Drama with Y carriage polymer bushing conversion

Y carriage conversion to RJMP-01-08 bushings

I didn’t have time to document this, but yesterday I removed the LM8UUs from the Y carriage and replaced them with RJMP-01-08 polymer bushings in printed blocks of my own design. I wrote yesterday that I was able to print a block with a bore right on 16.00 mm ID which is perfect (IGUS specifies an H7 bore, which means that the ID should be within +0.00 mm to +0.02 mm of the nominal shaft size), but what happened yesterday is that I ran out of filament and had to put a new spool in, and that spool has a different flow rate from the previous one and the blocks’ IDs came out far too large (about 16.20 mm ID). I wrapped the bushings in one layer of blue tape (about 0.10 mm thick) to bring it close to the bore size and fitted them to the printer. The bushings fit well, had no play, and ran smoothly and silently.

RJMP blocks loosening

One of the downsides of printing functional objects in PLA is that it likes to flow under pressure, and the bushing blocks from yesterday are no exception. I have noticed vertical play developing in the Y carriage just overnight and after printing a bunch of test objects, which means that the compressive force of the blocks has faded, so the RJMP bushings have gone back to being oversized.

I have three options from here:

Reprint the blocks with a smaller aperture and more perimeters, forcing the bushings into them and leaving them to flow overnight so that their expanded size is equivalent to the desired H7 bore.
Redesign the blocks so that they are tensioned using screws, which can be tightened without having to disassemble the Y axis.
Make the current oversized bores work somehow.

Option 3: Baking the bushing blocks

I was meaning to test oven annealing of PLA parts anyway, and since these blocks are useless as they are, but potentially invaluable if the oven annealing works, this is a pretty good opportunity to get the experiment done. What attracts me to this method is that the parts should shrink along X and Y, while at the same time becoming strengthened and rigid. The existing blocks would therefore shrink down to a more useful bore ID and also become resistant to future loosening.

Following Tom’s method in the video linked above, I gathered a collection of scrap parts (a hole shrinkage gauge with 5/10/15/20 mm ID holes, and 5 half-printed RJMP bushing blocks), cleaned them up with a deburring tool, took measurements of them as they are compared to their nominal size, and then put them in the oven at 100 °C for 1 hour on baking paper.

Table 1. Dimensions before and after baking. The actual size of the bore doesn’t vary much from the nominal value, but it is known that shrinkage becomes worse as the designed hole becomes smaller. The bore shrunk during annealing by an average of 0.85 mm, or 5.3 % of the nominal ID.

Dimension	Rep 1	Rep 2	Rep 3	Rep 4	Median	Mean
Nominal ID (mm)	16.05	16.05	16.05	16.05	16.05	16.05
Before baking ID (mm)	16.07	16.03	16.03	16.08	16.05	16.05
After baking ID (mm)	15.23	15.24	15.16	15.19	15.21	15.21
Print shrinkage (mm)	0.02	-0.02	-0.02	0.03	0.00	0.00
Baking shrinkage (mm)	-0.84	-0.79	-0.87	-0.89	-0.84	-0.85
Total shrinkage (mm)	-0.82	-0.81	-0.89	-0.86	-0.84	-0.85

Oops!

I went ahead and designed bushing blocks with 16.85 mm ID bores and annealed them and everything, but then I goofed while trying to fit heat-set inserts and I basically ruined the screw hole of one of the parts. I didn’t print extra and I wanted all of the blocks to be printed and annealed and installed as a single batch to keep them uniform, so I went back to option #1 and made new bushing blocks with 15.98 mm ID bores (because Rep 4 in Table 1 expanded by 30 microns) and more perimeters. They’re now installed on the printer.

The bushings had to be forced into the blocks with a vice and chased with a hammer and bar to seat them properly, but the blocks seem to be doing well. They could be pushed onto the 8 mm round rod with a bit more force and sticking than normal, and they were so tight at first that the Y stepper was skipping steps and grinding. But after moving the carriage around by hand for a few minutes and then running a few cycles of Y traversal to start breaking them in, the stepper no longer skips and the bushings run without shuddering.

They are, however, still tight on the rod. This is most evident by the shock that goes through the printer when the carriage stops, like the stiction just catches it and brings it to a sudden halt. I’m wondering whether I actually designed them with a 15.98 mm bore or if that’s just the most recent value in Fusion 360, and I actually printed a different size without saving it. I unfortunately don’t have the models I sliced, so I will need to print the first few layers of the g-code that I do have to measure what comes out. I’m betting that I made 15.80 mm bores, and maybe if these existing blocks don’t relax into a larger shape, I will need to reprint as 15.90 mm or something.