When we went to BSC6 to check the viewport for the ALS green beam, we've quickly found out that the beam was not hitting the EY, not even the primary.

The beam was at least 4mm too high on the secondary (picture will be posted by Cheryl). That's a huge number.

Since there's no lens/curved mirror between the ALS table and the secondary of the TMS telescope, and since the distance between the top periscope mirror of the ALS table and the secondary on the TMS telescope is something like 3.5m roughly, we're talking about 1mrad-ish number if it is something on the ALS table, and much much more if it is something on the TMS.

Vincent assured me that the ISI was good.

Hugh looked at the dial gauges of the HEPI and told me that the changes since May/07 are all 0.1mm or less. That's over 3 or 4mm distance. That's nothing compared to what we're seeing.

There was no change in the TMS bias voltages since we "finalized" the alignment a week ago, and though the OSEM values showed some drift it's about 1000 counts maximum, and that doesn't sound that big. Nobody worked on the ALS table since the last alignment. Picomotor driver cable was disconnected last time we finished working on that.

At this point, there are three potential cause of this:

1. Wedge of the viewport

BSC6 high quality viewport was removed from BSC6 for cleaning after our "final" TMS alignment, and then installed again.

Though this is "no-wedge" viewport, the spec (E1100267) says "up to 5 arc minutes", which is 1/12 degrees or 1.45 mrad. Assuming the refractive index of n=1.45 and using a small incident angle approximation, the angle deviation caused by this viewport is approximately up to

1.45 mrad * (n-1) = 0.65 mrad.

If the viewport was rotated 180 degrees after cleaning, that will cause up to 1.3 mrad difference.

Since the distance from the viewport to the secondary along the optical path is roughly 3m, 1.3 mrad roughly corresponds to 4mm shift.

2. Bias of the PZT mirror on the ALS table

We've found that one of the four inputs for the PZT mirrors is 14 Volt for whatever reason. I'm quite certain that the offset was there when we aligned the TMS. Indeed, when I disconnected the input cable from the PZT driver, the beam was totally misaligned and didn't even come out of the ALS table.

We still don't know why there's such a stupid offset, but anyway it's not impossible (though not very likely) that this offset drifted over time. The iris is much closer to the PZT mirrors than to the TMS secondary, so a tiny change in the PZT mirror angle could lead to a big position change on the secondary.

3. TMS mirrors

It's not impossible that the TMS mirrors got bumped. Again this is not very likely, because there are only two relevant mirrors  that are easy to bump, and it's kind of difficult to bump them to cause problem mostly in pitch only.