darkframe problem still QHY9

[Peteram2003]

hi Q,

I will do that, however can you please tell me how to read the average level of the overscan area. So far I took pictures in ccdcap or ezcap and then opened the fit file in Astroart to read the statistics.

Regards,


Peter

[Peteram2003]

hello Q,

I am doing the test right now in EZcap. I am using gain 10 offset 100 and temp -10. I will post some screenshots in a few minutes.


Peter

[Peteram2003]

Allright Q,

Seems you are on the right track. Here is a screenshot of the result. From left to right. 0 seconds 120 seconds and 0 seconds.
AVerage value of last 0 seconds is lower than  the first and also lower than the 120 seconds. I'll see if I can see the results of only the overscan area.

Peter

[QiuHY]

Hello,

       Yes your result is similar with me. And you can try some test after 15-30minutes *after CCD temperature reach a stable value. In this condition the 120sec and 0sec may get similar value.

I think the CCD sensor (the chip internal) has not reach themal balance at first.Although the temperature curve reach a stable value. Because the temperature sensor is not on the internal chip. It is on the copper pad beside CCD sensor(You can see this sensor from CCD widnow) So the internal CCD chip need more time to get thermal balance.

Normally you capture it at the sequece of 0sec, 60sec,120sec....etc. So the first 0sec 's temperature is higher than later long exposure images. So 0sec's average level is higher than longer exposure images. This is the reason.

You can do more test as this way. You can find it also.


Best regards,

[Peteram2003]

Ok Q,

I'll try to do that than.  So if I try say a 5 min darkframe first and then the bias it should turn out ok is what you are saying.
I will try a 5 minute first then a 2 min and than a 0sec to see what turns up.
I have attatched an image with the average values of the overscan area.

Peter

[Peteram2003]

Hello Q,

Here is a second set of testresults. I took a sequence of darkframes with EZcap. Settings gain:10 offset:100
I waited for the temperature to stabilize. Here is what I found from left to right top to bottom.
240 sec,  120 sec,  60 sec, 0sec, 5 sec and finally another 120 sec.
as you can see 240sec has an average of 412 versus 470 for 120 sec. Again reversed from what you expect. If what you are suggesting is correct, the chip was cooler during the first 240sec. The 60 sec is 412 again lower than 120 sec but same as the 240. then 0 seconds is higher again at 487! so highest so far! The 5 sec follows at value 401 lowest so far and much lower than the 0seconds. Finally the 120 seconds again and now at 412 or about the same as 240 and 60 seconds earlier.
There seems to be no consistency here as far as I can tell.
I am curious as to what you can come up with.

Regards,

Peter

[Peteram2003]

Hi Q,

I ran the exact same test after the first and this time the result seem more like it should be. What do you think?

[QiuHY]

Hello,

       The second test is correct.  First test is due to the CCD chip has not get real thermal balanced.
What's I said is  "And you can try some test after 15-30minutes *after CCD temperature reach a stable value".
There is two "after" .

Best regards,

[Peteram2003]

Hi Q,

Indeed this seems to be the case. So you suggest I let the camera temp stabilize at least 15 minutes or so before I shoot any correction frames, would that be correct?
Question: A bias frame of 0 seconds contains only readout signal. Shouldn't that be the same at different temperatures?

Regards,

Peter

[QiuHY]

Hello,

         Yes. It is best to wait the temperature stablize . This is best for both light and dark frames. 

For bias frame, It do contains readout signal . But is has relationship with a offset value.(Signal offset level) .This level is only a constant number for all pixels in a fixed temperature. It do not effect the noise performance. It just like you add a constant value with software.

         And the "Offset" setting is also such a thing. It do not effect any noise performance. For example, you can adjust the offset to 125, 130. And you will find the background average is changed, But the readout noise is similar. The image is also similar.

         Of cause big offset added into the image will waste the ADC dynamic range. So it is best to adjust the offset value to a suitable condition.

Best regards,

         

         

[Peteram2003]

Hi Q, so you are saying readout signal does change with different Temp.?
If you look in image 20, you see that the average signal of 240 seconds is actually lower that the 120 seconds that came after it. Does this then mean that the temperature of the chip at that time was lower than the one in the 120 second exposure?

Regards,

Peter

[QiuHY]

Hello,

         What I say is readout noise, not readout signal.


         Noise should use STD to test.

         And in addtions,  the readout noise do not includes the thermal noise. So you should use the bias frame to test it.  Or you can use the overscan area to test it.  So you can compare the STD of the overscan area of them.

For presion speaking, the readout noise can reduce a little after cooling. But not much. This is due to the signal amplifer on CCD sensor 's temperature become lower. Amplifer will procude less noise in low temperature conditons.



        For signal level, Some ccd outputs higher offset level in higher temerpature.  So the image20 of 240sec should be colder than 120sec.  I do not know what's the reason because I have not seen the detail experiment conditions
there. One guess is the cooling curves. If you see the cooling curves in auto mode. You will find the curve will go colder at frist then come back and reach the target temeprature after some vibration.

Best regards,


Best regards,

[Peteram2003]

Allright Q, I understand. When I took the sequence I did as you instructed. I used the auto setting for temp control and let the curve stabilzie at the requested temp before I shot the sequence, but appearantly that does not guarantee temp stability of the chip itself. So next time I'll wait a lot longer before starting the sequence.

thanks,

Peter