A while ago, Dmass made some beamscans of the PSL beam after it passed through two lenses. Here is a link to the Dmass Modematching elog.
We wanted to make sure that these measurements are consistent with what we think the incoming beam parameter is. I wrote a matlab file which works backwards from his data to find the incoming beam parameter using ABCD matrices. I used the focal lengths indicated on the lenses.
I found a disagreement of 20cm in the PSL waist location with a yesteryear elog entry Dmass checked to see if my computation matched with his, and got the same result.
The first thing that we thought might have been wrong is that the focal lengths need to be switched, but switching them didn't significantly our change our results (in fact, it made them worse).
We also suspected that the engineering tolerance of the lenses might allow for the difference. I tried changing the focal lengths of the lenses by 10% for both lens orientations, but the computations made only small changes to our original result, so this is probably not the culprit.