitations. So even a slow system can be properly designed if it has known limitations. If the limit on this system is 50ipm, and it can hold this during normal operation, then it's a properly designed system.

But as Stepper Monkey pointed out. What servo system does this machine use?

[scudzuki]

The servos are Reliance Motion Control, NEMA-17, 1/4hp servos. I asked the seller to try to figure out who makes the drivers and such. He claims that he only had issues when he exceeded 50 or 60 IPM, below that it would hold and repeat a thousandth. One item of note that surprised me, the controller (PC) does not register motion when the axes are moved manually. Since the only mill I ever ran that had manual controls was a retrofit knee mill with glass scales for feedback, I find this alarming, but in retrospect, any other machine that I jogged using the control could have been reporting without feedback also. Not like I would have known if the display was reporting actual position from rotary encoder feedback or just where it thought it was positioning... but I have no evidence of ever losing steps on those machine, either (TCs, VMCs and VBMs, not "industrial hobbyist" type machines).

Joe

[The Blight]

That sounds strange. What software is it running? When jogging, it should update its position.

[scudzuki]

Likely it does update when jogging. I said it does not update when you turn the handwheels. Now that I think about it, this could only be possible when the servos are off, therefore no feedback. This does not explain how it loses steps, however.

Joe