PDA

View Full Version : Taig Y-Axis binding?



CROSSHATCH
12-08-2008, 04:09 PM
I have been spending the last few days tuning the Taig. I have completly taken everything apart and have done a tuning on the machine. When it comes to the taig, I like to say that I know the machine pretty well.

Before we get to the problem. After all tuning was done, I recorded 0 backlash on the Z-axis. I found .002 backlash on the Y and a sloppy .004 backlash on the X (Need to retune this). All measurements were used with a .001 indicator.

The y-axis is the trickiest part to tune on the Taig. Mainly because of it's gib set-up. In any case, I adjusted the nut at the center of the screw to get that .002 backlash. I have been running the machine for over a year now with some decent CNC work. The problem I find right now is that when the y-axis is closest to the z-axis column and is being pulled toward the stepper, it's smooth as silk. When the table pushes from extreme to extreme, always on the push does it start to make a GRINDING noise, this bothers me because I know that something is loose or not right enough. However! When I apply a small amount of pressure (5-10 pds of pressure to help the table) the noise goes away. Meaning, that I help the table while the stepper pushes it.

I e-mailed Nick Carter and he thinks there is some vibration which tells me the y-axis is not tuned just right.


Has anyone dealt with this?


-Jason

TMaster
12-15-2008, 02:05 PM
If you can take some video of it happening that may help. I've been running my CNC taig for over 2 years now.

After I set the lead screw nuts I have .001 to .0015" backlash on the dials.

TMaster
12-16-2008, 02:42 AM
First of I would take the stepper off and manually slide the Y saddle casting back and fourth, feel it, wiggle it etc.. see if its loose or too tight even in certian spots.

Just a couple tips,

1. Make sure both the side screws that push the Y axis brass gib against the steel way are both have equal pressure or amounts. You shouldnt have to touch the lower screws underneith ever that could also mess up the gib area.

2. Make sure before you tighten the two 10-32 end plate screws to the saddle casting that the casting is all the way forward (where the lead screw nut is closest to the end plate). This way the leadscrew will be lined up strait all the way through the travel. If not you could have leadscrew binding in the nut because the leadscrew could be at an angle from slop when you tighten the screws.

This seems like a mechanical problem becuase the stepper motor is working fine in one direction, why would it be bad the other way?

Whats your software backlash set at and your speed % of?

CROSSHATCH
12-16-2008, 09:46 AM
It actually seemed to dissipate. The problem that I was having with my motor stalling because the Y-axis nut was so tight at the end of the leadscrew is now solved. I used my drill attached to the stepper and ran the table back in that area over and over. Oiled it it, no stalls, really...no more grinding.

I guess I was in a way lapping the table, but not really lapping...


-Jason

CROSSHATCH
12-22-2008, 03:51 PM
I have been losing steps even with my stock set-up. But I finally took the y-axis motor off today and checked the gibs. I re-tuned it all and am back in business. I ran many tests this morning and everything was cutting great!


-Jason

CROSSHATCH
01-17-2009, 07:55 PM
After a brand new upgrade with a Dell optiplex after everyone said how good it was. Things were looking good until again, without notice the machine skips again on the Y-axis!

At this point I took apart the y-axis once again, cleaned, tuned and tested by hand for tightness/looseness and put it back together. I also swapped Y-axis motor with x-axis motor to see if it's that.

I ran two file and they cut fine. But this is no indication of good news as this seems to happen whenever. It will cut fine one time and then lose steps all over again. Cut the file numerous time fine and then mess up. It only takes one mess up to ruin something very expensive. Something I cannot tolerate.

I will run a series of tests tomorrow in which I will cut the program repeatedly to see what happens. If it fails, like it has been doing so much.


I am going to look past mechanical/ computer issues and focus on the Parallel Cable/ Gecko540/ wiring/resistors.


I really can't take this anymore. I can't have programs fail on and off knowing that a customers part will be ruined.

!!!!!!!!!!!!!!!!!!




-Jason

TMaster
01-17-2009, 08:06 PM
Sounds like you may not have a binding issue, since your problem is off and on.

Could be:
what line of code is it failing on? Or is it rapid? make sure your radius' are more then the tooling radius. I had this problem when I first started machining. Example: radius of 2mm(4mmOD) used with a 4mm OD endmill when backlash is enabled will cause the stepper to stall. I mentioned this to a guy at makers of mach2,3 and he said he would look into it, but that was over a year or so ago. I've had no problems since I corrected this area.

Also try setting your rapid movement less then 25ipm. I found that durring code the steppers would rise 5 to 7ipm over what it was set for. Jogging was fine i stayed at what it was. I run 22ipm for the max rapid now. A dampener may help out with that problem But i have yet to make them.

CROSSHATCH
01-18-2009, 12:28 PM
Sounds like you may not have a binding issue, since your problem is off and on.

Could be:
what line of code is it failing on? Or is it rapid? make sure your radius' are more then the tooling radius. I had this problem when I first started machining. Example: radius of 2mm(4mmOD) used with a 4mm OD endmill when backlash is enabled will cause the stepper to stall. I mentioned this to a guy at makers of mach2,3 and he said he would look into it, but that was over a year or so ago. I've had no problems since I corrected this area.

Also try setting your rapid movement less then 25ipm. I found that durring code the steppers would rise 5 to 7ipm over what it was set for. Jogging was fine i stayed at what it was. I run 22ipm for the max rapid now. A dampener may help out with that problem But i have yet to make them.


It seems to fail at a certain point most of the time, I don't know if it's in a rapid or what. It will cut the first layer just fine, and I say Layer because this is a small pocket with a 2.78 mm cutter with a doc of .70 mm and IPM at 7. And it total depth of cut is about 3 mm, so usually the first layer is fine. Then it seems to gain steps in the positive y direction. It will cut the 2nd layer higher, the third and so. So it looks like there are steps in the material. Why it does this on and off? I have no idea...

You mentioned the whole radius thing. I think i'm fine in that department. But just for the sake. I am going to lower my stepover rate because this was a new file I had to replicate, so maybe my parameters are a little off. And maybe by lowering the stepover rate, I won't be pushing the steppers to hard. EVEN THOUGH, It's more then capable of running at these numbers. I have cut deeper, faster in the past and no problems. But I am going to give it a try. :)

My Rapids are less then 25, 24 ipm I believe. I made dampers, there not on right now. But it just seems all so odd to me for months and months and months I ran these programs and never ever ever dealt with an issue. I once tuned the machined after things were going south and it's been trouble since.

I have tuned the machine, tested for slop, tightness etc. I have spent a long time checking everything on the y axis. Turning the leadscrew by hand, with the stepper on and off. I can turn it by hand etc, I feel no "sticky" spots or anything.


I am going to, if my tests fail now with a swapped motor on the y-axis (stepper taken from x-axis) swap out the y axis controller in the 540 and replace it with a spare.


I'm looking for two things right now that would kidna make me happy, but still be a bummer. I would like to see the x-axis fail and skip steps; This would then focus the problem on the wiring/stepper motor. Or I would like to see it cut the tests fine over and over. If not, I am going to swap the board in the 540.

I have the new computer, so I feel comfortable there. Lets hope for the best, maybe Nick will come around and see this.




Oh and on the x-axis and Y-axis. I have .003 backlash. Not bad at all. I could get it lower, but won't bother until I fix these problems.

-Jason

TMaster
01-18-2009, 12:44 PM
I have my backlash speed 3% of max. I would put your backlash at .002" I never have my over that. Right now mines Y.001" and X.0015"


It will cut the 2nd layer higher, the third and so. So it looks like there are steps in the material. The problem is your Z axis leadscrew nut is wore out and there is up and down slop. Gravity is not enough to hold it down. The solution is to replace it. Or replace it with a Y axis leadscrew nut that is adjustable. I was having the same problem untill I replaced it with a adjustable nut. I haven't had to adjust it yet becuase the Y axis adjustable nut was already tight. I may be able to adjust it now though since its prolly broke in by now. You would adjust it just like the X and Y axis'. I was going to do a how to on nicks site, I just havent got to it yet. But I have pictures.

I would try and pin point exactly where your having the stall or missed steps. Run the same code a few times and see if it always errors on a certian spot. You shouldnt have to cut material, just run it open, no material, no tooling. Listen for the sounds. :D

CROSSHATCH
01-18-2009, 01:16 PM
It has nothing to do with the z-axis leadscrew. I call it steps in the material because the y-axis loses/gains steps in one direction. And on each pass of the z-axis it cuts a new layer in the different spot, y-axis messes up again and so on. See what I mean? It's 100% on the y and not the z or x.


I have replacement boards for the Gecko 540 so i will swap the y-axis one out!


-Jason

TOC Admin
01-18-2009, 02:16 PM
Jason.....you have probably done this but set your Y in Motor Tuning the same as X...

Should look something like this

CROSSHATCH
01-18-2009, 02:20 PM
I run EMC2, not Mach. I have tuned the motors. Going to run some tests today.


-Jason

TOC Admin
01-18-2009, 02:27 PM
OH ..thats right....I knew that too, Sorry

acondit
01-18-2009, 07:38 PM
Jason,

This is from the EMC2 Integrators Manual, :


BACKLASH = 0.000 Backlash in machine units. Backlash compensation value can be used to make up for small deficiencies in the hardware used to drive an axis.

It goes in your ".ini" file in each axis section for which you wish to add compensation. I haven't used it, so I don't know for sure whether you need a positive or negative value, but I would guess a positive value. Everyone that I have heard using it has had very positive comments. So give it a try and check your measurements.

So add:

For [AXIS_0]
BACKLASH = 0.004
and
for [AXIS_1]
BACKLASH = 0.002

Then use your dti and check the results.

Alan

CROSSHATCH
01-18-2009, 07:57 PM
Will do! And thanks for posting Alan.


I report back with Great news! I swapped Y-axis motors and X-axis motors. I also swapped out the Gecko540 y-axis board with a new one that I have. (Only the y-axis though).

I ran the file in a production test, where At first 15 pockets were cut. Test passed, then I re-surfaced the aluminum with a code I made and it came out nice and ran another test of 20 pockets. All a Great success! Pockets look great and no sign of missed steps! This is very exciting news!


Seems to me as if though the y-axis board could have been going faulty. Whatever I did cured the problem.

This also goes hand in hand with a final y-axis tune last night on the machine.


-Jason

acondit
01-19-2009, 12:27 AM
Will do! And thanks for posting Alan.

I report back with Great news! I swapped Y-axis motors and X-axis motors. I also swapped out the Gecko540 y-axis board with a new one that I have. (Only the y-axis though).

-Jason

Contact Gecko and send back the one you pulled out. They will test it and repair (often at no charge), or tell you that nothing is wrong with it. With all the problems that you have had with it, it sounds like some kind of intermittent problem, and those are the hardest to identify.

Alan

CROSSHATCH
01-19-2009, 01:32 PM
Boy am I happy. The backlash compensation works great, finished all customers this morning. I ran some tests, they all looked good and finally finished. Everything looks very stable and I am very very very happy!


My machine is up and running and I am no longer pissed about heading into the shop :) Thank God, Now I can take on new stuff and go back to having fun.


Thanks!


-Jason

TOC Admin
01-19-2009, 01:44 PM
Way to go Jason......Glad you got it Figured out

CROSSHATCH
01-19-2009, 05:03 PM
Yea me to, thanks god. I finally have my mill back and can enjoy playing with it!

Pretty excited :)


-Jason