Yesterday I spent the day rebuilding the HXTS prism gluing measurement apparatus in order to accomodate the HLTS sized optic. At the end of the day I finally glued the prism on. Today, Gerardo removed the fixturing and sees that the prism is 2mm out of spec in the pitch-direction. Ugh. The d-value direction (or distance up from the scribe line) has been corrected. Both of these dimension differences from the first round of gluing can be seen in the attached picture where the phantom footprint from the last glued on prism was. The new prism should line up with the phantom line on the left end of the prism, but should remain at the current horizontal position on the line. I estimate that this error would induce roughly ~40mRad of pitch error. Because this glue bond is still somewhat fresh (only an overnight room temp cure) and the position error is so large, Gerardo is placing the prism down into an acetone bath for the night to remove it.
Mulligan-again for tomorrow.
It is unfortunate that you cannot check any of the setup measurements once the prism is in the fixture and glued down. We *think* this happened because the measurement apparatus is not very good at staying alignment when traveling up and down (away from the optic) in order to adjust for focus. You focus on the scribe line on the optic during setup and then you must slide up to focus on the notches on the prism which are many mm of focus different. Likely this focal plane change has some sideshift (on the order of the 2mm error we now see). I'm not sure what a good fix for this is this. I'll try removing some of the brackets between the traveling bore-scope and the horizontally mounted height guage.
Playing with the measurement apparatus today, Gerardo and I discovered that the bore scope 1) wobbles up and down relative to g, and 2) runs downhill by ~3mm over a 300mm length due to the weight if itself hanging off the end of the horizontally mounted height guage (see pic attached). Both piggy-backed jacks have slop in this direction. I can remedy 2) via shimming the entire stack of joints, but I'm not likely to be able to improve 1) without rearranging the entire measurement motivation. Hopefully 2) was the larger contributor to Monday's error. The top portion of the apparatus is the same as what I used to glue prisms to MC2 and PR2, so we would have see some slop then as well, however since the optics were smaller, the travel (and subsequent shifting of load) must have been less noticeable.
Attached is the LLO gluing apparatus, shown with an HLTS optic. It also is stacked lab jacks.
