create new tag
view all tags


A signal on the order of pico or femtoseconds is too short to be measured by conventional electronic techniques. The only thing which is of the same temporal order as the signal to be measured is itself. Therefore you copy / split the signal and compare it to itself over a perceptible period of time.

An autocorrelator uses a 50:50 beam splitter to split the beam into two arms of a Michelson Interferometer. The beams are recombined in a nonlinear second harmonic generating crystal. The output of which is the autocorrelation signal. Dichroics are used to separate any remaining infrared from the second harmonic green light.

If the output beams from the Michelson Interferometer are spatially overlapped an interferometric autocorrelation is produced with the output signal oscillating as the pulses are scanned through each other. The oscillations can be fitted if the data is high enough quality - this requires an experimental setup stable to within $\frac{\lambda}{2}$.

If the output beams are not spatially overlapped, but parallel and overlap in the second harmonic generating crystal a third beam is produced in the crystal that is only the autocorrelation signal. The two IR beams can then be irised out. This setup is called an intensity autocorrelation and differs from the interferometric setup in that there are no interferometric oscillations and a smooth envelope is traced out. This can be fitted to a theoretical equation and the original pulse length calculated.

Autocorrelations do not give you specific information about the exact temporal profile of the laser pulse and you must know this to deconvolve the autocorrelation signal fully. Typically, a Gaussian or $\text{sech}^2%$ are assumed. Also, for laser pulse measurements it is common to quote the full-width half-maximum (FWHM) of the pulse and not %$\sigma$ or $4\sigma$. The FWHM of the autocorrelation is divided by the deconvolution factor of $\sqrt{2}$ or 1.54 for a Gaussian and hyperbolic secant pulse respectively.

VI Development

Autocorrelator V1

VI reads out full array of voltage samples from oscilloscope at 1Hz and records the maximum voltage which is proportional to the intensity of the autocorrelation signal. The photodiode is assumed to give a linear response. The photodiode was AC coupled so any DC background should not affect the maximum voltage. The problem with this version was that the oscilloscope VI would occasionally read out 200 samples instead of the full 5000 that cover the screen trace. The maximum value would therefore be noise. A case structure was added with a threshold level of 0.1V. This was mistakenly implemented in such a way that the no value was recorded. Given the autocorrelation technique relies on the data being spread evenly along the length of the stage, a curtailed data set would give a longer and inaccurate pulse length.

Autocorrelator V2-4

Tried to improve functionality in versions 2 to 4. These versions aimed to overcome issues with the stage not reporting its position but only a true value upon completion and hence knowing when to stop recording data from the scope. Customised versions of some of the supplied VIs to control the stage were made that let a parent VI progress whilst waiting for the stage to return a completion true value. A TTL trigger signal was used to trigger the scope so that even when the autocorrelation signal was low, data could still be recorded.

During the experiment / VI development the laser timing changed and the trigger signal used for scope sample acquisition had quadrupled in frequency which mismatched the laser pulses. This led to a lot of null samples in the data. This was due to someone changing the kicker timing signal from which our timing is derived.

Autocorrelator V4

V4 was the final version that was stable and accurate with no known flaws. It is located in "extlw\daq\laser\autocorrelator" and is called autocorrelator_low_rep_rate.vi. Additionally, the VIs required to move the Sigma Koki 500mm Stage at to a relative / absolute position or at a constant speed over a given range are also included for use in setting up the autocorrelator. In the autocorrelator directory there is a folder called "exe" which contains executable version of the VIs that will run on a system without LabView 8.6 installed.

This VI reads out a measurement from the oscilloscope which must be manually programmed first.

Data is recorded in a text file where the name is


where ### is the speed of the stage in $mms^{-1}$.

Latex rendering error!! dvi file was not created.

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions

Physics WebpagesRHUL WebpagesCampus Connect • Royal Holloway, University of London, Egham, Surrey TW20 0EX; Tel/Fax +44 (0)1784 434455/437520

Topic revision: r2 - 06 Feb 2014 - LaurieNevay

This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2023 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding RHUL Physics Department TWiki? Send feedback