In March 2012 I installed a 4-axis Electronica DRO with 0.0002" scales I purchased from DROPros. However, had I known then what I know now I probably would not have purchased it. The sales people at DROPros are great, the price is low, and the unit is easy to use, but my measurements have found an issue with the stability of the material used for the scales.
Making a long story short, a week or two after receiving the DRO in March 2012 I finished installing it. On a day when everything had sat for at least 8 hours at 20 oC, and when zeroed with respect to the mill's dial at one end of the travel, it read low by 0.0010-0.0016" at the other end of the full 24" travel. This is consistent with "perfect" DRO scales and a perfect-when-new lead screw on the mill that has ~0.002" of backlash at one end of the travel. Since this was good enough for my purposes, and since it was within the claimed specs of the DRO as well as the accuracy of the test I did, I did not try to check it better than that at that time
Twenty months later, in December 2013, and with everything equilibrated at 15 oC, I rechecked the scales using a centering microscope and the scales on the side of 18" Mitutoyo vernier calipers (which are spec'd to be accurate to +/-0.002"). To my surprise I found the DRO readout to be low by ~0.014" at full travel, i.e. well outside the specs of the calipers and 10x worse than when new 20 months earlier, and well outside the claimed specs of the DRO. I rechecked this using 25" B&S vernier calipers and got the same result so I used the deviation I found to input a linear deviation factor into the DRO's readout. Note this new "calibration" was only done using vernier calipers, and I didn't know whether or not the deviation actually was linear, i.e. if the scales had uniformly expanded over their entire length, or if the expansion was different in different regions, so everything would have to be checked more carefully.
Because of what I had found a few weeks earlier, in January 2014 I used my recently calibrated grade 2 Mitutoyo gage block set to check the scales again. This time at 20 oC I found the DRO readout was within no worse than +0.0004" over any 9" length of the full 24" travel. That is, with the new calibration constant I had entered a few weeks earlier, which was significantly different than the one it came from the factory with almost two years earlier, the DRO was working well enough for my purposes.
However, even though things were fine in January, this does not mean the scales won't continue to change their length in the future as whatever stainless steel alloy the factory used continues to age. So, during this period a third party put me in contact with an engineer at the Electronica factory in India (it only would have made matters more complex to try to deal indirectly with them through DROPros, which is only a sales outlet for them). In our initial exchanges the Electronica engineer was quite responsive. However, he went completely quiet after I provided him with information to indicate I have the expertise and instrumentation to be quite confident my measurements are correct. I have no idea why he went quiet because he stopped responding without providing any explanation.
It appears to me from my measurements that there is a fundamental problem with the temporal stability of the material they use for their scales. Whether this could be solved by them finding an accelerated aging process for the steel, or letting it sit for a few years before they apply the magnetic scales and calibrate it, only could be addressed by long-term measurements to determine when the aging has slowed down sufficiently so as not to be an issue.
Meanwhile, I basically have two data points on how the scales change their length with time. When new (presumably within a few months of manufacture) the fairly crude check I made at that time found them to be within a few thousandths of whatever length they had when calibrated at the factory. Twenty-two months later they had increased their length by ~580 ppm from the value they had when calibrated at the factory. This corresponds to a position error of ~0.006" out of 10", which certainly isn't negligible for many things made on a mill.
When winter comes and I again can make measurements at 20 oC I certainly will. My hope is that I will find the changes have slowed down. At this point, at least, returning the unit and starting all over again isn't something I want to contemplate. Meanwhile, these changes are a headache I will have to live with, that I wouldn't have to live with had I purchased a unit with glass scales.