What's new
What's new

Advantages of linear scales on a machining center?

snapatap

Aluminum
Joined
Apr 24, 2013
Location
south australia
What are the advantages of having linear scales on a machining center?
At work we a looking at a new machining center with 3m of X travel and a B axis rotating head (Doosan VCF850LSR with Heidenhain 640). One of the options we are considering is linear scales, just trying to work out if it will be a worthwhile option.
 
I work for Doosan but even if I didn't I would always order linear scales on a machine. Especially one such as the VCF850. Constant feedback from the scales means much better accuracy no matter where you are on the table.

Paul
 
Scales takes away the thermal growth differences between ballscrews and castings in most build designs. In a 5 axis application it helps the end user evaluate reasons for mismatch and such by removing most if not all of that main variable out of the picture.
 
In general the scale removes any mechanical error, by giving the control position data from the moving part of the axis. On an non-scale machine the feedback is from the motor, not the moving section.
 
In general, scales are a good choice.

Scales are not a cure-all for machine accuracy, but in most cases they help position accuracy. Adding scales to a poorly designed machine won't make it great, but may make it better or may not help at all.

Likewise scales can create issues of their own. Scales can mask mechanical issues like excessive lost motion. A machine with scales can look good while being tested with a ballbar or laser, but behave poorly under cutting forces.

Despite all the above, I always spec'd scales on machines I recommended for purchase. On a 5 axis purchase, be sure that when you spec scales that it includes the rotary axes. Some builders split the option so that one option covers the linears and another the rotaries. I got bit on that once......
 
I believe you mean the feedback comes from the encoder on the ballscrew, not the motor.

Depends on the design. There is a rotary encoder on the back side of the servo motor. Many machines just look at that. While you can add a second rotary encoder directly to the ballscrew, it's not required.
 
Temper your expectations when it comes to linear scales. It won't make the machine immune to positioning error caused by temperature fluctuations. Rather, it's an incremental improvement. Adding climate control to your shop, if you don't have it already, is still the lowest hanging fruit.

On a 3 meter travel machine, you'll likely see improvement over longer distances. Over shorter distances, e.g. 12" or 300mm, the machine is probably highly accurate even without scales, not leaving much room for additional accuracy.
 
I believe you mean the feedback comes from the encoder on the ballscrew, not the motor.

Although I admit to not having seen them all, I personally have never encountered an encoder mounted to a ball screw. Though come to think of it I'm sure it's been tried. Would certainly complicate things with another connection, sealing and mounting. Seems an interface inside a motor case is a much easier nut to crack. Regardless, with rigid couplings, the screw and motor shafts are practically one anyway.

Now you've got me curious as to what brands if any run encoders mounted on ball screws.

Sorry... don't mean to be hijacking anyones' thread. It's still all about position feedback idn'it? :-)
 
....Now you've got me curious as to what brands if any run encoders mounted on ball screws......

Older Okuma HMCs all had dual encoders standard. One on the motor and one on the screw. Have seen a few lathes with encoders on the screw when using a timing belt connection between the motor and screw.
 
Temper your expectations when it comes to linear scales. It won't make the machine immune to positioning error caused by temperature fluctuations. Rather, it's an incremental improvement. Adding climate control to your shop, if you don't have it already, is still the lowest hanging fruit.

On a 3 meter travel machine, you'll likely see improvement over longer distances. Over shorter distances, e.g. 12" or 300mm, the machine is probably highly accurate even without scales, not leaving much room for additional accuracy.

That's exactly right as scales are not made of invar - they too thermally expand and contract but also need temperature sensors in the scales and machine to enable better calibrated and "Computed" - compensations.

bi-directional repeatability along a single axis is certainly improved - all being well. [not the same as absolute accuracy.].

Sometimes the body of the scale (alloy) and the machine alloy and the work piece (ferrous) alloy can be of a similar coefficient of thermal expansion. So things naturally come together to be pretty damn accurate ; barring un-even heat build up in a machine.

long ball screws like what 2outof3 mentioned can cause real mayhem... [Without some sort of second reference system like scales ].

[ball screw (dynamic) shaft wind up + heat build up + pitch error compensation ]: Needs scales and maybe core cooling / either or depending on design. + pre tensioned ball screws + floating spring loaded bearing at one end to take up thermal expansion in some cases (Makino do that on some of their machines.).
 
Older Okuma HMCs all had dual encoders standard. One on the motor and one on the screw. Have seen a few lathes with encoders on the screw when using a timing belt connection between the motor and screw.

Hmmm... the belted lathe idea makes some sense. On the horizontals I'm assuming dual rotary encoders are used, and I have to wonder what the engineers decided on for the logic when interpreting the twin feedback signals. Master-slave, how/why and to what degree, or was one more or less there to eventually throw an alarm when the discrepancy got too large? That would sure be one expensive setup for a service reminder. Though it would certainly be one you could believe in if it came calling.
 
.... On the horizontals I'm assuming dual rotary encoders are used, and I have to wonder what the engineers decided on for the logic when interpreting the twin feedback signals. Master-slave, how/why and to what degree, or was one more or less there to eventually throw an alarm when the discrepancy got too large? .....

One is the velocity loop and the other is the position loop so they do not fight each other or disagree.
Bob
 
...... I have to wonder what the engineers decided on for the logic when interpreting the twin feedback signals. Master-slave, how/why and to what degree, or was one more or less there to eventually throw an alarm when the discrepancy got too large? ....

Don't know why Okuma did the dual encoders. This was on 80s to early 90s machines. No idea if they still do, but I doubt it. The CNC did compare the feedback of the encoders. Because Okuma encoders of that era were true absolute encoders there was a shift or offset value that had to set when you replaced one. I still recall the "DIFF OVER" alarm that would pop up when there was trouble with one.
 
Dr. Kitamura of "Kitamura" (machinery) fame has a few patents on dual ball screw drive mechanisms for machine tools.

If you think about it, it's almost impossible to grind two ball screws exactly the same + different pitch error compensation (tables) for each parallel dual ball screw system (close side by side arrangement).

I haven't dug up his patents yet but I guess one solution to that problem is offered. I think his aim with the dual ball screws is for really fast rapids on a box way machine. 'peeps" from the Makino quarter kind of scoff at that idea somewhat as such forcibly applied rapids esp. acc and dec would cause heat build up through some sort of hydrostatic action / friction from forcibly moving so fast on boxed ways.

____________________________


Curious as the Okuma M-460V 5ax has no direct scales , linear or rotational yet one engineer at Okuma Gmbh / Germany makes pretty astounding claims for accuracy (rotationally) and linear XYZ. [usual caveat of cut, in cut and air moves / measurements ].





In English ^^^

and




And in German ^^^ some additional technical detail and subtleties.

Personally I don't 100% believe his claims or assertions especially for machine that has no linear scales and no direct reading rotary encoders - but even if his claims are true THEN you start to wonder about long terms accuracy and repeatability as various contacting components start to wear (hypoid gears etc.) .

It's a nice machine for sure not bashing Okuma nor Jürgen Kläser - but seems a trifle overselling on positional accuracies that seem a little far fetched ?
 
We have 3 machines with heidenhain scales ranging from 1800mm to 4300mm travel, they all still have comp tables from laser calibration and despite having thermal sensors to account for temp variations in the spindle/ rams they still have thermal growth that is unaccounted for by the control. so yea, while they are much better than no scales they're still not perfect., these machines are around 20 years old though.
the MTB's implementations of the control will dictate how good the results are.

The nice thing about Heidenhain controls with scales is you can home them from anywhere on the travel range, the control only needs to traverse 2 distance coded reference marks and it knows where it is. about 20-40mm of travel.
 
If you think about it, it's almost impossible to grind two ball screws exactly the same + different pitch error compensation (tables) for each parallel dual ball screw system (close side by side arrangement).
......
Personally I don't 100% believe his claims or assertions especially for machine that has no linear scales and no direct reading rotary encoders - but even if his claims are true THEN you start to wonder about long terms accuracy and repeatability as various contacting components start to wear (hypoid gears etc.) .
?

I believe at least one member here has built dual screw drive machine tools with linear position feedback although not close side by side.
If you have two servos and two scales on one axis is that "quad loop"? Four counters on one axis, who do you love.

I like the Okuma five axis fine tuning with a probe and sphere. Seems a great idea to map out rotational errors automatically.
Wonder how long it takes to run that and how that map is handled by the software with direction reversal, axis loading and such. That would be an interesting project.

Having built grinders with scale feedback and even two scales (abbe correction) and one motor I'm not sold on it as worth the extra cost, maintenance, complexity and loss of servo stiffness.
Accuracy over distance for sure better with scales. Repeatability in a run maybe not as good as a no scale system.
Of course with linear motors scales a must.
Bob
 
Sub spindle on some Mazak lathes is belt driven and encoder is on the ballscrew. This is for accuracy and safety. If that belt breaks the machine stops and that keeps the turret from a nasty collision with sub spindle.
 








 
Back
Top