What's new
What's new

Ball bar test results - trying to interpret

Murdered Fiat

Plastic
Joined
May 27, 2019
Location
Ontario
Ball Bar Test.jpg

Hello all, I'm a new member of the forum, but have been visiting for years to find solutions to some problems. Hope you folks can help me out again.

I've attached an image of a recent ball bar test we had done to our Tos WH10. I did not arrange to have the test completed, nor do I fully understand it. My boss arranged it and is now asking me what the results mean. Hopefully you can clarify some things for me.

Looking at the values:

Backlash:
X has unequal backlash - damage or wear to x-axis drive components. Likely the case here as this machine is 20 years old and has a lot of hours on it.

Y backlash is relatively equal and minimal, not really a concern.

Reversal spikes:
X large reversal spikes - damaged or binding of moving axis components, inadequate lubrication, bad ball screw and or bearings

Y Not as bad as X (Y ways were recently reground and headstock was machined and tourcited). Ball screw problems?

Lateral play:
I don't know what this is telling me...

Cyclic error:
X and Y - Ballscrew and drive problems - more of the same information as reversal spikes?

Servo mismatch:
Time in milliseconds that one axis servo leads another. This is why forward and reverse runs are mirrored? Drives need to be tuned if possible?

Squareness: Actual squareness of X to Y in both vertical and horizontal planes? Level check of Y-X travel showed 0.005mm Y+...

Straightness: How straight the axis is, nothing to do with level or square?

Scaling error: How much error there is in the scale per meter. How is this corrected?

Positional tolerance:
191.3µm... holy! does this mean the machine doesnt have positional tolerance of 191µm over the 300mm radius of the test, or over 600mm? Is this a combination of all the errors? or what does this relate to?

Best fit radius: Basically the average sized circle that fits the data. But what does that mean? Wouldn't least square circle tell you better information ie. positional?

Circularity: difference between largest and smallest radius measured by the ball bar. 152µm... that's a lot but what does it really mean?

This is a really long winded first post. I'm very sorry. I hope I have posted in the right thread and I thank everyone for taking the time to read.. and hopefully respond.

Thanks

Zac
 
Put your values in Imperial measurements to make it easier for us to determine what you are talking about. And even then, may not totally understand what is being said. I assume this is a lathe? Ken
 
Put your values in Imperial measurements to make it easier for us to determine what you are talking about. And even then, may not totally understand what is being said. I assume this is a lathe? Ken


Hi Ken, thanks for the response. It is a CNC horizontal boring mill. As far as the numbers go, it doesn't really matter what they are. They are a measurement and that's the part I understand. What I don't understand is the headings for the values and what they are telling me.

If it makes it easier for you just pretend that it doesn't say micrometer and pretend it says thou.

Hopefully I didn't come across as a a** but converting it to imperial doesn't seem like it will help any.

Thanks

Zac
 
Thanks David, I have looked at that document and watched a couple of youtube videos as well. It doesn't really tell you what Cyclic error is or whether or not the squareness is over the travel of the 600mm diameter of the ball bar or the travel of the machine, if the squareness is in the horizontal plain like a tram of the head or the vertical plain or a combination of both? or what lateral play is etc. I do understand the backlash, reversal spikes, servo mismatch and scaling error.
 
..............................................................................................
If it makes it easier for you just pretend that it doesn't say micrometer and pretend it says thou.

Hopefully I didn't come across as a a** but converting it to imperial doesn't seem like it will help any.

Thanks

Zac

It helps a little for a senior that has worked in Imperial units all my life. Well, the last 12 years have involved both systems of measurements.

Generally speaking divide all be 25.4 and you have inches. But getting back to the interpretations of the chart, if it is not producing any inaccuracy's in the work it is doing, I would not worry about it. Chances are, what you are seeing is probably worn bearings or bearing preloads that need to be adjusted, of course, that means replacing bearings. And when I say replace bearings, primarly the one's on the ball screws. The ball screws them selves may need servicing, too. Stepping motors, unless they are run hard that would be the last thing to service. Slides and such, possible, depending if they are linear bearings or plain old frictional slides inserted with low friction materials. I would start with checking ball screws and go from there. And I'm no expert here, but been around stuff like this most of my life and just offering my thoughts of experience as a start. Anyone else, please add your thoughts, too! Ken.
 
Thanks Ken, there are major inaccuracies in the work.. positional tolerance on the chart 191.3µm (0.0075".. may as well be 191.3 thou) definitely looking at ball screws and bearings, likely re-turcite the saddle etc. The servo drives are old and I'm not sure if there's anyone around here that can tune them. This is a 20+ year old mill that has seen LOTS of hours.

Thanks again for getting back to me - Zac
 
As a first step you might want to check for lost motion on your ball screw. It is easy enough with dial indicator on the end of your ball screw, you do not want your screw to move axially. Among other things, this is a good check of the screws thrust bearings.
 
As a first step you might want to check for lost motion on your ball screw. It is easy enough with dial indicator on the end of your ball screw, you do not want your screw to move axially. Among other things, this is a good check of the screws thrust bearings.

Thanks M, I will check that. I read somewhere that a positive value for backlash means lost motion. I have one positive value on the X axis. This really helps, now I know what they meant by lost motion.

-Zac
 
Was this test run with the backlash comp turned off?
Does not look like it.
That is sort of the starting point for fixing things.
If it was on you don't have good information to work from. You only know how much to trust the existing setup.
Bob
 
Thanks Bob, I'm not sure.. we had outside service perform the test. Didn't get an explanation of any of the results..

Nobody thought to ask?

Get an indicator and something to lever against the table, in 10 minutes you could tell the backlash in your thrust bearings, and lead screws. Whether gibs need adjusting if it's got ways, or whether the rails/trucks need replacing if it's a linear way machine. That would go a long way to telling you if the machine needed parts replacing/maintenance etc.

I would have done that before spending money on a ball bar test.
 
Nobody thought to ask?

Get an indicator and something to lever against the table, in 10 minutes you could tell the backlash in your thrust bearings, and lead screws. Whether gibs need adjusting if it's got ways, or whether the rails/trucks need replacing if it's a linear way machine. That would go a long way to telling you if the machine needed parts replacing/maintenance etc.

I would have done that before spending money on a ball bar test.

I have asked, got a very general response, didn't really answer the questions that my boss was looking for. Again, I didn't arrange it, only was asked to explain the results...

Thanks for the info
 
A ballbar or grid test is a starting test.
You use it to find ways to remove errors.
Then you wash, rinse and repeat.
Your outside source should have given that info bit perhaps your inside was throwing good money after bad.
I don't like to say this but at this point you don't know shit so be careful on the limb you go out out on.
Bob
 
Was this test run with the backlash comp turned off?
Does not look like it.
That is sort of the starting point for fixing things.
If it was on you don't have good information to work from. You only know how much to trust the existing setup.
Bob


I am not disputing anything you say, I know nothing about doing a ball-bar test, or reading the results...

I was at a place that (thought they were hot shit ermm, anywho.. :skep: ) ran a ball-bar test on new machines after they were set up. Are you saying those results are no good? Like from a factory install of a machine? Just wondering (refer to first line above :D)
 
..... Are you saying those results are no good? Like from a factory install of a machine?......
The results simply tell you where you stand.
By itself a run tells you nothing about how to make the machine better if comps are on as you don't know what is machine and what is computer..
It is a valid quick test of a new or old machine although the older circle-diamond-square is perhaps more real word as it includes cutting forces and deflection.

The ballbar is oh-so handy tool but it's easy to misread what it is telling you.
You have to use it over and over as you make corrections not just take one shot.
Bob
 
That makes a lot of sense, the idea behind it here was to see where we stand and determine what to do to correct the machine. As far as whether or not compensation parameters were on, I don't know.. We haven't made any corrections yet as we didn't really get any report with it saying what it meant.
The only response I really got was that the reversal spikes could be causing a lot of the other problems and that we should see if we can get the drives tuned.

Anyway, we're getting it figured out sort of.. Thanks for all the responses
 
A ballbar or grid test is a starting test.
You use it to find ways to remove errors.
Then you wash, rinse and repeat.
Your outside source should have given that info bit perhaps your inside was throwing good money after bad.
I don't like to say this but at this point you don't know shit so be careful on the limb you go out out on.
Bob

Thanks Bob, I'm new to ball bar testing.. never seen it done before so I didn't really know what to expect. Like I said, I didn't arrange it and still trying to figure it out.
 
Thanks Bob, I'm new to ball bar testing.. never seen it done before so I didn't really know what to expect. Like I said, I didn't arrange it and still trying to figure it out.

imho your wasting your time with a ball bar test until you know the machine is in good mechanical condition first.

Pull back the way covers, check the leadscrews/thrust bearings, ways etc first.

For instance you could have a machine that is in good condition mechanically, but have an issue with the lube system, there'll be a difference in the ball bar test when a machine is getting adequate lube or not, assuming it has boxways.
 








 
Back
Top