BallBar done, now some ??

[Bob_Looney]

I have a 2 1/2 year old cnc. I have been real happy with it and havn't done anything but replace the front Y axis bearings. It is a linear guide machine and in the haas, fadal class. I paid 60K for it new.

 

I noticed it was cutting out of round and decided to do a ballbar to see exactly where I was.

The guy leveled and tramed the head to less than 2 tenths (.0002). The ballbar showed egg shape in Y. overshot equal amounts at 12 and 6, it was out .00055 total in Y and less than .0002 in x.

I was hoping for .0002 total, needless to say that didn't happen. He messed with the jack screws and lost motion params in the control and could get no better.

He came to the conclusion that pitch error was the problem and i should have re done the pitch comp after the bearing was replaced.

I never heard of needed to redo pitch comp after a bearing change, is this right?

also, is .0005 about what you would expect to see from a 2 1/2 year old cnc of this class?

[Guest CNC Apps Guy 1]

I'd say that the Pitch Error should be checked also. Any time you do anything with the physical drive system, even if it's just the thrust washers, it's best to get it checked, just to be on the safe side. How th bearing sit in the ball screw can have an effect. New should mean tighter so yes you should get it checked.

 

As for it being out that far, depending on how hard you run your machine... that's not bad for a machine of that class. Remember, that price of machine is in the "Throw away class", meaning every three to five years, it will need to be replaced if it's run anywhere near profitably.

 

JM2C

[deeherk]

i'm not sure how you do the pitch error, other than reloading software,but i would thimk anytime you change bearings or such,that should be done..what do you mean when you said he leveled and tramed..what is tramed...is that anything like having machine laser comped or what...just curious...if a person knew all the lingo out there,life would be easier...

dan

[MetalMarvels]

Note that this is for a vertical mill.

 

To tram a head is to move it (with shims or jacking screws - depending on the machine) so that the Z-axis (and spindle) is perpendicular to the X-Y plane of the machine (usually the table). Sometimes the table will also require shimming to ensure that the table surface stays "flat" throughout the X-Y plane. You should always check the X-Y table first - after leveling the whole machine.

 

Fadal publishes their procedures for performing an axis "survey" (pitch compensation). Basically you are checking and compensating for various degrees of error at several points along the axis ballscrew. For the manual method, you use a gage block setup that spaces several gage surfaces along a length (usually at whole inch increments - like 3.00000 inches or 4.00000 inch intervals). The gage setup is a fairly pricey piece, but it would be possible to "build" a setup using standard gage blocks. You then take measurements along the gage set to see if you get an actual 3.0000 inches travel for a 3.00000 inch space (always traveling in the same direction). You then enter the "deviations" as your pitch comp along the axis ballscrew. THEN you have to do backlash......

[Guest CNC Apps Guy 1]

Tram; make sure the head is paralell to the proper axes and perpendicular to the proper axes. Any time yo umake any changes to the drive system, it's a good idea to check it.

 

If a machine is level or not is completely and totally irrelevant. What matters is if the machine is square, and there is no sag, twist, crown or bow.

 

Pitch error comp is when the error/inaccuracy in the pitch of the ballscrew is compensated for. It is a set or group of parameters in the control. It's not software. It is checked through laser measurement. Then there is backlash.... , and then, and then, and then.

 

Ball-Bar testing is to see if your machine cuts circles, and how far out the circles are from perfectly round. Typically there is more error at the point there the servos reverse to go the opposite direction they were traveling, i.e. 0 Deg., 90 Deg. 180 Deg., and 270 Deg.

 

HTH

[MetalMarvels]

I will take a minor exception to the "level" of a machine.... It is perfectly true that to an extent, it does not matter if the machine is "level" with respect to local gravity. However, it is typically the easiest way to determine if you are introducing a "twist", "hump", or "sag" in a bed. Ideally, the machine would "float" with respect to the world - establishing its own "level". Unfortunately I have seen all too many machines that are not properly leveled and are consequently non-planar for X or Y. The biggest offenders usually occur when the machine is locked down (bolted down).....

[Bob_Looney]

James, I figured if I got 5 years I would be doing well. I run this machine hard all day. I cut steel for about 8 months, 700sfm and .004 per tooth. The rest of its life has been aluminum at 1000 rpm and 100-200 ipm. I am a captive shop so all my parts are repeat and I do thousands of parts / month.

 

I guess .0002 circularity was a little optomistic for a 60K cnc. My friend that has a 160K nv5000 thought it was optomistic even for his mori, they trammed his to .0003 when new and told him he was good to go.

 

-.000539 scaling mismatch

.000146, .000114 X backlash

.000142, .000146 Y backlash

.000055, -.000075 reversal spikes

15.9 uin/inch squareness

.00269 position tolerance

Position tols sure doesn't look good.

He swept the table in x and y, no visable sag or droop, maybe a tenth if that.

 

thanks for the info

[Guest CNC Apps Guy 1]

quote:

---

...I will take a minor exception to the "level" of a machine.... It is perfectly true that to an extent, it does not matter if the machine is "level" with respect to local gravity. However, it is typically the easiest way to determine if you are introducing a "twist", "hump", or "sag" in a bed. Ideally, the machine would "float" with respect to the world - establishing its own "level"...

---

True, it is much easier to check and adjust for twist, droop, etc... if the machine is level. Soooooo many people think that if their machine is level then it's good to go, I wanted to point out that erroneous thought.

 

That position tol is HORRIBLE!!!

[MetalMarvels]

I concur James. It is too easy to level the machine and think that you are good to go. Leveling it front-to-back and side-to-side is definitely NOT the place to stop. It is usually just the first of many steps to get it really "flat" and square. The "old-fashioned" way was to check level on the machine in a variety of locations and positions to ensure that it was "level" or flat throughout the X-Y travel. Laser checking is soooo much easier... and you can do volumetric compensating (if your machine control allows for it). However, between full-service calls, it is a quick and dirty way to determine if your machine is "moving".

 

Agreed, that position tolerance is intolerable... It should be less than .0002 on even a commodity machine. My linear guide Fadal is still less than .00012 after 4 years of use (2 shifts - 7 days). My backlash is starting to creep up a bit, but I check the pitch error and backlash comps about once every 6 months or so with a full checkup every year to year and a half.

 

Some possible contibutors to position tolerance problems would be:

 

- Lubrication issues (blocked feed lines or metering valve) contributing to stiction

- Loose guideways

- Loose or too tight slides

- Worn or loose motor/ballscrew coupling

- Bad measuring equipment or measuring technique

[Guest CNC Apps Guy 1]

quote:

---

...- Bad measuring equipment or measuring technique...

---

This probably accounts for most errors.

[cblythe3d]

Checking perpindicularity helps indicate the need to adjust your leveling pressure at various points in order to counter casting creep as the castings age. The larger the travel the more important this becomes.

 

Back in the day swiss jig borers were made from well seasoned castings that sat outside for years. Also they park the table in the middle of of the travel so as not to twist the casting, unlike some of todays manufacturers.

 

Laser calabration of of your pitch error will adjust your x,y,z positioning so that 10 inches in Y is the same 10 inches in X and Z

[Bob_Looney]

If the position error was anything other than measureing error wouldn't it have shown up in some of the ballbar tests? He ran 7 of them and nothing moved more than a cpl tenths as he was trying to get the scaling error down.

[Guest CNC Apps Guy 1]

quote:

---

...If the position error was anything other than measureing error wouldn't it have shown up in some of the ballbar tests?

---

Not necessarily. Pitch Error checks the FULL travel not just the travel of the ballbar.

 

It seems like you're still under the impression the two tests (Ballbar and pitch error comp) are testing the same thing. The two tests check completely seperate things. Re-read Metal Marvel's replies.

[MetalMarvels]

Absolutely, James.

 

To measure pitch error you have to measure exact points along an axis of travel - taking the measurements moving IN ONE DIRECTION ONLY. Typically you use a measuring gage bar laid on the table along the axis of travel and use a 10ths (or better) indicating dial test gage in the spindle. NOTE that you ALWAYS approach the measurement point moving in ONE DIRECTION only. These measurements are taken at several point along the axis of travel. Only the PITCH ERROR in the manufacturing of the lead of the ballscrew (or induced pitch errors due to wear) are measured by this technique (it does NOT measure backlash). Pitch error typically does not change much for the SAME ballscrew (except in cases of extreme wear). However, If you replace the ballscrew - whole new game.

 

For backlash, you approach the measurement point from one direction, advance until you read "0" on your tenths indicator, take the reading (should be "0"), retract the spindle, continue in the same direction, stop and reverse back to the measurement point and extend the spindle back down to take the second reading (the 10ths indicator will probably NOT read "0"). The difference between the 1st reading and 2nd reading is your backlash. You typically perform this measurement at both ends and the middle of the ballscrew travel. A worn ballscrew, worn ballscrew carriage, worn end bearings, worn couplings will all contribute to backlash. Some backlash MUST be present for the ballscrew to move - usually a 10th or less. This is the parameter that typically moves over time on a machine. It will definitely change if you change ballscrews, ballnuts, preload bearings, couplings, etc.. If you are having problems in this area, it will typically show up as "flat spots" at the four quadrants of a circle.

[Bob_Looney]

according to the tech, the "scaling error" is ballscrew pitch error.

I understand where this comes from and the backlash too.

I have assumed the "position error" is like a repeatability error. if that's the case, how does a ballbar check that?

[Guest CNC Apps Guy 1]

Repeatability is the ability of a machine tool to go from one position to another back to the first position within a certain amount of distance. Your best CNC Machine Tools will repeat between .000"1 to .0002". "Scaling error" ... never heard of pitch error comp called that before...

 

 

"Position Error" can be several things. It can be a combination of Pitch Error or Backlash or a combination of both.

 

A Ballbar Test will settle "Roundness" Issues not "Position" issues.

[Guest CNC Apps Guy 1]

Also, Ballbar Tests are the easiest to do. Thye require the least amount of skill and experience to complete and are decidedly cheaper to perform in most cases.

 

If the tech is getting jumpy results, ask him if he checked to be sure the thrust washers and bearings are tight.

[Bob_Looney]

thanks James

the software printed out the graph and listed the errors I listed above. Scaling error, back lash, reversal spikes, and position tolerance. The renishaw book said the scaling error was from ball screw pitch error.

I don't know how the software could measure position error accurately because after every test he removed the bar and sent the machine home. then he started his ballbar program which positioned the ball above the base, then moved over in X. then the bar was re installed and the test run again.