Atik Apx60

[gorann]

This thread is getting more an more interesting and I look forward to ATIKs explanation of how it all works. It would be very odd if they did not provide the HCG option.

[Tom OD]

25 minutes ago, gorann said:

This thread is getting more an more interesting and I look forward to ATIKs explanation of how it all works. It would be very odd if they did not provide the HCG option.

Yes its fun taking and plotting the data and seeing what the cameras does.

Is HCG normal in all other CMOS chips? I dont know, or is it only dependant on how the software reads the chips, sets the gain etc...

Tom

[gorann]

1 hour ago, Tom OD said:

Yes its fun taking and plotting the data and seeing what the cameras does.

Is HCG normal in all other CMOS chips? I dont know, or is it only dependant on how the software reads the chips, sets the gain etc...

Tom

It is something you get with the ZWO and QHY versions of the latest sensors (IMX 410, 455, 492, 533, 571) and seen as a sharp reduction in read noise when you move the gain over a certain threshold. Something kicks in in the amplifyer. Not sure if it is built into the Sony chip or in the amplifyer used by the camera manufacturer. I use that setting all the time with my ASI2600 and 6200 to be able to get more subs out of a give time period and assume it will significantly increase my S/N I get that night. Here is what it looks like for your sensor on the ASI6200, so everyone would use the gain 100 setting. ASI and QHY give these curves for all their cameras (e.g, https://astronomy-imaging-camera.com/product/asi6200mm-pro-mono). ZWO writes "When the gain value is set to 100, the HCG high gain mode is turned on. We’re not saying it’s magic…but we’d like to think it is pretty close. With the HCG-mode turned on, the readout noise is greatly reduced, and the dynamic range is basically unchanged."

 

Edited May 31, 2022 by gorann

[Tom OD]

7 minutes ago, gorann said:

It is something you get with the ZWO and QHY versions of the latest sensors (IMX 410, 455, 492, 533, 571) and seen as a sharp reduction in read noise when you move the gain over a certain threshold. Something kicks in in the amplifyer. Not sure if it is built into the Sony chip or in the amplifyer used by the camera manufacturer. I use that setting all the time with my ASI2600 and 6200 to be able to get more subs out of a give time period and assume it will significantly increase my S/R I get that night. Here is what it looks like for your sensor on the ASI6200, so everyone would use the gain 100 setting:

 

So a few things must be happening here.

Either the Atik amp works differently to ZWO, the most likely I would think, or the software is controlling it differently.

Given the chips are all from Sony, I would think that any specific silicon P/N gate signal or what ever it the correct term, will be the same. So any cliff in voltage signal should be the same for all further cameras manufacturers.

My data is not showing a cliff anywhere in the Std Dev. (We are taking the Std Dev in ADUs as the Read Noise here in Bias frames)

Now the Atik manual is very vague about settings that "PowerSave – This mode allows the camera to shut down non-essential circuits during exposures"

Who knows what this means. Maybe Atik are reading this thread and enjoying our struggle here

Tom

[symmetal]

1 hour ago, Tom OD said:

Is HCG normal in all other CMOS chips? I dont know, or is it only dependant on how the software reads the chips, sets the gain etc...

HCG is a feature in the latest Sony sensors. It's built in to the sensor circuitry itself as one of its readout modes. There is a LCG (Low Conversion Gain) mode too for use in bright conditions but astro cameras don't enable that mode. 😀

All the ADC conversion circuitry is built into the CMOS sensor chip too. Each sensor column has it's own ADC circuitry, so there are 9576 ADC converters on your IMX455 sensor, unlike CCD sensors where there is one ADC which is on a separate board to the sensor.

So the HCG mode is there if Atik choose to enable it.

As you use SGP Tom, like me, you probably aren't yet aware that the gain and offset can be set in SGP too which can catch you out. I just set it in the Ascom driver and don't let SGP change it but I do enable the gain and offset to be inserted into the image filename as a convenient check.  If you insert _%cg_%co_ as part of the SGP filename template you'll get a filename similar to this

M31_Lum_60sec_gain_100_offset_50_1x1_-15C_2021-09-06_frame44.fit

In the SGP control panel you can set gain and offset for the whole sequence here, though I leave it 'Not Set' so it uses the Ascom driver value. You can set higher gain when binned for quicker autofocusing if you wish.

In the Sequencer itself you can click on the cogwheel next to each event here

which will bring up another panel like this where a different gain and offset can be set just for that event.

SGP checks the event setting first and if it's not set there it checks the control panel setting, and if it's not set there it uses the Ascom driver setting. Hope this helps. 🙂

Alan

[Tom OD]

4 minutes ago, symmetal said:

HCG is a feature in the latest Sony sensors. It's built in to the sensor circuitry itself as one of its readout modes. There is a LCG (Low Conversion Gain) mode too for use in bright conditions but astro cameras don't enable that mode. 😀

All the ADC conversion circuitry is built into the CMOS sensor chip too. Each sensor column has it's own ADC circuitry, so there are 9576 ADC converters on your IMX455 sensor, unlike CCD sensors where there is one ADC which is on a separate board to the sensor.

So the HCG mode is there if Atik choose to enable it.

As you use SGP Tom, like me, you probably aren't yet aware that the gain and offset can be set in SGP too which can catch you out. I just set it in the Ascom driver and don't let SGP change it but I do enable the gain and offset to be inserted into the image filename as a convenient check.  If you insert _%cg_%co_ as part of the SGP filename template you'll get a filename similar to this

M31_Lum_60sec_gain_100_offset_50_1x1_-15C_2021-09-06_frame44.fit

In the SGP control panel you can set gain and offset for the whole sequence here, though I leave it 'Not Set' so it uses the Ascom driver value. You can set higher gain when binned for quicker autofocusing if you wish.

In the Sequencer itself you can click on the cogwheel next to each event here

which will bring up another panel like this where a different gain and offset can be set just for that event.

SGP checks the event setting first and if it's not set there it checks the control panel setting, and if it's not set there it uses the Ascom driver setting. Hope this helps. 🙂

Alan

Thanks Alan. I was looking at the main gain setting in SGP, but I was not aware of the Cog wheel second option, so thanks for that.

I'll add in the meta data to the files too, its good info to have.

I'll let the Ascom driver start with the Power save and Low, and go from there for starters.

Tom

 

[symmetal]

It would be interesting Tom to see if SGP reports the actual gain setting used in the filename, when you do a test image with Medium and High Gain presets. As it stands Medium and High seem to have a far higher gain than you'd ever want to use.

Alan

[symmetal]

I forgot to say that the fits header will show the GAIN and OFFSET values of the image, hopefully with actual values for the gain, and the EGAIN entry will show the electrons/ADU value from which the gain can also be determined. If you post those values for the LOW, MEDIUM and HIGH presets it will be useful Tom. 'Astroart' or 'Fits Liberator' will display the fits header. 😀

Alan

[Tom OD]

On 01/06/2022 at 17:21, symmetal said:

I forgot to say that the fits header will show the GAIN and OFFSET values of the image, hopefully with actual values for the gain, and the EGAIN entry will show the electrons/ADU value from which the gain can also be determined. If you post those values for the LOW, MEDIUM and HIGH presets it will be useful Tom. 'Astroart' or 'Fits Liberator' will display the fits header. 😀

Alan

Good thinking and yes the values are there. So even though the manual says the Gain goes from 0-35 the Medium and High have much higher values.

 

[symmetal]

Hi Tom,

Yes, those fits gain settings don't make any sense. The Sony datasheet for the IMX455 states max 36dB of gain. Can you confirm that the gain setting on Low is reported as 0 in the fits header. Also if you post the Low, Med and High EGAIN settings from the fits headers it will give an indication of the actual gain being used.

I assume Atik haven't replied to your email yet.

Alan

[Tom OD]

22 hours ago, symmetal said:

Hi Tom,

Yes, those fits gain settings don't make any sense. The Sony datasheet for the IMX455 states max 36dB of gain. Can you confirm that the gain setting on Low is reported as 0 in the fits header. Also if you post the Low, Med and High EGAIN settings from the fits headers it will give an indication of the actual gain being used.

I assume Atik haven't replied to your email yet.

Alan

Yes the low gain value is 0 as per Astroart header data. 
I haven’t written to atik yet as I wanted to get the tests done first so that I can word my mail correctly. I ll send it this weekend.  Tom. 

[symmetal]

I did a test of the variation in Bias Standard Deviation from gain 0 to 100 on the ASI 6200 to compare with your bias Gain/SD chart you posted a while ago Tom.

Here's the result which seems as expected. Gain 60 is actually 6dB and the range of min to max ADU changes from 204 at gain 0, to 418 at gain 60. This is roughly double, which corresponds to a doubling (6dB) of amplifier gain in the camera. Likewise the SD roughly doubles from 4.56 to 8.73.

These are from single frames but if you took the average of many frames they would be closer to actually being double.

Gain 100 is where the HCG mode kicks in and the ASI Ascom driver won't allow you to select a gain higher than 100, which is equivalent to 10dB. The ASI gain is in 0.1dB steps.

I then plotted your Atik figures you posted and overlaid the ASI figures on top. I changed the ASI gain steps to dB to keep the same x axis.

This suggests than your Atik gain steps as entered manually are not actually in dB steps as the increase in SD is far to small, and seems to be closer to 0.1dB steps like the ASI. If you plot the Atik 35dB SD value as 3.5dB the Atik and ASI graphs are of a similar slope. 

Tom, I think you need to query the actual manual gain step size with Atik, as well as the huge and unusable gain settings apparantly applied by the Atik Medium and High gain presets. Along with not allowing the built-in HCG mode to be enabled.

Alan

Edit: I've replotted the above graph assuming the Atik gain steps are 0.1dB and here's the result. The Atik is showing slightly higher noise overall for some reason, but the two graphs now agree closely on the change in SD with gain.

Edited June 5, 2022 by symmetal

[symmetal]

Any update on your Apx60 Tom, and any reply from Atik?

Alan

[Tom OD]

On 14/07/2022 at 02:10, symmetal said:

Any update on your Apx60 Tom, and any reply from Atik?

Alan

Hi Alan

I just started imaging leaving the camera on power Save, and Low as the settings. Sorry I still have to mail Atik, I got very side tracked, and started on a new mosaic.

I m not sure how to write a review of the camera, but I can say that the FITs are very clean. My old 11 meg had a lot of pixels, and bad columns, so each file of the camera is a joy from that point of view.

I have been taking 2hrs worth of Lum frames, where normally I would have taken 4hrs of Lum for each Lum pane. In my new 2hr frames the data is clean (so far without flats) and the data of the dim background dust is very evident.

I have not taken Ha or RGB frames yet, I ve been just running it for the Lum frames.

In short I m very impressed with the camera. The system still has tilt, so I cant give a proper account just yet, but in a mosaic the end downsized frames will not show the tilt issues, especially after the overlaps are blended.

The chip is o sensitive that taking flats is tricky. The Ha flat at 6 secs show s agoo even light frame. The lower 0.1s flats for the other filters, are giving bright to dark gradients across the chip that are unusable.

So you need a dim flat field box I think to get flats with this chip.

Tom.

[symmetal]

Glad to hear you're happy with the results Tom. Odd that you're having trouble with the flats as I just use the light panel at full brightness with no paper sheets or ND filter, and end up with around 0.01 secs for LRGB and around 0.5 secs for narrowband. They all look good, and calibrate fine. The newer cameras don't have the short exposure issues that some earlier cameras did.

Hope you get around to contacting Atik at some time. 😉

Alan

[GalaxyGael]

Good to see Tom, and nice form factor for this size sensor. Might even be possible to use this with a RASA8, as the 2600 and 6200 ASI versions are too bulbous.

The brightness of your light source might help to get the flat exposure times down to tenths of a second. These chips IMX 571 and 455 work very well with short exposure flats and bias (darks if you wish), rather than the longer flat darks and flats of the ASI1600 variants and some other camera.

 

side note: On flats, I have found that sharpcap's flat option is useful sometimes in one regard. You can obtain a single master flat, or save of each frame if your stacking software requires it. But, you can get the flat in monochrome, with CFA bias signal included. That really helped with color balance issues in flats and the bias is corrected with the master bias. normal flats still work fine, but I found the monochrome flats to be stable in every situation.

[Tom OD]

On 22/07/2022 at 02:17, symmetal said:

Glad to hear you're happy with the results Tom. Odd that you're having trouble with the flats as I just use the light panel at full brightness with no paper sheets or ND filter, and end up with around 0.01 secs for LRGB and around 0.5 secs for narrowband. They all look good, and calibrate fine. The newer cameras don't have the short exposure issues that some earlier cameras did.

Hope you get around to contacting Atik at some time. 😉

Alan

 

On 22/07/2022 at 09:04, GalaxyGael said:

Good to see Tom, and nice form factor for this size sensor. Might even be possible to use this with a RASA8, as the 2600 and 6200 ASI versions are too bulbous.

The brightness of your light source might help to get the flat exposure times down to tenths of a second. These chips IMX 571 and 455 work very well with short exposure flats and bias (darks if you wish), rather than the longer flat darks and flats of the ASI1600 variants and some other camera.

 

side note: On flats, I have found that sharpcap's flat option is useful sometimes in one regard. You can obtain a single master flat, or save of each frame if your stacking software requires it. But, you can get the flat in monochrome, with CFA bias signal included. That really helped with color balance issues in flats and the bias is corrected with the master bias. normal flats still work fine, but I found the monochrome flats to be stable in every situation.

Olly put the light panel on the scope, and at its lowest brightest setting. When I started taking pics, it was giving white pixels on the lower part of the chip through a gradient to black pixels at the top of the chip.

After adding in the paper sheets, I was able to take exp to 1.3s, and the gradient started to disappear and become uniform. The dust bunnies then wold show.

It took a few sheets of paper to get it to the point where the LRGB were a few seconds long,which gave a good gradient across the flat. Ha flat then become about 20s With these settings, the ADU was approx 20,000 of the full well depth, which I m happy with as 55k is the stated full max.

Tom