Mazak Ball Screws

Does anybody know, what is the accuracy grade of ballscrews used in Mazak machining centers like Variaxis. The catalogue mentions that the machine is High Accuracy, but no data of positioning/repeatability accuracy is given, neither the ballscrews accuracy grade.

Most machine tool builders use C3 grade ball screws for all linear axes. The accuracy specs of these ball screws allow for 24 +- 16 microns per meter of travel.

The idea is that machine tool builders can use less precise screws to save on cost, then use a laser interferometer to create a screw map or compensation chart, which is stored in your machines memory and used at all times.

Just for reference, all of the quotes and spec sheets I've gotten for good quality machines shows positioning accuracy of 4 or 5 microns and repeatability of 2 microns for each machine axis. If they use belts to couple the servos to the ball screws, expect at least twice those numbers.

Pitch and diameter + core cooling are other factors as well + scales if you have them + stretched / preloaded + double nut. Not sure if there is additional mechanical preloading to compensate for even bigger temperature differences ?


Will rummage around a bit.

_________________________________________________


I have to admit still does raise the question of cumulative errors for probing or using a machine as an inspection instrument, however for 5 axis GOOD and very accurate rotational calibration should (for rotaional axes and rotational errors of work piece) be possible even though linear moves may not have high correspondence to true distances and linear positions that are commanded.

_______________________________________________

Can't find specifics on MAZAK ball screws for 5 axis but some stuff on their verticals.

_______________________________________________


^^^
Food for thought as to how to "map out" a sh*tty ball screw to something that irons out pretty damn well (considering).

http://www.cbmind.com/linear/thk/pdf/Ballscrew Product Specifications.pdf

Accuracy charts vs grade for THK. ^^^ Lead screw accuracies on page K8.

PROBE said:

Does anybody know, what is the accuracy grade of ballscrews used in Mazak machining centers like Variaxis. The catalogue mentions that the machine is High Accuracy, but no data of positioning/repeatability accuracy is given, neither the ballscrews accuracy grade.

Click to expand...

I can't find specifics on that either.

Have you tired bugging application's engineers for MAZAK Israel ?


They should be able to source that info for you especially as you do things that help MAZAK machines "go" better.

Is this for some kind of weight matrix ? (weighted values for predicted or known errors to be thrown in the mix so the system is not over-constrained or under constrained ? When forced to converge on a set of most probable values ? ).

cameraman said:

Pitch and diameter + core cooling are other factors as well + scales if you have them + stretched / preloaded + double nut. Not sure if there is additional mechanical preloading to compensate for even bigger temperature differences ?


Will rummage around a bit.

_________________________________________________


I have to admit still does raise the question of cumulative errors for probing or using a machine as an inspection instrument, however for 5 axis GOOD and very accurate rotational calibration should (for rotaional axes and rotational errors of work piece) be possible even though linear moves may not have high correspondence to true distances and linear positions that are commanded.

_______________________________________________

Can't find specifics on MAZAK ball screws for 5 axis but some stuff on their verticals.

_______________________________________________

^^^
Food for thought as to how to "map out" a sh*tty ball screw to something that irons out pretty damn well (considering).

http://www.cbmind.com/linear/thk/pdf/Ballscrew Product Specifications.pdf

Accuracy charts vs grade for THK. ^^^ Lead screw accuracies on page K8.

Click to expand...

1. Couldn't see the stuff on MAZAK verticals ball screws.
2. Mentioning THK - does this mean that MAZAK is using their ballscrews ?

cameraman said:

I can't find specifics on that either.

Have you tired bugging application's engineers for MAZAK Israel ?


They should be able to source that info for you especially as you do things that help MAZAK machines "go" better.

Is this for some kind of weight matrix ? (weighted values for predicted or known errors to be thrown in the mix so the system is not over-constrained or under constrained ? When forced to converge on a set of most probable values ? ).

Click to expand...

A but more complicated then that. Can not elaborate now, will in the future.

PROBE said:

1. Couldn't see the stuff on MAZAK verticals ball screws.
2. Mentioning THK - does this mean that MAZAK is using their ballscrews ?

Click to expand...

MAZAK on their website do mention THK but whether they use THK ball_screws I'm not sure ?


I threw that up there as it shows the different tolerance bands for the different classes of Ball screw. in General. So given your application problem you might actually be able to reverse engineer what class the ball screws are ? Assuming you have access to a VARIAXIS machine. Bear in mind that i -500 may not be the same as i-600 etc. ?


There is MAZAK applications engineer that hangs out here on PM forum he might know...

Like you I have found that their brochures are not very fourth coming with certain technical details but I have been able to pursue things with MAZAK engineers to get the data I needed.


This is not the first time that you have made inquiry's about MAZAK accuracies and precisions... is understandable but needing to know specifics about ball screws is new one? I'm not a mind reader... ?????


As you know there are companies that regrind MAZAK ball screws but not sure that people get them mapped out properly after they are installed. Can't use old pitch error compensation on reground screws you can get a pretty wild / random response in combination ( summation of waves that can yield a seemingly chaotic response ).

PROBE said:

A but more complicated then that. Can not elaborate now, will in the future.

Click to expand...

That would be cool cuz, my perception of some of your work is that you are an expert on really elaborate probing and metrology type stuff as it relates directly to machines … So I've always imagineed the core mathematics that you do on 5 axis machines is basically an iterative least squares adjustments of sets of partial derivates composed from sets observation and condition equations derived from three dimensional affine transformations and that the "Weight Matrix" is of the form of a Jacobi Matrix (with estimated errors for the all the parameters that you enter to one sigma)... To correctly balance the system and have the ability to converge on "sound" mathematical solutions and values while iterating the residual errors are all brought to a minimum (hopefully) and hence final values and coefficients are statistically the most probable and accurate values... much more so than just averaging stuff (averaging simple transformations).

However needing to know the class of ball screws is one step more beyond "Need"... That's why that kinda sticks out like a sore thumb.

'Cuz they are not that accurate before proper calibration ?


Sounds like you got some knarly trouble shooting with a VARIAXIS machine???? Like a client is pointing the finger at "probing software" and you saying "Hey"... I think there's something up with your ball screws.

Mazak Ball Screw Repair

Seems at some point screws will need to be replaced or reground ?


I've also seen MAZAK videos (like factory tour) where they make their own ball screws too ?


________________________________________________________________________________________________________________


IME just keep bugging MAZAK for the info you need and they will give it to you... Just be persistent and sincere and transparent.
But maybe you don't have time for that as it can take a LOT of time to wake the "Sleeping Giant" to actually do something or source the information you need.

In agree that MAZAK could be better on that front but eventually they themselves can devise ways of sourcing info that even they have trouble sourcing from JAPAN.

Maybe MAZAK UK could be of more help to you ? European headquarters but UK is Brexiting ?

PROBE said:

Does anybody know, what is the accuracy grade of ballscrews used in Mazak machining centers like Variaxis. The catalogue mentions that the machine is High Accuracy, but no data of positioning/repeatability accuracy is given, neither the ballscrews accuracy grade.

Click to expand...

^^^ Seven minutes in (MAZAK factories tour*) they show their BALL SCREW PLANT ... So they seem to be manufacturing a lot of their own ball screws. And at 9 minutes 30s in you see ball screws being shipped on a pallet from their spare parts facility.

You could be justified for asking for technical data as to what standards that their proprietary ball screws are made to ?


__________________________________________________ __________________________________________________ _______________________


* I forget how automated and super BIG MAZAK really is...

__________________________________________________ __________________________________________________ __




European manufacturing plant ^^^ in Worcester in the UK. (that's your region perhaps ? (not sure if they build the VARIAXIS probably not but they can source the data you need from Japan. On the other hand maybe they are not very helpful at all and that is part of the troubles you are having ? . ).

cameraman said:

You could be justified for asking for technical data as to what standards that their proprietary ball screws are made to ?

On the other hand maybe they are not very helpful at all and that is part of the troubles you are having ? . ).

Click to expand...

You could ask, but their not going to tell you, why should they? The last thing they'd want to do is provide any type of data that could be used by a competitor, or provide it somebody who'll then publish it on the net, inviting all sorts of comments from people who know nothing about machine tools or leadscrews.

I think u need to consider age too, ball-screw wear will noticeably effect pitch errors, seen this cause real agro on a machine that had only used part of its travel for a sustained time. Came up when it was being laser'ed as part of a new accuracy drive, there was a very visible jump - quirky area then noticeably different error rates on the far more Virgin screw!

I was told, that MAZAK uses its own ballscrews, generally low quality (no specific grade was given), and pitch error compensation assures the requested accuracy.
I personally performed over 1000 pitch error compensation processes, and learned that the problem of low quality ballscrews is that except from overall inaccuracy the error is not linear. It means that along part of the screw its error is negative, and along other part positive. This fenomena fluctuates along the screw. As the nut is quite short comparing to the screw, the backlash fluctuates too. The backlash compensation amount, set before the pitch error compensation process is therefore not effective along whole length of the screw, and the results are clearly seen in parts produced.
I believe that there is no way to produce accurate machine using low quality ballscrews.

^ All depends on how fine the error compensation table is, these days theres no need to limit it, could be easily done, read and used at just thou increments.

I do agree, from what i have seen the errors are not like say 5 thou a foot, they go up and down and vary too on a given turn of ball screw pitch too. Theres a good reason the old jig bore'rs use to use ultra prescise screws and really long nuts for there error compensating, unlike the current fad of just a few fractions of a complete pitch of ball screw nut ball path with just a few tens of balls in point contact.

problem is more precise screws bump machine cost and you still have the real life wear issues, this is the point that it pays to only use the screw - servo for rough positioning and then do the real closing the servo loop with a separate wear free element like a linear glass scale. Like this you compensate for all the random thermal, screw, screw bearings and servo error loops.

triumph406 said:

You could ask, but their not going to tell you, why should they? The last thing they'd want to do is provide any type of data that could be used by a competitor, or provide it somebody who'll then publish it on the net, inviting all sorts of comments from people who know nothing about machine tools or leadscrews.

Click to expand...

Yup I totally agree and I suspect that's the reason they don't release that information or splurge it in their brochures.

However from personal experience as our core applications are in that dead zone between what most cnc machines can do vs. grinding we have had to dig deeper on how a lot of these machines perform. In a lot of cases I/we have had to forge good business relationships and gain mutual trust. In our case a near miss is still a miss and if we have to spend an additional $300K to bridge the gaps of a "Near miss" then that's a "fail"/ precipitous near ship sinking potential situation for us.


Full test sheets showing full plots and reversals (bi-directional) and decent statistical information should reveal what the positional capabilities of the machine are regardless of what (in this case) the ball screws are doing ? Admittedly that's usually for a NEW machine rather than in service field plots.

Generally I don't share or post test sheets etc. and in some instances I have signed NDAs ... The info is for US not for any other purpose. Even on line I try to be very careful about not presenting relative information where one could reverse engineer what tolerances bands various machines live in even by comparison.

Usually the very best MTBs do share and publish their metrology for a machine just as "Bragging rights", folks in the middle of the pack and below kinda hide in the "Murk"

PROBE said:

I was told, that MAZAK uses its own ballscrews, generally low quality (no specific grade was given), and pitch error compensation assures the requested accuracy.
I personally performed over 1000 pitch error compensation processes, and learned that the problem of low quality ballscrews is that except from overall inaccuracy the error is not linear. It means that along part of the screw its error is negative, and along other part positive. This fenomena fluctuates along the screw. As the nut is quite short comparing to the screw, the backlash fluctuates too. The backlash compensation amount, set before the pitch error compensation process is therefore not effective along whole length of the screw, and the results are clearly seen in parts produced.
I believe that there is no way to produce accurate machine using low quality ballscrews.

Click to expand...

Ok so let's say you and I are going to build a new machine tool... Wouldn't we completely map out the response of the linear movements ? So that the whole length of the screw and ("thread" unwound) would be completely mapped out (bi-directionally)? [@PROBE I'm also wondering if you are seeing predictive thermal compensation effects of the order of two tenths 0.0002" ? I.e. they map the machine out thermally and hence to compensate that machine "bounces around" up and down and up and down in a two tenths band ? ].

With stretched ball screws and double nut to help eliminate backlash you are saying that's not good enough.

I agree that a more accurate mechanical platform provides the basis and foundation for a better mathematical and digital set of calibrations and corrections... That is one of the cornerstones that my business is built on... However I don't fully understand why MAZAK would have five axis systems that rotaionally can position to a few arc seconds only to have linear positioning that is total "pants" ?

@Mr PROBE when you say 1000 machines are those lathes / turning centers and are all of those MAZAK and have you found difference with the Integrex*.

_______________________________________________________________________________________________


* Reason why I mention that is a MAZAK engineer (high up) kinda said that the Integrex is really high accuracy mill that happens to do turning and that typically such mills are built to a higher standard than a typical lathe. ?

_________________________________________________________________________________________________

* @PROBE can you suggest a test part that we should be having cut to demonstrate or reveal more precisely what you have observed in the field ? (Please and thank you ). Have you found that scales on MAZAK machines (like VARIAXIS) has the capabiltiy to iron / smooth some of that out ?

On a different note this also raises the question and approach to the build philosophy of lets say Makino versus Kitamura... One is "True Geometric accuracy" and the other is not but the "BAse" of the machine is good enough to mathematically and digitally / mechatronically close the gap very effectively.

@PROBE


If you were going to take a guess at the class of ball screw on specific machines like a VARIAXIS what class do you reckon they are and by how far do they manage to correct for that + other factors ?

Most of the ball screws on VARIAXIS are obviously much shorter than let's say an Integrex machine or similar at least on Z axis for intregrex.

__________________________________________________________________________________________________________


Also @PROBE check this out https://mazakfiles.blob.core.window...=Dcc+nZZ4d5vqFyTz0/JZP9V+tV7TqAuqngaRpzGGQOI=



^^^ Press release from Japan (september 5th) on the new Integrex AG (advanced gear machining) using the RENISHAW "SPRINT" system in process measurement for really accurate gear sets ? [Looks pretty sexy in gray].

Have you used the "SPRINT" system and are you kinda saying it's all "Pants" 'cuz the ball screws are no good ? They are claiming

"better machining compensation in one machine"

(not sure if they meant to say that but seems that's their assertion... ). They better have some serious "Metrics" for all that YKWIM ?

SPRINT™ technology

3d "scanning" probe system ? better for maching accurate gears, but can you use it to make more accurate other things ?

i.e. check free from surface machining accuracy. Wonder how that plays forward to make critical machine adjustments compensations (volumetrically/ three dimensionlly)?

__________________________________________________________________________________________________

So I'm gonna take a wild guess (plus some ideas from HuskerMcDoogle )that the Integrex AG (like the I-200 model) has linear and rotary scales all axes and also has very high accuracy grade ball screws and good interoperability with the Renishaw "Supa-scan" / "Sprint system".

I do wonder if thermal mapping will make things more difficult ?

When asking for the accuracy of the mazak in house made ballscrew you're most likely going to hear nothing. That can be some pretty sensitive data.

Speedie said:

When asking for the accuracy of the mazak in house made ballscrew you're most likely going to hear nothing. That can be some pretty sensitive data.

Click to expand...

The products of all known ball screw makers conform with standard JIS B 1192:2013. I believe that MAZAK, an ISO 9001 confirmed firm, is obliged to conform to this standard too. Kind of general rule is that the lowest quality screws used in machining centers is C3, below that the screws are rolled rather then ground. I wouldn't believe that Mazak uses rolled screws in their machines, so probably they conform to C3.

Accuracy with this stuff ? Well, as mentioned before, no promises made in advertised material.

PROBE said:

The products of all known ball screw makers conform with standard JIS B 1192:2013. I believe that MAZAK, an ISO 9001 confirmed firm, is obliged to conform to this standard too. Kind of general rule is that the lowest quality screws used in machining centers is C3, below that the screws are rolled rather then ground. I wouldn't believe that Mazak uses rolled screws in their machines, so probably they conform to C3.

Accuracy with this stuff ? Well, as mentioned before, no promises made in advertised material.

Click to expand...

@Probe you can get test sheets for these machines... Usually their broader tolerances are 2X that of ISO 230-2 type standards and higher than JIS standards and a lot of the machines test out well within certain margins.


I'm wondering in your case whether you are working with machines that are "Clapped" out and should have had their ball screws replaced a long time ago ?



I admit it IS a time consuming pain in the are to have to jump through hoops to get this information...


The thing that's puzzling to me as it seems that you already know what the in field positions and accuracies are for various machines?

Here's the thing on MAZAK broshures for example on the Integrex I 200 etc. they will show a plot for 3 micron circularity in the ZY plane.... So something has to be going right ? YKWIM ? It can't all be a total sh*t show 'cuz they don't tell you what ball screws they use ? I have to assume the engineers at MAZAK at not brain damaged devolved squirrels ? Why would they unnecessarily cheap out on substandard ball screws (for w few thousand dollars) just to compromise the performance of the machine [Obviously a totally false economy that is only counterproductive ?]. That's why I say I think maybe there's something up with thermal compensation techniques?

_______________________________________________________________________________


Here's the thing some of the MTBs DO tell you what class of ball screws they are using but it doesn't necessarily tell you how accurate the machine is going to be...

Personally for the stuff we are doing I like to see fine pitch ball screws but even MTBs like Okuma on the M-460 V 5ax which is an improved (in house) clone of the MU does not use fine pitch ball screws for the sake of speed/ productivity + longevity.

A bit off topic, but did anyone notice how freaking cool the entire sheet metal "house" is dropped right down over the Quick Turn lathe frame??

Now, that is some impressive design and process engineering!