Apart from the detector read-out noise, images taken with the CONICA array contain coherent noise. The extra noise is present in all images, but the amount varies with time. It is usually well below the read-out noise, but, on some occasions, can have a similar strength. This page describes the noise, presents examples and lists the dates when the noise was strong.

The noise patterns include:

1. Randomly appearing horizontal stripes that seem to be some form of electrical interference. They were seen in the NACO frames between August 13 and September 5 and then gradually disappeared.
   

2. 8-pixel noise. It disappeared after the last year intervention but has now returned and affects NACO frames randomly. It was first seen on August 15 and abruptly disappeared on September 20. Due to these high frequency variations the readout noise (RON) values increased in case of every readout mode (with Double_RdRsRd appearing to be the one that is most affected), but returned to normal level after September 20. For reference see NACO trending plots and control charts. The 8-pixel noise appeared to have uncorrectable random pattern.
   

3. Dense pattern of stripes that are uniformly distributed over the whole frame, called the 50 Hz noise. It was found that it is caused by the fans in the front end electronics of IRACE. The lines of the 50 Hz noise became stronger in mid-August, for reason which is not clear yet, and they have been randomly seen in NACO frames since then.

The 50 Hz noise and the horizontal stripes can be successfully removed with appropriate processing. Click here for an example of such algorithm.

General timeline of events:

• Mid 2002 CONICA Intervention.
  • Weak 50 Hz
  • Weak horizontal stripes
  • Weak 8-pixel noise
• August 2003 CONICA Intervention
  • Weak 50 Hz
  • Strong horizontal stripes
  • Strong 8-pixel noise
• October 1st CONICA warm up and cool down
  • Strong 50 Hz
  • Weak horizontal stripes
  • Weak 8-pixel noise.

Examples of frames affected by different noise features are shown. Since the noise structures are at the level of only a few counts they are best seen in the difference frames. What follows are NACO calibration frames with their corresponding reference data subtracted.

This is an example of irregular horizontal stripes in a difference frame of a dark from August 16, 2003, DIT=20 s, FowlerNsample readout mode, HighSensitivity.

A closeup of a center of the dark frame from August 18, 2003, DIT=4 s, Double_RdRstRd readout mode, HighSensitivity, showing 8-pixel noise.

This dark from August 19, 2003, DIT=180 s, Double_RdRstRd readout mode, HighSensitivity, is an example of a 8-pixel noise superposed on the 50 Hz noise type stripes.