Using beamsplitters for co-axial light paths

Question

Good afternoon all, I have been racking my 'optics novice' brain on how to use the same optical path for both the transmit and receive optics for a simple laser range finder. For simplicity in the diagram below I have removed the collimation asphere, and beam expanders that I use to get the divergence under control.

To achieve that shown in the diagram I thought I could take advantage of the fact that the laser diode is predominantly polarised in the plane parallel to the laser diode emitting surface. Pass it through a polarised cube beam splitter, through a quarter wave retarder to circularly polarise the light. Then when it is reflected it would pass through the retarder again, after which it would be 90deg to the original and therefore would hit the polarised cube beam splitter and be passed out at in the manner shown below.

But... It dawned on me that it was very likely that the polarisation would not be maintained upon reflection from real world surfaces, and at varying angles of incidence. After a bit more reading I found that the polarisation after reflection would be random. This kind of ruined my idea.

So my question is, how can I achieve the scenario above, can it be done with a basic plate beam splitter. Ideally, I would like as much of the transmit light to hit the reflection surface, but if a small amount was deflected I could actually use that light for a timing trigger. I would also like as much of the reflected light to hit the APD but again the power level of this light would be so small that I don't think it would hurt the laser diode too much if a fraction went back to the laser diode.

Can anyone give me a pointer to where I should direct further reading and experimentation?

thank you

Answer 1

Your polarized laser, polarizing beamsplitter, and retarder plate approach is probably the best that can be done in a simple system. True, light reflected from rough objects gets de-polarized, but not completely.

Answer 2

The "ideal" design here would be to have your laser diode followed by an optical isolator to protect it from any back propagating light.

Then you can either use a polarizing beamsplitting and polarizing optics as you've described to try to get maximum light going to your APD assuming your object-under-test has the right polarization behavior. However, there's a chance your object-under-test rotates the polarization in a way so as to minimize the light going to the APD so you have to be careful.

The alternative is to just use a 50:50 beamsplitter, then you know you'll always receive 50% of the returning power independent of polarization. You of course need to design your diode, TX/RX optics, APD, etc. to ensure you have sufficient power SNR overhead to sense something.

There will always be design tradeoffs in this sort of optical system. The 50% power loss from beamsplitter will result in like 3 dB loss of SNR. Is 3 dB limitting for you? Likely the range finder will still work at some range. What range do you need? Is it physically possible with the power you have? Improving the specs will cost $$$. Maybe better to just buy a more powerful diode. These are the sorts of engineering trade offs one has to make.

edit: Actually I take back what I said about the object changing the light in a worst case scenario way. If your object are things like rocks and trees then the likely behavior is depolarization as has been pointed out already. But if your objects are reflective, or go through things lke car windows you might run into more subtle polarization issues.