Laser safety reminder
All lasers are considered eye-safe (both visible and telecom)
Bending modes discussion
PSF does a good job correction X translation
TMS seems to be overshooting in Y
On one hand, moving mirror around on actuators
On the other hand, moving the mirror around on a steel structure, where changes are not picked up by PSF
TMS should help correct some of these unsensed structural perturbations
W/ Y Translation, there is a similar behavior in Rx (Rx about 5X more motion than Ry)
Collimation model variation w/ elevation is seen in Y offset + Rx offset, which is corrected with software
Need collimation model calculation with change in Y + change in Rx which should be subtracted out
PSF already compensates the LUT for the collimation model, but the TMS does NOT incorporate the LUT, yet
Andrew suggests using the collimation model with TMS, as well as PSF
Blue line is PSF + gravity flexure
TMS sees the change that the collimation model was built to correct, but doesn't correct it
John: "PSF data is the collimation model."
As the hub rotates, we would immediately run out of range in Y as elevation changes
Instead, we use a LUT in X and Y to rotate about a coma neutral point, providing for best collimation at all elevation
Over range of elevation, the collimation model is correcting something like 5 mm of Y change
Yang's code is not currently applying the LUT Collimation Model to the TMS data readout
Code has some comments about applying the correction
"could eliminate correction by moving correction from collimation model to the telescope pointing module"
We need to accommodate the commanded change from the LUT correction
The correction algorithm DOES appear to be applying the LUT correction after the measurement
The big up and downs at 4:40 and 5:05 are where the jumps happen (100 um spikes, plus and minus in Y)
As we apply SA correction, we change the goal posts, and we aren't fully accommodating the change
John reminds us that the SA correction should make a step, not a spike, as observed
Between 4:40 and 5:00, during the three superfoc series (see ladder in Z translation), is where the Y spikes are observed
No manual inputs were applied on PSF, though these could be at different positions on LBC, but this is irrelevant
Seems like a linear increment per superfoc, but it's too big to be thermal
This might be caused by moving the mirror on the hardpoints while running TMS
As w/ Rx, Ry, we have a Z change of elevation with the LUT
We are seeing a difference between the LUT and the uncompensated, raw measurement
TMS code only has to take care of pointing problem from running out range for correction
From last week, Heejoo showed the different channels changing size at different rates
This week's data might explain some of the variable changes in channel length
Andrew suggests that this is mostly attributable to not applying the LUT offsets to the TMS data
The hub is not only changing focus position, but is also rotating
This causes the different legs to change by different amounts
Good exercise to apply the LUT offsets to TMS data to see the effect
Also, may check the health of the LUT collimation model
Yang has been working non-stop for Al's OSCO project, and has not yet had time to finish the TMS software update
Has time tomorrow, 10/16 to work on TMS
Based on Andrew's testing, the algorithm we have come up with seems to work well in Mathematica
Yang set to implement the algorithm and set up a test to see if it works well
Andrew working on SPIE paper for TMS to present in December at Astronomical Telescopes and Instrumentation
"A laser-truss-based optical alignment system on LBT"
Goal is to develop a paper where someone starting from scratch could have a good start
Recall, the system works by contrast from usable return from the fiber tip and the return from the retro
eTalon is developing a new gen. with a more powerful laser and Maksutov-type retro allowing for 10 mm beam walk
New Maksutov-type setup has low return, so won't work unless there is a more powerful laser
Andrew suggests talking with Mark at eTalon about this issue to talk technical details
Laser safety reminder All lasers are considered eye-safe (both visible and telecom)
Bending modes discussion PSF does a good job correction X translation TMS seems to be overshooting in Y
On one hand, moving mirror around on actuators On the other hand, moving the mirror around on a steel structure, where changes are not picked up by PSF TMS should help correct some of these unsensed structural perturbations W/ Y Translation, there is a similar behavior in Rx (Rx about 5X more motion than Ry) Collimation model variation w/ elevation is seen in Y offset + Rx offset, which is corrected with software Need collimation model calculation with change in Y + change in Rx which should be subtracted out
PSF already compensates the LUT for the collimation model, but the TMS does NOT incorporate the LUT, yet Andrew suggests using the collimation model with TMS, as well as PSF Blue line is PSF + gravity flexure
TMS sees the change that the collimation model was built to correct, but doesn't correct it John: "PSF data is the collimation model."
As the hub rotates, we would immediately run out of range in Y as elevation changes Instead, we use a LUT in X and Y to rotate about a coma neutral point, providing for best collimation at all elevation
Over range of elevation, the collimation model is correcting something like 5 mm of Y change
SEE JOHN'S EMAIL FOR COLLIMATION MODEL SPECIFICS DXPMLBCCollimation.zip
Yang's code is not currently applying the LUT Collimation Model to the TMS data readout Code has some comments about applying the correction "could eliminate correction by moving correction from collimation model to the telescope pointing module"
We need to accommodate the commanded change from the LUT correction
The correction algorithm DOES appear to be applying the LUT correction after the measurement
The big up and downs at 4:40 and 5:05 are where the jumps happen (100 um spikes, plus and minus in Y) As we apply SA correction, we change the goal posts, and we aren't fully accommodating the change
John reminds us that the SA correction should make a step, not a spike, as observed Between 4:40 and 5:00, during the three superfoc series (see ladder in Z translation), is where the Y spikes are observed No manual inputs were applied on PSF, though these could be at different positions on LBC, but this is irrelevant
Seems like a linear increment per superfoc, but it's too big to be thermal This might be caused by moving the mirror on the hardpoints while running TMS
As w/ Rx, Ry, we have a Z change of elevation with the LUT We are seeing a difference between the LUT and the uncompensated, raw measurement TMS code only has to take care of pointing problem from running out range for correction
From last week, Heejoo showed the different channels changing size at different rates This week's data might explain some of the variable changes in channel length
Andrew suggests that this is mostly attributable to not applying the LUT offsets to the TMS data
The hub is not only changing focus position, but is also rotating This causes the different legs to change by different amounts
Good exercise to apply the LUT offsets to TMS data to see the effect Also, may check the health of the LUT collimation model
Yang has been working non-stop for Al's OSCO project, and has not yet had time to finish the TMS software update Has time tomorrow, 10/16 to work on TMS
Based on Andrew's testing, the algorithm we have come up with seems to work well in Mathematica Yang set to implement the algorithm and set up a test to see if it works well
Andrew working on SPIE paper for TMS to present in December at Astronomical Telescopes and Instrumentation "A laser-truss-based optical alignment system on LBT"
Goal is to develop a paper where someone starting from scratch could have a good start
Recall, the system works by contrast from usable return from the fiber tip and the return from the retro eTalon is developing a new gen. with a more powerful laser and Maksutov-type retro allowing for 10 mm beam walk New Maksutov-type setup has low return, so won't work unless there is a more powerful laser
Andrew suggests talking with Mark at eTalon about this issue to talk technical details