Forums

Hi Thomas,

Thank you for submitting this post on the E2E site today.  I am working on finding you some answers and helpful links.  In the meantime, can you tell me what resolution DMD size you're working with?  This information will help me narrow down some specific answers for you. 

I look forward to hearing back from you.

Justin Lemon

Hello Justin Lemon,

Thanks for your quick feedback, but since it is not used for the projector, there is no resolution spec.  Sorry for it.

Sincerely yours

Thomas

Hello Thomas,

First, welcome to the DLP section of the TI-E2E Community. 

I think I can help understand the numbers you are looking at. 

The DMD acts as a diffraction grating when interacting with light.  The diffraction causes light to scatter into orders based on wavelength.  For a given collection cone of light (typically a cone 1/2 angle of 12°) some light will be scattered due to diffraction outside of the collection cone.  This is the diffraction efficiency.  Other light is lost between the mirrors (i.e. fill factor) and reflectivity of the mirrors themselves.

The diffraction efficiency quoted in the "Features" secion on page one are for teh conditions listed in Note #12 on page 21 of the datasheet.

NOTE:  The datasheet specifies the maximum illumination for wavelengths less than 420 nm to be 2 mW/cm².  These chips are not intended for use in UVC .

To answer the questions in a general way:

1. DLP3000 datasheet says array diffractive efficiency has 86%.  What's the defination of this efficiency value?  I want to know which caculcation for the total efficiency is correct:

 Total efficiency = (single passing windows efficiency)^4*(array diffractive efficiency)  This is not correct.

or Total efficiency = (single passing windows efficiency)^4*(mirror reflectivity) *(array fill factor) *(array diffractive efficiency)  This is almost correct.

The window efficiency only needs to be squared not raised to the 4th.  I think the confusion comes from note #12 on page 21 of the datasheet (item d under the third group) which says, "(d) Window transmission: nominally 97% (single pass, through two surface transitions)"  The parenthetical at the end is discribing a single pass which consists of two surface transitions (i.e. entering and exiting).  The 97% is for both transitions combined.  Therefore:

Total Efficiency = (single pass window efficiency)² * (mirror reflectivity) * (array fill factor) * (array diffraction efficiency)

2. How many micro-mirrors are needed to get the maximum 86% diffractive efficiency shown in the datasheet?  I want to know the mimimum beam size required at the DLP chip surface.  

About 10 mirrors diameter will do fairly well to produce the diffraction effects.  However, this may or may not be the maximum efficiency, only that it will have converged to that value.  For smaller areas you would have to model it.

Fizix

Thomas,

I wanted to add, that the discussion in my previous answer is just the DMD contribution. and the diffraction efficiency depends on your illumination and projection conditions.  All other losses in the system, geometric - (i.e. overfill) and transmissive/reflective from other optical elements will add to the total system losses and affect the total system efficiency.

Fizix.

Hello Fizix,

Thanks a lot for your answers.  That's much helpful for my design.
One more question:
One early post says that the pitch of DMD should be 6 times larger than the wavelength.
But for my application, the pitch to wavelength is only 4.9.  Do you know how much diffraction efficiency penalty it will induce?
Thanks a lot.

Thomas

Hello Thomas,

You realize must have meant light in the region of 1.53 um to 1.56 um (rather than 1.53 nm to 1.56 nm) since in nm this would be in the .  You are working in the IR.  In this region the datasheet specifies < 10 mW/cm² for wavlengths longer than 700 nm.

In the range you are referring to it sounds likely that you are using telecom wavelengths (such as 1550 nm).  In this region, the diffractiion penalty will depend on the incident angle, the collection angle and the specific wavelength.

Fizix

Hello Fizix,

Yes.  I refer to the telecom wavelength 1.55 um, not 1.55 nm.  Sorry for the mistake.
So based on your feedback, for 1.55um wavelength, the diffraction efficiency of DLP3000 can also reach 86% under the correct incident angle and collection angle with 10 or more mirrors diameter beam size.  Is this statement correct?

Forgive me for one more question about 10 mirrors diameter beam size.  It must need 10 mirrors in each direction or it is ok if the total beam size is 100.  For example, if my beam size takes 4*25 mirrors, could it have the similar diffraction efficiency with 10*10 mirrors if other conditions are perfectly matched?

Sincerely yours
Thomas