149
edits
Zjs log: Difference between revisions
Jump to navigation
Jump to search
no edit summary
No edit summary |
No edit summary |
||
|
Today i made some changes to the electronics.
* '''Added 50 ohm impedance matching resistors to the photodiode inputs.''' I had been hesitant to do this on account of it shrinking our input signal but decided to try it on account of noise; perhaps the low impedance would be less susceptible from noise coming from the photodiodes or noise from some ground loop through the photodiodes. Indeed the error signal shrank somewhat, but was still plenty big enough to do feedback.
* '''Added a new passive fast feedback branch that goes directly to the laser diode.''' I talked to Larry yesterday about our issues trying to feedback through the laser driver. He said that that's problematic for same two reasons we had been suspecting: ground/noise issues and the fact the laser driver mod input has very high gain, so you need very low noise. He suggested a fast feedback branch directly to the diode and sketched out a little circuit. Turns out we had a lot of little discarded fast feedback boxes lying around so i cannibalized one of those used the layout suggested by Larry. His circuit is nice and simple and passive, so we also don't have to worry about powering it. I stuck it inline immediately before the laser. As for an error signal pick-off, i added one inside the lockbox, picking off inside the fast feedback branch, immediately after the first op-amp buffer. This makes it so the phase is correct, and not inverted relative to the slow feedback branch. I didn't disconnect the existing fast feedback but left it in place. We could try sending that directly to he diode as well. Below is a PDF of the circuit.
[[File:Passive_fast_feedback.pdf]]
It consists of two sections. There is a filter on the main laser driver line: two 820uH inductors as well as a 10uF cap and 3ohm resistor to ground. This is to prevent fast feedback from going back into the laser driver. Also, the low pass may clean up the laser driver output. The second section is the fast feedback branch. There are low pass and high pass branches, so the gain is actually less in the middle of the frequency range. From Larry: apparently you want to suppress the gain a tad in the middle as that's where the laser will naturally respond. The high frequency gain is to fix phase response, so that there is not too much phase accumulation at high freq. Also, he recommended the whole circuit being preceded by 10dB attenuator. I ended up putting in 16dB as that seemed to work well. It's interesting how you really don't need any gain in the fast feedback branch.
The lock was much better when utilizing this fast feedback branch, the transmitted IR was very flat in output intensity, fluctuating less than 10%. Lock is a little more finicky, i.e. photodiodes seem to get unbalanced more frequently, this probably due in part to the 50inputs. Hopefully we're finally getting a handle on these issues.
--[[User:Zjsimmons|Zjsimmons]] ([[User talk:Zjsimmons|talk]]) 18:33
| |||