Xiao's 2011May run

May 8th

1. find the best focus for luminous 5 and 6 We mount a visible fiber to one stage and shoot visible light in from the other end. Moving z axis manually to make sure the spot bounce off the parabola mirror maintain similar size to a long distance, which means we find rough focus position. Then we search XY position by letting the spot go through an alignment telescope, and look at the spot size and shape on wall. Then we adjust the alignment telescope to infinity by placing a mirror in front of it, and find the two clear grid image. Now we try to adjust z axis of luminous stage manually again to find a finer focus by looking at the spot through the telescope (BE care to reduce light first before looking in to the telescope)

2. re-find focus position for beam 1 AND 2 AND 3 Since we replaced all the actuators on luminous stages with new ones, we have to find the focus positions for all the stages. For beam 1, 2 and 3, because they still use the same infrared fibers, the fiber should stay at the best focus positions as before, which means you don't need to adjust the pickoff mirror. We shoot laser into the MIRC vgroove and align it with CHARA beam on the alignment telescope. First place a beamsplitter in between MIRC light and CHARA beam, and a retro near the beamsplitter:

                              mirror  ==>   alignment telescope
                                  /\
                                   ||
              MIRC light = = >   beamsplitter         <= = CHARA beam
                               ||    /\
                                \/   ||
                                retro

Second, place the CHARA beam in the center of telescope by adjusting beamsplitter, retro, and mirror. Third, overlap the CHARA beam with MIRC light by moving XYZ axis through FE gui.

Be care when looking into the telscope, reduce the laser or waring a goggles.

May 9th

1. change the fiber on luminous stage 4 shoot laser into MIRC vgroove. Moving XYZ axises to overlap with CHARA beam using the method above. then change to new infrared fiber, and align again. Hopefully the best positions are close

2. mount luminous stages 5 and 6 shoot laser into MIRC vgroove, and move the pickoff mirror to overlap with CHARA beam. Then change to the MIRC infrared fiber, and do this again by moving XYZ of luminous stages.

By now, we have six infrared fiber plugged in and well focused, and roughly aligned with CHARA beam.

May 10th

1. assemble new MIRC hardware pieces Only assemble the MIRC parts: MIRC lens array and beamsplitter and spherical mirror. With all the above steps now MIRC luminous stages are well aligned with CHARA beam, but it drifts. So ones needs to redo the aligned a little using the method above to get most light through.

With CHARA laser on, one should see light come out of mirc vgroove. By directly looking at the bright fiber tips and the MIRC lens array, one can roughly match them. Then check the spot size for focus, and rotate the lens array to make sure the spots are vertical. Then move the whole stage so that the light bounce off the spherical mirror goes through black dot on the target, doublet and camera.

We did this on beam 3 and 4, which are the furthest two. We see beam 4 light on camera with the whole detector view, and we don't see beam 3. The new RTschedule doesn't work with the whole detect mode, need to be fixed. We run the old MIRC system and astropci, but FE has some problem with the whole mode, need to be fixed.

May 11th

Tried to figure out the new MIRC system problem: segmentation error when configuring the whole detector. But failed. I have a feeling that it has something to do with the data type, because the index could exceed the limit and become some unexpected number. Such that there is an memory leak.

1. figured out the FE problem. The code is right, it is just the row/col offsets apply to the four quadrants so that if we are using 252 X 252, the maximum offset will be 2.

Meanwhile I spent some time on PAVO alignment and observing for the whole night.

May 12th

Find fiber position for beam 4, 5 and 6. I used old MIRC system to monitor the whole detector, using old fiberexplorer method. Their spot sizes are similar, but not over lapped on detector.

I also tried to find beam 1 fiber position using the same method, but no success. I will try again tomorrow.

May 13th

find fiber positions for beam 1,2,3. I used old mirc system and old fiberexplorer method. The camera configuration is rows = 252 cols= 252 row off=2 col off =2 Nreads=5 number of frames=100 frames per reset=100 cohere coadds=10 number ps coadds=15

I also find a way to shoot laser backwards without taking off anything. I find a mirrow in cabinet 8 from MIRC cabinet, and place it at the position where the prism is, and place the real laser near it.

May 14th

polarization alignment!!! 1. I checked the polarization alignment on beam 1,2,3, they all look well aligned. I put one polarizer in front of the pickoff mirror, the other in front of the camera. I measured the intensity ratio with and without the polarizer in front of the camera. I measured at three position: the two parallel polarizer, the one in front of the camera rotate right, and left. Apparently, the ratio when two polarizer are parallel is maximum.

2. Then I align beam 4,5,6 to beam 1,2,3. The two polarizer are placed at the same position as above, but this time they are parallel. I rotate the fibers to maximize the intensity ratio. I measured at least six points for each of them. The data is in my notebook and I haven't analyzed yet. But I can already see where the peak is.

The process is very slow because we are reading the whole detector, and noise is high. Later we install the old MIRC cylindrical lens, and we only need to read a small range of pixels. The process is much faster.

I also find the problem causing the new MIRC system to crash at full detector mode. It is some initialization of xtransform and ytransform. Not sure how to fix it yet. As I comment them out, the new MIRC system runs smoothly. In fact I used the new MIRC system to do polarization alignment.

One important thing is that background has to be taken each time I took and or put in the polarizer.

I also worked a little on 6T fiberexplorer, I made a little change, and i seems to run well.

May15th

I have analyzed the data taken yesterday, and here the data points and best fits. The angles of beam 4,5,6 are beam4: 195 degree; beam5: 250 degree; beam6: 110 degree;

Media:Beam4.png

Media:Beam5.png

Media:Beam6.png

note that in beam5, I ignored the two left points. I dont' know what happened, maybe I forgot to re-take background after putting in or taking out the polarizer. And these angles are rough and could have 5 degrees error. But this is OK because 5 degree means 0.4% less of coherent light.

Anyway, I confirm these best estimation of rotational angle of fibers by two methods. One is using two orthogonal polarizers, one in front of pickoff mirror, the other in front of camera. The light gets into the detect is very low. I can not really estimate the percentage because of the background variation. I also rotate the fiber by a small angle (~5 degrees) to see if the light will go away, but it doesn't, and seems to maintain the same light intensity level. Like I said, at this point background variation dominates, so it should be good enough.

Another method to using two parallel polarizers, and putting them in the same position as above. By rotating the polarizer in front of camera a small angle to see if the light intensity increases. Again, it doesn't really help because of background variation.

I just have an idea of doing polarization alignment of fibers if I hadn't aligned them yet. Instead of rotating the fibers each time and using fiber explorer to find the XY position, I can rotate the polarizer in front of camera. In this way, I don't have to do fiber explorer each time, which is time consuming. Once I find the angle that the maximize the light through put, then I know how much the fiber has to rotate.

May 16th and 17th

We got MIRC aligned, and got fringes on W1(beam3) - W2(beam4)!!!

1. Align MIRC. Beam 3 and 4 are used because they are the farthest apart. First, we need to get CHARA laser into MIRC fibers. Then we can visually check the laser coming out the MIRC lens array. There is a faint circle around the center bright spot, make sure the circle is even. We also check if the laser beams are horizontal by taking off the spherical mirror and looking at whether the laser beams do up or down as they propagate. To check the focus of lens array, we put a paper at the slit where the focus of the spherical mirror should be, then by adjusting the distance between MIRC vgroove and lens array to make the image on paper at sharp as possible. Once this is done, we steer the spherical mirror to reflect beams right to the center of the doublet.

If all the above procedure are well done, we should be able to see white light on detector once we switch it on. then by turning the actuators on spherical mirror, we can get the beams on detector to lower left quadrant. If the beam 3 and 4 are not overlapped on detector, then there are two cases. If they are not overlapped up and down, we can rotate the goniometer of mirc lens array. If they are not overlapped left and right, that might be due to the minor fault in the positions of fibers in vgroove, we can doing nothing about that.

Another thing is that we find beam3 and beam4 are not focused at the same time, meaning that the mirc lens array is probably tilted relative to the vgroove. so we have to tilt vgroove (we don't have that adjustment for mirc lens array) so that the surface of len array and vgroove are parallel. So now we have all the beams well overlapped.

Media:beam1_aligned.png

Media:beam2_aligned.png

Media:beam3_aligned.png

Media:beam4_aligned.png

Media:beam5_aligned.png

Media:beam6_aligned.png

2. fringes on W1W2 with lab source. Once we got all beams aligned, we search fringes by putting a retro on beam3 and 4. The position of pickoff mirrors are

beam3: 53700

beam4: 96915

And the retro position is 4.5103 mm. Here is the fringes image.

Media:fringes_beam34.png

We find a small problem in fringes that the fringes is not balance on detector, meaning the fringes is tilted relative to detector rows. So we have to rotate the mirc lens array and vgroove with the same angle to match the detector rows. And here are the images of fringe cross section after alignment.

Media:frignes_cross_section1.png

Media:frignes_cross_section2.png

Media:frignes_cross_section3.png

May 18th

We find MIRC fringes on 4 more pairs of beams: b4b5, b5b6, b1b2, b2b3.

Here are the ealing position and retro position:

b4b5: b4: 96915; b5: 30500; retro: 5.4mm

b5b6: b5: 30500; b6: 65300; retro: 10.38mm

b1b2: b1: 90350; b2: 78825; retro: 2.9mm

b2b3: b2: 90325; b3: 53700; retro: 6.8mm

the fringe quality are

b3b4: ~50% b4b5: ~25% b5b6, b1b2, b2b3 are all above 90%.

note: two problems. 1. the intensity of b5 is about half of the rest, don't know why. 2. using 450 grism mode, we couldnt bring the beam down to the lower left quadrant because the actuator of the doublet hits the limit. needs to move the whole mirc stage and realign tomorrow.

[[Media:beam12.png]

[[Media:beam23.png]

[[Media:beam56.png]