So I've owned my Mendel90 for almost exactly two years, but only began to really enjoy the printer lately. When I realized that my opinion of the Mendel90, actually 3D printing in general, had changed, I did some soul-searching.
Most of the problems are my own, and not issues with the Mendel90, but there was one issue that I can't take the blame for: A little bit of play between the hot-end and the Wade's block. The result of this was an off-set on the bottom layer (probably because the nozzle was dragging due to another problem) and that in-fill didn't quite make it to the perimeters. It didn't really cause prints to fail, but quality suffered. Luckily, I managed to print a Wade's block that fit my version of the Mendel90, which happened to fit quite tightly, and those two problems disappeared. Happily, the new version of the Wade's block as used in the latest version of the Mendel90, connects to the hot-end in way where this can't be an issue.
I should point out that the problems that I ran afoul, have largely been addressed in the current version of the Mendel90.
By far my biggest problem was described
here. Nut-shell: The coupling between a lead-screw and Z-motor managed to work-itself loose and eat away the plastic in the coupler, loosening the connection to the lead-screw as it progressed. This was almost definitely my error, probably a result of over-tightening the coupler, but did it ever cause a lot of frustration. I had to recalibrate the Z_HOME_POS and bring the X-axis into parallel with the bed constantly. Many prints failed as a result. Since figuring this problem out and replacing the problematic coupling, I've calibrated the machine exactly once. To be honest, other than attempts at 0.1mm layer-height, I don't think I've had a failed print with this problem fixed.
The rest of my issues were related to learning about plastic, especially how to get it to stick to the bed, and to understanding the design constraints one works under when using 3D printing for production. I also made the mistake of buying really cheap plastic, a painful mistake for a beginner. Before the coupler problem became too bad, I could print PLA fairly reliably. Switching to ABS was quite a learning experience. Learning about ABS-juice, then Kapton tape, and finally PET tape, helped, but with my machine constantly out of calibration due to the failing coupler, failed resulted more often than not. With the coupler sorted out, I tried Kapton tape, but found the plastic to stick to it too aggressively. PET is just right, holding tight while the bed is hot (120C), and releasing when it cools. Learning about the design constraints when manufacturing with a 3D printer was straight-forward, but with all the other problems I was running into, was just another straw on the camel's back. With the other problems out of the way, it doesn't bother me that much.
I suppose another change has been how the open source tools have come a long way in the last two years. My experience with commercial CAD tools made it hard to accept Blender and OpenSCAD. I like both of them just fine for certain things, but the lack of a constraint- and sketch-based tool like SolidWorks or Autodesk Inventor was painful. Even SketchUp is a step up from Blender and OpenSCAD for most tasks. Chance had me run into FreeCAD recently. I had tried it a few years ago, when I had access to Solidworks and Inventor, and therefore was disappointed with FreeCAD's lack of constraints in sketches. But FreeCAD has come a long way since and now has a decent constraint and sketch system. FreeCAD is still young, but is becoming quite useable. LibreCAD has also become part of my toolbox of late, and deserves mention.
Only a month ago, I wouldn't bother to design anything with 3D printing in mind. Given the large chance of failed prints and the quality issues I was having, it just wasn't worth the trouble. Slicing an STL downloaded from the usual sources is something that I would do grudgingly, but only if there was a real need. Fast forward to last night: I was about to run out of ABS and my only roll of PLA had been "relaxed" to a spool with a small diameter spool. I needed to come up with a spool that could fit in my oven and have a large diameter barrel. After considering making something in cardboard or wood, I realized that a simple part could be used to drastically increase the barrel diameter of some of my empty spools. It took all of about 15 minutes to draw up the part, and fire up the printer. 90 minutes later I had eight instances of the part installed on a spool. This morning, the oven is cooling to the right temperature and the filament should be ready for printing this evening. My whole outlook on 3D printings has changed for the better.