QHY 174 GPS Calibration LED issue // USB Traffic weirdness

[ptg66]

Hello
I am myself a QHY174GPS user and I have a question:

can you explain me which is the effect of the "stabilisation code" added to this version of sharpcap? Is it expected to improve the consistency of the GPS timing...?

best regards
Paolo

[procyon12]

Hi,

Robin, I agree.

Another question:
Do you know / can you imply the currently empty data in the GPSlog (there is also no such data in fits headers)? The altitude is particularly relevant for occultation work.

Thanks. Christian

[admin]

Hi Christian

You need to upgrade the firmware on your camera in order to get the additional data (this is quite an involved procedure and involves upgrading both the FPGA and the firmware of another component of the camera – I have the cables and tools but I haven't dared do it yet!). I think qhy will be able to provide the necessary instructions for you.

Cheers, Robin

[admin]

Hello
I am myself a QHY174GPS user and I have a question:

can you explain me which is the effect of the "stabilisation code" added to this version of sharpcap? Is it expected to improve the consistency of the GPS timing...?

best regards
Paolo

Hi Paolo,

the new code adjusts the voltage of the internal oscillator in the camera, to try to keep it as close as possible to 10Mhz. This will improve the accuracy of timestamps on subsequent frames if the GPS signal is lost.

Robin

[wunni]

Hello all,

With further tests I got even stranger effects, maybe because I had also tried different QHY drivers.
So that we all start from the same place during the tests:
Is SC 3.2.6212.0 still the latest version?
Which driver version do you use to work with SC v3.2.6212.0?
For a clean new installation, is it enough to uninstall the old drivers and SC first, or should I also delete the directories and clean the registry by hand?
Then install the QHY system driver (which one?) without a camera connected, then install SharpCap v3.2.6212.0 (?) and nothing else (no QHY patch...).
In SharpCap, switch off "Restore camera settings automatically".
Is anything wrong or missing?

Thanks,
Niko

[procyon12]

Hi

from my experience only:

SC latest (so far): 3.2.6212.0 17/01/20 - Pro version not required but recommended (e.g. Frame buffer)
Driver: https://www.qhyccd.com/file/repository/ ... dition.zip
No patches!
No additional SW required.

SC and driver overinstall should / can work. If you find problems / want to be completely safe, you can uninstall and clean something.
See also above for further hints.

Christian

[admin]

Hi Robin,
thanks for your work.
In the meanwhile I have a more precise diagramm. I (there are no other results known for me) found - and only with my system and within the reported ranges:
The ExpStartPos depends only on the USB traffic (linear).
The ExpEndPos is depending on the exp. time (linear).
I tested only for 16bit, with 8bit there are other results. It seems that gain, offset, area, binning have no influence.

It looks like I may need to extend the calibration range to allow for longer exposure calibration in a future version.

I agree. But there is a question for me, because QHY wrote that the Cal EndPos will not increase with the exp. time (?):
https://www.qhyccd.com/index.php?m=cont ... =30&id=190

I need to make the adjustment easier perhaps by having three sliders find/medium/coarse rather than the current two.

Maybe. The calibration in the current state is bit tricky..., 6 sliders could be too much, however? Moving with the keyboard arrow keys might help.

Another piece of information that would be useful if you have it to hand is how sensitive the recorded exposure time is to the exact value of the calibration parameters. By this I mean that if we want the exposure time to be correct within 1 µs, how close does our calibration value need to be? +/-10 ? +/-50? +/- 500? If for instance I knew that we didn't need to be closer than +/-100 then it would help me work out the best way to set up the sliders to give good control and accuracy.

For me, there are some open points.
How or where can I see the result of the succesful calibration (apart from the LED)?
I observed that in the case of an exact calibration the exp. time shown in the GPS status window is not exact in the range of about +/- 100 µs. To get it exact I must move the calibration slider(s) - but this is destroying the calibration found.

Example in calibrated state (ini):
Enable Live Broadcast=Off
USB Traffic=3
Offset=0
Amp Noise Reduction=Off
Frame Rate Limit=Maximum
Gain=480
Exposure=100
Calibration End Pos (Fine)=9361
Calibration End Pos (Coarse)=7500000
Calibration Start Pos (Fine)=14140
Calibration Start Pos (Coarse)=0
GPS Calibration LED=Off
GPS Freq Stabilization=On
GPS=On
Timestamp Frames=On
Contrast=0
Brightness=0
Gamma=1
Temperature=27.9
Target Temperature=0
Cooler Power=255(Auto)
Banding Threshold=35
Banding Suppression=0
Apply Flat=None
Subtract Dark=None
#Black Point
Display Black Point=0
#MidTone Point
Display MidTone Point=0.5
#White Point
Display White Point=1

The chosen exp. time is 100 ms, in the above calibrated state the GPS status window shows an exp. time 99.936 ms. The counter was precise to about +/- 6 µs, with frequ. stab. on. If I decrease the the CalStartPos from 14140 to 9361 the exp. time in the GPS status window is now 99.999 ms but the calibration is lost (?).
In both cases the measured exp. time of a recorded SER was 100.12 ms / 9.987 fps, no dropped frames.
This could give you also a feeling about the accuracy.

Christian

Hi Christian,

I tried some experiments with my own qhy 174 to see if I get the same results as you for the calibration parameters, and sadly it seems that I get different results. My camera has not yet been updated to give correct GPS info in 16 bit mode, but the calibration LED does work in that mode yet I get different results to you.

For instance, at 20 ms and USB traffic 50, 640 x 480 in 16 bit, I get a calibration start position of 1,047,460 and an end pos of 2,529,670.

If I repeat the same settings except for eight bit mode, I get a start position of about 867,000 and an end position of about 2,354,000.

Staying in 8 bit mode and leaving the exposure and USB traffic the same but changing to full resolution, I get a start position of just over 4 million and an end position of 5,550,000 or so.

So, it looks like the dependency of these values on the camera parameters is more complicated than we first thought. If we can start getting the same results from different cameras for the same settings then at least we can perhaps eliminate the possibility that the whole calibration does need to be camera specific and then proceed in understanding the full dependency pattern.

Cheers, Robin

[procyon12]

Hi Robin,

very interesting, thank you.

Because it's late, I could make some quick tests only. I can confirm _all_ St/EndPos regarding to your 3 set-ups (showing that your missing update obviously doesn't matter - and that our cameras work at least similarly), however, with traffic 50 I get very bad fps (e.g. for 20ms, 640x480, 16bit: ~29fps only).

My tests were all made with usb-traffic values < 10, see my picture d8.png above. Can you make some tests with settings like in d8.png? Maybe that for higher usb-traffic values my above conclusions are wrong (?).

Tomorrow I'll have a deeper look into.

Christian

[procyon12]

UPDATE

I checked (for 20ms, 640x480, 16bit, 1x1) the dependencies of CalSt/EndPos on the USB-traffic. Depending on my system (see above, pic d8.png), for USB-traffic > about 25, the fps reached, shown at the bottom of SC, decreases significantly. And then we see an increase of both CSP and CEP. For USB-traffic values < about 25, my conclusions from above (see pic d8.png) can be valid, however, there are other exp. times to be involved.
All this is preliminary.

Christian

[admin]

Hi,

Yes, I can confirm that I get reading similar to yours for the same USB traffic and exposure settings. It looks like the start position has two separate linear parts of the graph that join at a dogleg. Some quick experiments indicate that the position of the dogleg varies – lower exposure times move it to lower USB traffic values, higher exposure times seem to move it to higher USB traffic values.

If we can extract the gradients of the two sections and a way to calculate the position of the dogleg then we should be able to predict the starting position value.

The ending position looks more like a constant up until a certain USB traffic and then linear beyond that – once again we need to work out the value of the constant (which may depend on exposure), the dogleg position and the gradient of the linear portion and then we may have a formula for the end position.

This is looking rather promising



Cheers, Robin