The power we are getting at the dark port is of the order Watts, not microW. If we change any parameters like the radius of curvatures or the lengths of the cavities, this affects the stability of the cavities as well as mode matching and that is why it’s not desirable to tweak these paramters . What should be done to get the power down in this case?
History
| # | Date | User | Information |
|---|---|---|---|
| 395 | 5 years ago | (current) | |
| 394 | 5 years ago | (original) |
3 comments on “Regarding dark port power”
Leave a Reply
You must be logged in to post a comment.
As you have realised, that’s not supposed to be the case! The power at the output port is influenced by many things, but the simplest one to check is the differential ETM detuning. If you are doing Task A, you might find you have to apply a detuning to the ETMs in order to create some DC power at the output port. If you’ve got too much power, try reducing this detuning.
If that doesn’t help, then I would check the power in various parts of your interferometer makes sense in the case when you do not apply any signals to mirrors. First add some
pds in central interferometer (e.g. looking at port 1 of the beam splitter) and check that the power is what it should be. It should be within a factor of a few of the input power times the power recycling cavity finesse. Then check the arm cavity powers – these should be roughly the power in the central interferometer times the arm cavity finesse (it will be slightly smaller if you have applied a differential detuning). They should be equal to at least the ~1% level in both arms. Another thing to check would be in which frequency the light at the dark port is, using anad. If it’s in the carrier frequency (f=0) then it’s possible your interferometer is not set close to the dark fringe condition and you must adjust the relative phase of your arms (too much differential arm cavity detuning, for example). If the power is in another frequency, such as one of your modulation sidebands, then check the relative distance between the ITMs and the beam splitter is what it should be (what this distance should be, depends on which task you are doing).After trying differential detuning the minimum power we got is order of some hundred of micro watt but the expected should be around 10 micro watt. So whether this will work or should we try to bring it to 10 micro watt
Sean’s advice is good – you should be able to use these to get around (or even below) 10uW. If you’re still struggling, and all that power *is* at the carrier frequency, one additional thing you could check is whether the power at the dark port is in the 00 mode or in higher order modes. If it’s in HOMs, that’s a sign your mode matching could be improved.
There’s several possibilities there, e.g. the P/SRC could be impefectly matched to the arms, in which case you should re-optimize the mirror curvatures.
Another possibility is that the beam modes from the arms could be slightly different to one another by the time they interfere at the beamsplitter, so e.g. PRX!=PRY. In that case, the wavefronts of the beams from X and Y wouldn’t be identical, so rather than getting a purely dark spot at the dark port, you’d get a ringed pattern (see e.g. the pattern in this video, though the video doesn’t explain the rings itself) – i.e. you scatter light from the 00 into HOMs. You could try differentially varying the *macroscopic* lengths (i.e. L, not phi) between the BS and ITMs by a small amount to try and minimize this X/Y mismatch. Bear in mind that changing this length might affect how any sidebands you have will propagate through your detector too.