Rubin Calib 2023

from Tucson visit April 21 2023:

Stubbs tasks: 

• think about imaging output of flat projector at f/4. 
• think about beam projector at f/4 with top hat intensity distribution- pair of OAPs? 
• masks with 2 inch diameter
• RASA imager to catch full field, full aperture
• design multi-cell CBP monitor
• Zemax engineering
• CNS mask fab
• talk with Parker at some regularity
• think about our staffing level
• work on Parker's commissioning plan
• think more about mask aperture sizes
• field distortion mapping for RASA and for CBP
• CBP for AuxTel
• Energetiq source calculations
• SOLIS source for f/4 beam with Axicon
• dual aperture baffle for screen measuring
• full aperture solar cell monitor. CBP aperture is 250mm. Solar cells are 120 x 120mm so a 2z2 array does what we need. just stopping down pupil slightly would work. Ring of 5 would certainly work. 
• Ask about CfA temp enger and technician avail, and willingness to move to Tucson. What about at Mt. Hopkins? 
• Uniform fiber brightness source 
• off-campus overhead rate and proposal for Elana
• rate sheet for engineering
• paper on donuts and screen uniformity spec
• Add shutter to projector design!

Screen monitor:
April 23, 2023 CWS

Hamamatsu S2281 has sensitive area diameter of 11.3mm. https://www.hamamatsu.com/us/en/product/optical-sensors/photodiodes/si-photodiodes/S2281-01.html 

Fits inside SM1 lens tube: 

If we stop this down with a 10mm aperture, then for this to subtend 3.5 degrees, which is the range of angles off the screen we are interested in, a lens needs a focal length of sin(3.5) = 10mm/ FL, and so FL=10mm/0.061 = 164mm. 
We basically want the fastest lens we can get with this focal length, and the region on the screen being sampled by the diode has a diameter of D_lens+L*sin(3.5). For L=5 meters this is D_lens+305mm. For a smaller sampling area, get closer. 

we can also stop down the lens. Angular acceptance is determined only by the focal length not the aperture. So for finer spatial resolution (at the expense of flux) then reduce the lens aperture. Aluminum foil over the front will work for that. 

Here's a candidate Fresnel lens from Amazon. Since all we want here is a relative measurement, this should be fine:

For a 140mm focal length we want an aperture covering the diode of diameter 140mm*sin(3.5) = 8.54mm. 
140mm focal length is 5.51 inches

Now we need some kind of housing with 150mm ID. Carbon fiber tube comes in metric sizes. Here is an option: 

So we'd just glue the Fresnel lens into this tube at one end. base plate on the other end to hold diode using SM1 threads 

support lens on Lab Jack for gluing. 

Focal length of Fresnel lens is 140mm or just over 5.5 inches. Let's assume we want a 1.5 inch back focal length off the back side of the mounting plate. That means we end-to-flange distance of 4 inches.

Full Aperture CBP monitor: 

We just need to cram the CBP beam onto a single solar cell. This is going to be better, well at least faster, than trying to make multi-solar-cell monitor. 

Bought 300mm diameter and 700mm focal length lens on eBay from China. They say it's K9 glass, allegedly equivalent of BK7. Confirmed with the vendor that there is no AR coating, and it is indeed K9 glass. 

Material for holding this lens - 

This is better and flatter than any reflective mirror we might use. 

can line the inside with Acktar lambertian film. inner circumference is 12.75*25.4*pi mm = 1.017 m. Almost exactly one meter. 

Lens focal length is 700mm. Solar cell is 120mm on a side. Converging beam diameter is 300mm*(L/700) where L is location of solar cell compared to focus. If we set beam size to 50mm then L=50*700/300 = 116 mm. 

Might as well set the solar cell at the focus, since this didn't gain us much. We might  get away with using a Hamamatsu diode there, depending on amount of spherical aberration. 

We could use Omegon OTA for a test collimator.  

Projector optics

From Parker on Slack, April 26, 2023:
-------------------------------------------------

Is this for the white light/LED or the laser projector?
The LED design is still being optimized but we essentially have it complete.
The distance from the reflector to the source is 3200mm.
The current design has the f/# at 3.7
-------------------------------------------------

So at that distance, reflector spans a distance of 3.2m/3.7 = 0.865m in diameter. 

Rubin obscuration is 3.42m diameter. Primary is 8.4m in diameter. So in radius the obscuration is 3.42/8.4 = 0.40. That means the beam fan illuminates from r=0.40*0.865 = 0.346 to 0.865m.
That amounts to beam half-angles of atand(0.346/3.2)= 6.17 degrees to atand(0.865/3.2) = 15.13 deg. The midway angle is the average of these, or 10.65 deg. So we want a beam that spans 10.65 +- 5 degrees. 
Crude estimate is an axicon that introduces 10.5 degrees of deflection into a beam that has a divergence of 10 degrees. A beam divergence of 10 degrees (full width) is the same thing as f/5.67.
If we're doing this with a 1 inch beam, that's a focal length of 5.67 inches or 144mm. 

Another option for white light source is a halogen bulb at the focus of an elliptical reflector. See https://www.edmundoptics.com/p/113mm-dia-x-272mm-fl-alum-coated-ellipsoidal-reflector/23410/ 

There is an f/4 beam version of this, about 2 inches in diameter with 1 inch hole. 

We did a quick prototype of this axicon illuminator using 8mm focal length reflective collimator. Annulus was highly non-uniform, way fainter at the edges. That's a cos^3 dilution effect. 
Here's a plot of illumination at the edge vs. center, as a function of f-number, using cos^3:

So if we want 10% uniformity we need to collimate with a beam slower than f/8. 

For cos^4 it's this:

so f/8 is OK at the 10% level. The tradeoff is efficiency, which has light collection going as 1/f^2. 

f-number goes from 2 at the top left to 10 at bottom right. 

I think our system has cos^4 behavior: two powers of cosine for 1/r^2, one for projection effect at source, and one because illuminated area on screen for a fixed solid angle has a projection effect as well. 
Unless they have corrected for this, the variation of normal to reflector has strong dilution at edge of reflector, as well. 

Here is our initial lab test:

Ignoring blue edge that we can clip, outer surface brightness is around 100. Max is around 180. Ratio is 0.55. Collimator had 8mm beam dia and focal length =33mm. So its f-numer is 33/8 ~ f/4. Expect ratio of 0.7. So not that far off. 

April 30 2023: 

Thor labs apodizing filter has this transmission profile: 

Assuming off-axis distance converts to angle, we have the product of two angular dilution factors. Assume we'll use an axicon to divert the beam. Beam to reflector has a half-angle of (15.13-6.17) = 8.9 so call it 9 degrees. Surface brightness at edge of disk normal to optical axis is down by cos(9)^4 = 0.95. For the apodizing filter the center-to-edge transmission ratio is 0.83/0.1 = 8.3. What f-number of collimator does that correspond to? we want (1/8.3)=cos(theta)^4 so theta~54 deg. Full opening angle is 108 deg so that's a super-fast collimator!

Lab tests with top hat beam:

Top hat put near fiber output, axion put 6 inches away (not optimized at all).

Seems promising if we optimize the spacings...