Showing posts with label webcam. Show all posts
Showing posts with label webcam. Show all posts

Saturday, January 4, 2014

Legacy Series, General Thoughs on Imaging

General Thoughts on Imaging

CCDs are sensitive on visual, ultra-violet and infra-red wavelengths. Lets forget about UV by now! IR, quiet interesting, has a very different focal plane with refractive optics than visible light (it's all about the Snellius-stuff, e.g. chromatic aberration), thereby opening two possibilities:

  1. Block visual completely (a waste of light since IR is pretty well attenuated through our atmosphere)
  2. Use pure reflective optics (i.e. Newtonians).

I am still playing with the thought of the second option for the future. The problem here: The available telescopes of this kind having a reasonable size and a not too long focal length (about 70mm aperture and not more than about 500mm focal length) are usually of extremely cheap quality. To get some decent images the mount of the telescope (and it's tracking) has to be rather stable, usually then the telescope on a quality mount is again much bigger (i.e. focal lengths about 2m), making the field of view rather small, also the f-number ususally shifts to „darker“ (aka slower) values. All these things are supposed to be teleSCOPEs, optimised for visual applications usually.
Still one option to go for, a „cheap“ scope on a good mount. But remember, focal-reducers are no option here, these would include „chromatic“ aberrations again. Thus, a pure reflective telescope (i.e. Newtonian design) with absolutely no refractive element, the greatest possible aperture and the shortest possible focal length would be the intrument to go for, preferrably with a parabolic mirror (most of the cheapoes have spherical primary mirrors).

Presently I am using two basic setups for imaging, both including an IR-cut filter. The first setup is the relatively cheap, computerised refracting telescope ETX-70 by Meade, having an aperture of 70mm and a focal length of 350mm (making it f/5, a rather fast setup). The ETX-70, meant to be a beginners level scope, has quiet inaccurate tracking, thus exposure times are limited to about half a minute (still recording stars fainter than 14th magnitude!), field of view (FOV) is less than about 1°. The second setup consists of a webcam and a photographic objective (have a look the Bellow-Cam MK-II page for details). The setup is tracked by standart "cheap" hobby material, namely an EQ-2 mount (usually provided with very very simple telescopes) and the appropriate right ascension (RA) motor. Compared to the focal length (mostly 50mm) of the system this mount tracks well enough to expose for quiet some time. Drawback on this system: the camera fitting best mechnically (QC4000pro) is not as good as the one used together with the ETX-70. Advantage though: even faster optics, the 50mm lens, for example, is f/1.8, an IR-cut filter (not the best, better than nothing) present in the base of the CCD... Even faster lenses are available (e.g. on ebay) and, besides the webcam and the motor, everything in this setup was obtained via ebay for a real bargain total amount of money.
The alternative setup to the ETX-70 is a SK8035 (SkyWatcher 80mm 350mm f/4.4 achromatic refractor). The newest addition to the family is a SK15075 (SkyWatcher 150mm 750mm f/5 achromatic refractor), which performs really nice; more starry nights needed!

More on filters? Yes, there still is something to mention, I would recommend filters of all kinds cutting out Na- (Sodium) and Hg- (Quicksilver) lines. The visual impression might be disturbed, the photographic will be fine.


Legacy Series, Another Kind of HDR Photography

M42


Finally, some time and sky to work on long exposure webcam-astronomy... The remaining clouds allowed an open view on Orion. M42, certainly one of the more prominent objects, easy to see and easy to record, gives a perfect light source for experimenting as it includes a very bright open cluster as well as a dark cloud and a bright nebula.


Instrument
ETX-70 equatorial mode, #494 autostar
Camera
ToUCam pro PCVC-740, Baader IR filter
Data acquisition K3CCDTools, 10s exposures
Data registration
RegiStax 2, K3CCDTools
Frame stacking
RegiStax 2, K3CCDTools
Postprocessing
RegiStax 2 & IrfanView, iMerge

Some resulting images (based on the same recorded AVI file - have a look at a compressed WMV version) were obtained using different sets of postprocessing parameters (contrast, brightness, saturation, gamma curve, etc.) in order to respect the different aspects of the complex object. All image tuning steps were carried out on all pixels equally (no area selective tinkering).



The Results



Image resulting from registration and stacking, no postprocessing



The above "raw" image postprocessed using IrfanView



Light postprocessing using RegiStax


Massive fiddeling with RegiStax and IrfanView postprocessing


Registration using K3CCDtools linear scale


Registration using K3CCDtools logarithmic scale


All above combined using iMerge


The Observatory

Have a look at my hyper-professional setup to catch the hole in the clouds (greyish stuff above the roofs in the background). Even though the mechanical setup looks pretty solid, the RA motor of the fork produces some jitter occasionally. Consequently I performed a manual selection of the frames to be registered, resulting in a loss of about 30% of the 10sec frames.


Legacy Series, General Thoughs on Imaging

General Thoughts on Imaging

CCDs are sensitive on visual, ultra-violet and infra-red wavelengths. Lets forget about UV by now! IR, quiet interesting, has a very different focal plane with refractive optics than visible light (it's all about the Snellius-stuff, e.g. chromatic aberration), thereby opening two possibilities:

  1. Block visual completely (a waste of light since IR is pretty well attenuated through our atmosphere)
  2. Use pure reflective optics (i.e. Newtonians).

I am still playing with the thought of the second option for the future. The problem here: The available telescopes of this kind having a reasonable size and a not too long focal length (about 70mm aperture and not more than about 500mm focal length) are usually of extremely cheap quality. To get some decent images the mount of the telescope (and it's tracking) has to be rather stable, usually then the telescope on a quality mount is again much bigger (i.e. focal lengths about 2m), making the field of view rather small, also the f-number ususally shifts to „darker“ (aka slower) values. All these things are supposed to be teleSCOPEs, optimised for visual applications usually.
Still one option to go for, a „cheap“ scope on a good mount. But remember, focal-reducers are no option here, these would include „chromatic“ aberrations again. Thus, a pure reflective telescope (i.e. Newtonian design) with absolutely no refractive element, the greatest possible aperture and the shortest possible focal length would be the intrument to go for, preferrably with a parabolic mirror (most of the cheapoes have spherical primary mirrors).

Presently I am using two basic setups for imaging, both including an IR-cut filter. The first setup is the relatively cheap, computerised refracting telescope ETX-70 by Meade, having an aperture of 70mm and a focal length of 350mm (making it f/5, a rather fast setup). The ETX-70, meant to be a beginners level scope, has quiet inaccurate tracking, thus exposure times are limited to about half a minute (still recording stars fainter than 14th magnitude!), field of view (FOV) is less than about 1°. The second setup consists of a webcam and a photographic objective (have a look the Bellow-Cam MK-II page for details). The setup is tracked by standart "cheap" hobby material, namely an EQ-2 mount (usually provided with very very simple telescopes) and the appropriate right ascension (RA) motor. Compared to the focal length (mostly 50mm) of the system this mount tracks well enough to expose for quiet some time. Drawback on this system: the camera fitting best mechnically (QC4000pro) is not as good as the one used together with the ETX-70. Advantage though: even faster optics, the 50mm lens, for example, is f/1.8, an IR-cut filter (not the best, better than nothing) present in the base of the CCD... Even faster lenses are available (e.g. on ebay) and, besides the webcam and the motor, everything in this setup was obtained via ebay for a real bargain total amount of money.
The alternative setup to the ETX-70 is a SK8035 (SkyWatcher 80mm 350mm f/4.4 achromatic refractor). The newest addition to the family is a SK15075 (SkyWatcher 150mm 750mm f/5 achromatic refractor), which performs really nice; more starry nights needed!

More on filters? Yes, there still is something to mention, I would recommend filters of all kinds cutting out Na- (Sodium) and Hg- (Quicksilver) lines. The visual impression might be disturbed, the photographic will be fine.


Legacy Series, how to analyse long exposure AZ data




Analysing Data Recorded with Azimuthal Setups

It is quite obvious that an azimuthal mount causes field rotation, when not used on the North- or the South-pole. For real long exposures (i.e. minutes to hours) this is utterly devastating for every image taken.
Here the abilities and power of CCDs are coming into play. With exposure times of a couple of seconds, field rotation does not play any role for the single frame. As long as the CCD amplifier ensures that the charge in a single pixel is high enough to result in a signal greater than the detection threshold we can integrate over several different frames. Now field rotation will return when not properly taken care of .

IRIS however is capable to compensate for rotation between individual frames during alignment. Lets see what is to be done, on a step by step basis, assuming the raw fits series (after conversion) is called r#.fits (for red), g#.fits (for green) and b#.fits (for blue). Furthermore assume that the data set contains 50 frames.
  • Get a “Display commands window” first.
  • If not using IRIS in the first place you most likely have to convert your AVI-file into a plurality of FITS-files. When going for color, every frame will be present in a red, a green and a blue channel in separate FITS-files. That means, that you will have to perform all following steps on a respective color files individually. Conversion done by: [click: File -> AVI conversion... ].
  • Now you will have to select two suitable objects (stars) by marking.
    For this you have to get one frame into the main window, preferably the first of a series: [type: load r1].
    Now mark two medium-bright stars which are not too close to each other: [click: Analysis -> Select Objects => with the funny mouse pointer click on two medium-bright stars].
  • If everything went fine so far, you will be ready for registering the frames yet: [type: rregister r rr 100 50].
    A new file-set with the name rr will be created. Forget about the third parameter for the moment, you will learn to use it with some practice.
    With the time it will become clearer which registration objects will give better results and what combination will not be so good.
  • Stacking registered frames, nothing easier than this: [type: add2 rr 50].
  • Now you certainly would like to save the result: [type: save red].
  • Do the same with the other two colors...
  • To combine the three images you could do all sorts of tricks with IRIS. A simple first glance could be: [type: trichro red green blue].
    Save your result by: [type: savebmp myimage].
Have a look at an example analysis done on M52 data. Both results came from the same data set, analysed in different ways.





M52, ETX-70, analysed with K3CCDTools, dark frame subtracted
M52, ETX-70, analysed as described above, no dark frame subtraction



Legacy Series, long-exposure ToUCam PCVC-740K

ToUCam PCVC-740K long-exposure modification


Background Story

Please have a look at the superb pages available all over the globe. Many thanks to Steve Chambers for sharing his findings with the internet-community! Please have a look at Steve's page to understand that it will not be allowed to use the information provided for purposes of profit.
Three pages to be mentioned as being a very good source for rock solid informations:
(1)   http://www.pmdo.com/wintro.htm
(2)   http://members.bellatlantic.net/~vze29wzh/toucam740mod.htm
(3)   http://www.aludobson.de/CCD/umbau_einer_toucam_pro.htm


Please note that whatever you do to your camera is up to you. What I disclose on my pages is based on personal experience. Other setups (OS, software, soldering skills) might lead to other results, don't blame it on me.

Schematics

This is just a tiny little part of it. For convenience I decided to used different NAND gates than shown in (2). Reason being that the ugly style / dead bug / Manhattan method appeared to be the appropriate design choice to me for an incomplex circuit like this one. In dead bug it is easiest to bend pins together... pin 1 and pin 2, pin 3 and pin 12. Half the work done w/o any effort.



Warning

If this is your first SMD-sized project, please do not start directly... do some practice before. A webcam certainly is ruined very very quickly!


Photographs...








Overview of the battle-field side of the toucam's PCB, cuts already being applied







Practice cut.... get rid of additional light produced by the green SMD-LED






These cuts are essential (see link (2) mentioned above), cutting too much of the mass-plane does not harm, but... watch out for the other leads! I used a sphere type diamond-tool to mill away the leads to be disconnected.



Wires do not need to go around the PCB in any case. Here are two vias which can be used to solder [Pad8] to pin 1 of the 4011 (lower wire) and [Pad10] to the potential switch (upper wire). I used 0.2mm "magnet wire" (enameled copper wire)...
In these places, the wires are soldered to the PCB-vias. BTW: finally I did not add a switch to the cam... what for anyway?





The leads (0.2mm magnet wire) connecting "Pin8" to pin 11 (4011) and [Pin10] to the 100k resistor. Both wires use "foreign" vias to channel to the side of the PCB (as long as the enamel varnish is not scratched at the tunneling portion of the wire, no problem can be encountered).






Both sides of the story.... the blue wire is the power supply for the 4011. (Sure, I could have searched for another place to get 5V from.... but, time is money, isn't it?).





And... on the other side of life....  Dead bug "mounted" (cyan-acrylate) 4011 with 100k resistor. In the upper quarter of this image the connection between [Pad8] and pins 1 and 2 of the 4011 is visible. The lower quarter show the "tunneled" leads to pin 11 (4011) and to the 100k resistor.







Dead bug 4011.... the two blue wires are the link to the parallel port connector.



Switch?

You will need the switch only if you intend to take still-image with the camera in the common way. For video the camera will still be working fine with [Pin10] of the 16510 being connected to the 100k resistor.
Personally I omitted the switch...



Have a look at an observation using the camera recording the Great Orion Nebula (M42) in combination with an ETX-70.


Legacy Series, the Bellow Cam Mk 2 SC GT



Bellow-Cam Mk 2 GT

 Being used to GOTO-telescopes, I also liked my wide-angle camera to be equipped with this feature. OK, it is not a big deal at all to handle right ascension and declination setting circles, pointing the scope/camera by α and δ looked up from catalog or star chart data, but, goto is sooo convenient! After some looking around, I got myself a used MEADE DS-127 mount with an AutoStar #957 computer. Perfect!
Fitting Bellow-Cam Mk 2 to the mount is so simple that I would not like to waist words on this issue. It took a piece of hardwood, two bolts and two nuts to get it all sorted.

Bellow-Cam Mk 2 GT, the wide-angle camera setup on the “autostared” DS-mount, in operation.

Analysis

It is quiet obvious that this is an azimuthal mount, resulting in field rotation. For real long exposures (i.e. minutes to hours) this is utterly devastating for every image taken.
Here the abilities and power of CCDs are coming into play. With exposure times of a couple of seconds, field rotation does not play any role for the single frame. As long as the CCD amplifier ensures that the charge in a single pixel is high enough to result in a signal greater than the detection threshold we can integrate over several different frames. IRIS offers a very good possibility to register and stack frames which are slightly rotated between one another.
A very very short introduction how an analysis like this could be looking like can be found here.

Result

The image of M45 was taken 2005 January 9th. Again I was to lazy to subtract a dark frame. Note the image tilt, that's the sacrifice for compensating field-rotation when using an azimuthal mount. To be noted on the image is amplifier-glow on the upper left corner. The QuickCam appears not to be the best camera, even with the use of non-raw patch, ear-like artifacts still occur.
It is amazing, I think, that the nebulosity of M45 can be recorded using a f=50mm SLR lens and a webcam from a light polluted place like South Holland, on a night with light overcast and quiet poor transparency...

Lagecy Series, the Bellow Cam Mk 2 SC



QC4000pro LX (SC2)

The QuickCam 4000 pro had to go through a long exposure  modification following the "SC2" approach as described by Martin Burri. There is nothing much beyond Martin's remarks to be mentioned. The modification went pretty straightforward. No pins are to be lifted in this way of modifying, cutting PCB traces does the deal. I soldered magnet wire (I have tons of this around for my radio projects) to the pins of the ICs. The individual leads are connected to a kind of "patch panel", the tiny bit of bread-board PCB glued to the camera opposite the USB and microphone connector. Blue wires are making the final connection to the 4011 hidden below the whole mess. The red and the yellow wire (stereo phono cable) are serving as connection to the parallel port of a computer.
BTW, in contrary to the ToUCam, this camera holds an infrared cut filter inside the objective lens bearing, thus no additional IR-filter is needed in a setup like shown here.










Camera front-side, no lens mounted.
For reasons of experimentation, the camera is built in tilted...


Additionally I spent money and obtained a
Ra-motor for the EQ2. Decent tracking is
a necessity for the long exposure times the
camera was modified for.


LX-astro images

You are invited to compare with the images recorded using Bellow-Cam MK1 and Bellow-Cam MK2. Messier 44  gives some impression about the benefits of longer exposures. Exposure time was chosen to be 10s/frame. Conditions on March 13th 2004 when I was acquiring data for the following image were fair (for this site), with the unarmed eye, I could see Orion's sword.


M44, 50mm f/1.8

Legacy series, the Bellow-Cam Mk 2

Bellow-Cam Mk 2

Another Camera

The Vesta fits quiet OK, but not really well.... Philips, for what ever reason, appears do design funky shapes for webcams... Logitech obviously shows some more heart for tinkerers, doing so they designed a ball, long ago, and sticked to that design. Certainly a ball fits better into the bellow-cam... Have another look... you might recognize a QC-4000pro in un-modified state.
The webcam is held by self-adhesive window frame insulation rubber, the circular openings of the bellow are perfect to fit the QuickCam-ball in....
In this stage I have chosen to mount the camera on an EQ2 mount. The Ra and Dec axis allow easy aiming when the coordinates of the object of interest are known, furthermore the mount enables manual tracking.


side view with 50mm lens
resulting field of view about 4°

side view, 135mm lens with "dew cap"
resulting field of view about 1°30'

front view with 50mm lens

back view

front mount detail, no lens

rear mount detail

the "victim"

the "supporter"


M45 (Pleiades)  recorded with the above setup

Legacy series, the Bellow-Cam

Relatively old stuff from my original web-page... 
 

The Bellow-Cam

A Simple Wide-Angle Camera

Guess what, I am lazy.... So, I tried to find as much "completed work" as possible. Resulting in a rather simple design. All parts can be easily found on fleamarkets or rallies... The basis of all are a photographic bellow, a Vesta pro and a couple of lenses (M42). Cardboard finished the first design, but just have a look...


Vesta-Cam using a 50mm f/1.8 lens (no long-exposure!)



M44 (Praesepe) recorded with the above setup, unguided camera on a simple tripod,
the bright spot on the lower left edge is Jupiter