Collimator for a Short arc Lamp (380nm to 780 nm)

  • 2 August 2022
  • 3 replies


I am working on a collimator system for a commercial lamp. The lamp has following specs (from optical design point of view):

Wavelength range: 380nm to 780 nm

Working Distance (W) / Focal plane of the elliptical reflector: 60 mm 

d= 3 mm or 5 mm (Spot made by the lamp at 60mm)

Beam-to-axis angle (α): 30 deg.

Reflector diameter (D): 82.5mm


  1. I am not getting the collimated output for this system. Since spot made by the lamp could be of diameter 3mm or 5 mm what could be the possible solution when defining field data for this optical design? I have taken an object of 3mm for the starting design. Kindly comment.
  2. Object to first lens distance is taken as 60mm. Is it right approach or this distance could be changed. Source is very high power so to keep optics from damaging =, collimating optics is placed exactly working distance away. Pls comment. 
  3. What could be the optimization criteria for such collimating systems? I have taken angular radius as the criteria to optimize.
  4. Are there any more points to consider for this design (apart from above points)?
  5. Below are some of the images from the design:




Best answer by David 8 August 2022, 02:35

View original

3 replies

Userlevel 5
Badge +3

Hello Naresh,

The spot produced by the elliptical reflector of the lamp is essentially an extended source, and it is not possible to perfectly collimate an extended source. The reason for that, and the mathematics that define the limits on collimation, can be found in the concepts of etendue and, for imaging systems, the Lagrange invariant. 

These concepts are too complicated to explain in a post. But the Zemax website offers an excellent tutorial on illumination system design here.

I especially draw your attention to the section titled “Background theory of illumination design.” But it’s well worth the time to go through the entire tutorial. 

Dear David

Thanks for your comments.

After going through the knowledgebase articles, it it clear that a perfectly collimated beam may not be possible.

How do we calculate the collimation for Lamp data provided in the question.

What could be the possible design changes you would suggest??




Userlevel 5
Badge +3

Hi Naresh,

This application likely requires an aspheric condenser lens. To collect and collimate light with a 30 degree cone half angle requires a lens with NA = Sin(30) = 0.5. These are common parts for this application. The exact diameter and focal length will depend on the beam diameter you wish to produce. The shorter the focal length the smaller the diameter and the greater the divergence of the beam. That’s the etendue trade off.

There are other important considerations in working with arc lamps. These lamps produce a lot of IR and UV.

  1. It is important not to reflect IR back into the lamp because that can overheat the lamp.
  2. It is often important not to send IR down the beam because that can overheat the apparatus making use of the visible light.
  3. It is often important to strip the UV out of the beam, because extended exposure to the UV can deteriorate the apparatus.
  4. It is very important not to expose people to the UV. It can damage eyesight. Engineers working with prototypes often wear safety glasses that do not pass UV.

Many of these design issues are dealt with using hot mirrors, cold mirrors, and filters in the optical path to divert unwanted wavelengths to an internal beam dump.

The fundamental design of illumination paths like this is often done in sequential mode, with quantitative analysis performed in non-sequential mode