Index of sections
Contents
the New RF System
Overview
- The Upgraded 40 Meter will have an entirely new configuration of the RF system. That is mostly because:
- 1. the length of the short cavities will change (getting longer);
- 2. the interferometer will be set in a broadband sensing configuration and a new locking scheme and signals will be necessary.
- This is how the new RF system will look like.
A PDF version of the diagram can be foundhere.
Here's an older version of the plan made with yEd. Descriptions of the single components in the diagram are shown just by passing over them with the mouse. Clicking on them links to the correspondent data sheet. The numbers included in the diagrams describe the limit case of the highest modulation depth that we may use at the 40m (gamma=0.3).
Requirements
- Need control of the modulation depths from the control room. Either we keep the current RFAM Stab. boxes or we introduce our own remote control attenuators.
- The system needs to preserve the low phase noise of the Wenzel crystal and multiplier (~ -160dBc/Hz).
- Care should be taken to avoid reflections. There should be an amplifier at each split to avoid reflections between the loads from cross-talking.
- Wherever we have high level signals, we need to use the Heliax cable. Each Heliax cable must be strain releieved at each end. The connection between the Heliax and the components shall be the same semi-rigid RG-174 as we have now or some equally low radiation type.
- All connections shall be SMA or type N. Absolutely no BNC allowed.
Main Features of the new RF Scheme
Change in Length and Frequency
- The change of the length of the recycling cavities will impose different sideband frequencies.
- (It is actually the other way around, since we're changing the cavity lengths in order to accommodate lower modulation frequencies).
f1 = 11065399 Hz and f2 = 5xf1 =55326995 Hz .
Sideband generation
The modulations will be generated by a single broadband EOM , thus no Mach-Zehnder any more .
There will be a single master oscillator to generate the main 11 MHz frequency.
All the other signals, with frequencies multiple of the main one, 2x, 3x, 5x, 10x, 15x respectively, will be generated starting from it, rather than by independent oscillators.
f1 will be generated by a crystal oscillator
- f2 will be obtained by f1 by means of a 5-time frequency multiplier. The two signals will be electrically combined before the EOM
- The designed modulation depth for the Upgrade will range somewhere in between 0.1 and 0.3.
- The New Focus KTP 4064 boradband EOM has an efficiency factor of 13mrad/V. Therefore the driving voltage should range between 30 dBm and 40 dBm. However the frequency multiplier will output 20 dBm, so 20 dBm will have to be gained by the signals before the EOM in order to match the requirements on the modulation depth.
- The step will be covered in part by the Multiple-resonant EOM in part by an low noise, high power amplifier placed in between.
- After the multiplier and before the driver, the f5 signal will be have to be amplified
Demodulation
Concepts of the Upgrade
- Use of the same signal from the main oscillator for the demodulation
The same demod boards as the old 40m: LIGO D990511.
- Installation of surface mount low-pass filters in correspondence of U5 (as in the schematic):
- Installation of high pass filters to the RFPD input of the demod boards (e.g. SHP-25 for 55 MHz demod board)
Installation status
Find the actual setups here : LSC dmoedulators
Frequency Generation and Distribution
Installation status
This page describes the frequency generation box.
This page describes the frequency distribution box.
Hartmut's Suggestions
Grounding inside the frequency generation box. If we choose to mount the single components inside the box on to a conducting surface, they should make good contact with it. Bad contact can come from oxide forming on the surface, or by mere unevenness of the surface. for this reason at GEO, to make sure they wouldn't incur into this issue, they mounted the components on a plastic breadboard.
- Testing the system. Bad connectors, bad cables, bad soldering can all deteriorate the modulation/demodulation signals and eventually produce phase noise that goes into the demod mixers. One way to measure the presence of phase noise is to pick off with a coupler a demod signals from just before the modualtions' combiner and use it as a local oscillator in the demod boxes. Then we can connect audio earphones to the demodualted output and listen for phase noise. That would simulate the 'closed loop' made by the IFO and the actual phase nosie present in the modulation/demodulation chain. We could test single connections or cables, inside the frequency generation box or in the frequency distribution board by touching/shaking them and listening to the earphones.
Modulation/demodulation system's SNR calculation
See Alberto's thesis 40m svn
RF Photodetectors
This page follows the state of the work in progress on the RF photodiodes upgrade
Baseline RFPDs and LSC Signals
- For locking, the baseline plan is to use the same error signals of AdvLIGO:
DOF
Signal
CARM
REFL_I1
DARM
AS_DC
PRCL
POP_I1
MICH
POP_Q2
SRCL
POP_I2
In case that 3f signals will be used for lock acquisition, as a result of Optickle simulations, we propose the following sensing matrix:
DOF
Signal
X
POX
Y
POY
PRCL
REFL_32I
MICH
REFL_31I
SRCL
REFL_32Q
Photodiode properties
More specs for the diodes (as diode shunt resistance) can be found inthis.
This document by Rich Abbott contains lots of measurements of PD capacitance and resistance.
- We're going to adapt the current photodetectors for the new signals as described in the following table:
Photodetector
Adapted From
Changes
Parts Needed
Modeling
Measurements
REFL11
C34
tunable capacitor: SD3012-ND 9-120 pF
L4
tunable inductor: CoilCraft 143-10J12L
L5
tunable inductor: CoilCraft 143-20J12L
POX11
C34
tunable capacitor: SD3012-ND 9-120 pF
L4
tunable inductor: CoilCraft 143-10J12L
L5
tunable inductor: CoilCraft 143-20J12L
POY11
C34
tunable capacitor: SD3012-ND 9-120 pF
L4
tunable inductor: CoilCraft 143-10J12L
L5
tunable inductor: CoilCraft 143-20J12L
REFL55
diode
2mm InGaAs diode (C30642GH)
C34
tunable capacitor: SD3012-ND 9-120 pF
L4
tunable inductor: CoilCraft 164-02A06SL
L5
tunable inductor: CoilCraft 164-04A06SL
AS55
diode
2mm InGaAs diode (C30642GH)
C34
tunable capacitor: SD3012-ND 9-120 pF
L4
tunable inductor: CoilCraft 164-02A06SL
L5
tunable inductor: CoilCraft 164-04A06SL
REFL33
SPOB66
TBD
TBD
TBD
REFL166
REFL166 (PD11)
TBD
TBD
TBD
AS166
AS166 (PD1)
TBD
TBD
TBD
POP22
PO133 (PD7-8)
TBD
TBD
TBD
POP110
PO199 (PD7-8)
TBD
TBD
TBD
MCREFL
MCREFL
TBD
TBD
TBD
Upgrade Schedule
- As of Sunday July 25th, the remainder of the RF System Upgrade is organized in blocks as below.
The daily schedule can be found in this RF System Upgrade Google Calendar
(2 weeks: Sun Jul 25 -> Fri Aug 6): RF Generation
(2 weeks: Mon Aug 9 -> Fri Aug 20): RF Distribution
(3.5 weeks: Sun Aug 22 -> Wed Sep 8th): LSC PDs
(2 weeks: Thu Sept 16 -> Fri Oct 1st): /schedule/Commissioning
Plan reviews
- In addition to the requirements listed above, Rana suggested the following changes/additions to plan B:
replace the directional coupler with splitters - replace LIGO LSC Frequency Distribution splitters with 8-way power splitters from Mini-Circuit
- include cable power losses in the calculations
- abandon the 2 Omega I and Q Demodulator Board
- buy a voltage controlled attenuator to test it
- check what's the modulation depth we need for the mode cleaner and evaluate whether we need a high-power amplifier for the 29MHz as well as for the 11 and 55MHz
- redesign the cabling: check out Intra-Flex for custom made cables
- a band-pass filter before each demodulator board
References, Some diagrams and Links
Tables and Diagrams
References
Documentation of the design, construction and characterization of the RF System is now available in the LIGO DOC# T000461.
Coax cables Coaxial Cables