Info I have here says:
"absolute rotary encoders on B and C axes and absolute glass scales on XYZ axes"
"The X, Y and Z-axes have absolute type, linear glass scales that do not require referencing at power on. These
virtually eliminate the impact of slides, ball screw thermal expansion, and help to enhance repeatability. This
improves the ability to hold tight part tolerances with fewer offset changes."
Pretty Interesting....
These
virtually eliminate the impact of slides, ball screw thermal expansion, and help to enhance repeatability
More ^^^ over generalized brochure-like statements (perhaps);
So Slides or linear rolling element slides and trucks etc. ~ basic build straightness should not be affected by reading scales. [even when talking about Abbé errors ]. Some machine builders will claim a true "geometric" accuracy and others are a little sketchy or vague as to whether they correct orthogonal "squareness" digitally and dynamically.
A machine should be able to cut straight and square without any scales or corrections unless a given machine is especially asymmetric for a given set of applications / good reasons.
That's why laser plots that indicate straightness or at least videos of DTI against a granite test bar / plate can be useful.
So
scales will thermally expand and contract and so will the frame of the machine and so will the ball screws unless some attempt is made to thermally regulate those - like - core cooled ball screws.
Some machines fair quite well without scales up to a point depending on thermal management strategies and quality of build geometry.
Most good machines have laser calibrated pitch error compensation (+ tables) for ball screws. [In some cases pre-tensioned ball screws can be an advantage.].
Some MTBs will have
scales that have temperature sensors built in to monitor / measure the anticipated thermal expansion of a given linear scale and will even have cooling points within the machine such as critical bearings and parts of various castings and spindle / spindle housings and other sources of heat.
Each MTB has their own style and way of closing in on all that.
Scales are helpful when integrated properly into a larger scheme of overall "metrological" management of the machine.
Usually process reliability can be improved by
linear- scales -part to part- accuracy but the individual part accuracy might not be improved that much. Card carrying mold machines with long cycle time precision work need every trick in the book to keep everything on course, different need / approach.
_______________________________________________________________________________
Rough mental mnemonic for arc seconds,/
getting a handle on this " Arc second " business...
Roughly speaking, (1° degree of arc) divided by 3600 = 1/3600 = 0.00027777° ... or = 1 arc second or written (confusingly) as 1")
One arc minute being 1/60 = 0.0166666° so 60 arc seconds in one arc minute. 1 arc minute being written as 1'
In terms of practical tilt,
1 arc second is 1 micron (40- "Millionths" [0.00004" ]) tilt over 8.0" (roughly).
+/- 2.5 arc second (5 arc second spread) ≈ 5micron tilt over 8.0" (inches) ≈ to a "two tenths" - spread over 8.0 "
+/- 5 arc second (rotational error/ positional uncertainty) ≈ +/- 0.0002" or 0.0004" spread over 8.0"
+/- 10 arc second positional uncertainty (like the quoted B axis on a universal 5 axis machine ) +/- 0.0004" or 0.0008"
^^^ basically edging in on a "thou" for positional uncertainty (on only one rotational axis) when referenced to the spindle which is about as good as tramming a head in on a 50 year old manual Bridgeport.
It's a lot of money to pay for something when the "Universal" 5 axis machine
idea is that it can function as a good 3 axis machine and can do more demanding positional work.
All errors are additive so smash in your other 22 or more errors beyond tilt errors.
That's why I feel that 5 arc second rotational tolerance +/- 2.5 arc seconds to 2 or 3 sigma confidence level + genuine machine sensitivity of movement to near sub arc second adjustment would really serve a lot better to actually obtain the mythical 'Two tenths" with a machine that has tilted axes. [The idea is (perhaps) to eliminate surface grinders and specialized jig borers and multiple set ups, scrapped parts and too much time in inspection etc. ? + precision fixtures ].
]. 
