what is the light intensity delivery threshold for DMD

The biggest concern would be the flux density and the resulting heat generated.  If the power is focussed on a very small spot, the localized heat may be enough to damage the mirrors.  If the energy is spread out over a large area, damage is less likely to occur.  In Digital Cinema applications, the DMDs are routinely exposed to 10s of watts per cm^2 of visible light for many 1000s of operating hours.  Tests with lasers confirmed no damage at flux densities >100W/cm^2.  The actual damage threshold limit has not been fully explored at this time.  Can you provide more details about the spot size you want to use and the resulting flux density?  We can try to match that with some lab tests.

Thank you for the information.

We will expand laser beam to about 1cm^2, even though the femtosecond laser still make very strong peak flux to DMD (the peak flux we need is around the order of 100000W/cm^2). The laser pulse is about only 100 femtoseconds long, then after about 10 nanoseconds drak duration, there will be the next pulse.  

I think in the dark duration, heat will be disspated (because if averages the power with the dark duration, the average power density is only 1W/cm^2 which for sure is in the safe range), what I worried is if DMD can survive in 100 femtoseconds pulse duration with 100000W/cm^2 flux density?

ShengWang said:

Thank you for the information.

We will expand laser beam to about 1cm^2, even though the femtosecond laser still make very strong peak flux to DMD (the peak flux we need is around the order of 100000W/cm^2). The laser pulse is about only 100 femtoseconds long, then after about 10 nanoseconds drak duration, there will be the next pulse.  

I think in the dark duration, heat will be disspated (because if averages the power with the dark duration, the average power density is only 1W/cm^2 which for sure is in the safe range), what I worried is if DMD can survive in 100 femtoseconds pulse duration with 100000W/cm^2 flux density?

Hmmm, that's a tough question.  I'll have to look into that one.  I know there have been some tests with high power lasers but not at that level.  You are correct that the average energy is very small so the heat should be dissipated in the dark period.  It probably will not be a problem...but you may just have to try it and see what happens.  Let me look into available information and someone will reply to your post shortly.

I also calculate the average power of what you are describing to be at about 1W/cm^2, which if the beam were continuous at this power is  within the ability of the DMD to dissipate.  This agrees with your numbers.

The DMD should be able to handle the load; I estimate each pulse in the operation you describe to be around 10nJ/cm^2, which is not huge, but no one has actually tried this pulse rate.  

Some tests have been performed on DMDs that were not operating at the time (mirrors flat) with pulse energy densities that were significantly higher than this, without damage, but at a much lower repetition rate.  This suggests that your operation should be ok, however, there is no guarantee.

I would be very interested in your results with the DMD in operation.  How are you driving the DMD?

I'm considering a related problem (DMD illumination power threshold for lasers).

We want to apply a mask pattern to a line-shaped laser beam. We're working with a 400 mW CW 830 nm laser beam (for Raman spectroscopy). We will expand the collimated laser line to fill the long axis of the DLP device we use (we haven't purchased a kit yet). I'm considering the 0.55" Discovery DLP4000 kit. Obviously, we wouldn't want to kill the DLP device, however I haven't been able to locate tech specs on the DLP chips (quite surprising, I have never had difficulty finding specs on the MSP430).

I looked at the wavelength chart for the UV , vis, and NIR windows, but I couldn't tell whether the visible or NIR window is more efficient at 830 nm.

Also, I'm wondering approximately how much light we can expect to lose to diffraction? I've used the holoeye spatial light modulators in the past, and I'm quite aware of the 'disco-ball' diffraction patterns that result (pretty, but not very useful). We intend to collimate light from one line shaped multimode fiber optic bundle (100 micrometer core diameter fibers) to reflect off the DLP, and focus onto a second line shaped multimode fiber optic bundle (same as first).

Thanks,

Francis

Hello Francis,

Welcome to the DLP section of the TI E2E Community.  We are pleased to have you here.

The power density of the system you have described is not unreasonable.  Specs on the Power Handling of a DLP(R) Chip for Laser applications are currently being developed. 

You are correct that at about 830 nm the windows efficiency for the Visible and IR coatings are about the same. 

As far as diffraction efficiency, this will depend on optimizing the order of interest in the "disco-ball" diffraction pattern.  Please refer to "Using Lasers with DLP(R) Technology" at http://focus.ti.com/dlpdmd/docs/dlpdesigncenterdetail.tsp?sectionId=61&tabId=2273.

Again,  welcome.

Hello TI-Team, 

I would like to use DLP in high power laser apps. pulsed and CW.
Is there any update on the laser damage threshold tests you are running?
Or can you provide first results?  

Specially for the Nd:YAG wavelengths(1064, 532 and 355nm) in ns-pulse range

Thanks!

Andreas,

I am sorry that your questions seems to have slipped by. 

At this time there are no updates to give. 

Francis,

How good to hear from you again. 

The reflectivity is a little bit more complicated than just the fill factor.   Moreover, there are also diffraction losses that are not driven by fill factor alone. 

If you are able to achieve 80% that would be outstanding, but I think that the diffraction losses in addition to what you have already accounted for will push it down a little lower.  Assuming that you are intending to only collect the blazed order.  

If you are going to collect multiple orders then you will be able to recover light from some of the other nearby orders as well and this will allow better throughput. 

I hope this helps.

P.S.  I do not know what happened, but it submitted multiple entries when I hit "Post" 

All,

Sorry something glitched when I hit post and it made a post for each word in the last line of my entry.  Please ignore the multiple posts!!

Hi Fizix,

I'm only collecting the primary order. Can you give me a ball-park estimation of what the maximum efficiency should be?

Thanks again,

Francis

Francis,

I see my post yesterday did not take.  I don't think this thread likes me.

As a very rough ball-park I would expect somewhere around 70-75% collecting a single order.

Thanks! That's very helpful.

Francis