All the linear stages were removed because they will have to be carefully aligned during assembly, however the frame and servo drives remain together.
I also placed some of the additional components on a nearby coffee table. This picture doesn't look particularly intimidating but there are about forty bags in that Amazon box.
I started by flipping the printer on its back and attaching the Z stage. You can't see it from this angle, but the Y stage has been attached as well (its cables can be seen cutting through the center plate).
The Y-Axis linear stage will be used as a fixed reference because it is not easily adjustable. Because it is my reference, I will not move or adjust the Y axis in any way. Instead every other component of the printer will be positioned around it.
The first alignment I will be measuring is Z axis squareness. This alignment is measured by placing a granite square on the Y axis and a test indicator to the Z axis. I then slide the Z axis to the upper limit of its travel and check how much the test indicator has changed.
The indicator has moved about a quarter of the dial between the last two pictures, which tells me that the Z axis is leaning forward 0.18mm (0.007") over 200mm of travel. This is an error of 0.1%.
This error isn't particularly material because a squareness error in a 200mm tall part is likely to be swamped by other printing artifacts. On the other hand I'm trying to do a good job with this printer so I'll take a stab at fixing it.
There are a couple ways to fix this alignment. The most permanent option would be to scrape or lap the frame into alignment, but that is a very complex process. An easier method would be to cut a set of steel shims, but that is also fairly inconvenient. I chose the easy way out - which is to shim it with Kapton tape.
A single layer of Kapton tape is 0.04mm (0.0015") thick so to make my Z axis square I should need about four layers of tape. I don't recommend shimming with tape in applications that will experience a lot of force, but for something like a 3d printer it should be sufficient.
I decided to play it safe and started with three layers of tape which slowly taper out to a single layer near the middle of the axis. I then reassembled the axis, zeroed out, and checked the error.
My Z axis is now leaning forward 0.05mm (0.002") which is almost exactly what I expected based on the amount of tape used. One more application of tape should get it close to perfect, so I took it apart and threw on a fourth layer.
Reassembling to test confirmed that the Z axis now has a total error of about +-0.01mm over its full travel. Pretty impressive for a bit of tape.
So far I've only checked the back-to-front squareness of the Z axis, but I also have to verify the side-to-side alignment. This is a very similar process, just with the square rotated ninety degrees.
The Z axis is out about 0.04mm (0.0015") in this direction which is actually surprisingly good.
Since the Z axis has some freedom we don't need to shim it. Instead the axis is aligned by carefully tapping it with a rubber mallet. This might sound silly but it works extremely well.
Start by loosening the screws and slowly tap it into alignment. As you get closer, gradually tighten the screws so it moves less with each hit.
I won't bore you with more pictures of the alignment in progress, but tapping it in is much faster than shimming because you can adjust in smaller increments and immediately check the results without having to reassemble.
With this alignment finished my Z axis is now square to the Y axis.
Even though currently measured squareness is within +-0.01mm, in practice I'm expecting that error to increase over time. The alignments will move once the X axis is added to the printer, and frankly this design is simply not rigid enough to hold +-0.01mm under load.
Even 1/2" aluminum plates start to look pretty flexible when you are counting microns.
After aligning the Z axis I made sure all the fasteners were appropriately tightened and introduced the sheet metal we will need for the next step.
No comments:
Post a Comment