QHY128C Gain Question

[GoFish]

How does the brightness of the image vary if you go between 8 and 16 bit mode without changing DG or gain?

Image brightness stays constant.

The only (peculiar) difference is that the synthesized white curve on the h-gram has a much lower peak in 16-bit mode. The location of the white peak along the horizontal axis is the same, and the R, G, and B humps are virtually identical. But the white hump is just lower in 16 bit mode, by a lot.

[GoFish]

After what seemed like an hour of sensor analysis, it entered the "measure relative gains" mode. All went well until it reached gain=4001, at which point it entered a death loop of trying 12.4, 17.0, and 19.3 ms exposures. Over and over. Then it said "sorry". Failed to find an exposure value.

I'm not in a very good place right now with this camera or with SC sensor analysis

[admin]

Hi,

sorry to hear that the analysis is giving trouble. Usually the 'loop and die' behaviour you are seeing happens when the exposure that is being set is not actually being faithfully applied by the camera. SharpCap expects that if it increases the exposure by 5% then the brightness will increase by 5% (at least once any pedestal level is taken care of). Sometimes I find cameras where the SDK rounds the exposure in some way, so that small changes have no effect, and then another small change has a much larger than expected effect (once you get past the next rounding threshold). There is at least one that rounds exposures above 1s to the nearest 0.2s if I remember correctly

You may be able to get around this by a couple of means

1) brighter light levels during the 'measuring gain range' phase. During this phase, if the exposure gets too short (too near the minimum value), SharpCap will decide to limit the gain range it is going to measure rather than try higher gains.

2) darker light levels during the measureing relative gain stage - that will increase the exposures being used and perhaps make any issues with the exposure not being 100% accurate small enough to not matter.

In both cases, you should aim for constant light levels during that stage, but it's not critical that the light levels during the stage are the same as those for the preceeding or following stages.

I am not quite sure what to make of the histogram effect you are seeing - could you post screenshots of both, as maybe that will make things clearer.

cheers,

Robin

PS. Also, make sure you have the very latest SharpCap 4.0 (or beta 4.1) when running the sensor analysis - there have been improvements to the reliability of it in the last few months.

[GoFish]

Still no luck after a few more attempts at sensor analysis. Varying the light levels did not allow me to complete the analysis successfully, but did affect where along the way the analysis failed. At some point each time it got stuck in a death loop trying to find an exposure value to satisfy the desired exposure level.

What I learned, or at least think I learned, is that the exposure levels (as indicated by the 0-65000, or 0-100% range in the histogram) change in large discrete chunks at particular threshold values of the exposure time.

For instance, at 30.4ms and below, the exposure was 20,800/31.7%. At 30.5ms,the exposure jumped to 38,900/59.3%. At all exposure times between 30.5 and 91.5ms, the exposure stayed at 38,900/59.3%. Then at 91.6ms the exposure made another large jump.

This was a repeatable sequence at a few other exposure values around 183ms, 274ms, and 205ms. There are probably many others.

I made this problem "go away" once by making the red sampling box cover the entire camera sensor at full resolution. But then I couldn't repeat this and keep the problem gone.

Of course, the algorithm in sensor analysis bombs when changes in exposure time don't result in changes in exposure level. I can also see several green crosses stack up on themselves in the graph.

Does this give you any new ideas? Thanks.

[admin]

Hi,

the only idea that springs to mind is to try switching the camera into live mode (if you are running the analysis in still mode) or switching it to still mode if you are running in live mode. If you are trying live mode, make sure the control 'force still mode' is off.

The two modes use different ways to drive the exposures, and maybe one of them has the steps that you are encountering and the other does not.

Hopefully this will help,

Robin

[GoFish]

No luck with switching it out of live mode.

My light source for these tests has been an LED tracing panel. I've just measured the flicker frequency to be 20kHz, a period of 0.05ms. My belief is that this is essentially flicker-free for the purpose of testing this camera. The shortest exposure I noticed during the sensor test was 0.1ms.

Am I wrong about that?

[admin]

Hi,

at 20Khz flicker, you may see horizontal banding in the image at short exposures, since the camera has a rolling shutter. If that is visible during the analysis procedure then the best way to avoid problems is to make the selection region tall and thin so that it incorporates several bands, which tends to average out the effects. However, I would be very surprised if this is the cause of the analysis becoming stuck.

I think the best thing to do is to send me a SharpCap log, captured after a failed analysis attempt - that should let me see what stage the procedure because stuck and if there is anything else that might help out.

cheers,

Robin

[GoFish]

Thank you, Robin. When there’s an opportunity to remove the camera from the scope I will give it another go.