I recalibrated the build table, then the projector. But this time to 70 micron XY. I then tried the calibration print using the PC. Nothing appeared on the build plate. So I tried with the Mac, and it half-worked. I will not try with the PC again since its never had a successful print. The B+ side of the print is ripped, thin, and fails partially. It fails just like this guy's failures: viewtopic.php?t=4411
I ended up trying the calibration 6 or 7 times, with 4 or 5 re-calibrations of the build table and projector. The reason I tried so many times is because each time I would think of some new reason of why it would fail (such as lens dusty, non-stirred resin, old resin, maybe damaged vat). And so with the last failure, I used new stirred resin, new vat, perfect calibration, etc... but in every case, the print turned out exactly the same.
I now think it has to be one of these two problems:
1) The balance is off. I guess this should have been obvious from the prints, but the printer was factory calibrated to have 0 balance and 0 fade, and it did have one successful print before.
2) The build table calibration is not happening correctly. I noticed with both vats when I looked from below that when it should have been flush with the PDMS, it was only contacting one side of the table. If I put some pressure on the side that was too high, the table would contact the pdms, but then revert back away from it when I removed the pressure. A couple times I tried to hold it in the flush position while tightening the screws, but this didn't change the printing result. I noticed that the red table doesn't move smoothly when the 4 screws are loose - the movement is rough when I do it by hand, so maybe that results in the build table not being calibrated correctly.
One last IMPORTANT bit of info: during each print, I looked from below at the build table. For every calibration print, when the projector was projecting the calibration grid onto the PDMS and when the table was close to the PDMS, I would notice that the B+ side would be more RED. Even on layer 8, when the projector was projecting the calibration image, I would notice the squares would be more RED. I'm not sure if this means that the build table was closer to the PDMS on the B+ side, or if it means that the projector light is stronger on that side.
Pictures to follow.