Coil driver upgrade for Optomechanics experiments @ 40m
Motivation
From noise-budgeting of the DRMI, and measurements of the current suspension coil driver electronics noise, it is projected that this will the be the limiting noise source in DARM strain sensitivity at frequencies > ~100Hz. Since we'd like to measure ponderomotive squeezing, we need the limiting interferometer noise at the frequencies of interest to be quantum (shot) noise. Therefore, we need to reduce the coil driver noise in the frequency band of interest. Ideas/measurements/data towards this end are collected on this page.
Characterization of the current situation
Measurement of coil driver electronics noise, with input terminated to ground via 50ohms: elog13015.
Projected noise contribution with current de-whitening electronics, for various choices of series resistance for current coil driver topology: elog13146. Note that these plots were made with MICH in mind, so for DARM, they will have to be scaled by a factor of sqrt(4/3).
Impact of increasing the series resistance, and hence the reduced actuation range available: elog13417. Blech; this statistical analysis isn't accurate enough. We need an actuation budget to find out how much force is needed for each part of the spectrum in which state.
Current noise budget of MICH in DRMI configuration: elog13428.
Note that in the current incarnation of the coil driver board, there are actually two series resistances whose Johnson noise we have to consider - one for the bias path and one for the fast DAC signal path.
The schematics for the suspension electronics, and associated modifications that have been implemented, are available at the 40m DCC tree.
Ideas for modeling and prototyping
- Current topology, with increased series resistance.
- How high can we go without compromising on actuation range?
aLIGO RFPD (D1101124) test circuit approach, with transistor acting as nearly ideal voltage-controlled current source.
- What kind of dynamic range does this give us?
- What is the right kind of transistor to use? MOSFET? JFET? BJT?
What are the current noises of the various types of transistor? How do we calculate this? This document seems like a pretty good summary of the noises.
This is an example of low noise current source with FETs (http://ieeexplore.ieee.org/document/612447/)
- Current topology, but add a series inductor in series with the series resistor, for a frequency-dependant transimpedance.
- What inductance do we need in order to reduce the current noise between 100Hz-1kHz (say)?
Plan of action
- We should rethink this whole thing. We don't want to have to make a crazy-super current source, if it would be better to redo something else like changing the number of windings or changing the magnet size. For sure we want to get the low displacement noise, but maybe there's a better way to do it.
- First, we need to make models of all 3 approaches and compute the contribution of coil driver noise to displacement noise - the requirement is to be lower than quantum noise level by factor of 10?
- If one approach in particular seems promising, prototype it, and confirm the modelling by making a measurement of the current noise.
- Steve Vass suggestion of two parallel paths, one that is low noise but low actuation range (engaged once IFO is locked) and higher noise but higher actuation range path (for lock acquisition). We would need a relay to switch between these paths. Rana suggests mercury relay. I wonder why we cannot use the optically triggered solid-state relay technology?