Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Wayne on September 18, 2007, 11:06:41 AM
-
Hi
I started with Mach 3 a few weeks ago and haven't been able to resolve an issue, so I am looking for some help.
I have a Gantry Torch cutting setup and am using the plasma profile. When I do a test like X30 Y30 it is very very accurate. When I do a 4 in. circle my Y axis is +0.187 in. and X axis is -.0625 creating a very symetrical oval. All motor movements are very smooth.
Here is what I have tried so far; Exact stop mode enabled
Eliminate any compensation modes
absolute and incremental modes
G12 mode
changed ramping rates, this did help, X axis has 2/1 ratio, Y axis is 1/1
Backlash on both axis is less than .015
Switched Gecko drives, no difference
installed new shielded cable thru the entire system
1.8 Ghz computer with 1gig ram (ouch!)
I have tried everything I can come up with, maybe someone can spot a simple fix I have overlooked. Any help is much appreciated !
Very Frustrated
Wayne
-
Hey Wayne, can you post your code?
Brett
-
me too
i have a little test g-code that cuts a 1.0 inch circle and a 1.0 inch square in the same run.it cuts them so i can
mic them from the outside.the circle cuts oval and the square x =1.006 y=1.009
i also have checked my machine all over.last friday i started going back to previos versions looking for the change
have not found anything yet.
bill
-
Here are 2 examples and setup files.
-
Sorry didn't get them all in that post. I'll try again ;)
-
Have you checked your steps per unit? also do you have much for backlash?
-
Hi
Yes, steps per inch are correct and when I do a straight line test the measurements are right on. When I make a square or a circle ( something enclosed) my Y axis is long and my X axis is short, usually Y is about
2 % long and X is about 1 % short. Backlash is .010 to .014. I did try programing in for backlash, it did not help. The error amount is proportional to the dimension of the item. aka. 2" circle off .125 4" circle off .250. I was thinking maybe noise or the Gecko drivers were off, so I switched x & y drivers, it had no affect. Then I replaced all cables with better shielded ones and put each driver in a separate metal encloser to reduce the chance of noise or false triggering, it did not help either. I am using a Campbell Iso board and a new Printer Port cable. The Iso Board is also in its own metal encloser. I've tried everything I can think of and am totally stumped !
Thanks
Wayne
-
When you say the straight line measurements are spot on, how spot on, what are you using to check the distance, and over what length.
If the error is proportional then the steps per inch/mm must be wrong. Are your backlash figures inches or metric?
when you check the distance moved have you removed the backlash first?
Graham.
-
My straight line measurements were 48 inches long, both x and y axis, measured with a 12 inch digital caliper and manually held by hand to take up backlash. I am using ground and polished acme threaded rod for travel. My measurements are all in inches. My x axis motor has a 2/1 reduction. !.8 degree stepper and 5tpi acme threaded rod and 10 microstep gecko drive -----20,000 spi. My y axis is 1/1 ratio with 10 microstep drive, 1.8 degree and same 5 tpi threaded rod-----10,000 spi. Testing is done at normal cutting speeds 8 ipm to 17 ipm.
Thanks
Wayne
-
if it is noise try this as a file
G00 X0.0
G4 P1
G00 X6.0 Y0.0
G12 I3 F100000
M47
Let this run for 20 min and check to see how far off it is from where it started with a dial indicator. I would set the inicator at X 0.0 and check it during the 1sec dwell that you have in the file.
thanks
Brian
-
OK, sounds good. I'll give it a try tomorrow AM. Thanks for the idea.
Wayne
-
Well I tried the code, ran it for 1/2 hr. Checked it during and at the end, it came within .001 inch of starting position every time. Truly confusing how it can be so deadly accurate, but my circles are elongated. ????????
Thanks for the test, I am open to ANY other ideas or suggestions. I use this in my Welding and Fabrication business so it is very important to me to get it right.
-
Well... I think I need to see what we are working with. Do you have a Pic of the machine and what you are using for a drive on the machine? Also can you tell me if the rate that you go around the circle changes the amount of ovaling that you have?
-
Well the only thing I can think of - and it's a bit of a long shot, is that the X and Y axes aren't at 90 degrees to each other. So a square would be machined into a losenge shape and a circle would be made into an oval. Straight line movement would remain accurate though. Might be worth checking. All the best, Martin.
-
We tested the mill here and we are not seeing the problem that you are having... Could you please check that the axis is square.
Thanks
Brian
-
I had the flat spot on circles for a while now. I didn't think that if X & Y axis was a little off square it wouldn't matter. After trying everything else I tested the Square & it was off just a little, Squared it up & that's all it took.
Wouldn't think a small thing like that could cause so much of a problem.
Thanks
Ron
-
One of the things I picked up on ebay is a nice granite angle plate that is 9"x9"x9". I just set it on the table butted up against the axis way ( stationary table design) and I can check all 3 axis with a dial indicator in just a couple minutes. Handy for When I crash!