Multi-Frame file slope problem

Multi-Frame file slope problem

I first noticed this problem in th 2c data set and found it also in the 2d and 3 data sets. These are some of the first files collected using more than 1 frame in a file. When I plotted the spectral track darks the first frames dark was low and all the following darks were higher and on top of each other (this was 600 second data). When I looked at the 2d and 3 darks at 0.1 and 0.35 second data it was the same. You also it is in the between track darks. Then I looked at some light scans collected for the wavelength cal #2 and the light scans do the same thing.

Email from Mark to Alan.

Can you explain the following PIXIS CCD camera operational characteristics:

We are concentrated here on background images (though the effect is also present in multi-frame "light" image files). I

In all cases the CCD temp is -75 deg C, the accumulations are set to 1. We are using speed 100khz and gain 3. We see the problem described below at integration times ranging from 0.1 to 600 seconds. We are in the process of checking the other speed/gain combinations for the problem.

If we take one image frame per data file we get a an image that has consistently lower ADU vs. a data file with more than one image frame per data file.

In these multiple frame data files, we consistently get a first frame that is about 10 ADU lower than the rest of the frames in the file.

In this "bad" dark frame, we observe that the dark ADU per pixel consistently increase from pixel 0,0 to pixel 1000,1000. The higher ADU level of the first frame's 1000,1000 pixel matches the mean ADU level of the remaining dark frames. Image frames number 2 through N have a consistent ADU level.

This effect seams consistent. But functionally, it means we cannot do an adequate background correction. Correcting a multi-frame "light" image file using a single frame background image, generates screwed up results. Acquiring and tossing dummy images is one workaround but we cant afford the extra time for this.

So,

Why does it take a full image for the readout circuitry to "settle" and behave consistently?

Is this a problem with our CCD unit or is this issue inherent to the PIXIS design?

Could this be a WinView issue?

Any help on the would be much appreciated.

All of our PIXIS evaluation data is available at:

http://data.moby.mlml.calstate.edu/moby2_testing/sudirlsit.html

In all the tests at the site above, the CCD camera is attached to a VPH imaging spectrograph having 14-track fiber optic input.

Mahalo,

Mark

Response from Al Lichty, Princeton Instruments

Subject:	RE: CCD function
Date:	Fri, 11 Dec 2009 11:34:42 -0500
From:	Lichty, Alan
To:	Mark Yarbrough

FYI
Alan

from engineering. Here is what is going on with this.

The idle operation of the camera is not the same as when it is in readout. The camera does “pre-sequence” cleans before it starts acquiring frames, however the clean routine as opposed to reading the pixels is different. Essentially, the preamplifier that is on the CCD is reacting differently at the start of the sequence, which is causing this first frame to look different. Of course, it would be nice if the camera could be operated like a digital scope where it is always sending data to the computer and you just trigger when you want to store the incoming data.

If it is at all possible, they can take the second or a follow-up frame as the dark reference to be subtracted, the problem will be taken care of.

Finding: the "first frame issue" is a characteristic of the PIXIS camera. It is known to the manufacturer and may be correctable in firmware.

These are the 2c data, plotted are the tracks (Pixels are along the wavelength direction). Note that in these graphs, the first image/frame from each file (red and cyan line) are lower than the following images! The black line is the 1b dark, which was collected wth just one image/frame in the file. The question is why is the first frame of the darks lower. Is this a cleaning issue?? Ie would doing a clean between frame help? Notice also that is affect decrease from track 1 to track 14?? So there is a slope in frame one which does not exist in frame 2-5.

The lines labeled 2c_5BI_600s_01_1 are the 5 frames from the first file and 2c_5BI_600s_02_1 are the 5 frames from the second file.

These are the 5 frames from the 5b light data (0.35 sec), plotted are the across the tracks.

These are the 5 frames from the 2d data (0.1 sec), plotted are the across the tracks. I it is a little harder to see the problem in the light scans.

pwd: C:\zflora\mldata\MOBY2\Intital_testing\091208
Date: 14-Dec-2009 09:14:17
Created from multi_track_frame_problem_(1)