Laser Projector for Speckle

[thomastt]

Dear guys, I am working on a research project for defect detection. The idea is to project laser speckle on a plastic plate, and analyze the temperature map on the plate to localize the defect. Right now I have two questions to set up the system. Any suggestion is appreciated.

1. Do you have laser projector recommended, which could generate uniform dot patterns like follows?

2. For defect analysis, each laser dot need to heat up that local area 1 celsius (best within 0.5 sec). The plate is made of plastic, 1 meter by 1 meter. How much power do you think is required for the laser source?

Thank you very much!

Thomas (from California)

 

[diachi]

How much power do you think is required for the laser source?

Too much

 

[LaserMicroscope]

Yeah, the short answer is probably “too much”. The long answer is that if you want uniform heating, you’re in trouble. While the pattern you’re trying to achieve could basically be done with a 10 cent diffraction grating, the way diffraction gratings work mean that the beamlets do not have uniform power, and the process is lossy - you need to dump HUGE amounts of energy in for the beamlets to have substantial power. In terms of your project goal, I’m not sure lasers are the right answer. Even IF it were possible, a heatmap from such a treatment would look like little hot bullseyes surrounded by cooler plastic and then very quickly an area of background temperature between them. I’m not a defect detection expert, but wouldn’t you be better served by uniform hearing of the material? (By the way, the scattering caused by diffraction gratings would likely slightly warm the material evenly if you put enough power through one to achieve 1c/0.5s). If point heating is critical though, and you feel you MUST use lasers, and you’re willing to do a little more engineering, the best I could recommend is either setting up X/Y sliders to move a laser from point to point to test them one at a time, or (and this is pure theory, I’m not even bothering to do the math to figure out what kind of power this would require) perhaps using scanners to move a pulsed beam rapidly from one spot to the next.

 

[Cyparagon]

Why are you looking for defects in plastic plates? Where I'm from, plastic plates aren't know for quality. Are you a manufacturer? Trying to throw lasers at a problem suggests to me you're not. So why do you care?

How are you going to "analyze the temperature map"? How does this translate to a defect?

Why are the plates square? Bit unusual.

The dot pattern you've linked is anything but "uniform". What do you mean, and why have you selected that pattern?

How have you ruled out all other ways to heat things, and jumped to the most expensive and dangerous one? Because it sounds cool?

 

[thomastt]

Thank you guys for your prompt reply. For defect detection, previous method is to use laser line to heat up the object, and measure temperature by cooled IR camera. To scan the whole object, either move the object on a x-y platform or move the laser line by rotating lens. Following is a illustration. SYNRAD 10W CO2 laser was used to heat the object (line) 1 celsius up in 0.025s.



The problem of previous method is requiring moving mechanism. We would like to improve by 2D speckle projection instead of 1D line projection. Note that the it is no need to heat up the whole surface. Only heat up that speckle dots 1 celsius then the data is enough for analysis by multi-Gaussian calculation.

Hope the above can clarify the research background.

 

[Cyparagon]

Why can't this be done with a $5 halogen flood?

 

[dylpicklechip]

Why can't this be done with a $5 halogen flood?

Agreed, perhaps a tungsten source similar to an ellipsoidal stage light with a very even field, focused at your specific plane + some sort of thermal imaging to compare to a known non-defected measurement?

 

[LaserMicroscope]

Thank you guys for your prompt reply. For defect detection, previous method is to use laser line to heat up the object, and measure temperature by cooled IR camera. To scan the whole object, either move the object on a x-y platform or move the laser line by rotating lens. Following is a illustration. SYNRAD 10W CO2 laser was used to heat the object (line) 1 celsius up in 0.025s.

View attachment 68544

The problem of previous method is requiring moving mechanism. We would like to improve by 2D speckle projection instead of 1D line projection. Note that the it is no need to heat up the whole surface. Only heat up that speckle dots 1 celsius then the data is enough for analysis by multi-Gaussian calculation.

Hope the above can clarify the research background.

It does help clarify, I had a feeling based on the pattern you showed you'd be using some kind of Gaussian calculation based analysis. Unfortunately, the issue still stands. As far as I'm aware, without highly specialized optics (that would take more investment of money and time than would be worthwhile), it's going to be a huge challenge to split a beam that many times and achieve uniform heating of the spots. Can I imagine theoretical ways to correct for this? Maybe, but just developing a method to do this would be a publishable research accomplishment and would take more time than I think makes sense for a technique you're trying to actually apply to analysis.

Based on the schematic you posted, the only improvement I could suggest would be a reiteration of something I mentioned in my first reply, using scanners rather than moving the object. Again, not an expert, but my approach would be to program a scanner to move a beam in a raster pattern across the work-piece, and set up a pulsed laser source. If the scanner moves the laser point by a known distance per second and you know the pulse rate of the beam, you can achieve a grid of point illuminations without having to move the object.