QHY9 Offset and Gain starting point

[Manav]

Hi guys - I'm trying to get around to setting up my QHY9 after a long time and had a few questions.

I followed the instructions from the following PDF CLICK HERE (http://www.astropixel.gr/uploads/7/8/3/5/7835053/scientifically_determining_ccd_gain _and_offset.pdf).

For the first test i.e. 0 Gain and 100 offset I get the following result:

Exposure(sec) - Mean Pixel Value
0.02 - 50714
0.04 - 50808
0.08 - 50835
0.16 - 50858
0.32 - 50865
0.64 - 50873
1.28 - 50859

Not sure what I'm doing wrong? I'm testing at night in a room which is dimly lit. Should I not be using the 100 offset as a starting point?

Any help would be appreciated.

[tonez]

that whole thing is a fairly long winded procedure and i never really got anything repeatable out of it for some reason

i ended up going with craig stark's method near the bottom of the page here:
http://www.stark-labs.com/help/blog/files/GainAndOffset.php

[multiweb]

Gain 1 Offset 113 for me @-20c

[RobF]

I'm on Gain 10, Offset 100, temp -25 and only use unbinned, unless framing or plate solving. 2x2 binning doesn't work very efficiently for this camera and is somewhat pointless for quality imaging data, but ok if you need fast download (as per examples above).


My simplistic explanation for gain/offset below. Gain is effectively slope and offset is intercept in an algebraic sense:

Offset - This value is added to all pixels. You need to make sure you never get negative or zero values read out - actually value doesn't really matter too much - anything above 100 or below 1000 is probably fine.

Gain - the multiplier you apply to all pixels. Too low and you won't "spread out" your incoming data over the full 16 bits (values of 0 to 65535), too high and you'll "clip" at 65535. You brightest stars should never be over 60000'ish in long exposures or you will lose colour saturation. Shooting a bright ceiling with a quick exposure is another technique.

Temperature - the reduction in noise drops off significantly once you're down to -20. The camera can do about 50-55 degrees celsius below ambient. Pick a temp that you will always be able to reach even on the hottest summer nights. For that matter, if you'll be shooting darks during the day in a dark cupboard, might be worth taking daytime temps into account too.


Once you get gain/offset in the ballpark, there is no reason to fiddle. Make sure you never have "fast downloads" selected in the driver - LOTS more noise.

[Manav]

Rob and Marc - I have checked the QHY forums the values everyone comes up with all over the shop i.e. high amount of variance between QHY9 builds.

Tonez - Might give that process a try tonight.

Its a pity I couldn't use the attached process as it seemed fairly well explained. I'd like to take guess work out of doing things.

Edit: Just found this at QHY website may try this method too (CLICK HERE (http://www.qhyccd.com/ManualQHY123.html)))

[Atmos]

From memory I used something along the order of Gain 21 and Offset 115 but it's been a while.

The idea is to increase your exposure times until the brightness levels stop increasing (full well capacity) and then increase the Gain until that is near 65535 ADU.

Even at 0.02s it looks like you're at full well so you can just use that. Increase the Gain until it gets to 65535. Then you want to do Bias frames and set your average Bias ADU around the 500-1000 mark.

As Rob says, it is best doing this with the cooler on as the bias levels do change with sensor temperature.

[billdan]

Yugant, this is the procedure if you are starting off first time -

Place the dust cover onto the camera, set the gain to 0 and then take a 1msec exposure. Adjust the offset number so that the RMS ADU reading is somewhere between 500 and 1000 ADU. In my case with the QHY12 the offset was 126 to get a 700 ADU bias reading.

Then do a full well test, remove the dust cover and point the camera to a wall that is not too bright. Keep increasing the exposure times and create a table as follows
1sec - 39600 ADU
2sec - 57600
3sec - 61900
4sec - 63340
5sec - 63340
10sec - 63350
30sec - 63390

As you can see the ADU count has flattened out after 4 sec's, so we now set the exposure to 4secs and increase the gain until it reaches 65,000 ADU.

In my case the gain was set to 10, and then you can retest the bias level again which for me was reducing the offset to 125.

Hope this helps
Bill

[Manav]

Colin and Bill - Thanks that makes sense; I think doing the bias frame will do the trick. I'll do this tonight when I get home and see how it goes.

And yes will turn the cooler on this time :)

[Manav]

Step 1

So I basically did the whole thing again this time set the exposure in Nebulosity to 0.001 (1ms) for the bias frame. Adjusted offset until the min value of image was just under 1000 i.e. 740 in my case. My Offset value was 115.

Step 2

Followed Bill's step above with Open Shutter/ and got the following results.

Exposure(sec) - Mean Pixel Value
2 - 32022
3 - 46119
4 - 52701
5 - 54265
6 - 54349
7 - 54362
8 - 54384
10 - 54370
30 - 54561

Based on above I used the 5sec marker as the exposure length and adjusted the gain. This time I looked at the max value of the image until it capped to 65,535 which for me was 16%.

I did the whole test at -10C; I'll repeat it at -20C tomorrow and then leave it as is. I'll also check if ASCOM vs Base Driver for QHY9 make a difference.

Anyway thanks everyone for the help!

Regards
YB