Atik Cameras

Author Topic: Horizontal Banding - Color Horizon - Still a problem!  (Read 21294 times)

CraigG

  • Full Member
  • ***
  • Posts: 189
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #15 on: January 08, 2019, 08:48:24 pm »
bwa,

Thanks so much for your comments! I'm still holding out hope that Atik can implement a software fix for this banding issue however your experience with other 1600 cameras doesn't bode well at this point. Before your positive comment about the ASI1600MM Pro I wondered if one of the reasons that ZWO discontinued their 1600MC Pro was in part due to banding. Anyway, as I commented to David, I'm communicating with Atik on the issue and my fingers are still crossed that a EAA friendly fix can happen! I've found Atik to be VERY responsive to my questions so far. Maybe our shared concerns will prompt a successful outcome!

Best wishes,
Craig

atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #16 on: July 29, 2020, 06:40:11 pm »
I'm and electrical engineer, and I helped a friend by repairing the power connector on his Atik Horizon.  The power cable was pulled sideways during a mount slew and one of the connector pins was sheared off at the pc board interface.  Replacing the connector solved that problem.

To the thrust of this post.

When I had the Atik Horizon disassembled, I took pictures.  One bad thing I noticed is the heavy power leads to the Peltier cooler run right above the flat cable connecting to the sensor.  This situation is a perfect setup for electromagnetic coupling from the Peltier drive signal into the video signals.  Please see the attached pc board photograph.

In all likelihood, the high-current, low-voltage Peltier drive is obtained via pulse-width modulation of the 12V supply.  Note the power transistor and inductor near the pc board termination of the Peltier drive wires that could form such a low-loss high-current low-voltage down converter.

Pulse-width modulation involves high frequency square-wave switching between the 12V supply and ground.  Square waves are rich in harmonics, which can wreak havoc on other circuits if not properly managed, shielded, and/or filtered.

At a very least, the Peltier drive wires should be twisted.  Ferrite beads placed on these wires will also probably help, as will filtering on the prime 12 volt power.

I no longer have access to the Atik Horizon, as we now live an 8-hour drive apart, so I can't test this hypothesis or do spectrum analysis using a small "sniffer" pickup coil held above the operating pc board to search for problematic electromagnetic signals.

atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #17 on: July 29, 2020, 06:49:23 pm »
I'm and electrical engineer, and I helped a friend by repairing the power connector on his Atik Horizon.  The power cable was pulled sideways during a mount slew and one of the connector pins was sheared off at the pc board interface.  Replacing the connector solved that problem.

To the thrust of this post.

When I had the Atik Horizon disassembled, I took pictures.  One bad thing I noticed is the heavy power leads to the Peltier cooler run right above the flat cable connecting to the sensor.  This situation is a perfect setup for electromagnetic coupling from the Peltier drive signal into the video signals.  Please see the attached pc board photograph.

In all likelihood, the high-current, low-voltage Peltier drive is obtained via pulse-width modulation of the 12V supply.  Note the power transistor and inductor near the pc board termination of the Peltier drive wires that could form such a low-loss high-current low-voltage down converter.

Pulse-width modulation involves high frequency square-wave switching between the 12V supply and ground.  Square waves are rich in harmonics, which can wreak havoc on other circuits if not properly managed, shielded, and/or filtered.

At a very least, the Peltier drive wires should be twisted.  Ferrite beads placed on these wires will also probably help, as will filtering on the prime 12 volt power.

I no longer have access to the Atik Horizon, as we now live an 8-hour drive apart, so I can't test this hypothesis or do spectrum analysis using a small "sniffer" pickup coil held above the operating pc board to search for problematic electromagnetic signals

vince

  • Atik Staff
  • Sr. Member
  • *****
  • Posts: 283
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #18 on: July 30, 2020, 04:07:09 pm »
Hello Bruce,

Thanks for the feedback which is of course always welcome :)  I believe Andrew has got back to Richard about this.

Vince



Hello Richard,

Thank you for your email, Vince did mention that he had recently read Bruce's original forum post.

The signal coming out of the Horizon's CMOS sensor is at a frequency that is of orders of magnitude greater than that used to control the peltier module. Therefore, its affect on the quality of the final image is relatively small. However, because we are always trying, where possible, to reduce all sources of noise, we have tackled this issue in a novel way.

During the period when the image data is passed from the sensor to the main PCB, the dynamic cooling control is momentarily suspended so that what was a small issue is now completely removed.

One of the main sources of noise in such complex and compact designs, is crosstalk between the very high speed clocks operating in different parts of the circuit. One of the principal concerns of a camera designer, is to prevent this sort of problem becoming a significant issue. In fact, it is almost an obsession, and requires a great deal of experimentation. Sometimes, a seemingly logical solution actually generates unexpected or undesirable results.

The original Horizon camera was our first foray into designing a CMOS based camera, as opposed to one utilising a CCD. We believe that we did a very good job, Inevitably, since that time, we have gained more experience and have developed new ideas, some of which have been introduce into the Horizon II camera. This is all part of the gradual evolution of  hi-tech products.

I hope that you find my explanation useful

Best Regards,

Andrew

Atik / QSI / Opus Service
Unit 8 Lodge Farm Barns
New Road
Bawburgh
Norwich
NR9 3LZ
UK

susan-parker

  • Jr. Member
  • **
  • Posts: 65
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #19 on: August 01, 2020, 02:23:49 pm »
The problem here is that the ADCs are in the CMOS sensor, not separate like with a CCD, so any high-frequency punch-through will be from the Peltier to the sensor.

Additionally, Peltiers are non-linear devices and really don't like to be switched with a PWM signal - although obviously it does mostly work if it's kept within a tight temperature feedback loop.

There is an inductor (marked 2R2) which is next to the MOSFET transistor however I don't think it is part of the Peltier supply (despite being a lot chunkier than the other regulators) as there is a test-point marked 1V.

If it isn't already implemented in the Horizon (I can't see what might be on the other side of the board) my recommendation would be to introduce an LC filter the same way as one does for Class-D audio amplifiers, to smooth out the switching waveform to something more resembling DC. It doesn't have to be super smooth, just not switching the full 12V.

atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #20 on: August 01, 2020, 04:37:35 pm »
Susan,

Thanks for the response.  I didn't trace out the circuitry, but obviously could.  Here is a picture of the PC board backside; resolution reduced for manageable file size.  The heat sink pins obscure part of the PC board. 

I've designed Peltier cooler controls, and did use switching power supplies.  There are simple tricks to separate the high frequency feedback from the regulated output feedback, so the Peltier device only sees DC.

This system has a very wide range of signal voltages, from the PWM power supplies, to the low level signals compromising the 12- bit precision video signals.  The banding occurs in the darker areas of the image.

Regards


atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #21 on: August 01, 2020, 04:57:58 pm »
Andrew,

Thank you for a useful and informative explanation.  Yes unexpected experimental results can be startling, but also informative.  I wish you luck in your quest.

As I said, I am an electrical engineer who worked for a General Motors advanced research facility for 37 years.  In the last 15 years of my tenure, I was sent all over the corporation to solve electronic problems, many of which involved frequencies and interference like those under question.

Sometimes, problem solutions are very obscure.  A case in point was a piece of military electronics on a light armored vehicle (LAV) that malfunctioned sporadically and randomly when the 25mm cannon was fired.  I studied that problem for many months, attending firing exercises all over North America. 

I finally found the problem was direct sunshine illuminating a laser sensor created a photodiode bias voltage, which was blocked by a coupling capacitor, so the bias voltage was blocked into further electronics.  It turned out that shock from the cannon firing compressed the glass-epoxy pc board enough to change its capacitance and create a voltage: I = d/dt(CV).  This displacement current created a voltage in the associated circuitry.  Changing and thickening the pc board material solved the problem.

Yes, I would expect the Peltier module control switching frequency is significantly lower than the CMOS sensor scan frequency.  With pulse width modulation (square wave) used to control the switching regulator for the Peltier cooler drive wires, high frequency harmonics are present, albeit somewhat filtered.   As you know, inductors have capacitance, and they only function as inductors below their self resonant frequency.  Short high frequency current paths are also very important.

I would agree that the Peltier PWM switching frequency is probably many orders-of-magnitude below the video clock frequencies.  You also have the opposite orders-of-magnitude difference between the switching regulator waveforms and the video output signals (before on-chip A/D conversion).

Since the banding is somewhat suppressed, i.e, only affects lower level signals, even more attenuation of interfering signals to the video signals may be needed.

When I see a potentially problematic electromechanical design, like the Peltier drive wires in a big open loop directly above low level video signals, I get concerned.

There are simple tests that can be done to prove or disprove this conjecture.  The simplest of which is to raise the Peltier drive wires away from the video flat cable and press them together to minimize coupling area.  Compare the banding in before and after photos.

Subsequently, add a shield over the video flat cable area.  Such a shield can be a small piece of single-sided 0.31" PC board material. The ground connection should be as short as possible.   Retest, as before.

If there is an improvement, then the above conjecture is valid.

You have a tiger by the tail, judging from the blogs complaining about the problem, i.e., disappointed customers who  bought an expensive camera.

I hope you find a solution soon.

Regards,

susan-parker

  • Jr. Member
  • **
  • Posts: 65
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #22 on: August 01, 2020, 09:35:44 pm »
Hi,

Just to repeat, the CMOS sensors have the ADCs on the chip, so there are no analog signals coming out of the sensor, so the banding cannot be coming from this.

However there are analog signals going into the sensor, in the form of the various power supplies. Modulation on the power might cause an artefact in the chip, particularly something that affects the ADCs reference levels.

Or the banding may be simply differences in the ADCs and gains of the sensor as there is stuff going on in the chip doing analog twiddling things.

I know that overloading a full well causes banding from brightest stars, which is obvious full width streaks.

As an aside I would note that I am imaging with strong background light pollution (Bortle 8+), which apparently is masking this effect to some degree. I usually ruin at high-gain in Dusk and my images come out of stacking with high background levels.

Taking a look at the Panasonic sensor specifications there is a lot going on in the chip, unlike a CCD sensor:

https://industrial.panasonic.com/ww/eol/semiconductors/imagesensors/broadcasting-digital-still-camera/models/MN34230

... and the character of the CMOS sensors will be similar to other parts.

To get the best out of the Horizon one should really do white-light calibration, which should sort out some of these issues. I have a flat-light box, need to get round to using it!

Best,
Susan.

atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #23 on: August 02, 2020, 10:02:26 am »
Susan,

Twice now you have related that the CMOS sensor has the ADC on chip.  If interference is being injected, it really matters in what way it evidences.  Yes, noise on digital lines, if low, can be ignored.  Noise on other lines may not be able to be ignored.  One must look at all possibililties.

Atik recognized the noise in the Peltier switching power supply was causing problems, and the solution was to gate off the PWM switching during some CMOS sensor operations.  What about transient die downs, and switching noise from other power supplies?

The Peltier cooler is directly in back of the sensor.  The banding problem is a low-level phenomenon, i.e., affecting dark image areas.  I think EMI interference is related to the problem. 

The Panasonic MN34230 16 MP CMOS sensor requires three supply voltages: 3.3, 1.8, and 1.2 volts.  This product is being discontinued, so I can only find a cursory data sheet.

How are these voltages being generated?  What is the effect of ripple on them (synchronous, non synchronous).  As I said in a previous post, the devil is in the details.

Atik has now introduced the Horizon II.  Does it have the banding problem?  Does it use a newer sensor?

Atik has disappointed customers who bought an expensive camera.  Don't ignore them.

susan-parker

  • Jr. Member
  • **
  • Posts: 65
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #24 on: August 02, 2020, 03:31:27 pm »
Hi,

I am not ignoring the Atik customers, I am one myself with an original Horizon so I do have skin in the game.

I have been doing electronics for over 50 years, including analogue and digital design profesionally for instrumentation as well as a few switching power supplies - mainly for HV generation for flashlamps and that sort of thing. I am more of a generalist than a hyper-focused specialist so I am perfectly aware that there are lots of things I don't know in depth or have direct experience of. I am also aware I have forgotten a lot of stuff too... what was that you were saying?

I am making the point that in CMOS that the ADC and the analogue circuitry is all on chip, so outside of Atik's control, compared to a CCD where that is all up to the designer. I am currently working on a 1Khz rep rate CCD line array spectrometer, so I am aware of the differences.

Yes, one can clearly see three switching supplies in SOT-23 packages, as well as the bigger one, in the middle of the board which power the digital sensor controller circuitry. There are also five linear regulators which I assume power the sensor. And yes, noise induced into the linear supply rails/ground could impact the ADCs and analog side as I previousl mentioned.

How is the dark-level banding showing up?

1. Does it move around, randomly on a frame by frame basis?
2. Does it move, but aligned with brighter stars or light sources?
3. Is it static and always on the same lines irrespective of the image?
4. Or... ???

I need to do some testing, but unfortunately don't have the time this weekend to do so as I have some work deadlines to deal with.

Best,
Susan.

atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #25 on: August 02, 2020, 05:02:57 pm »
Susan,

Thank you for a cogent reply.  I await your further investigation and comments.

I don't own an Atik Horizon, so I can't do any electronic testing.  As I said in my initial post, I repaired a friend's Horizon which had a broken power connector.  My friend has moved away, and is now an 8-hour drive away.

Atik was less than friendly to him regarding troubleshooting and repair.  I think the power connector, and its mounting are inferior, especially if the connector tabs can be ruptured so easily. 

I have had problems with pin-and-sleeve connectors on Orion SkyView Pro mounts.  Not outright breakage, but intermittent connections that cause mount reset when the power cable gets moved.  I also had a problem with bad power connector solder joints on one of the SkyView pro mounts.

On my Orion Atlas Pro mount, the power connector is an all metal threaded-together connector, and is not disrupted when the supply cable gets pulled.

Now, all power and data connections to the mounts are through coiled cords, which are easily made by winding the cable around a 1/4" wood dowel, then heating it with hot air gun until the cable insulation just starts to get shiny, then letting it cool.  I did this modification on my friends Horizon power cable.  See photo.  The transition from coiled to straight cable is fixed to the mount leg with Velcro.

I have 61 years experience as an electrical engineer, and now have an engineering consulting business.  Like you, I have seen it all, and have experienced difficult to solve problems.

Again, I await your engineering investigation results.  I hope this problem has an easy solution.


susan-parker

  • Jr. Member
  • **
  • Posts: 65
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #26 on: August 02, 2020, 09:14:12 pm »
Thanks, when I get the time I will give an update.

Meanwhile I would note the same sensor is used in the ZWO ASI 1600MM/MC series:

  https://astronomy-imaging-camera.com/product/asi1600mm-cool

So probably worth looking at their forum chats to see if there has been a similar issue reported.

atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #27 on: August 03, 2020, 11:52:40 am »
Susan,

Thank you for the information that the ZWO ASI 1600MM/MC camera uses the same Panasonic sensor.

Oh ZWO's forum, there also are comments about horizontal banding.

Sam, who lists his title as 'ZWO Founder' blogs 'We saw this problem sometimes and we suspect this is because of magnetic field effect caused by power supply cable'.  Moreover other bloggers relate the banding is worse when the Peltier cooler is cooling down (high current drive), and gets les when the cooler reaches set temperature.

Magnetic field coupled noise is another way of saying EMI.

Aluminum is a poor magnetic shielding material for low frequencies.  If indeed the coupled interference is coming from the power supply cable, then the aluminum camera housing is not providing enough shielding.

atikuser

  • Newbie
  • *
  • Posts: 8
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #28 on: August 13, 2020, 06:43:43 pm »
nikkinemo95, interesting comment. 

You can use Photoshop's magic marker tool to separate the sky from the foreground.  You probably will have to experiment with the tolerance level.  You may have to modify the transition line, which you can do by holding down 'shift' while using the 'lasso' tool to add to the selected area, or hold down 'alt' while using the same tool to subtract from the selected area.  I generally have 'anti-alias' turned off.

How does your comment relate to horizontal banding?  Do you see horizontal banding that is related to the sky/foreground transition?

On the subject of banding, Atik might investigate what sensors Nikon and Canon use.  Their latest cameras use CMOS sensors.  I take astrophotos with either a Nikon D5200 or D500, principally the D500 now.  I have never seen banding in images taken.  Attached is a recent D500 image of M81 (30-second, iso 25600) on a C11, Orion Atlas Pro mount.  Unguided.

Atmospheric turbulence is an ever present problem.  It smears out star images from pinpoints in time exposures, as can be seen in the M81 image.
« Last Edit: August 13, 2020, 06:48:56 pm by atikuser »

Moonrider

  • Newbie
  • *
  • Posts: 2
Re: Horizontal Banding - Color Horizon - Still a problem!
« Reply #29 on: March 27, 2021, 03:03:43 am »
Yep, it is not only a problem for the color sensors...