QHYCCD

Low temperature firware/ezcap testing

QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Low temperature firware/ezcap testing
« on: January 08, 2010, 11:14:59 AM »
Hello, JP and JuhaFin

This is only a test version to confirm if it works .Please try to download http://www.qhyccd.com/vspace/EZCAP19x64.rar

And the driver http://www.qhyccd.com/vspace/QHY9DrvBETA.exe


Please use the version check tools to check after the driver installed. (Old driver should be removed at first and delete the driver in device manager)
The correct version should be 10-1-8.

Try to adjust the No.1 FineTone value to see if it helps.


Note: This is only a test software for low temperature. Not fully released version.



Best regards,
Qiu Hongyun

Re: Low temperature firware/ezcap testing
« Reply #1 on: January 08, 2010, 02:02:10 PM »
Thanks,

This might solve a sympton but it's not a cure.

If offset value is drifting and ambient teperature has a direct impact  to it, makes image calibration impossibe even current situation can be corrected by tweaking some values.
I have noticed, that the setted offset value drifts in a not extreme temperatures as well. If it really does do that, it means, no calibration library can't be build. I newer met this problem before with any other camera. Setted offset has allways been a constant, as it should be.

I think, Ivan has a good idea here, http://qhyccd.com/ccdbbs/index.php?topic=1792.msg11145#msg11145


QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Re: Low temperature firware/ezcap testing
« Reply #2 on: January 08, 2010, 08:38:02 PM »
Hello,

       Yes offset is drift with temperature of CCD sensor.But the important is that the CCD sensor is under temperature regular.So it will be no problem of the calibrations. The problem you meet in Autumn is because the temperature need more than to get stablize than the curve of temperature display. Because the temperature displayed is the point of the temp sensor. The sensor is in front of the ccd . There is no way to put the sensor
inside the really silicon chip of CCD . The really silicon chip of CCD is bigger than temp sensor so it need more time to get thermal balance.   

       Remeber the temperature control need some time to meet the enviroment changes if changing fast. I do not know if the temerapture is really get balanced when you take camera from 20C to -20C enviroment and do the messurement within some minutes.


       Calibration library is easy to do .Just wait more time afte the temperature curve reach stable.  Normally speaking, 15mintues is better.Then begin to do dark or bias image.

       Another point is : the offset is only a constant thing.Like you add a constant in whole image.

      This new software /driver is to solve the unstable and zero image in low temperature. No relationship with the CCD sensor offset changes with temperature.




      What's Ivan said is a good idear. DSLR are using it ,called Optic black calibration. This can be done by software.
     
Best regards,
« Last Edit: January 08, 2010, 08:49:21 PM by QiuHY »
Qiu Hongyun

Re: Low temperature firware/ezcap testing
« Reply #3 on: January 09, 2010, 03:38:51 AM »
Ok,

I will look for this new driver and test it today.
What worries me, is that by my experience and tests the actual offset is moving by ambient temp, even CCD-temp is regulated to fixed value.
Before I tested camera inside and out, it was permanetly installed to my observatory and behavior was same, offset value was moved by ambient temp.
I know how to take good galibration files, it takes up to 25 hours to shoot them, so I'm not going to do it/can do it for every imaging session.

Gain and Offset values has to be same all the time, otherwise calibration files will be obsolite.
They have to be same for all binning modes too, since there is no way to re adjust
them for each bin mode while doing automated imaging sessions.

Re: Low temperature firware/ezcap testing
« Reply #4 on: January 09, 2010, 07:43:23 AM »
I have done some more tests.

Set up with a new driver.
1.- First test inside +22 C, CCD-temp -30C, 30 min. wait for good stabilized temp before tests.
2.- Second test outside -18,5 C, CCD-temp -30C, 30 min. wait for good stabilized temp before tests.
- MaximDl V5.07, Windows XP, latest Tom's  ASCOM drivers

1, +22C. Without any tweaking to any value, an average background value  of BIAS-frame binned 1x1 is 650.
Bin modes 2x2 and 3x3 gives a ZERO value, bin mode 4x4 shows some noise with lots of Zero values.

2. -18,5C.  Without any tweaking to any value, an average background value  of BIAS-frame binned 1x1 is 2550.
All other bin modes are ZERO.

This alone shows that CCD-temp has nothing to do with this problem!
Ambient temperature has a direct impact to a background value measured from a Bias frame.

This camera has a random element effected by a ambient temp.
Yes, by tweaking some Gain, Offset and etc... values it can be made working, but it has to be done again if temperature changes.
After that this is no astro camera though, since it will be impossible to calibrate images.

This is clearly visible with large temperature delta but it'll be there with smaller deltas as well. This makes good constant  calibration impossible.

NOTE!
Gain and Offset for every Bin mode has to be preset and locked down by manufacture.
This is important, since when automated, no tweaking can be done. (Other manufactures don't even have a possibility to
tweak Gain and Offset, since there really should be no need for that!)


I'm really a very frustrated man at the moment!



QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Re: Low temperature firware/ezcap testing
« Reply #5 on: January 09, 2010, 08:03:23 AM »


       Which FineTone setting do you use for this?
Qiu Hongyun

Re: Low temperature firware/ezcap testing
« Reply #6 on: January 09, 2010, 08:09:03 AM »
I did not use any, since there is no point for that and I couldn't find any settings from Ezcap SW.
As I mentioned, this might be solved by settings, but when temp changes, it has to be done again.

QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Re: Low temperature firware/ezcap testing
« Reply #7 on: January 09, 2010, 08:22:55 AM »
Hello,

        Not sure . Please try it.  It is in the menu->favorite.

        The reason I say it is not sure , because I did the experiment on one of QHY9. And the result is

#1 FineTone      Before adjust,  start from -10degree, the bin4 become zero even at gain=255. bin3 begin to get noise image and become darker and darker.
                        After adjust,  the bin3 and bin4 is solid and no any fault after -20C(enviroment temperature)

Best regards,
« Last Edit: January 09, 2010, 08:25:24 AM by QiuHY »
Qiu Hongyun

QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Re: Low temperature firware/ezcap testing
« Reply #8 on: January 09, 2010, 08:28:50 AM »
I have done some more tests.

Set up with a new driver.
1.- First test inside +22 C, CCD-temp -30C, 30 min. wait for good stabilized temp before tests.
2.- Second test outside -18,5 C, CCD-temp -30C, 30 min. wait for good stabilized temp before tests.
- MaximDl V5.07, Windows XP, latest Tom's  ASCOM drivers

1, +22C. Without any tweaking to any value, an average background value  of BIAS-frame binned 1x1 is 650.
Bin modes 2x2 and 3x3 gives a ZERO value, bin mode 4x4 shows some noise with lots of Zero values.

2. -18,5C.  Without any tweaking to any value, an average background value  of BIAS-frame binned 1x1 is 2550.
All other bin modes are ZERO.

This alone shows that CCD-temp has nothing to do with this problem!
Ambient temperature has a direct impact to a background value measured from a Bias frame.

This camera has a random element effected by a ambient temp.
Yes, by tweaking some Gain, Offset and etc... values it can be made working, but it has to be done again if temperature changes.
After that this is no astro camera though, since it will be impossible to calibrate images.

This is clearly visible with large temperature delta but it'll be there with smaller deltas as well. This makes good constant  calibration impossible.

NOTE!
Gain and Offset for every Bin mode has to be preset and locked down by manufacture.
This is important, since when automated, no tweaking can be done. (Other manufactures don't even have a possibility to
tweak Gain and Offset, since there really should be no need for that!)


I'm really a very frustrated man at the moment!










"1, +22C. Without any tweaking to any value, an average background value  of BIAS-frame binned 1x1 is 650.
Bin modes 2x2 and 3x3 gives a ZERO value, bin mode 4x4 shows some noise with lots of Zero values."

Seems this is not correctly. Do you remeber I send you the 1*1,2*2,3*3,4*4 imagings before I ship the new camera to you?  This is also tested under +20C.

Best regards,
Qiu Hongyun

Re: Low temperature firware/ezcap testing
« Reply #9 on: January 09, 2010, 08:35:06 AM »
That is what happend after installing this new driver.
With a latest QHY driver it works fine in a room temp giving me images like you send.

Outside it goes nuts, giving Zero values other bin modes than 1x1 and 2x2 but very noisy. (Previous test -25 celsius)
NOTE!! CCD- temperature was well stabilzed to -30 Celsius, about 30min in all tests.
« Last Edit: January 09, 2010, 08:36:53 AM by hercule1939 »

QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Re: Low temperature firware/ezcap testing
« Reply #10 on: January 09, 2010, 08:37:23 AM »
Hello,

       This shows the FineTone is taking effection and it is just in a wrong position. So you can adjust it to let it get correctly working. To make the progress faster,I prepared a same firware version QHY9 here.  I will do the experiment parallel with you and we can discuss what's found. Please contact with my yahoo messenger: qhy_astronomy@yahoo.com   I think this problem can be solved very soon.


Best regards,
Qiu Hongyun

QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Re: Low temperature firware/ezcap testing
« Reply #11 on: January 09, 2010, 08:40:00 AM »
And JuhaFin please try this and give me the result of testing.


Best regards,
Qiu Hongyun

Re: Low temperature firware/ezcap testing
« Reply #12 on: January 09, 2010, 08:42:38 AM »
I'm not in my observatory any more. I have some family issues here too...
I'll look for this again tomorrow.

What do you think, did my test show, that ambient temperature has a effect to camera operation, even the CCD temperature was fixed to -30C all the time?

Re: Low temperature firware/ezcap testing
« Reply #13 on: January 09, 2010, 10:15:50 AM »
Hi,

Here's some test results (minimum values, gain 30 offset 90), sensor temp -40C, temperature inside observatory -18,3C.

Finetone 0:
bin 1: 4724
bin 4: 5264

Finetone16384:
bin 1: 4736
bin 4: 5346

Finetone32768:
bin 1: 4805
bin 4: 5652

Finetone49152:
bin 1: 4748
bin 4: 6093

I noticed earlier that you find problems faster by NOT using high speed with binning 3*3 in EZCAP. Its behavior is almost like maxims, so here's that test also:
Finetone0: minimum and maximum 0
Finetone16384: mimimum 0, average 2,77 maximum 3863
FT32768: minimum 0, average 6017, max 8262
FT49152: min 0, avg 6326, max 7314

Juha
« Last Edit: January 09, 2010, 10:17:59 AM by JuhaFin »
Juha

EQ-6 SynTrek, QHY-9, Atik314e, C8 SCT, APM 80/500, QHY color wheel
http://kepastro.blogspot.com/

QiuHY

  • *****
  • 5000
    • View Profile
    • Email
Re: Low temperature firware/ezcap testing
« Reply #14 on: January 09, 2010, 10:58:03 AM »
Hello,

         More test here, And the result is :

1.The noise performance will not change with the ambient temperature (If you have set the #1FineTone to 3,or 4) If in position1,noise performace will become worse in low ambient temperature and zero condition(even gain=255) may happens
2.The average background level do change with the temperature but not so large. And the delta(image area - overscan area) are not changing almostly. From +20C to -20C, the 1*1binning average background changes from 840 to 1232.  The overscan area changes from 783 to 1169.   It shows the changes is only the whole image offset. The  delta(image area - overscan area) is almost fixed.  If it is need to be correctly really, the method of optic black calibration can be used in driver or in the image process.
3.Bin2,3,4 has more bigger shift than bin1.But the delta(image area - overscan area) is fixed.
4.I tested with QHY8PRO  .The average background value still changes also.  Bin1 from 2156 to 2353.  (A little smaller than QHY9,maybe the system gain is not the same)
5.I tested all noise performance(use the STD to test it). It is almost not changed with the ambient temperature.


The conclusion is :
1.From 20C to -20C,there is big temperature, all electric component has temperature drift. For example, the reference voltage of the ADC, the offset DAC output used to adjust the offset, the OPAMP's DC bias level. So it will effect the absolute value of output. But the relativity signal (which is just the really CCD signal) delta(image area - overscan area) is fixed , and the noise performance is also almost fixed.   Not only QHY9, but also QHY8/PRO etc are all have this drift.

2.The best way to solve it is to do the optic black calibration. The optic black calibration is using the optic black pixel(the pixel mask with a metal shield with block the light) or overscan pixel to calibration the whole image. And it will give a pure CCD signal with offset free.  This way can be done in capture software , it can be done also in image process software. We will consider if need to add it into capture software.

    The image process method is :

     a.capture the dark frame, bias frame flat frame and light frame as normal (of course during ccd sensor temperature stablize) ,No matter what's the ambient temperature.  Starcking them as normal.
     b.Check the overscan area average value.
     c.Subtact the whole image with this value(In maixmdl you can use pixel math) for stacked dark, bias ,light and flat frame
     d.Process and calibration them.
   

3.There is a quesiton that if really need to do calibration and capture seperately in a cold enviroment and a warm enviroment.


Best regards,
« Last Edit: January 09, 2010, 11:39:17 AM by QiuHY »
Qiu Hongyun