Motor Tuning for IH Mill

[SMW Precision]

I have heard tuning your Gecko's makes a large difference in mill operation. Thought I see what others have done in tuning Gecko drivers to servo motor. Particularly without using an O-scope.

Thanks,

Ken

[Cruiser]

Doing your tuning without a scope is easier than you might think. It will just take a little trial and error till you get it. what to look for is smooth snappy responce yet quiet and maybe a gentle purr while sitting idle. You will set the currant limiter all the way high for the z but only bout 55% to 60% for x & y, this is pretty much the way Aaron of IH expained it to me and when i did it I was happy afterwards. This will need some tuning in Mach as well so everything is ballanced. Mach 3 settings will be 20,000 by 100 by 6 for tuning and you'll see from there.

[SMW Precision]

Originally Posted by Cruiser Doing your tuning without a scope is easier than you might think. It will just take a little trial and error till you get it. what to look for is smooth snappy responce yet quiet and maybe a gentle purr while sitting idle. You will set the currant limiter all the way high for the z but only bout 55% to 60% for x & y, this is pretty much the way Aaron of IH expained it to me and when i did it I was happy afterwards. This will need some tuning in Mach as well so everything is ballanced. Mach 3 settings will be 20,000 by 100 by 6 for tuning and you'll see from there.

Thanks for the directions, Cruiser.

[youngfg]

When I first got my mill I tuned my drives the way Cruiser described. About a year later I thought I would go to the trouble of using an Oscilliscope to tune the drives. After a few hours of work with the scope I came up with the same settings I had previously.

[SMW Precision]

Originally Posted by youngfg When I first got my mill I tuned my drives the way Cruiser described. About a year later I thought I would go to the trouble of using an Oscilliscope to tune the drives. After a few hours of work with the scope I came up with the same settings I had previously.

I am hearing with properly tuned drives repeatable accuracy is in the range or 0.0001 to 0.0002 of an inch. About 5 to 10 times what Aaron had shown on the website. What kind of results were you able to get after tuning motors?


Thanks,

[youngfg]

I have never checked that.
I was having servo fault problems, that's why I though I needed to retune the servos, but in the end it was that I had the acceleration rate set too high in mach.

[BobWarfield]

Originally Posted by SMW Precision I am hearing with properly tuned drives repeatable accuracy is in the range or 0.0001 to 0.0002 of an inch. About 5 to 10 times what Aaron had shown on the website. What kind of results were you able to get after tuning motors? Thanks,

I'll bet you'd have to have everything extremely well adjusted and map the leadscrews to get there. That would be quite an exciting result!

I'm planning to map my leadscrews, and this is one reason why I bothered to install a cheap DRO on a machine I intend to CNC anyway.

Best,

BW

[SMW Precision]

Originally Posted by BobWarfield I'm planning to map my leadscrews, and this is one reason why I bothered to install a cheap DRO on a machine I intend to CNC anyway. Best, BW

Bob,

How are you planning to input your results and what CNC program are you going to use. I am working on form of this tech and wonder what you have planned.
IH mill kit has Nook ballscrews which greatly helps in accuracy versus leadscrews.

Thanks,

[SMW Precision]

Originally Posted by youngfg I have never checked that. I was having servo fault problems, that's why I though I needed to retune the servos, but in the end it was that I had the acceleration rate set too high in mach.

Same problem I have with computers "They do what I tell them to do; Instead of what I want them to do!". If you get a chance check the repeatable accuracy between a couple of parts. I talked to someone that was doing high quality molds with an IH mill. He said it was all in getting Gecko's tuned properly.
Interesting that hand tuning versus scope tuning yielded identical results in settings. I am very encouraged by your comments.

Thanks,

[youngfg]

One thing I did was put a wrench on the end of the ball screw and try to turn it with the axis energized but stopped. You can watch your error voltage and see how it compares to your actual error and you get an idea of the stiffness and torque of the drives.

[BobWarfield]

Originally Posted by SMW Precision Bob, How are you planning to input your results and what CNC program are you going to use. I am working on form of this tech and wonder what you have planned. IH mill kit has Nook ballscrews which greatly helps in accuracy versus leadscrews. Thanks,

Ken, this is a pretty straightforward function in Mach 3. You basically build a map of corrections at various locations on the ballscrew. I think Mach will take as many as you want to enter for each axis, and then it will interpolate between them, which is pretty slick. I'll probably build a map on 1/2" intervals for full travel of each axis. It's just a matter of spending a tedious afternoon commanding jogs to a location, reading the "real" location off the DRO, and writing it all down. When you are done, you key it into Mach 3.

Good call on the ballscrews. I'm not sure leadscrew mapping would be worthwhile without ballscrews. Seems like backlash could overwhelm any screw error.

Now, here is the interesting question for a fella like you working with the IH folks. There are folks who have gotten Mach 3 to read linear scales. You can look that up or ask about it on the Mach 3 boards. There are a couple of cool things you could do from there, if you sell a DRO together with some nice Mach 3 add-ins.

First, it would be straightforward to write a script that automates the ballscrew mapping. You would simply automate what I previously described:

- Home the axis.

- Sit in a loop jogging 1/2", 1/4" or whatever floats your boat.

- After each jog, have your script input the correction to Mach 3's table.

- Stop when you get to the end of the axis.

Seems like you might be able to compensate for wear over time and it might be desirable to perform the calibration quarterly if you have it automated.

Second cool thing would be to do something similar to what Rogers Machine offers with their encoder board. They have some simple macros that will force a Feed Hold if the position reported by the encoders gets too far from where Mach 3 thinks it is. If you had that functionality based around linear scales it would be particularly nifty. For example, it would identify when thermal expansion was creating a problem because you'd been running the machine for too long.

Just a couple ideas.

Best,

BW

[Cruiser]

Your making a simple error in your interpretation of what Aaron was saying on his website, He didn't say "accuracy" he said "resolution" two differant things and altho resolution is part of accuracy it is not the final word. And, the resolution he was referring to was with the optical limit switch he was testing for repeatability. As far as functional accuracy, I find that with everything done as well as i could get them, then interpolating a circle, that the measure showed my circle within a .001" tollerance. and the linear with my 3" indicator showed a possible error of .0015" ! I'd call this pretty good for this system. I noticed a gain loss gain repeat of .oo1 on the linear of the ball screw itself and again i'd say good enough. The biggest problems so far i've found would pertain to the mixing of electronics and adjustments of mechanics then the issue of castiron particles in spindle bearings i'm assuming were in the tube when assembly was done, which of course failed the bearings. Now i'm also finding out that with the limits pulled off the combo board and on a seperate port that my drives are occasionally hit causing lost position ! i'm still working on that one, and my hitachi l200 is occasionally faulting out on shut down. So i may slow it down some more. Its been one thing after another and i hope it gets stable pretty soon.