QHY128C Gain Question

[3ricJohanson]

Sadly, it didn't work. I tried a few times with different settings on my flat master - no love.

Logs and screenshots here:

https://www.dropbox.com/sh/r3sp4gqsnzxq ... KNqCa?dl=0

[admin]

Hi,

thanks for the screenshot – that made the problem exceptionally clear

I was expecting the camera to be called "QHY128C", but it's actually called "QHY128PROC", so these special handling for the colour gain controls was not being applied to your camera.

I'll put a fix into SharpCap 3.3 for the next update.

Cheers, Robin

[3ricJohanson]

They are actually two slightly different cameras -- we have one of each (but the non-pro is in the shop right now). Is there some way I can get this run sometime soon by overriding something? I'm headed out shortly to utah to shoot the sky and I'd love to be able to profile my sensors.

A deeper question is what's the difference: I'm not sure there is any difference between the Pro and non-pro, other than the label.

[dghent]

They are actually two slightly different cameras -- we have one of each (but the non-pro is in the shop right now). Is there some way I can get this run sometime soon by overriding something? I'm headed out shortly to utah to shoot the sky and I'd love to be able to profile my sensors.

A deeper question is what's the difference: I'm not sure there is any difference between the Pro and non-pro, other than the label.

The difference between the original and Pro model is that the 128-Pro has improved noise reduction and better cooling. QHY will occasionally introduce a Pro variant of an existing model which denotes that it has some upgrades and improvements that are mainly evolutionary in nature. For example, the QHY367 Pro introduced an improved sensor pcboard design and TEC-sensor mechanical interface which enables better cooling over the original design. QHY typically makes these upgrades available for existing cameras, but you'd have to talk to them about what that entails for a given model.

I'm the "owner" of the native QHY driver in NINA and caught the conversation you were having with another contributor last night regarding the strange gain issue, where you referenced this thread. I talked to QHY and the per-channel gain on this camera (and perhaps the others like it) is a peculiar to stream mode only. In single exposure mode, gain should still act and function like normal/typical gain, but there might be an issue with this specific model in the SDK that prevents it from functioning properly and not changing the exposure, as you found. QHY will look into it and get back to me on it.

Robin, shall I share whatever info I get with you in case you need to make any adjustments in SC?

[admin]

Hi,

thanks for the information about the colour gain channels – please please share with me any new information that you get! I love QHY because they make some really interesting cameras that are a bit different from the rest of the crowd, but they sure do make supporting them in software an interesting experience!

@3ricJohanson - based on the information above, in theory if you put the camera into force still mode then the current sensor analysis should (correctly, although by accident) control the gain!

Cheers, Robin

[GoFish]

Hi Bruce,

Does your camera have red/green/blue white balance controls available to adjust? If it does do they have the range of 0 to 4027 suggested by qhy? I have a horrible feeling that with this camera they have put control of the gain behind the white balance controls and then added (for some strange reason) a digital gain behind the normal analogue gain control! I must admit I wasn't expecting that!!!

Cheers, Robin

Hi Robin

Yes- to both of your questions. When I open the camera, the gain defaults to 10 and the R/G/B white balance is at midpoint. If I slide them to the far right they reach 4027. Changing the gain control to any other value does not effect the R/G/B max value of 4027. If I understand what they are saying, I need to keep the digital gain at 1 and then slide each R/G/B white balance slider to a specific setting???

We are similarly struggling to understand how to interpret the digital gain and the RGB gains in SharpCap for the QHY128ProC camera.

Just for the sake of having a concrete example, let's say my goal is to use the camera at "unity gain," which according to the specs from QHY, is at gain = 3300.

Since the default digital gain seems to be 10 in the driver, does that mean that setting RGB gain = 3300 and digital gain = 10 will give me unity gain? Or do I use digital gain = 1 with RGB gain = 3300?

Thanks in advance for any help.

Jim

[admin]

Hi Jim,

it's hard to be sure, since I don't have that particular camera to test with. My approach to try to work it out would be as follows

* Set the R/G/B gains to minimum and the camera into 8 bit mode
* Show the histogram, arrange exposure to give a fairly broad spread of brightnesses, so a broad peak in the histogram
* Adjust the digital gain, and try to find a value where there are no odd gaps or spikes in the histogram graph - this may be the default 10, or may be 1.

This should give you the 'neutral' digital gain value - The gaps and spikes come from the fact that digital gain just multiplies the pixel values by some constant (like 1.2 say) - If you multiply whole numbers by 1.2 and round then you will never get a value of 99, since 82*1.2 rounds to 98 and 83*1.2 rounds to 100 - that gives the gap in the histogram.

Once you have your neutral digital gain sorted, then you can check out the e/ADU vs gain graph on this page - https://www.qhyccd.com/astronomical-cam ... 28-qhy367/ . From that you can see that the minimum gain e/ADU for the camera is 4.5, so at unity gain the image should be about 4.5 times brighter than minimum gain. If gain 3300 gives that then you are sorted (ie gain 3300, exposure=100ms is the same brightness as gain=0, exposure=450ms).

Hope this helps - do report back your findings if you can.

Robin

[GoFish]

Thanks, Robin. We’ll give this a try.

[GoFish]

My results.

First, I found the histogram gap/no gap boundary to be digital gain (DG) = 16 and 64 for raw16 and raw8 modes, respectively.

The histogram was not smooth in raw8 at many values of digital gain, but the full height gaps first appeared at DG = 64. In raw16 mode, the histogram was smoother, and it was easy to see that the gaps appeared at DG values above [edit] 16.

I conducted tests in both raw8 and raw16 modes. First raw8.

At R=G=B=1 (minimum gain) and DG = 64, exposure of 3ms gave a midrange histogram. At R=G=B=3200, the equivalent midrange histogram needed around 0.6ms. This is reasonably close to the expected multiplier of 4.5, and within error bands for the coarse test I used to judge equivalent exposure.

Next raw16.

At R=G=B=1 and DG = 16, exposure of 16ms gave a midrange histogram. At R=G=B=3200, the equivalent midrange histogram was reached at 3.4ms. Again, very similar to the 4.5 multiplier expected.

I don't completely understand all of this, but should I take away that the neutral value of DG is 16 in raw16, and is 64 in raw8? And if so, should we just adopt those values of DG for all imaging?

Finally, is there a way to determine the well contents (e-) of a pixel in SharpCap?

[admin]

Hi,

that all sounds reasonable, the only thing I am not sure about is the DG=64 for 8 bit mode. I wonder if the A to D for 8 bit mode is actually 10 bit, so at DG=16 you would be taking the high 8 bits of 10 and at DG=64 you would be taking the low 8 bits. How does the brightness of the image vary if you go between 8 and 16 bit mode without changing DG or gain?

On the e- count, yes, you can do this. First you need to run the Sensor Analysis procedure from the tools menu that works out things like e/ADU and read noise for various gain levels. Once you have done that, use the histogram with a small selected ROI. Hover the mouse over the histogram peak and it will show the pixel value, e- count, etc.

cheers,

Robin