so when you "Set-up" and verify a digital micrometer or caliper, you have gauge block at varying lengths and you check that the tips measure that accurately. yes, this is a "go or no go" because there is not a lot of "fix" to a digital device, but, there are some electronic "fixes" to set "True zero" (not with just the "Zero / ABS" button and verifying that the "Flats" of the caliper match the "tips" (if you didn't know, only the tips are certified accurate - using the "flats" of a caliper is not per the standard...
a "master gauge tech" CAN pull the head on a caliper and "Adjust it" for an accurate reading over the length of the caliper, but most trash them and buy a new "calibrated" device... as
@Mike Davis mentioned most of these devices are double or triple the cost of one you or I would purchase with a COC (certificate of conformance)
I understand the "tips" but even my cheap ones are darn square when up to the light. As far as accuracy over length, that is defined by the pattern of the etched strip under the scale just as a dial is based on the accuracy of the tooth grind and a vernier based on the accuracy of the etching. Unlike a dial, a digital that is dirty could get a false trigger or skip across the scale as the basic technology is a 2-phase capacitive sense. Any de-lamination will cause a localized error. A dial could drift due to gear wear. All of them can only be accurate at a given temperature. The biggest error is more likely any play in the jaw against the body as the tiniest rock will be amplified at the tip. Basic a reason a mic is more accurate than a caliper. A simple go-no block can identify that. I wonder how many of those high end brands folks claim to use for 50 years daily have worn enough to cause a bit of head rocking? Do they verify every now and again with a gauge block set?
The "zero" button only changes the display in the processor. It does not change the sensor position, so maybe, maybe there could be some adjustment for the first half a thou, but I would need that explained by the design engineer of the technology. The one I tore apart was glued in place. This initial position could be the difference between low end, accuracy of .001" ( but should really be +/- and digit to be an accurate spec) and the mid to high end reading .001" but accuracy .0005.
A possible feature low end are venerable to is as a battery just starts to fall off, epically an alkaline that drops slowly vs a silver or lithium, you might get processing or sensor error smaller than when you notice it getting flakey. My Wixy (CR2023) went from reliable to erratic very quickly so I noticed it, but my cheap calipers with L44 were harder to identify when they got funny. I guess the S44s would be more obvious but they have half the energy of an alkaline.
From the mid-price up, a NIST traceable cert looks like it costs about $60. I don't know what annual verification costs. When I was in industry, we had our own met-lab so the cost was not identified to the project or contract. I would hope that from mid price up, they are verified to a standard at end of line. I doubt the $12 ones are though I could imagine an automated process with optical reading and a little bit of AI could do the job if one were to build millions, Too expensive for thousands. Next time my Fluke battery goes dead, maybe I'll use it as an excuse for a new one as the range switch is wearing out. Come to think about it, one of few electronic devices I have had that works fine for 30 years and on only the second battery. Yea Fluke.
For my uses, consistency and relative difference are the measures that are most important. Not absolute. The one I ordered, due to day, is a CR2023. I buy batteries in 10-packs from Amazon. Had good luck with their brand big packs of AA and AAA too.
Hmmm. My tringles are probably 50 years old and don't have chipped up edges. Good Post ones. Just tossed in my desk drawer. It may depend on the material. Mine I think are Lexan, not plexi. But for setup, as they are clear, not as handy as metal machinist squares as backlight is easier to see.
I found another measurement error last night. I bought a flat tip for my dial indicator to make setting up my jointer knives more accurately. ( but put in a Byrd head anyway). Used it to verify my TS. It is not flat! As I slide across the corner of a tooth , it changed by half a thou. Cheap tool. Back to the conical tip.