LSC Photodiode Bias Feedback Tuning

Transcription

1 LSER INTERFEROMETER GRVITTIONL WVE OBSERVTORY LIGO MSSHUSETTS INSTITUTE OF TEHNOLOGY LIFORNI INSTITUTE OF TEHNOLOGY Document Type LIGO T D 0/0/00 LS Photodiode Bias Feedback Tuning Rana dhikari Distribution of this draft: xyz This is an internal working note of the LIGO Project... alifornia Institute of Technology Massachusetts Institute of Technology LIGO Project MS LIGO Project MS NW Pasadena, 9 ambridge, M 09 Phone () 9 9 Phone () Fax () 0 9 Fax () 0 E mail: info@ligo.caltech.edu E mail: info@ligo.mit.edu WWW:

2 Overview This document describes the procedure used to modify the current version (D990 0 B circa July, 000) of the LIGO Length Sensing & ontrol Photodiode (LSPD) to reduce/remove an observed nonlinearity in the RF response. The LSPD is used at the various sensing ports of the interferometer (IFO) to convert optical signals into the electronic error signals used in the various length/frequency control loops and is thereby the device which converts the gravity wave s optical signature into an electronic signal. Various versions of the LSPD are also used in the Input Optics (IO) and Pre Stabilized Laser (PSL) subsystems for similar purposes. s the performance of this device is so crucial to the sensing and control of the IFO, a detailed characterization is important. The Nonlinearity and its onsequences During the characterization of a sample batch of prototype Pds (photodiodes), a nonlinearity in the RF response was noticed: the RF response has a dependence on the D photocurrent. For typical designed power levels (~0 0 mw incident), the magnitude reponse at the resonant frequency of the tuned circuit on the head board may vary by as much as 0% over the range. More serious, however, is the phase lag of 0 0 deg over this range. The current signal readout scheme requires absolute phase orthogonality between the I & Q demodulated signals of deg. and stringent requirements on RF phase wander. In particular, once the relative RF/LO phase has been set, it cannot be allowed to drift by more than deg. over the full range of power. The reason for this being that the error signals corresponding to various linearly independent length degrees of freedom can only be discriminated by their RF phase. During detection mode operation, the seriousness of this effect is less, since with the entire IFO locked and the laser power stabilized, the fractional power fluctuations will be small. However, during the commissioning phase, the laser power varies due to alignment fluctuations and is also intentionally varied for diagnostic purposes. More importantly, during lock acquisition, the power at the various ports of the IFO will change by large amounts. In order to retain good discrimination of signals, the RF phase response should be a fairly flat function of power.

3 The Measurement. Required Materials/Intruments The following is a list of the equipment necessary to conduct the measurement: > 00 mw Nd:YG laser (e.g. Lightwave Electronics NPRO xxx) () ¼ wave plate () Polarizing cube beamsplitter () RF network analyzer (e.g. HP 9) () Broadband amplitude modulator (M) (e.g. New Focus 0) In addition, the setup will require various lenses and mirrors to get the beam size correct through the M and at the LSPD face. recommended accessory is a ½ wave plate/cube polarizing beamsplitter (PBS) combo to adjust the power incident on the LSPD in a way that does not influence the laser.. Experimental Setup Shown below is a schematic of the setup on an optical table at MIT and which was also repeated at LHO. Beam ½ PBS Mirror Dump 00 mw Laser Lens ¼ mplitude Modulator Lens Mirror 0 degree RF Splitter Beam Dump Mirror 0 Ohm Term PBS pin D sub D Out RF Out Source Out Input R Input B LSPD ontrol Voltage IN RF RF Network Network nalyzer Voltmeter Voltage Source LSPD Portable Interface Box Fig. : onceptual Drawing of the tabletop setup to characterize the RF Response

4 . Measurement Procedure The setup of the ¼ wave plate/m section should be such that with 0 volts on the M, 0% of the incident power goes to the beam dump on the PBS immediately following. This is done by rotating the ¼ wave plate so as to bias the polarization. The light incident on the PBS is now in the linear regime. So the power incident on the LSPD has a component at the modulation frequency who s amplitude is proportional to the level of the modulation drive. For the New Focus 0, when coupled with a 0 Ohm terminator, a drive level of approximately dbm is sufficient. The lenses should be chosen so as to ensure a small beam size through the M (~0. mm dia. is good for the 0) and that the beam diameter on the PD is 0. mm / 0. mm. The first step is to operate the Network nalyzer in the network mode and take a swept sine response from 0 00 Mhz. With ~0 mw incident on the PD, the resulting response will allow one to calculate all useful parameters of the resonant circuit (@ xω) and the notch (@ xω), where ω is the relevant modulation frequency. The frequencies and Q s of these should be verified to correspond to spec for that particular PD. Next we evaluate the performance of the PD. djust the sweep range to a few Mhz around the resonant frequency and measure the magnitude and phase of the response. lso note the D Out voltage. In a typical configuration, the D transimpedance of the circuit is 0 Ohms, so assuming a PD responsivity of 0. /W, the 0 mw of power should give ~0 mv. This measurement should be repeated with varying power, with the parameters of Vdc, RF mag, and RF phase, recorded for each value of the power. The number of points which are required to get an accurate picture of the response depends on the particular PD, but typically 0 pts. is sufficient, between 0 to 0 mw or from 0. to. Vdc. The following are plots of data taken in this manner from a "typical" LSPD. The first two show data taken as explained above. The third will be explained in the section.

5 Fig. : Data from the initial test of LSPD s/n #. Bias Feedback Tuning There is an input to the LSPD box, referred to as Vc djust which allows one to sum into the bias voltage (Vb) on the PD s cathode. By adjusting this voltage by hand as the incident power is increased, the phase of the response can be easily kept to within a small fraction of a degree. By repeating the above measurement, but this time tweaking Vb via Vc to keep the phase flat and recording Vc, the positive feedback required can be determined. The next step is to adjust the circuit to apply the correct positive feedback. One can see from the LSPD schematic that a circuit has already been implemented to supply a positive feedback to Vb. In fact, the bias voltage (Vb) can be represented by the formula: Vb = Vi Idc (R RB) (R/R) Vc where Vi is the initial (no photocurrent) bias voltage. This can be taken to be V. So the third plot in Fig. shows data representing what Vb was required to correct that particular diode. linear fit extracts from the data the additional feedback required. Then changing the value of R, changes the gain in the feedback loop.

6 fter doing this, the measurement of section should be repeated in order to determine that:. Nothing is broken.. The required phase flatness has been achieved.. The final broadband frequency response MUST be recorded and archived for every single LSPD as this will be used in detector calibration. This last step is vital not only for the listed reason but also to ensure that there has been no significant change in the bandpass or bandstop frequencies. hanging the bias voltage across the PD may change the resonant response of the tuned circuit whether it is by changing the capacitance of the PD or through some other mechanism. Typically the phase flatness which is achieved amounts to a shallow phase lag of ~ deg. as one ramps the power from 0 0 mw.

7 J DB BULKHED MLE ONNETOR J ID- V 0 0.0uF L 0uH 0.0uF V FLT IN() OUT() V D 9 0 V RW RW Vc DJUST (Vc DJUST REF) TSENSE (TSENSE REF) D OUT (D OUT REF) ENBLE IN (ENBLE IN REF) STTUS/OTEMP OUT (STTUS/OTEMP OUT REF) NOT USED 9 0 V U LM09H Vin V FLT IN() OUT() D OUTSIDE THE ENLOSURE INSIDE THE ENLOSURE -ONDUTOR RIBBON BLE ENBLE IN must be HIGH (i.e. ) to enable the photodiode. FOR SIGNL RETURNS FLT IN() OUT() FLT IN() OUT() OUTSIDE THE RF GE WITHIN THE RF GE V 0.0uF Vin L 0uH U V LM0H- 0.0uF V V FLT IN() OUT() FLT IN() OUT() FLT IN() OUT() D b.SH RF FENE, pins B D OUT J SM FEEDTHROUGH FLT IN() OUT() Vc DJUST TSENSE D_OUT V V RF LID B FLT9 IN() OUT() ENBLE IN STTUS/OTEMP RF OUT J SM FEEDTHROUGH D OUT SM RF OUT SM TEST INPUT J SM FEEDTHROUGH TEST INPUT SM RF Box Date Last Modified: -Jan-00 Title LIGO Laboratory LS Photodiode alifornia Institute of Technology Massachusetts Institute of Technology LIGO Date: -Jan-000 Size: B D Number: D90-0- PB / SH Revision: B / Engineer: J. Suina D. Ouimette Time: ::09 File: S:\Systems\LS\Photodiode\PB Ver B\Sch Ver \000b.prj Sheet of

8 IRUITRY ON THIS PGE IS WITHIN THE RF GE L See Note -Omega Bandstop Filter 0.uF D B TEST INPUT SM R 9.9 % Output scale is 0.0 mv/ e.g., 0.V = V Vc DJUST V R0 TSENSE.00K % 0.uF R9 R.00K % 9 V 0.uF R0B.9K %.K % R.K % Multiple resistors divide the distributed capacitance between TEST INPUT and RF IRUIT U D0BR 0.uF R.00K % R.00K % R R 00 %.K % 0V U 0.0uF 0V R U0 LMH V.00K % R.K % OUT 0.0uF R.00K % luminum Thermal Block Temp Sense R0 00 % R.K % THERMBLOK LUMINUM THERML BLOK screw holes connected to R.K % V R K 0.uF 0 0V U OPF E/S.0-0pF D See Note SE 0.uF Photodiode urrent Positive Feedback Loop, GIN = -omega Bandpass Filter shield pins connected to R 0 % /W RB 0 % /W V L See Note U 9 0.0uF LT 0.uF 0V 0.0uF R See Note 0.uF L.uH V R 00 % R 9.9 % 0.0uF 0V 0.0uF LT R 0 0V nf U MX0ES R % 0 nf 0.uF R 0 R 0 RF Out, Gain = 0 R 0 D_OUT D OUT SM U9E RF OUT SM U9F D B 0V 0 SNH0D SNH0D R9 9 % R0.K % Photodiode Bias Voltage, Vi =.9Volts 9 V Vout DJ Vin 0V U LMH 0.uF V NOTES: ) Value depends on one-omega frequency required. See Table for frequency and value. ) Value depends on two-omega frequency required. See Table for frequency and value. ) R is not used at this time. U9- U9- R9 K % 0.uF U9 E/S = TTL logic HIGH E/S ENBLES the OPF Op mp Date Last Modified: -Jan-00 Title LIGO Laboratory LS Photodiode alifornia Institute of Technology Massachusetts Institute of Technology LIGO Date: -Jan-000 Size: B D Number: D90-0- PB / SH Revision: B / Engineer: J. Suina D. Ouimette Time: :: File: S:\Systems\LS\Photodiode\PB Ver B\Sch Ver \00b.SH Sheet of ENBLE IN SNH0D R K % U9B SNH0D R K % U9 SNH0D R K % U9D 9 SNH0D R K % STTUS/OTEMP