ZWO ASI178MM Cooled CMOS camera

Affiliation
American Association of Variable Star Observers (AAVSO)
Wed, 02/10/2016 - 05:34

Just curious if anyone has used this camera, or another ZWO camera for photometry? At around 700 dollars it is on the lower end of the price scale but according to the specs it should do a fair job. Since you can use CMOS chips in DSLRs I suppose this camera could be used as well. I am using a loaner camera right now, so am keeping an eye out for a replacement camera. I would like to use the loaner on my LX200 eventually.  The IMX178 spectral response seemed to be similar to a CCD chip in efficiency. I do have a color ZWO120 MC (color) for planetary imaging but have done a couple of deep space objects with some success. I have attached an image of M 42 taken with the ZWO 120. It was shot using AVI with an AR 102 F/6.5 telescope. I also took several images of 30-40 seconds in JPG and got pretty good results from using the camera in a way that would be better for photomery. Here are the specs. 

Camera Type
Single Frame Camera

Color or Monochrome?
Monochrome

Camera Series
ZWO ASI178

Sensor Manufacturer
Sony

Sensor Model
IMX178

Sensor Type
CMOS

Sensor Dimensions (mm)
7.4mm x 5mm

Pixel Size (µm)
2.4 um

Pixel Array (pixels)
3096 x 2080

Camera Cooling
Built-In Fan, TEC Cooling via Convection

Minimum Exposure
32 microseconds

Maximum Exposure
1000 seconds

Software Included
Yes

Computer Interface
USB 3.0

Includes Filter Wheel?
No, Not Necessary

Camera Weight (lb.)
14 ounces

OS Compatibility
Mac, Windows, Linux

File Upload
Affiliation
American Association of Variable Star Observers (AAVSO)
I'll take that as a no!

A review came out today in Astronomy Technology on the ZWO 120 series camera. Hopefully it will provide some insight!

Affiliation
American Association of Variable Star Observers (AAVSO)
ZWO camera

It has a Pregius Global Shutter, which means that it does not have a physical shutter. Could explain the price. Not sure how much that would effect photometry?

Paul

Affiliation
American Association of Variable Star Observers (AAVSO)
After thinking about it...

Jeff Hopkins wrote a book on using the Meade DSI for photometry. Though I never used the info for the AAVSO database it did do pretty solid photometry. It is a non-cooled camera with no shutter. So I think the ZWO could provide low cost service for those who can't afford a more expensive camera. By the way the article in Astronomy Technology is on building a low cost All Sky camera. It is pretty good but does not answer the question that I posed. 

 

Paul

Affiliation
American Association of Variable Star Observers (AAVSO)
Photometry with ZWO ASI174MM

Paul,

I am doing photometry with a ZWO ASI174MM. I have the uncooled version (so take regular dark frames) but there is also a cooled one. Both have global shuttering and there are color versions. Minimum of 12-bit required for photometry (Howell 2000) and important that FWHM of star images cover 2-3 pixels. ASI174MM achieves both these criteria (I use it with 150mm f/5 Newtonian reflector and 70 mm f/7 refractor). My friend has a ASI1600MM-Cool and is also using an ASI034MC on his guide telescope. We both find ZWO cameras easy to use and operate reliably. (I currently use SharpCap for image capture but plan to try Astroart in the next couple of weeks.)

Attached are B-V and V-I transformations I obtained on 26 January 2017 using this camera and BVI filters. I am using a Xagyl filter wheel (5 position, 248g) but have just bought a ZWO EFW (5 position, 300g) which I have tested; it also works well. The ASI174 has a full well depth of 32ke- and I have found it to have good linearity up to that (can send graph of checking of linearity if required). Happy to send further information via email and answer any other questions. I live way down under (Tasmania) but have good Internet bandwidth so can Skype if preferred.

Tex

Affiliation
American Association of Variable Star Observers (AAVSO)
Tex,

Tex,

Have you found that the ASI174 calibrates just like a CCD camera? Do you do the regular Bias/Dark/Flat calibration sequence?

I'm testing a ASI178 and am finding some strange behavior. For instance darks wilth averages greater than the bias shots. I've read recommendations from other forums that suggest that bias can be ignored.

Thanks,
George

Affiliation
American Association of Variable Star Observers (AAVSO)
CMOS cameras for photometry

I'm also making a transition from CCD to CMOS photometry with a ZWO ASI178MM. This thread is a year old; are there insights and lessons learned in the meantime that you can share?

A couple things that I've learning:
  - The camera has a dual nature: It can take timed exposures and it can also act like a video camera. Don't understand yet the transition.
  - The ASCOM driver allows tuning of the gain. What is the best setting if doing photometry.
  - The camera has a native driver and an ASCOM driver. Is it worth finding software that uses native driver? SharpCap does and also warns that that using the ASCOM driver reduces performance. MaximDL does not support the native driver. I've not thought of SharpCap as a photometric package. Suggestions?

Thanks,
George

 

Affiliation
American Association of Variable Star Observers (AAVSO)
Gain for photometry

What is your average seeing (FWHM), and what is the FL of your telescope?  I'll see if I can calculate at least a starting place for the gain setting which could be optimized by linearity testing.

Phil

Affiliation
American Association of Variable Star Observers (AAVSO)
ZWO CMOS gain setting for ASI178

Phil,

Thanks for the offer to explain how to optimize a gain setting! I would have started with a gain set to 0 to get the greatest dynamic range.

My setup is a 12" Meade with a focal reducer bringing the focal length down to 1700mm. In the past if I can capture images with FWHM of 3" I was happy; rarely could do better than 2.5".

There is a lot of interest in the ASI178MM camera. A couple others have purchased one and the AAVSOnet may add some to the BSM network.

I have my camera on the optical bench right now, trying to characterize it. Getting some strange results: The darks don't seem to change much with exposure. And I've seen darks with lower averages than the bias shots. So there needs to be some study on how to calibrate this camera for photometry.

Thanks,
George

 

 

CMOS dark current

Hi George,

The CMOS sensor dark current issue you mention is well know in DSLR photometry, see Appendix A of the AAVSO DSLR Observing Manual.

I calibrate DSLR images with master bias, master dark and master flat frames. The dark frames must be recorded with the same exposire time as the science (light) frames because they cannot be scaled like CCD darks can be. Cheers,

Mark

Affiliation
American Association of Variable Star Observers (AAVSO)
Gain for 1178

George,

That brash offer came from my budding interest in trying to understand  if/how these CMOS cameras can work in photometry, but is was based on the mistaken assumption that on-chip binning was possible with CMOS.

I have a smart, enthusiastic guy in the CCD2 class now with an Altair 290M monochrome CMOS camera (12-bit, 15ke- FWD, 1920x1080, 2.9 micron pixel).  He has been doing pretty pictures for a while, apparently with some success.  I think he typically varies his gain to get a good SNR's in the different  RGB images, but if he mantains a constant gain for photometry in B and V he has trouble with a very limited dynamic range.  

It is also possible to adjust the bias level with his camera.  Some people in his astrophotography group set the bias to 0.  Perhaps that works for pretty pictures, but I don't see how that would work for photometry.

I admit to being pretty confused by CMOS. I'm hoping Mark and Tex can help understand this better.

Phil

 

 

Affiliation
American Association of Variable Star Observers (AAVSO)
Lucky Imaging

Hi Phil,

I had a look at similar cameras recently to help a friend who was looking to make photometry with it as well as other purpose (IMX 174 and the like ). Those cameras are in fact designed for lucky imaging, by the way the guys use very high gain (and short exposure). At such gain they saturate strongly on the analog to digital converter, in addition that converter is only 12 bits, that's a little short for photometry. Certain of such camera are not able to set a gain low enough to cover the full well of the sensor (the one of the 174 is very good 32500 e- ). That high "0" gain makes those cameras unusable for photometry ! In addition the size of the sensor is not large implying a short focal for a given FOV and by the way a small optical aperture: fewer photons, low SNR...  The 178 has somewhat better noise level but is far too small I think. 

They usually provide the calibration factor electron / ADU, simple, knowing the ADC capacity we can calculate the max electrons we can get: in one example 0 gain was 0.49 e-/ADU and 12 bits. That means the ADC saturates at 0.49 x 4096 = 2007 electrons ! Instead of the 32500 e- sensor full well, no way to use it. 

Bias: yes we need a systematic bias that encompass the negative part of the noise distribution, more gain, more bias needed. Without it it could be difficult to calculate the black level. 

Clear Skies !

Roger

Affiliation
American Association of Variable Star Observers (AAVSO)
It has a 14bit ADC

Hi!

I have the uncooled version of the ZWO ASI 178mm , and yes, the sensor is a bit small, but the ADC can resolve 14 bit for long exposures. It has some fast video capture modes that work in 10 bits to keep the bandwidth manageable, but for photometry with single exposures, you can use 14 bits.

CS

HB

https://astronomy-imaging-camera.com/products/asi-cooled-cameras/asi-17…

 

 

Affiliation
American Association of Variable Star Observers (AAVSO)
CMOS cameras for photometry

Hi All!
By the way...
Did anyone try APT (https://ideiki.com/astro/Default.aspx)?
I tried this low-cost software, but working very well and mading the plate solves with "platesolve2" made by Planewave. Compatible with Ascom, it allows many cameras in photo mode and video mode,
and can save in fits.
On the other hand, the photometric treatment is done with MaximDL.

JBD

Affiliation
American Association of Variable Star Observers (AAVSO)
Bad weather

I posted this thread quite a while ago with the intention of acquiring a ZWOASI mono camera and giving it a go. I did purchase a used ZWO ASI120 MM (Mono) CMOS and have used it as an autoguider with some success. Unfortunately, I am now living in Missouri (job change) and we have had about 12 clear nights since last March. I did get my 11" Celestron CPC setup, along with a SBIG ST-8E. Unfortunately, I am still working on getting the Meade 8" LX200 working. So have not had a chance to try out photometry with the 120MM yet since the LX200 would be the scope to be used. Having done some imaging with the 120 MM on a ST80 refractor, I think it would do excellent photometry. The big negative is doing dark frames and bias frames. You have to cover the aperture physically to take them. Since these are not cooled cameras you would have to shoot more dark frames if the temeperature shifted up or down. This is not a big issue unless you are doing multiple targets, with multiple exposures using a robotic setup. I think that these types of cameras can work quite well if you understand the limitations. 

Affiliation
American Association of Variable Star Observers (AAVSO)
DSLR dark currents

I am surprised from some comments about the dark currents in DSLR CMOS. In fact the best technique to cancel such "dark pulses" that occur at few of the pixels of DSLR CMOS is to use a single strong master dark and rescale it based on the measurement of the level of dark current in each sky image,  analyzing it in the pixels known to have extra dark current. This technique provides much better results than any other. I have implemented it in my own software years ago and the result of a lot of experiments is very clear: the  classical way making darks at each observation is far worst than using what I call the "adaptive dark technique". The classical technique is never very good as the temperature of the sensor changes quickly during an observing session, even making two dark sets, one before, another after, is not perfect. Instead the  "adaptive technique" eliminates perfectly the pulses in any case. In addition it provides the information of the level of the dark current for each image, and then the variations during the session are obvious ! 

This technique is usually available in various software under various ways but it seems very few amateurs use it. If someone interested I can describe it into more details.

Another aspect of the adaptive technique is the information it provides: in fact, most of the time, the level of dark currents of the recent DSLR CMOS is very low and possibly negligible. A couple of electrons per minute is common, far below the random noise. The correction is really needed only by warm weather.

The other aspect of dark currents is the "normal" level that corresponds to the design parameters of all pixels. In today DSLR CMOS there are few pixels affected by the extra currents making pulses into the image. In a 24 Mega pixels sensors we have only a few thousands  having extra currents. All are more or less at the "design" level and that level is usually very low. Such "normal" dark current just uniformly elevates the black level of the image, but this is not visible in DSLR as the sensor compensates for it. What remains is the shot noise corresponding to that electron count. This is the true reason for cooling CCD as nothing can cancel such random noise. In most DSLR CMOS that normal dark level is very low, much lower than CCD (I say "most" as it's generally true but I have found few cases not that good !... ). The clear result is that the cooling is not so useful to recent DSLR CMOS and we should remember the power dissipation of CMOS is extremely low compared to such of the CCD technology (it has been invented for !... think to general electronics that uses CMOS in most cases). 

I  also say "DSLR CMOS" as it's not so obvious in other cases. Those sensors are restricted to the DSLR manufacturers and generally not accessible to astro camera makers. What is today accessible is the "IMX" range from Sony. Those sensors are designed for industrial and security applications, some are very good, some less, but most are very small and for me it's an issue for photometry ( limited electron count, and SNR, for a given FOV ). There is one large 4/3 being available (not Sony), used in past DSLR, but somewhat old ( technology twices perfs every 18 months... ). 

Clear Skies !

Roger

 

Affiliation
American Association of Variable Star Observers (AAVSO)
Further Details

Hello Roger

You wrote:

"I am surprised from some comments about the dark currents in DSLR CMOS. In fact the best technique to cancel such "dark pulses" that occur at few of the pixels of DSLR CMOS is to use a single strong master dark and rescale it based on the measurement of the level of dark current in each sky image,  analyzing it in the pixels known to have extra dark current.

 

Can you give more details on this technique.  I would like to try it, but I am not quite sure what you have done.

 

Gary

Affiliation
American Association of Variable Star Observers (AAVSO)
Software implementing ADT

Greetings Roger,

Thank you for the process details!

Do you know of any software applications that implement this ADTnprocess for CMOS image processing?

I see in an old post ( https://www.aavso.org/comment/29761#comment-29761 ) that you do much of your coding in APL. Back in the day I was an APLer too. Have you developed any process code that can be shared?

Thanks,
George

Affiliation
American Association of Variable Star Observers (AAVSO)
APL code

Hello George,

nice to read you are an APLer ! Long time ago I was also coding with C/C++ in addition to APL, but I have 100% switched to APL more than 10 years ago and most what I do now is for astronomy. APL is so effective for experiment, massive data analysis, thinking... that choice is obvious for an APLer (but we know so well: not for non-APLers ! ).

But ok, my APL code has never been intended for distribution, it is just for my own experiments and as an APLer yourself you probably know the way we work directly under the work-space without much GUI. Then I do it 100% for EOS DSLR, it should be simple to switch it to other camera but it's to be done (.fit is also available but not integrated in the photometry code.). Anyhow no problem to share such code. It works under Dyalog APL 11 (with some script / auxiliary under Windows), not a very recent one ! I plan to convert to Dyalog 16 but it's not done yet (I don't expect much problem). 

I think the best is to talk off-line, you could contact me through the AAVSO "contact".

Other software ?  In fact I discovered this principle in IRIS, in it it's supposed to be an "optimisation" of the dark process,  then I extended it under APL to make it 100% automatic and more general. I saw functions in several other softwares  that should enable to implement it (ie PixInsight... ).

Clear Skies !

Roger

Affiliation
American Association of Variable Star Observers (AAVSO)
CMOS dark currents

Greetings Roger and thank you for the detailed response. I would like to learn more about the adaptive dark technique. What software implements it? With a detailed description I could try to implement it. Since I own a ASI178 now I want to make the best of it!

Thanks
George

Affiliation
American Association of Variable Star Observers (AAVSO)
Stacking and how it affects SNR in VPhot

Good day, everyone.

Phil, thank you for that kind introduction: "I have a smart, enthusiastic guy in the CCD2 class now with an Altair 290M monochrome CMOS camera"

I am pleased with the results I've obtained with my camera in the CCD1 class -- targets have been a three-hour session covering one period of YZ Boo in "V", and the Standard Fields M67 and Melotte 111. Now in CCD2 with the introduction of the "B" filter my work is becoming more challenging.

With a max FWD of 15ke- at the lowest gain setting I can achieve three magnitudes of dynamic range in a single frame with SNR > 100 for several comparison stars and the target. RR Boo is definitely possible with B-V of 1.5m but RU Vir is out of the question with B-V of 5m although I don't give up easily. Somehow I'd like to find a way to pull a rabbit out my hat.

This was a roundabout way of getting to my question which I am sorry is a bit off-topic. I was hoping that stacking could come to the rescue by averaging perhaps 9 100-second frames. According to my understanding of the math, a stack of 9 should give me a three-fold increase in SNR. Unfortunately when I upload a stack of 9 to VPhot it says there is just a minor improvement, far less than what I expected. Furthermore when I upload a stack of 46 I get these results:

"Star 1" went from SNR 208 on a single frame to 217 on a 46-frame stack. "Star 2" went from SNR 93 to 106, and "Star 3" went from 87 to 102.

I tried two methods for creating the stack:

  1. Calibration, Registration, and Stacking all within CCDStack2.
  2. Calibration and Registration in CCDStack2. Stacking in VPhot.

Same results.

I am not familiar with the algorithm that VPhot uses to calculate SNR but it seems to be relatively unresponsive to stacking.

I would be thrilled to be at fault in some way. Can anyone shed some light on this problem?

Brian

Affiliation
American Association of Variable Star Observers (AAVSO)
SNR calculation

Hello Brian, 

I have no idea how VPhot works, I am a DSLR guy and use my own software. But ok I suspect some problem: in photometry, under usual condition, the SNR is nearly 100% determined by the signal and its own shot noise. The camera noise (Nyquist-Johnson and possibly dark shot noise) is fully dominated by the signal shot noise, and negligible. Usually the photometry SNR is approximated by simply taking the square-root of the signal. BUT that works ONLY if the signal is expressed in electron-count cumulated in the photo-diodes of the measuring aperture, this is the physics. If that calculation is applied to ADU the result has no meaning. To make such calculation we have to make the conversion, determine the e-/ADU calibration factor (often said "gain" by the astronomers even if it's somewhat the inverse !). If the data sent to the softtware are ADU and the calibration factor is not provided the software has no way to calculate the SNR (even if the full CCD equation is used). If VPhot do it assuming e-/ADU = 1 (very common in photometry cameras) and your stack is a mean or normalized by some way, it would find a very similar SNR for a single image and the stack:  wrong in both cases ! 

The gain scale found in the lucky imaging cameras is in fact non-linear (more or less dB , log scale) it has little physics meaning, different from brand to brand, the 100% usually corresponds to a very high gain. One of the brand reaches 0.001 e-/ADU at 100% and 0.45 e-/ADU at 0% ! Just crazy, then the dynamics is limited at 2000 e- at 0% (12 bits ADC). We should make our own measurement of the calibration factor and then work about e-/ADU =1 to get the full dynamics. A lower gain (e-/ADU = 2~4) often permits an higher SNR if the signal is strong enough but no more dynamics. 

It's essential to always make noise / SNR reasoning, calculation, in electron-count. Same for the dynamics, ADUs are just an illusion, a simple multiplication of the electron count, the physics is at the sensor level: electrons.

Clear Skies !

Roger  

Affiliation
American Association of Variable Star Observers (AAVSO)
Frames from CMOS cameras work

Frames from CMOS cameras work differently than from CCD and such issues are common. It took many attempts before I understood what settings to put in Muniwin. I use ASI178MM-c and had Altair IMX224 before, both have similar problems. I understand the problem is with so low noise level. Mostly photometric accuracy in modern CMOS cameras is about scintilation, so difference after stacking shouldn't be large. Not a surprise with gain 0, where the noise is at extremely low level. I use that advantage to record with higher gain, so I can "manipulate" with lucky imaging things too (because I need shorter exposures). I set usually between 30-50% instead of 0, where noise level is still at >quite low< level. Actually, lowest gain has the highest read noise value. A typical 0.005 mag transit of 11 mag star (using 107mm lens) with 3-5s single frames that reaches maximum SNR (before saturation level) isn't a problem at all. I don't really recommend to use gain 0 unless you really need to do that.

Try to check that with barely visible objects or with higher gain. And possibly, use fainter object for that. The difference after stacking should be more significant then. Currently what you have done is mostly just averaging seeing stuff (in ~90%).

Regards,

Gabriel Murawski

Affiliation
American Association of Variable Star Observers (AAVSO)
ZWO ASI178 cooling notes and disconnect problem

The disconnect problem:  While testing I was plagued with the camera disconnecting from Maxim at random times. I found some hint in the ZWO forums about the power supply being involved and tried replacing the 12V/3A DC supply with a regulated supply ( Amazon: TFDirect Regulated 12 Volt 3 Amp Power Adapter, AC to DC, 2.1mm X 5.5mm Barrel Plug, . Cost $13). Since then, no disconnects!

Cooling. I've attached some profiles I ran on the camera. It was not able to achieve it's 40C advertised differential and would keep trying until drawing 100% power, even though the heat generation at this level was warming up the camera. It could achieve a 35C differential. So, keep an eye on the temp settings if you are near the limit.

George

 

Affiliation
Variable Stars South (VSS)
ASI183MM

I'm not sure about the ASI178 but here is an interesting article about the ASI183MM and it's use in science imaging.

http://www.astrosurf.com/buil/asi183mm/

It is partly in French but google translate works well.

I'm using one for spectroscopy and it seems very good. Little pixels but this can be offset somewhat with binning.

The CMOS sensors are certainly improving.

Terry