- Cooling a
       telescope mirror

    - Flatfield lightbox

    - Dewcap

    - Flip mirror  system

    - Modified Canon

    - DSLR shutter cable

 - SNR
Click on the thumbnails to see a larger picture


Besides telescopes, cameras and observatories, there is still a lot of astronomical stuff that you can make yourself.
In this chapter you'll find several little devices who are easy to build and who are indispensable for an amateur astronomer.


The biggest problem while observing with my Coulter-optics Cassegrain, was the ever present turbulence due to the thermal imbalance between the mirror temperature and the ambient air. A mirror 12.5 inch diameter and 2 1/4 inch thick cools very very slow.

First light with the telescope was an observation of saturn. It was end march 2009. It had been a beautiful day with lots of sunshine. The temp. inside my observatory reached nearly 18°C. At about 8 pm i started observing the planet with a magnification of 200x.

In the meantime the ambient temp. dropped to 11°C. The disappointment was huge. Saturns rings were dancing around. The cassini-gap was hardly to see, and the shadow of the rings on the planet was moving up and down like crazy. Next i tried a bright star, and moved the focus in and out, to see the diffraction rings. It was just as i looked into an open fire. The air turbulence made it impossible to see anything usefull.

Dissapointed i gave up and went to sleep. But i left the observatory open, and set my alarm clock at 3 am. Once again i looked at saturn. Oh boy, what a difference. Immediately i saw the cassini-gap, and the rings and the shadow of the rings on the planet were reasonably crisp. So it was obvious that i had to cool that thick glass cake more rapidly. For sure i didn't wanted to wait until 3 pm in the morning every time, to watch the stars.

Some googling around brought me to the websites of Alan Adler and Bryan Greer. These folks did some great work on this problem.
Everyone who has thermal problems with his mirror, should read what Alan and Bryan wrote about this.

So i modified the telescope with 7 axial fans. Three were placed at the back, blowing air against the back of the mirror. Four were placed at the side of the mirror, blowing air across the mirror surface. Two fans are blowing the air into the scope, the two others are blowing the air out. The back- and side fans can seperately be switched on and off.


Of course the result was better. I mean the time to wait until observing was a lot shorter, but on warm spring- and autumn days, it still took a few hours to cool the mirror down to an acceptable level. So, i was not very satisfied, and looked for more.

The internet is the greatest source of information. So after a while i discovered  the site of Anthony Wesley.
He also encountered the problem of slowly cooling mirrors. His solution is one with forced Peltier cooling. Since my trashbox contains several peltiers, i decided to go for a peltier-cooled mirror.

I did not make a copy of Anthony's system. It is completely different. In my system the cold air is produced in an seperate device. Four peltier elements are cooling two large heatsinks mounted in a duct. A radial blower forces air trough this duct, and blows the air to te back of the telescope via a flexible hose. The cooled air is sucked into the telescope, and to the mirror by the three fans in the backplate. This way the mirror compartment, and of course the mirror itself is cooling off. The used air is still cold enough. So it would be a sin to waste it. Therefore it is via a ventilation opening (2 side fans), sucked into the radial blowers intake, and brought back to the peltiers coolblocks. And the cycle begins again.
The heat produced by the 4 Peltiers is removed by 2 heat-exchangers who are watercooled with a coolfluid-pump.

You can see a schematic drawing of the cooling system here.


The blower is a radial type with a 230 Volt asynchrone motor.


The 4 Peltier elements and the heatsinks.
The peltiers are 5 amp. types at 12 volt. ( 60 Watt)
That makes the need of a power supply that can deliver 20 amps / 12 volt.
I use 2 old computer power supply's 12 volt 15 amps. Each supply feeds 2 peltiers.


Here you can see how i made the heat-exchagers. The principle is the same as by the Cookbook camera.
The brass nipples are glued into the coverplates with 2-component epoxy glue.
(Thanks again to my friend Johan Coussens for working out the black alu-blocks with his milling machine)

          plate        h-ex      h-ex-compl

Testing the unit on the workbench.
I found out that if the blower is running at full speed, the air slowly warms up. The blower works as a compressor, and
compressed air gets warmer ! To prevent this, i had to slow down the airstream. This is done by feeding the blower with
a lower voltage. Instead of 230 Volt, it runs now at 127 volt by means of a transformer.


The controlunit.
As you can see in the schematic, there are 3 temperature controllers.
The Jumo shows the outgoing cooled air temperature.
The Philips KS94 shows the difference in temperature between de mirror and the ambient air around the mirror.
The Omron E5CN is a real
temperature controller. Via a relay it switches de peltiers on and off, according to the preset temperature (green display).

I glued a Pt100 on the mirror and sealed it with a cap made of polyurethane.
A Pt100 is a platinum resistor who has exactly 100 ohms resistance at 0°C.
Its resistance value varies with changing temperature.

          Pt100a       Pt100b       Pt100c

The hoods and the flexible hoses for the cold air intake, and the used air outlet.

           intake          outlet

The Coolwaterpump.
The tank is filled with a mixture of 80% water and 20% ethylene-glycol (car antifreeze)


This is the completed system.
I placed the coolunit on an old TV wall bracket, mounted at a wall of the observatory.
So i can turn it away when not needed. That gives some extra space behind the scope.

            systeem        system1


The system makes it possible to keep the mirror at a lower temperature during the day. 
My intention is to keep the mirror, at a predicted night temperature, during the day. I think it will take some time to find out what the right setting will be.
The first tests during winter 2009 were very good. Ik kept the mirror 4°C below day ambient, and that seemed to work very well.

Of course it is impossible to use the cooling while observing. It has to be shut off and removed first.
But this works very quickly. It takes 2 minutes to shut down the power, remove the hoods and swing the coolunit away.
The mirror temperature seems to keep up well with further dropping ambient temperature, with the working
internal fans only.
I also don't see any effect on the image whether the fans running or not, so i keep them on.

After a lot of observing now, i can say for sure, that a telescope mirror with a thermal imbalance > 0.5 °C will not show good images.
And that counts especially for a cassegrain. Because light has to pass tree times the turbulent air.

Of course I will have to evaluate this system further, but I am already sure now, that this cooling system made a very big improvement in observation time.

Every time you take an astronomical raw image there are some unwanted things in the result. Due to dust particles on the optical parts you'll see so called "dust doughnuts" in the image. You can also have vignetting, caused by slight differences in the sensivity of the ccd chip etc.. To eliminate all this, you have to take flatfield images. These flatfields are used afterwards during image processing. The thing is, to have an evenly illuminated surface, from wich you take an image. Some people use the inside wall from their observatory or even the open dusk sky. I think it is more simple to use a light box. The box is then attached to the front of the telescope, and off you go. My lightbox is made of 2cm thick polystyrene plates, and has the form of a cube.


In the 4 inner corners I mounted 4 small 12 Volt light bulbs. The lamps are shielded, so there is no direct light going to the scope. The lamps are connected is series, and with a 12 volt source they give just enough light to do the job. The back wall of the box is the illuminated screen, and at the front I placed a diffuse plastified screen, to scatter the light. The result is a very uniform flat image.


My dewcap is made from pliable plastic covered with black felt inside.


The plastic is the same that bricklayers use to prevent brick walls from rising moisture.


There is a lot of software on the internet to do autoguiding with a Philips ToUcam. The program I use, called 'PHD guiding' is freeware. You can download it from the site Since I had such a camera lying there for a few years unemployed in a drawing, I decided to try out the autoguider. As a guidescope I use a cheap chinese 7cm refractor. I installed PHD on my laptop, and pushed the camera into the drawtube of the 7cm. And.... it worked allright.

The program is a simple "no nonsense" application without much features and gadgets, but it works reasonably well.

The only drawback in the system was to find a descent star on the chip. And the backlash of the chinese focussing system was also a thing that irritated me. So, The best solution was a flip mirror system. I removed the focusser from de tubus and replaced it by a solid alu tube from an old 5cm toy-refractor. To fit in the 7cm tube I had to turn a flange. The other side also got a flange for the ToUcam adapter. The flip mirror sits in a housing in MDF.

Here you can see the parts before they were glued together, and the finished device.

                       delen        flipdoos

The flip system is spring loaded as you can see here.



On Ebay i bought a Canon 350D secondhand for 200 €.
Of course my goal was to remove the original IR-blocking filter in front of the CMOS image chip.
That filter blocks all the interesting infra red radiation. And since i wanted to do astrofotography on some deepsky objects, that filter had to be replaced by another one.
So i ordered a
DSLR Astro Conversion Filter (baader ACF2) at 



On the internet there is really a lot to find about modifying a Canon.
Therefore it does not make sense to give a discription here, on how i did the job.
I simply followed the instruktions on the screen.
I prefered the english one by Ashley Roeckeleyn.

Here are a couple of URL's (these three are in dutch) ;

This one is in english ;

My experience is, that modifying a Canon is an operation that everyone with a steady hand, and a little skill in handling small pieces, can do.
However, some things one has to keep in mind.

Take care to work in an absolute clean area.
In my house we have a little hobby room which i vacuumcleaned carefully, and removed everything that was not needed there.

Take anti-electrostatic precautions.
Don't ware clothing in wool, and remove all plastic bags, foils etc. from your working area.
ware shoes with leather soles.
Make sure to connect yourself and your soldering iron to the earth. (via a 1 megohm resistor)
I use a special anti-electrostatic wristband.  (Every electronic store has it).
I even covered my working desk with aluminum foil wich was earthed too.

This all may sound ridiculous, but the CMOS circuits in most electronic devices are extremely sensitive to electrostatic discharges.
And remember, this modification is allways at your own risk !!! If your camera is destroyed, you are responsable and you can only blame yourself.

Make sure to have everything (camera, tools, small boxes for little parts, description etc. ) within hand reach.

Now take a deep breath and start working.

SUCCES !!!!!!!

This is a picture of NGC 7000 (North America nebula) with my modified Canon 350D
The picture is a stack of 20 exposures, each 1 minute long on 1600 ISO.



This cable is made for use with the program 'DSLR-Shutter' found on

Via this cable you can control the shutter of the CANON 300D and 350D.
These 2 camera's can't control their 'BULB' function via a pc with the Canon-software 'EOS utility'.
The software stops at shutter speed 30". The 'BULB' function has to be controlled via a 'remote-control'.
Since my pc uses its serial port as a link between a planetarium program and the DOS-pc (Bartels-scope),
and both USB ports are used for the mouse and the ToUcam, there was only one port left: the parallel port.
Fortunately 'DSLR shutter' has the opportunity to use this port.
So, i made a little interface that works very well. This is the schematic;


The simple circuit is build in a little plastic housing.
Use always shielded cable between pc and Camera.



One of the most important parameters in deep sky astrophotography is the amount of noise in an image.
The greater the ratio between the amplitude of a star signal and the amplitude of the noise, the better the image will be. To do astrometry, and even more to do photometry it is indispensable to have absolutely pure pictures, free of noise. Simply said, The challenge is to make, that an image of a star, a comet or an asteroide is not "drowned" in the background noise. A good rule of thumb could be that an object on an image has to have a SNR of 10 to perform astrometry, and a SNR of 100 for photometry.
A practical example;
Suppose an image with a sky background of 500 ADU (Analog Digital Unit). Now consider a small part of the image. Let's say an area of 10 by 10 pixels. Suppose that the difference between the brightest and the darkest pixel is 50 ADU. Consequently the SNR at this area is 500/50 = 10.
It will be very difficult to distinguish an object of i.e. 550 ADU. Because it will be drowned in the noise, who is 50 ADU itself. Considering above mentioned rule of thumb, we can say that a starsignal must have an ADU of 1000 for good astrometrie (1000/50 = 20  and that is 10 ADU more than the basic image SNR, who is 10), and an ADU of 5500 for efficient photometry.
Sky & Telescope febr. 1993 published a good article about SNR. The Author Anthony Mallama made a computer program to calculate the SNR. The program is written in basic, wich was a regular programming language at that time. Nowadays there are few people who has a machine to run this. Therefore I have transformed his program into an excel spreadsheet.

To use the spreadsheet you have to fill in some parameters. First of all, you have to know the "sky brightness" of your observing place. Type it in the upper blue cell. Subsequently fill in the "readout noise", the "thermal noise" and the pixel dimensions of your CCD camera into the yellow cells. After that, excel needs to know in the gray cells, the aperture diameter en focal lenght of your scope or telelens. Finally the last things you have to put into the purple cells, are the desired magnitude, the exposure time of one image, and the total exposure time of all images you think you might need. After entering the last time, you will see the number of exposures to make in the green cell, and the expected SNR in the red cell.


 HERE You can download the Excel spreadsheet

The "Sky brightness"
If you have an SQM (Sky Quality Meter) you can measure this parameter directly were you observe.
If you have not, then use the following values.

22   In moutains and deserts far from civilisation.
21   In the countryside far from city's and with very poor streetlights.
20   Not very far from city's and factory's with moderate streetlights.
19   In the outskirts of litlle towns and villages.
18   In the outskirts of big towns.
17   In the middle of a town

Sky signal constant
This is a constant, determined by the specific light sensivity of a standard CCD cell trough an objective lens of 1 inch diameter. Anthony Mallama calculated a value of 2,000,000 electrons par second. this value must decrease when filters are used. Sky & Telescope gives a value of 300,000 if one is using a V-filter. "V" stands for Visual. This means that we are dealing with a filter with a bandwith almost equal to the human eye.

Readout noise, thermal noise en pixeldimensions.
This data is part of the documentation that comes with the camera. But also on the internet you can find everything if you know the type of CCD sensor.

Detection cercle in arcseconds
3 arcsec is a good value here.