40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  40m Log  Not logged in ELOG logo
Entry  Tue Jul 29 19:41:38 2008, Yoichi, Update, PSL, FSS loop transfer functions 7x
    Reply  Tue Jul 29 23:04:34 2008, Yoichi, Update, PSL, FSS loop transfer functions OpltfPZTOnlyRaw.epsOpltfPZTOnly.eps
Message ID: 758     Entry time: Tue Jul 29 19:41:38 2008     Reply to this: 761
Author: Yoichi 
Type: Update 
Category: PSL 
Subject: FSS loop transfer functions 
Last night I measured a bunch of transfer functions on the FSS loop.
All the loop gains were measured with the common gain = 30db and the fast gain = 18dB.

(1) The first attachment is the overall open loop transfer function of the FSS loop. I put a signal from the Test IN2 and observed signals from IN1 and IN2.
The UGF is about 180kHz.
By increasing the RF amplitude going to the EOM (i.e. increasing the sideband power), I can further increase the gain of the servo.
However, it made the PC drive immediately crazy. Probably it was some oscillation.

(2) Then I locked the ref. cav. with only the PZT actuator. I did so by simply unplugging the cable going to the PC.
Surprisingly, the cavity locked with the *same* gain setting as before. The second attachment shows the open loop transfer function measured in this configuration. It seems wrong, I mean, it should be unstable. But the cavity locked. A mystery.

(3) The third plot is the measured TF from the Test IN1 of the FSS board to the fast out (output to the PZT).

(4) By dividing the TF measured in (2) with the TF of (3), I got the response of the PZT times the cavity response. This is shown in the attachment 4.

(5) We can guess the open loop TF of the PC path by subtracting the TF in (2) from (1). It is shown in the attachment 5.

(6) The filter shape of the PC path is measured by injecting signal from the Test IN1 of the FSS board and observing it at the PC output. Since it is a high voltage output, I reduced the common gain to -8.5dB during the measurement. The attachment 6 is the measured filter shape. The gain is corrected to show what it should look like when the common gain = 30dB.

(7) By dividing (5) with (6), I plotted the response of the PC times the cavity response in the attachment 7. Since the 1/f cavity pole and the response of the PC, which is proportional to f, should cancel out, we expect a flat response above the cavity pole frequency (38kHz). You could say it is a sort of flat, if you have obscured eyes.

The measurement of the PZT open loop TF is very suspicious. According to this, the PC path has a very large gain even at very low frequencies (there is no cross over above 1kHz). This cannot be true. Maybe the cavity's optical gain was low when it was locked with only the PZT. I will re-measure it.
The plot (4) is also strange becaues it does not show the low pass feature expected from the cavity pole.
Attachment 1: OverallOPLTF.eps  24 kB  | Hide | Hide all
OverallOPLTF.eps
Attachment 2: OpltfPZTOnly.eps  24 kB  | Hide | Hide all
OpltfPZTOnly.eps
Attachment 3: PZTFilter.eps  22 kB  | Hide | Hide all
PZTFilter.eps
Attachment 4: PZTxCavityPole.eps  23 kB  | Hide | Hide all
PZTxCavityPole.eps
Attachment 5: OpltfPCOnly.eps  23 kB  | Hide | Hide all
OpltfPCOnly.eps
Attachment 6: PCFilter.eps  23 kB  | Hide | Hide all
PCFilter.eps
Attachment 7: PCxCavityPole.eps  24 kB  | Hide | Hide all
PCxCavityPole.eps
ELOG V3.1.3-