Welcome to the SeeMeCNC Community

I cannot seem to get my x and y dimensions in printed parts to come out symmetrical. When I print a 50 mm calibration cube or a 100mm calibration cube, my X measurement will be within .05 mm of the expected dimension of the cube most of the way across. When I measure in the y dimension in the center of the cube, it will be ~0.4 mm smaller than the expected measurement. What is more odd is that left side of the cube is consistently longer in y direction than the right side of the cube. The variation is by around 0.4 mm as well. This means the y-axis a third of the way in may measure 49.8mm, the center might measure 49.6mm, and it might measure 49.4mm a third of the way from the right. Meanwhile the X-axis will measure pretty close to 50mm all the way across. The difference doesn't appear to scale for a 100mm cube. It stays approximately the same.

For a long time, I thought it was the cheap T-square that I used to level my towers or some other build error on my part. I finally got a digital level that tells me that all of my towers are less than 0.2 degrees off of vertical. I'm inclined to think this is acceptable, but perhaps it is not.

Still scratching my head, I printed an object from thingiverse (http://www.thingiverse.com/thing:240962) thinking perhaps my individual delta radius could fix the flaws. It turned out my x tower was a full 3mm farther away from the object than my y and z towers. I adjusted this in the EEPROM by changing my Delta A radius to 3. After that, the object printing an equal distance from all 3 towers. Unfortunately, I got the exact same offset in the x and y planes.

Does anyone have a clue what could be causing this discrepancy?

Thanks

Bumping this message to see if anyone can respond.

It staying approximately the same leads me to think that it may be that extrusion width. Have you got any objects that have a clear view of a layer, say a good flat bottom? It may also be that mis-aligned hotend Geneb mentioned in another thread, with the misalignment torquing causing it to torque slightly, producing your bottom to top variation.

Is the distance between the top and bottom plates exactly the same for all towers? When you home the machine then move the extruder nozzle down to near the base of each tower, in order to set the zero-adjust screws for each carriage, how different were the Z distances? The Rostock MAX v2 design (haven't seen the other models) depends on screws in the top and bottom tower clamp pieces hitting the ends of the tower extrusions to set this distance equal for all three towers. In my kit, whoever filed down the sharp edges off of the extrusions got... a bit over-enthusiastic, and the screws are hitting a filed-down area (about 0.050"!) instead of square ends of the extrusion. This throws off the accuracy of the machine, until you adjust to allow for it.
I would have returned the parts if I had figured this out before finishing mechanical assembly.

I'm working on changing out a few parts to see if it gets fixed. I have adjusted the extrusion width previously, but I think that it should not vary in the x and y directions unless it is moving at a different speed due to thinking it is moving a longer or shorter distance than it is moving.

I recently got a new effector platform, so I am switching that out to see if it helps. I have to cut off those annoying nubs (not the main circular ones, the half circles) to make the axle adapters fit flat. Seriously, why are those there? I'll report back if that fix helps.

Update:

I think I have it fixed now. The 50mm calibration box is now off by less than .05mm in the x and y directions. I think that is well within my margin of error for trying to line up calipers on layered plastic.

I changed out both the carriages and the effector platform. My previous setup was the wooden style cheapskates, the V2 platform, and trick laser 300mm carbon fiber arms. I now have the injection molded carriages, the newest style effector, and the same arms.

As mentioned earlier, I cut off the nubs that would have made the axle adapters sit at a funny angle to the platform. This new setup seems to have solved my issue, but I cannot say whether it was the cheapskates, the effector platform, or the way the arms were previously attached.