(David H, Matt H)

So a silly assumption that I made has caused me to make what I consider a rookie mistake and thus we are having a few problems with TCSx alignment.

The story thus far. We are trying to set the beam height to 4 inches in height to be flat across the table. This was accomplished up to the polarisers after the Beckhoff controlled waveplate (see layout) and for the ISS photodiode alignment, the quadrant detectors. Now the silly assumption I made (and I kick myself as it was hammered home to me all throughout my phd to never assume anything, and thats been my mantra at LIGO, until this one time...and its bitten me..*sigh*) is that because the two polarisers and the flipper mirror (which is next in line for the central heating beam path) were all in fixed mounts that if the beam coming onto them was parallel with the table at 4 inches, then the beam after it had reflected off all these optics would still be parallel with the table at 4 inches.

With this assumed knoweldge and because the optics had all been laid out in roughly the right position in advance making the table cluttered (no excuse I know), to check if each mirror, beamsplitter, etc I used in the path after these optics  was projecting the beam correctly I placed a target as far away as I could and projected the beam onto a target and altered the vertical adjustment of each mount until the beam came to 4 inches on the target. Now this technique only works if the beam incident on the optic you are testing is horizontal with the table, if its not you are just setting the beam height to be 4 inches at that one spot and the beam path could actually be going up and down. What I should of done is what I always normally do with optics and check the beam height up close and far away and see that it stays at the same height...lesson learned....*sigh*.

I had aligned the central heating path with a little trouble but got a solution so that the beam path would go be able to go through the irises we had marked as our optical axis to get the beam up the periscope and into the chamber and onto the cp correctly (this we had done the other day with the HeNe beam....Im sorry on the layout I forgot to mark the iris location but they are between the last gold mirror (at the top of the page) and the periscope). The trouble came when we tried to get the annular heating path to also go through these irises, we couldnt find a solution to do so. David came up with the brilliant solution to shine a HeNe through the irises the other way and see what the beam path should roughly be for the annular heating. What we found is that the HeNe beam shot over the top of the mirrors used to steer the annular heating beam onto the beamsplitter meaning no solution was in fact possible with what I had laid out. 

So whats the problem. The problem is in my assumption. With everything in its current layout, we used a combination of CO2 beam and HeNe beam to look at what the reflections of the polarisers and optics in the flipper mirrors are doing. What we found is that each of the two polarisers after the HWP both kick the beam up, as does the first flipper mirror after the polarisers, and the 2nd flipper mirror kicks the beam down (we are talking numbers like 1/4 inch over say 15-20 inches). I checked the two flipper mirrors set aside for the TCSy table and they both kick the beam up as well.

This did not bite us with the central heating beam (for the moment, it would of when went to put the mask in and wondered why the beam was not at 4 inches) as the otics are all nice and close to each other, so even if the beam is going up and down a bit, they still hit roughly the center of the optics and can find a solution to get the beam through the irises we set to align the beam heading  towards the periscope. It turns out that the optical beam path off the second flipper mirror towards the beamsplitter that recombines the central heating and the annular beam path, is on a downwards slope (because of the flippers/polarisers throwing us out of plane). It bites us for the annular path because its such a long lever arm that to match the optical path set by the central heating optics we would need such a steep angle of the beam that the mirrors we used to steer the annular path onto the beamsplitter cant provide this.

Sooooo whats the solution. Well first...never make assumptions again. But in all seriousness there are solutions and I throw out options without having really thought about it or having discussed it with the TCS folk like Aidan. Options include, using flipper mirrors (like those I believe that were used in eLIGO) that have vertical adjustment so that we can correct vertical alignment, putting in an extra mirror in a bow tie type arrangement to try to correct out the vertical kick up or down, change the layout (but keep the length the same) of the annular path so that at least one of the steering mirrors is closer to the second flipper and so we can try to match the downward trajectory of the beam path set by the second flipper for the central heating route. We could try to remount the polarisers better. Again this is all just being thrown out there for now.

Hopefully you arent all thoroughly confused by now, and I really am sorry we didnt catch this earlier. Kicking myself all night for such a rookie error and feeling wasted peoples time. If nothing else its given us fair warning to look out for this on TCSy and when we do the two livingston tables, so it actually could be a time saver by finding this problem now on the first table and coming up with a solution.