Another status update, and a fix that will help anybody whose printer isn't printing round holes:
Status update -
I only recently found out that you're supposed to scuff up the Wham Bam PEX surface with either a green Scotchbrite pad or #000 steel wool before using it the first time. This does 2 things, it vastly improves surface adhesion, and the parts now have a satin finish instead of s glass smooth finish.
The FIX -
I also recently discovered that my printer was not dimensionally accurate at larger sizes. Below 40mm it was reasonably accurate, but above 100mm discrepancies became REALLY apparent. I spent 2 days using my usual method of calibrating the printer, printing 20mm cubes and/or a 40mm wireframe cube and using the slicer's scaling to try and correct this - and got no where. The real problem was that while all 3 axis were printing at about 98% of their actual size, they were all different. X was 98.6%, Y was 97.4%, and Z was at 98%,
I decided to use "old school brute force" to correct this. I used a method I discovered/developed for my first 3D printer, over 10 years ago, when it refused to print round holes - instead it printed ellipses. What I did was to use a pair of M commands that the Marlin software recognizes, M92 and M503. The M92 command lets you set the steps/mm settings for the X, Y, and Z axis, and also for the extruder if you need to. Since I didn't have the source code for my printer I needed a way of finding out what the default values were on the EPROM. That's where the M503 command comes in. When you directly connect the slicing software to the printer, you can send the M503 command to the printer. The M503 command tells the printer to print the settings it has loaded in memory to the computer's screen. This gives me base-line values to use with the M92 command.
Since I didn't want to have to directly connect to the printer and "FIX" the steps/mm settings EVERYTIME I turned on the printer, I needed a way to do that automatically. Enter the custom g-code "On-start" script. I added a line to that script that sends an M92 command to the printer with the new steps/mm settings every time I load a g-code file. There are 2 down-sides to this, first - the printer sees the M92 command, gives me a warning message that this will overwrite the printer's settings, and asks if I'm REALLY sure about doing this. Second, each filament can print differently, even different spools of the same filament can react differently. To be REALLY accurate I SHOULD test every spool of filament. To do that, I need a different. and faster, method for checking dimensional accuracy. (My 20mm cube takes about 1/2 hour to print, and my 40mm wireframe cube takes almost an hour to print.) I need to check at LEAST in X and Y, since Z doesn't get anywhere as much use as X and Y - so the error doesn't have as much chance to accumulate.
What I came up with was my 100mm "calibration square", this gives me a quick check on the X and Y axis. This is a square with 100mm outside dimensions, 2mm wide walls, and is less than 1mm thick - and it takes less than 2 minutes to print. The first check, using the factory defaults, printed with X at 98.6mm and Y at 97.4mm - I'm not concerned about Z yet. I made corrections to X & Y and tried again, this time I was at 99.5mm on both X and Y. I made one more tweak and tried again - this time both X and Y were at 100.1mm, +/- 0.05mm. I decided that was close enough for Gub-mint work, averaged the X and Y settings and applied that value to Z, then printed a 40mm wireframe cube. An hour later, this also was declared good enough for Gub-mint work.
Soooo.... In less than 2 hours I was able to accomplish what I was unable to do in the previous 2 days using my previous method.
Don
