This routine incorporates a data slice into a run of data and updates the Allan variance of the run. The Allan variance is calculated for a range of simulated integration times for which artificial samples are calculated from the input data. The Allan variance for a particular integration time is calculated from:-

$$variance=\sum _{i=1\to n-1}\frac{{\left(sample\left(i\right)-sample\left(i+1\right)\right)}^2}{2n}$$

(58)

where n is the number of artificial samples available.

CALL SCULIB_MAP_ALLAN_VARIANCE (DEMOD, N_BOLS, N_SAMPLES, N_ALLAN_BOL, ALLAN_BOL, KMAX, SUM, N_SUM, ARTIFICIAL, N_ARTIFICIAL, ALLAN_VARIANCE, ALLAN_QUALITY, STATUS)

DEMOD (4, N_BOLS, N_SAMPLES) = REAL (Given)

N_BOLS = INTEGER (Given)

N_SAMPLES = INTEGER (Given)

N_ALLAN_BOL = INTEGER (Given)

ALLAN_BOL (N_ALLAN_BOL) = INTEGER (Given)

KMAX = INTEGER (Given)

SUM (KMAX) = REAL (Given and returned)

N_SUM (KMAX) = INTEGER (Given and returned)

ARTIFICIAL (KMAX) = REAL (Given and returned)

N_ARTIFICIAL (KMAX) = INTEGER (Given and returned)

ALLAN_VARIANCE (KMAX) = REAL (Given and returned)

ALLAN_QUALITY (KMAX) = INTEGER (Returned)

STATUS = INTEGER (Given and returned)

If status on entry is good the routine will:-

loop through the data samples to be incorporated -

loop through the bolometers whose data is to be averaged into this sample -

if the bolometer was measured - if the measurement had good quality -

add the measurement to the average for this sample

end if end if

end of loop

if any bolometers had good data for this sample -

calculate the sample average

loop through the Allan variances -

add the sample to the buffer used to calculate the latest artificial sample at the simulated integration time of this variance

if all the data has been obtained for the latest artificial sample -

calculate the artificial sample

if there is only one artificial sample for this simulated integration time we can’t calculate the Allan variance, so set its quality to bad

if there are 2 artifical samples then the Allan variance can be calculated from -

$$variance=\frac{{\left(artificial\text{\_}2-artificial\text{\_}1\right)}^2}{2\times 2}$$

(59)

if there are more than 2 artificial samples

the sum of the squares of the differences between the previous samples is recovered from the current value of the Allan variance

the square of the difference between the current artificial sample and the previous one is added to the sum

the Allan variance is re-calculated

end if

the current artifical sample is stored to be used as the ‘previous’ sample next time round

the variables used to calculate the artificial samples at this simulated integration time are reset

end if

end of loop through Allan variances

end if

end of loop through samples in this dataslice

Copyright ©1995,1996,1997,1998,1999 Particle Physics and Astronomy Research Council. All Rights Reserved.