Electronic noise pickup is one of the major uncorrected Electronic Problems seen in the 2MASS data so far. Electronic Problems contains a summary of the problem as well as links to numerous examples of affected images.
T. Jarrett has written software to analyze all processed coadds for electronic noise pickup that is now being run operationally. After calculating the median source-removed row value for each half of the coadd, the software performs a Fourier transform on the column of those median row values to find the dominant period, and then analyzes the median row values to determine the actual amplitude near that period.
That noise program has now been run on 6 southern nights from 980514s to 980519s and on 7 northern nights from 980325n to 980404n. We present here a summary of the results from those runs, characterizing the period, amplitude and frequency of occurrence of the electronic pickup in all bands.
The previous analysis of 971116n, scan 116 showed that the observed one sigma error in calculating the amplitude of the pickup is 0.019 DN at J and 0.032 DN at H, very close to the values expected theoretically from the observed noise in each band. The value couldn't be determined for K from that night because there was observed electronic pickup on both halves of the array. We expect that the value at K would be similar to that of H.
Analysis of Photometric Noises for 2MASS Galaxies shows that as long as the electronic noise is less than ~0.05 DN, the galaxy photometric error is increased by less than 10%. At an electronic noise of 0.10 DN, the photometric error is increased by 30% for some galaxies at J, and by 13% at H and K. At an electronic noise of 0.15 DN, the photometric error is increased by 60% for some galaxies at J, and by 30% at H and K.
We have plotted the values of the period vs. amplitude for each band for each half of the array separately for the north and the south.
For the north, there is clearly electronic noise pickup for this set of nights on the left hand side of the array in all three bands with periods of 50-75". For the south, there is a small amount of noise pickup for this set of nights on the right hand side of the array at J with periods of ~90-110". H band shows excess noise on both sides of the array with periods of 70-110". However, it is difficult to tell if this excess noise is due to electronic pickup noise or to some other cause such as airglow.
For these nights, the maximum electronic noise pickup is as follows, for each side of the arrays separately:
| Band | Maximum Electronic Noise Pickup (DN) | |||
|---|---|---|---|---|
| North | South | |||
| Left | Right | Left | Right | |
| J | 0.18 | none | none | 0.17 |
| H | 0.12 | none | ?? | ?? |
| K | 0.20 | none | none | none? |
The values for the north are in good agreement with the previous results for 971116n, scan 116.
The cumulative number of scans with measured noise above a given threshold is shown for both north and south. The difference between the left and right side of the arrays is immediately apparent for all numeric entries in the above table.
Also apparent now is the much larger tail of large periodic noise values in the south compared to the north. Since previous analysis of the J Coadd Noise Vs. Background For North And South showed that the mean values of the noise in the north and south are comparable at J, there is clearly a significant source of additional infrequent noise in the south. It is possible that the south has electronic noise pickup on both sides of the array at some level in all bands. However, until these nights are determined to be photometric, any such conclusions are premature. In any case, further work needs to be done to verify this speculation.
(To be added: plot of % of scans vs. galaxy photometric error, which is easily determined from the above.)