HeliosCap: Hard Drive Sun Tracker Heliostat for Fixed Telescope (Part 1)

If you love Astronomy, you love the night.. but.. what to do during a long boring day...wait until the next sunset? Not necessary, we can see also a star during daylight, the closest one.. our Sun, only 8 minutes light away. But how can we really enjoy Sun observation without the burden of using an expensive telescope?

Use whatever binocular, small refracting or reflective telescope in a fixed mount, you read ok, fixed!! Align it once to polaris star, and put your imagination to reuse whatever spare parts you have at home.

The present instructable has been my entertainment during this "looong" lock down (hope everyone reading this is sound and safe!). Furthermore, living in a small island in the middle of Atlantic ocean (La Palma) got worse the availability of internet shopping, as the island was really isolated (no flights, no ships for weeks).

But a good maker always look for parts anywhere.. so.. I asked myself.. what can I do with a crashed hard drive? I opened, dismantled it a bit, and saw a wonderful set of pristine flat mirrors. Yes indeed, they are really flat, (to avoid crash with the reader magnetic head), so they were polished on purpose ( I don't know yet how many parts of lambda of yellow light, but will check soon), and they are coated.. (don't know either the chemical composition) but read in literature that its reflectivity is more than 60% of the incoming flux. That is really poor, in comparison to telescope mirrors 80-90% of reflectivity, but.. wait.. when we look to the Sun (or even greatest planets), they have overwhelming luminosity. In fact with the Sun, normally we have to filter out most of the light, to prevent overheating.

Eureka!, it happen to be that I have several useless hard drives and I suddenly have converted them into my "great coelostats" collection... What's this? A coelostat is a mirror that one can aim with a steering holder to deviate light from one direction into another one. In the case of the Sun, we use the term Heliostat that help to reflect the light coming from our Star into a fix telescope. Why is good to have fix telescope installation. Because you don't need a huge tracking mount for it, so you can build a long focal telescope with permanent installation and use an Heliostat to track the Sun and bounce the light to your fix optical tube. Furthermore, then you can attach heavy cameras, spectrographs.. or even an Spectroheliograph to that optical tube, without having to steer them. What are all this "spectr... " terms? Don't panic.. that will be explained in second and third parts of this instructable.. but now.. let's go back to the basic. Let's build our Heliostat based on a hard drive mirror.

Step 1: Step 1: Look for All Necessary Parts

Being a "lockdown" project, you have to use "what you have", so look for available spare parts at your garage or bench room. What are the primary items to look for? (amazon links attached)

1. An optical device. Telescope, binocular, lense, parabolic mirror...(i.e. Skywatcher 90mm Maksutov **)
2. An out of use 3.5" hard drive (caution: don't expect this to work again after the project :-)
3. 3D printer with filament
4. Two stepper motors with drivers (28BYJ-48 stepper with ULN driver
5. Two GT2 20 tooth pulley
6. An arduino like MCU (e.g. NodeMCU)
7. Step down current limited regulated buck converter
8. 6v, 9v or 12v 3a power adapter
9. 6 wheels (3 Kossel k800 scroll wheels, 3 MR105ZZ vslot wheels)
10. The necessary screws and nuts.
11. A tripod or any other fix mount for your telescope.

Let's explain each part and its preparation on different steps

(** Note: each telescope model has different dimensions and the 3d design parameters should be updated accordingly before printing them)

Step 2: Step 2: Which Optical System to Choose

In order to use the present project you need to look for a spare telescope or binocular, preferably with the largest focal length possible, and an aperture around the size of your hard drive (the larger the better, but do not use much larger one).

In my case, I had a spare 90mm aperture, 1200 focal length Maksutov Cassegrain telescope. The quality is not the best but was very cheap on eBay (Seben brand but there are also Celestron ones). With the Sun, as there is plenty of light, we can enlarge the image as much as we like and still getting sufficient flux, the only limit is resolution.

The image will be blurred if you enlarge too much with a small aperture, so 1200 mm is enough for the effective aperture, that is really limited by the flat mirror (more on this later).

The second important characteristic of the choose telescope is that the tube (OTA) should have a smooth rounded cap, as our design is based on that cap to support the tracking rotation. Than is, if you find a telescope with a squared tube.. don't use this design.. it won't work. ;-)

In case you find a different optic size, you need to adapt the 3d design, what is not difficult as the provided code is parametric and you can adapt it at wish.

As the disk of a hard drive flat "mirror" has a central hole, they are better suited for reflective telescopes with a secondary obstruction. However, the best quality solar telescopes are base on a refractive lens optic, as the lack of obstruction result in a better contrast images, specially importan for Sun, Moon or Planets.

Step 3: Step 3: Looking for a Flat Mirror

Don't even think about it.. No, common flat mirrors at home (bathroom like) are not suitable for astronomy. Firstly because those flat mirrors are coated on its back side, in order to survive many cleaning, so they produce double reflection, one due to its back aluminium coating and other the from the front glass surface. This effect is more evident in dim light like as the one of an astronomical object, resulting in a double blurred image.

Secondly, because they are not really flat. They are produced with an industrial process that do not warranty its flat accuracy, as is irrelevant for home usage.

So, we need a really flat mirror and coated on its front surface. Surprisingly, they are more difficult to manufacture than a parabolic or spherical mirror of the same size, and so, more expensive. Why, because when you grind a glass with another as a tool, the one on top become concave and the bottom one convex.. but is very difficult to achieve a flat surface.

But.. we are lucky. Hard drive disks, as said before, are manufactured with flatness as requirement, as a magnetic head is going to fly over it trillions of times, and any surface feature would endanger the collision, lowering its live time.

Just, pick up a out of use, spare hard drive! 3.5" ones are preferible because size matter ;-)

We need to tear down the hard drive. It is not difficult at all. Just find 7 or 8 screws heads hidden under the plastic sticky label (see picture). You will need a Torx T6 screw driver.

After the cover, we can tear down its L shaped magnet over the flying heads shaft. It is very strong but it came out with brute force.

Later, we unscrew also the central aluminium disc, populated with another 6 screws and then, we reach our treasures! The hard drive platters. You will be surprised to look at them, as your image reflection is perfect. DO NOT use your naked fingers. Use dispensable gloves (sadly very popular nowadays). Our skin is always loaded with fat oil and we will leave our finger prints with a minimal touch. CSI guys will love them and they are permanent witness signatures of who touch your astronomy gears without your permission.

3.5" hard drive platters (technical name for hard disks) are actually 3.74" (95 mm) in diameter, so they are suitable for telescopes below that aperture.

So, dismantle them with care. Use only the first platter during initial tests of your parts, and leave the remains one on place until the product is finished, as they can be damaged very easily.

That is one of the most important innovation on this instructable. Front coated flat mirrors for free!. Sure you will have plenty of ideas of where to use an almost perfect 3.74"/ 95 mm flat mirrors!!!

Step 4: Design Mirror Motion to Track the Sun

Why do we need to steer the flat mirror? well, we are NOT giving an extensive astronomy lesson this time, but we will just summarise the requirement saying that the Sun is a moving target on our star background, basically because the earth moves around it, spinning daily tilted 23.5 deg around it every year.

As a result, we know that the Sun elevation over the horizon at noon varies plus 23.5 deg on the summer solstice (maximum elevation) and minus 23.5 deg in winter solstice. The exact Sun position against the star background is defined by what we call Right Ascension and Declination, and change every second of the year. The mean sun elevation at noon of the spring or autumn equinox depends on your location:

Sun elevation at equinox = 90 deg - your location latitud

So, we need a play of twice 23.5 deg (47 deg) around the squared angle of the earth spin axis. That implies that if we install our telescope align with polar north or south, we will need to steer the flat mirror 47 deg from a 45 deg of inclination from our telescope lens.

That derives our first design (black tube with animated gif). It was a hollow tube with a transparent ovale hole that will allow to receive light from the Sun every day of the year, depending on the Sun declination.

It will have two independent motions. One on the Z axis for daily Sun motion (earth spin) and the second for Sun declination for each day.

Being a circular mirror, the platter from a hard drive, is not the optimal shape, and will result in an elliptical illumination of our telescope. This is not good, but just acceptable. In the future, I will replace this flat mirror for an elliptical secondary of a newton telescope. It will cost much more, but that shape is the one we need to get a even illumination of our optic.

Step 5: Design and 3d Print Your Parts

To simplify the design, only the structure of the mirror has been created, leaving the holding tube for later phases. The structure consist on three main parts:

- Flat mirror (platter) holder with declination gear

- RA bearing (Right ascension axis) including declination motor holder

- RA motor holder around OTA tube.

All the parts has been designed in OpenScad. Why? I know that there are many tools around. I am coded oriented (more than graphic designer), and OpenScad is parametric, besides is opensource and free.

The resulting design is not the best, but it is functional and, furthermore, can be adapted by anyone by modifying variable values.

The first iteration (gray piece) was good to solve declination axis. It is very light, and includes two spare bearings that I have from my 3d printing hobby. You can choose any other and modify the OpenScad parameters, but I used 625zz ones.

But the first iteration fail on RA design. Reason being that I want to use the telescope OTA edge as support for that axis, and I was too optimistic on the friction of a PLA plastic piece over a rubber surface of my specific telescope tube. FAIL!

I even try to insert a PVC slide to soft a bit that friction. But .. no way!.. its too much for the torque produce by my chosen motors (more on this later).

In a second iteration, I wanted to solve RA issue by both, enlarge the diameter and number of teeth of RA gear (now 240 teeth) to get better torque, and also add bearing wheels all around.

Ideally one would require a large ball bearing with a inner clear diameter equal to the size of your telescope lens. Those large ball bearing exists.. but they are rare and very expensive (several hundred bucks).. so, that was out of the budget and lockdown situation (covid19 season).

So it comes the second main innovation of this instructable. How to design a large ball bearing with a 3d printer. There are many tries using small balls (plastic or metal) that I did not have available, so again, a look for at my garage and found several vslot wheels from 3d printing building activity, and.. they were perfect. Small, soft with rubber cover and integrated in a ball bearing.

We have chosen two models, smaller ones for radial loads are three MR105ZZ vslot wheels, and the 3 more for the axial load (Kossel k800 roller scroll wheels of 19 mm diameter), but any other can be used, if conveniently adapted in the parameterised OpenScad code.

The design allows the wheels to roll over the telescope cap edges, in radial and axial direction, so the flat mirror can turn freely and softly behaving as an very large ball bearing. This design can be use to generate any size ball bearing interface to any telescope, or other projects. Only limitation comes if you live near the equator, as the polar north is very near the horizon and the telescope tube is almost horizontal, so theoretically there is a possibility that the mirror frame could slip out of the tube. In our case, there is a further restriction due to the pulley of RA motor covering the RA printed gear, so slip is possible for any latitude.

To print the parts you have to download scad (OpenScad files) from github HeliosCap repository, activate each part (commented line with asterisc), and generate STL files. Then, use your favourite slicer software and.. print it.

This is a short cut of the scad file parameter that you may modify, adapting for your materials:

use ;

h1=30; // height of cap adaptor ring

tod=110; // telescope tube outer diameter.

toh=35; // telescope tube last cylinder height

d2=98.5; // telescope original cap insercion diameter

w1=2; // cap wall thickness

w2=2; // mirror baffle thickness(black);

mod=95; // flat mirror outer diameter.

mid=25; // flat mirror inner diameter

a=sin(360*$t); // parameter for animation (-1..1)

dec=23.5*a; // Ecliptic inclination 23º 27' 8" (47" less every 100y)

mt=1.26; // mirror thickness

ma=45-dec/2; // mirror angle

ms=80+h1; // mirror spacing from telescope

h2=ms+mod*0.6; // baffle height

bm=3.0; // beam margin. free space around beam

bid=5; // 625zz bearing inner diameter

bod=16; // 625zz bearing outer diameter

bt=5; // 625zz bearing thickness

brd=7.8; // 625zz bearing rotating diameter

bw=7; // GT2 belt width

dec_teeth=150; // GT2 teeth on DEC gear

ra_teeth=240; // GT2 teeth on RA gear

ra_d=2*ra_teeth/PI; // ra diameter.

l=ra_teeth*2 in GT2. l=2*PI*r

r=ra_teeth*2/(2PI)*2

wd1=15.25;

wd2=13;

wt1=5;

wt2=8.96; // wheel thickness

wh=4.9; // bearing hole diameter

wbd=10; // bearing diameter

wht=3; //wheel holder thickness

PLA is ok, and support is only necessary for the Dec motor holder overhang ring, so piece is relatively fast.

Yes, you are right, bright green colour shown in pictures is not the best, but it the filament that we have during the lock down. Recommended colour is black, as will avoid any reflection, specially in flat mirror central holder.

Step 6: Electronic Configuration and Programming

This step is in very early stage, although it barely allows to test the project and track the sun. We have used a NodeMCU ESP8266 (what we had) and not an arduino, because we intend in part two, to control the setting from our cell phone via wifi. The code is very simplistic. We just use library AccelStepper as is a very robust and flexible code to manage stepper motors like the inexpensive 28BYJ-48. We just configure it with 4 duponts wires using D0,D1,D2 and D3 gpio pins of NodeMCU, and connect them to the ULN driver 1-4 pin header.You need also to power both, NodeMCU and ULN drivers (RA and DEC) with 5v, using the output of the step down buck converter that lower the voltage of the 12v power adaptor.

The pictures are taken fro,m an excellent tutorial about this motor here that we recommend you to read, and we are not repeating.

The .ino code is included also in github HeliosCap repository, where it will be upgraded with more controls to point, track and hopefully even guide in later versions of this instructable.

In the code, there is a very simple calculation of the delay among RA steps, taking into account that we need to use HALF steps every 1766 ms taking into account that:

- The motor has a very complicated gear reduction ratio (see picture) of about 63.68 times, that multiplied by 64 step per rev provides 4075 (aprox) steps per revolution.

- RA gear has 240 GT2 tooth (to allow the use of GT2 20 tooth pulley).

so taking into account that in average the Sun turn once every 24h (this is an approximation that will be upgraded to real Sun daily RA as it change every day), will result in an step every 1766 ms.

We encourage you suscribe for updates to the code and contribute if you like with new features ;-)

Step 7: To Be Continued..

This is a really nice project to make, it combines astronomy, with optics, mechanics, 3d design and printing, electronic and software development, so there are many areas of improvement. This was just a working model to test the capability.

There are many ideas to develop a permanent installation for museums and planetariums allowing them to show the live Sun image, static projected by a HeliosCap with a fix telescope, or even install heavy spectrograph to show live Solar spectrum and... large list of dreams...

We have just scrap the surface of a larger project.. come in and join us with your ideas!