Affiliation
American Association of Variable Star Observers (AAVSO)
Fri, 03/19/2021 - 17:20

Have emailed QHY several times with questions and not gotten a response

 

anyone using the QHY268M for photometry?  what mode are you using..?

what gain and offset settings..

getting what looks like blooming....on some images...starlight from one bright star streaks to another star...etc..

any comments?

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M

Hi Mike,

I have a QHY268M in the lab for testing.  It will go on the sky this weekend, so I should have some comparison values for you soon.

Arne

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M

great news!!!! hope you can keep us posted as to your tests...as to when to use what mode and settings of gain and offset...hopefully can understand the "overscan area"...

not sure I posted it here but I had different pixel counts...on calibration frames...think it was flats...so none of software would process them...

 

Mike

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M

Arne

Have you figured out why different types of images(flats, darks light frames) have different numbers of pixels.? something to do with the overscan area?..

As I noted below, My 120sec light frames..and my 0.01sec flats are 6252 x 4176..while 1 min and 2 min darks are 6280 x 4210?  I am using APT..and I get option to check a box for overscan...but it will not let me check it?...possibly that is only available using certain modes?

Cheers,

Mike

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M experience

Hi, Michael:
Yes, I have a QHY268M that I'm just starting to use for LPV photometry. (I'm still doing integration and testing, so I haven't yet settled into a repeatable rhythm.) What I've learned so far:

  • I've been using two "profiles": one for faint work and one for bright stars (such as an LPV near its peak through an Ic filter). For faint work I'm using mode 1, gain 56 (offset 5). This is the low-noise point on the gain curve that keeps a good dynamic range. For bright stuff I'm using mode 3 (correlated double sampling), gain 0, offset 5. This doesn't provide any more dynamic range than the 1/56 setting, but keeps exposure times more reasonable.
  • I haven't fiddled with USB Traffic setting, but intend to. I'm seeing some slight horizontal banding that's supposedly a symptom of a non-optimum choice for this setting.
  • I generated a set of color transformations (for B, V, Rc, and Ic) the other day using about 160 stars from M67. The transformations look reasonable, but there's a little more scatter than I'm used to. That could be because of all the dust I got onto my filters during installation into the QHY filter wheel. My flats look embarrassing. The filters now need a good cleaning.
  • I'm using Astrodon 1.25" filters with the QHY filter wheel, and have run into some annoying hassles with that. I ended up removing the filters from the Astrodon filter cells and putting them into new cells that I fabricated on my 3D printer. That's working much better, but a little fine-tuning is still needed. These filters don't cover the entire sensor; I'm cropping about 600 columns from the left and right sides.
  • With careful attention to USB cables, computers, and the home ethernet, I'm seeing much faster download/transfer times from the camera to my in-house computer than I ever could achieve using my old SBIG ST-9 camera.
  • Linearity checks that I've done on the QHY268M show it is far more linear throughout the entire transfer curve than my ST-9 was. I'm hoping that translates into better photometry (eventually).
  • The dramatic change in pixel size from 20 micron (ST-9) to 3.8 microns (QHY268M) has led me to choose to bin 4x4 with the QHY268M for the time being (I'm at 2,500mm EFL). I'm finding I have to be really careful in my software to ensure that any saturated unbinned pixels don't accidentally get "hidden" during the binning process.

So far, I having seen anything that I would call "blooming".

- Mark M (MMU)

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M

Mark

My flats also look terrible. andstill getting too much light with twilight flats(tee shirt) at 0.01 second exposures.!

have you determined a good exposure duration for flats?...my sensor has lots of dust bunnies...hundreds!! am going to try to blow so air directly on sensor..and re-do them 

my target was 30,000 or 60% of full well depth in mode 1 (51,000)

as I referred to earlier...am getting different pixel counts on some calibration frames...using Nebulosity....got 6252x4176 bias....2 minute Darks 6280 x 4210 pixels!!

flats  6252 x 4176..pixels..

my software would not process them for this reason...

 

do you have any info on exposure times?...I tried 120sec on an exoplanet mag V 12.2...and when I went to process in software...got error that target was near saturation.

I am imaging at 2000 FL...10inch Meade ACF f8..

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M Experience

Mike:

Almost all of my software is written by me. (I use a few pieces of IRAF: APPHOT and DAOFIND.) All the camera-control code goes through the QHY SDK, which gives me pretty close to complete control over what goes in and out of the camera. If I ask the QHY SDK for a full frame, I get back a 6270x4210 image. Every time. (Camera doesn't know the difference between a dark, flat, and science image.) That said, I'm using subframes so frequently that I made my calibration software "subframe-aware" so that I can properly calibrate subframe science images from full-frame darks and flats (so that frames don't have to have identical height/width to be calibrated, as long as the science image is a clean subset of the dark/flat/bias references).

The short exposure times are kind of intriguing. I haven't tried it yet, but I'm wondering whether there's anything to be gained by doing daytime flats using millisecond-range exposure times while pointing at blue sky, instead of conventional flats at night using my lightbox.

- Mark, AJ1B

 

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M

good for you...but I am not a programmer...and guess I will just have to figure out how to use EZCAP...or sharpcap...ezcap is very confusing asks for same settings multiple times...I need something simple like Nebulosity..

may have bitten off more than I can chew...

do you work any 20 meters?  may run into you sometime.....have upgraded my novice call to vanity K5SMA...

 

73,

 

Mike

Affiliation
American Association of Variable Star Observers (AAVSO)
QHY268M for exoplanet transit photometry?

Does anyone have more experience with using the QHY268M for photometry, especially for exoplanet transits?  Does binning 2x2 or 3x3 cause any issues?  Preferred settings?  I would like to find a good substitute for a SBIG ST-8XME.

thanks!

Paul

Affiliation
American Association of Variable Star Observers (AAVSO)
Re: QHY268M for exoplanet transit photometry?

Paul:
I'll offer an update on my experience so far...

I've tried doing a direct, head-to-head comparison of my ST-9 photometry against QHY268M photometry. My residual errors (accuracy, not precision) are a little bit worse with the QHY268M, but that really doesn't bother me at all at this point. I spent 15+ years optimizing the system for the ST-9, and only a few weeks working on QHY268M accuracy. One thing that makes the comparison a little unfair is that I gained significantly more FOV with the QHY268M than I had with the ST-9, and images show far more off-axis coma than I ever saw at the edges of the ST-9 camera. Since this causes the PSF to vary across the image, I'm sure that this is one of the sources of error with the QHY268M that is relatively easy to correct. The larger FOV, combined with the availability of much shorter exposure times, has changed my approach to making flats. I have no doubt that this is also something that needs more optimization. The improvement in both QE and linearity has been wonderful! In particular, residual errors are noticeably more random when plotted against SNR compared to that same plot for the ST-9 (where detector nonlinearity caused systemic photometric error as a function of star brightness).

I haven't yet done any time series with the QHY268M, but see no reason why my experience with an EB or exoplanet transit will be any different than I'm seeing with my LPV and Landolt fields.

I continue to bin 4x4 exclusively for photometry, and it works really well. (You do have to take care with understanding/managing saturation and onset of nonlinearity when binning, since non-linearity happens on a per-camera-pixel basis, not on a binned-pixel basis.) All my Landolt field measurements have been binned 4x4. Raw images from the sensor (unbinned) are in 16-bit FITS format; I shift to 32-bit FITS (compressed) as soon as the binning is done so that I don't lose any dynamic range.

Early on I had been using a mix of two mode/gain combinations, depending on star brightness. I've now simplified that, using gain=56, mode=1 exclusively. I found no benefit to using mode 3, even for bright stars.

On the downside, I'm no longer using the QHY268M. What started off as an annoying thermal quirk that would disable the camera electronics under combinations of high cooler power and high ambient temperature has worsened to the point that the camera is simply unusable on a warm June evening. I've started the process of pursuing repair/replacement through QHY.

- Mark M