Printable View
We have a lathe with Fanuc 20-TA control. We thread with it 100 percent of the time. The lead on the threads is creeping up to areas that are out of tolerance. Looking in the maintenance manual I found pitch error compensation, my question does this have anything to do with thread pitch?
It can, but I would check other things first, like backlash amount, and spindle encoder belt/pulley.
Backlash was out .004 for both rapid and feed. It is now 0. I bet that will help, I checked it a while ago and it was ok, but things change. Spindle encoder looks to be direct drive, not real sure without opening up headstock. Thanks, Dave
Well that didn't help much. I cheated untill I can learn more. Changed the feed rate on threading by .0004 and it put lead on 0.
when threading you are machining mm/rev or inch/rev, so the spindle encoder, belt and pulleys need to be good.
backlash comp is for the ballscrew nut, if the support bearings at each end are good and the belts/ pulleys are correctly adjusted then the backlash can be adjusted up to 0.004", if more the ballscrew is defective.
pitch error comp is quite different, it is a laser measured pitch check and data is entered into a comp table to optimise the axis positioning, it is of no use in your case.
How can you possibly say it's of no use in your case? Pitch error is used in threading, as all axis motion. I don't think .004 backlash is acceptable in any CNC machine, and electronicaly comping for it is not a fix. To properly set up pitch error, it should be done with a laser, but a step gauge has also been used before lasers for years.Quote:
Originally Posted by cncserveng
Now, back to your problem, I believe you are seeing wear on your Z ballscrew or the thrust bearings need to be changed or both. Set up an indicator on the end of the ballscrew. It should have a center drill mark in the center of it. Place a ball bearing there (grease helps to stick it on) and place indicator tip on the ball bearing. Now in MPG mode, move the Z back and forth a few thou and see if the ballscrew is walking. Should be less than .0001" movement.
Please don't make such sweeping statements, laser measuring is only of any use if the machine is mechanically good, if you believe otherwise you are wrong.
laser measurement is usually done at the time of manufacture, I am a very experienced electronic CNC commissioning/ service engineer with 30 years experience.
do not make statements you cannot support without qualifying those statements, you obviously know less than you think you do. I work for a huge Swiss machine tool company manufacturing high end 5 axis machine tools for the aerospace industry. And I offer sound advise.
:nono:
You obviously didn't read or didn't understand what I wrote. And your the one who said it wouldn't help in his case. I told him right from the beginning to check mechanicals. You don't know from sitting on a computer if it will help or not, neither can I. I simply said laser or step gauge is used for compensation, not that he should do that unless he does more tests that I outlined before. I don't care who you work for, I work with the people who INVENTED the ballbar, and THE LASER SYSTEM ! We laser our high accuracy machines every year, and he never said who actually built the thing. Not all companies laser check at the factory !
So you are an expert because you know people who have invented things, how does that help ?
I on the other hand attend customers every day to investigate eletronic and mechanical issues on CNC Machine tools costing up to $3 million.
Only last week I replaced a 4 metre ballcrew having diagnosed that a loss of preload had caused cross check errors on a Siemens 840D system with safety integrated. The customers ballbar check showed nothing and the machine had active laser compensation.
Neither the ballscrew nut or support bearings showed any backlash that could be measured using an indicator, the issue was that the motor encoder and axis linear scale were not synchronous during axis motion and intermittently there were standstill and moving errors. Everything pointed to a drive or measuring system error, but my experience in these matters made me investigate the basic mechanicals first, that was and is my point in this case.
n.b.
I trained for 5 years at a Machine tool manufacurer where I worked for 10 years on the build. I then went on to University and gained a BSc in electronics. I have worked all over the world on a whole range of CNC machine tools including 3 years in California.
I was only trying to help this guy where on the other hand you tried to "big yourself up" by rubbishing and discounting my advise.
You may not care who I work for, but I am an expert in my field and offer my advise in good faith, I wonder what your motivations are when you critisize an honest attempt to help ?
(wedge)
3 million is chump change compared to some of the stuff I work on now. I was a service manager for two major machine distributors in CA for 20 years. Factory trained at several of the largest builders in the world. I now do metrology because i got tired of being on the road. I'm also not here trying to sell my services to people. You basically mimicked what i already told him to check, then told him pitch error comp wouldn't help him. How can you say that unless you are standing in front of the machine? So what, you changed a ballscrew, BFD. And the whole laser feedback system was invented here and used since the 70's. We routinely hold tolerance in the micro inch range.
Possible causes of threading pitch errors.
defective spindle encoder; cable or connector (coolant inside ?)
spindle encoder bellows coupling is loose or damaged
spindle encoder belt has missing teeth or is slipping relative to spindle
ballscrew endfloat; due to worn or loose support/ thrust bearings
ballscrew backlash; due to worn ball nut i.e. >0.002" (0.050mm/ 50mu)
uneven wear along the ballscrew,
slippage in ballscrew drive train, gears, belts etc; during reversal
defective axis motor FBU or axis encoder feedback
incorrect pitch compensation values (unlikely)
defective nc axis control card
The best way to check backlash is dynamically, that is to say by writing a program to drive the machine onto and off the indicator, put dwells in to enable reading of the indicator.
Using jog/incremental for a backlash check is not the best solution and many manufacturers do not recommend type of check, It is much more accurate using a program.
Perhaps another poster can list some more possibilites, I'm sure there are some very clever and brilliant engineers waiting to offer advise.
:devious:
Yea I used MDI to move the machine on and off the indicator.
We purchased a Poreba knock off (Buffalo Machine) so far a good solid lathe. We lost an X axes drive belt in July. After all was fixed I took the book and did the backlash check on both axes, Wow the Z on a 19 month old lathe is out .006. And the lead was never bad on this lathe, so I see now that backlash really has little to do with thread lead, as long as the screw is loaded in one direction.
Put a 125 backlash comp number in 1851 and 1852. our repeatablity got a lot better. I have been talking to the dealer to get a quote on a new screw. We have not crashed it hard. The oilers all work and we keep it clean. Can a ball screw and nut be repaired, or mechanically adjusted? I saw that the nut is a two pieces with a shim in the middle.
I have only been around CNC's for 10 years, been in the oil field threading industry for 30, mainly manual lathes, when the lead is off on them we replace lead screws and half nuts. So with most the posts even thought there is a computer doing the work, common since still applies. If the mechanics are out there are out.
0.006" is too much (0.002" max) and some designs of ballscrews do allow for nut adjustment by grinding a spacer or peeling off shims from the centre nut joint.
Only remove by trial and error though, so as not to put on too much preload.
Maybe start with 0.005" and remember to zero your backlash comp. parameter and adjust again later.
I have a machine with a 21i TB controller. We bought it new and discovered it would not turn diameters correctly. If you programmed it to turn a .5, then a .75, 2" and so on the error would show up as the diameters got larger. If I remember if the .500 dia. was on the 3" dia would be off approx .002. I had it laser checked so I could prove to the manufacturer it was an inferior machine. The machine mfg came in a checked some of the obvious like spindle alignment, even replaced the x axis ball screw to no avail. They replaced the machine with a good one and offered us a deal on the one with error we could not refuse. I talked to Fanuc about it and they said they had had issues with the pulse coder....replaced it and no improvement. Question.....can I adjust the x axis error with the pitch error parameter. If yes I can get it better and try it for awhile then bring in a laser and put it spot on.
Thanks
I am no expert on the pitch parameter. If you have a machine just like the one check the all the parameter just to see if they match. Good place to start. I have a mill and one of the parameters were change from a 1 to a 2 and it made the z move in a radius or half what the dro said. Weird!!
Dave
Most posts die and we never learn what or if there is a fix. It appears most lathes do not have pitch error comp enabled. Most mills do. Our new machine had a pitch error we were fighting and we ended up having it activated. Then we had a laser on it and it is spot on! This allows the machine to be calibrated or adjusted the full length of the screw. I’m just sayin……..
Pitch Error Compensation is an option. Do you know how much it cost? ....just curious :)
To anyone wondering if they have it, you can tell if it's enabled by looking where parameters are punched out and there is also an item to punch out Pitch Error Compensation numbers :-)
We had to get the pitch error through the machine tool builder and Fanuc installed it. If I remember correctly it was around $1500.00Quote:
Originally Posted by fordav11