I found the idea of "bastelfreud" to 3D-print a USB-microscope great - besides the dovetail. In a first attempt I modified the angle of all dovetails to 60° each (better to file?), but was not convinced about the results - even filing with great accuracy it results simply having too much tolerance and as results parts are wiggling... So I found a simpler solution, resulting+ with much more precision: A drill usually has a great hole-precision, without the need to file anything... But printing holes are normally resulting smaller (about -0,2mm), so ø2mm-holes can be drilled relatively exactly to ±0,01mm diameter (minus the material-flexibility)...
And using bike-spokes with ø2mm diameter this would result in a carriage with greater precision. But: How to make this rails going smooth? ⇒ Using a trick: If we drill-enlarge the printed holes (here with a ø2mm-drill) a bit greater than 2mm (drills normally are a few hunderts mm smaller as there nominal diameter – f.ex. a ø2mm-drill has normally about ø1,97 to ø1,98mm) – we may use this drill as a file (in longitudinal direction !), to increase the hole a bit, until the spoke slides gently through → "filing" only a few hundreds of mm are normally enough, prove this manually upto slide smoothly... ⇒ Glue PLA-parts with superglue! (here: the 2 "MidTables", "GearRack" with "HolderArm" and the 4 RubberPads under the groundtable). ⇒ The 4 (PLA-)bushings to hold both M3-Carriage-Screws are rounder than printing this holes directly, so drilling them with a ø2mm drill and pressing into the corresponding crosstable ø5mm-holes will result in more precision than printed directly, because M3-screw-diameters have approx. ø2,75-2,8mm...