Author Options:

I want to model rays of light bouncing off a parabolic dish with technical drawing, How do I do it? Answered

Basically I want to model the sunlight as it bounces from both a parabolic and a compound parabolic dish. I want just to use technical drawing. It is no problem to model if the dish is pointed straight at the sun but I want to model the rays if the dish is NOT pointed at the sun.
On a curve, I will gestimate a tangent to an arc and light bouncing off goes off it at the same angle as it hits it.
NO problem it the light arrives parallel to the plane of the dish but most of it will not arrive like this.
How do I do it? Just with elevation plan and sideviews and sections through the dish.
Thanks Brian It is in relation to


. Your "tangent method" should work with any angle of incidence. Whether or not the light source is "on axis" shouldn't matter. . Not 100% sure, but no one else has answered. What happened to make you think it won't work?

Anyway, I finally did it with art of illusion software and I did it for parabolic dish, compound parabolic dish, hemisphere and winston w dish. (All different dish curves). None of the dishes are great!
I believe that if people put in a bit of effort, we can design a dish with a 2 3 4 or 5 hour acceptance angle for solar cooking or solar electricity generation.
That provides better concentration for a known time. I think that video shows the way forward. It will be shaped like 2 leaves joined at the back of the focus, The basic dish curve would be compound parabolic or winston w but there will be a little extra up curve close to the edges of the leaves. The leaves up curve will increase near to the cooking pot and may even form a tube close to it.
I think it will be really wort it to desigh this new dish shape.
I have done a combined trough alternative to it in the following video

why make a dish with a 2-5 hour acceptance angle? If you use a regular parabolic dish you can heat salt (which is very cheap) to store that energy. And molten salt as you may or may not know, will slowly dissipate the stored heat for hours. I believe the salt needs to be heated to 1000F to get to this molten stage. No problem with my dish..ive melted copper(1984.32F) 

Your dish is a precision instrument that requires constant tracking.  I agree it is an option. But it is  pretty high tech.  If you design a dish with a 2 hour acceptance angle, that can be good enough for a meal or to solar pasturized a large batch of water, that is a much more financially viable option. Carbon offsets are being used in India to provide parabolic solar cookers now. People are even paid to use them!
But users are complaining about having to constantly readjust them. And  every time you readjust, you risk flash injury to your eyes.
Providing a dish and pot with a 2 or 3 hour acceptance angle would be so much more practical, if your object is to reach as many people as possible.

you're absolutely right. And I am currently working on a program that will find the optimum shaped dish for the longest acceptance angle. But the problem with the acceptance angle dishes is that the solar energy gets dispersed more over a bigger focal point. which is why you can do nothing more than boil water. and thats great for people who cant get/have that luxury. But over here in the states, especially during this economic slump, 100s of thousands of people wont have heat this winter.
From all the research that I've done I can get a single axis tracker for $35 add that to the cost of supplies for my dish/heater that I got at Home depot. you're looking at maybe $80 + a window fan. I'll tell you something, I'd much rather pay $80, than the thousands I typically would spend for heating oil.
i dont mean to sound cocky when I say this, but I believe I can make a solar tracker for even cheaper than the $35 listed. Thats my next project after I finish that compound parabolic dish calculator/program.
Practical, you say? call me selfish, but i can already boil water. I guess its all a matter of perspective. I must say, i AM glad that there are people like you who are willing to help people outside the US.

I think your heating a material hot will also be useful for room heating.  You can have your Parabolic dish heating up the insulated salt or iron block all day, then at the end, wheel in the block, take off the insulation, cook your meal on the hot top, and then just let the heat radiate to keep the room warm. 
I made my "compound parabolic" dishes but later testing showed that they did not quite achieve what I wanted. 
The Winston w dish did best in my crude software simulations. (And since then, another guy has confirmed this with other software) but nobody has tested the winston w dish for real as far as I know!
If smarter people than me work on it a little, I think a 2 or 3 hour dish with good concentration can be achieved pretty quickly.

I found a great website just now that has a list of concentrators
and whats better is it has the potential output linked to each one.
When you design your winston dish it makes sense to me at least, to have it at one end, an extreme parabola ((2x^2) and then have it gradiate to a less extreme sloped parabola(.5x^2) on the opposite side. It may look a bit silly and you may or may not have to shift the focal line on one side but its my only logic on achieving a much higher acceptance angle time.

Theres only a couple ways to get a longer acceptance angle.... have more mirrors or a shape like the aforementioned one. with your application it  doesnt seem Ideal (for lack of a better word) to have more mirrors. and if you have that odd shaped one you can at least absorb the sunlight for longer at the expense of losing a bit of heat. I dont know...let me know what you think.

maybe this idea seems a bit out of the box, but i was playing around with 'snap switches' on my last solar oven. and what happens is when the heater gets to a certain temperature the switch (comprised of the same metal coil found in old thermostats) "snap" as to complete or disconnect the electronic circuit.
well if you take the same concept but apply it to a parabolic lens....you have x-shaped mirror and when the switch activates the x-shaped mirror snaps open to a newly shaped mirror to get those extra 2 hours or so of sunlight in. granted you will have to manuanlly close the dish in the PM so come morning time....
and honestly...having to adjust a mirror, who's stupid enough to stand between the mirror and the sun.... theres always a critic. :)

It is hard to actually measure the angle of incidence on a drawing except at the edges and then only if the light is going parallel with the paper. And I have to draw a tangent with the same problem too! I have decided to make a small model from cob with mylar stuck to it. Then I can shine the laser level down on it to simulate the sun. I can put the model on various angled mounts to simulate the sun moving across the sky and I can also mount it to simulate it being pointed not correctly. A bit of a pain though! Brian

The sun's arc across the sky stays in a single plane (that plane moves over the course of the year, but just pick a day :-). You should model that plane. If you're using a parabolic dish (circular), then it doesn't matter how the plane is oriented -- the dish is symmetric. You just need to specify in your drawing what orientation you've assumed.

My problem is that I want to model it as the sun is moving across the sky and the dish is not pointed exactly at the sun. Therefore the rays will not be striking the dish symetrically. The dish might be pointed up to 30 degrees off target (off the plane). That would be the very maximum and not something I want to do, but useful as an example. Brian

I think that doesn't matter, as long as you're using a circular dish. A plane-wave (parallel ray) source that is off axis to a true parabolic reflector will come to a focus at the same angle on the opposite side of the axis. If your dish isn't aligned with the plane of the sun's path (you don't have it pointed at local noon), then the focus will trace out an arc above the dish, on the side opposite the sun. The extreme points of the arc will be at sunrise and sunset, and those focal points will be on the east and west sides of the dish, respectively. The focal point at (local, not time-zone) noon will be the center of that arc. If you can mount your dish on an axle, then you can at least keep it pointed at local noon, and the arc will degenerate to a diameter.

The people at "art of illusion" have been extremely helpful. They have basically set up a "starter" file that you can just download (SunEngine2.aoi) that is already set up with a reflective surface and the sun in the right general area. and one of them has even a link to a tutorial he did! https://sourceforge.net/forum/forum.php?thread_id=2915545&forum_id=47782
It is a free download java program (not applet) that runs in windows and linux, even works on my acer1 in "limping" linux! If you want a parabolic dish, draw your parabola, and then use the lathe tool to spin the line into a surface!
I have just been playing with it so far. (Have not done any real work with it yet)
Finally some modeling software that I feel I can handle!

Thank you, I have blender and yafray on my computer and they can do all that I want them to do. (so say the software experts on blenderartists forums) and they show delightful tidbits of what it can do. Trouble is, ME! I have totally failed to figure out blender or any other 3d package! I will leave that to other people. Hopefully my pics will interest someone who is into that stuff. I will continue with physical models. I need to get a way to easily measure the widest angle where the dish gives a full reflection of the ball and also do some laser dot stuff to show where stray rays go and to figure a way of directing them to the middle. Perhaps next will be a winston double parabolic (like in the other peoples work section in the instructable) but spun around to be a dish like what I did. Brian

I would again strongly recommend using Rhino. It'll show you completely accurately in two minutes what two hours making a model will tell you roughly. I can show you (or anyone else interested) how to use the software to do this, it really is incredibly easy. As in, my grandmother (no disrespect) would get the gist of it in about 5 minutes. If you're already somewhat comfy with tech drawing, it'd probably take you a minute and a half. In fact, the exact process is this: Open Rhino. Select 'parabola' from the curve tools. Select the focus and drag out how big you want the curve to be. Select the curve and select 'revolve' from the surfaces menu. - you now have a parabolic dish. Click the 'reflect curve (or whatever it's called)' tool. Select the parabola and draw a line at it. - you now have a reflected line. And then you can start messing with the shape of the parabola and instantly see what effect it has on the optics. If you've got Skype or somesuch I can talk you through it all.

I have downloaded rhino. Long term though, I want other people to take ownership of the whole compound parabolic dish thingy. It is a very big area for research and the "big players" are just not interested. Thanks for modeling. why not upload a pic of what you saw? Brian

I was going to but the laptop on which I did it is currently on the fritz, and may well stay that way. I'll be shortly making a parabolic dish, which will require me to get back into rhino to develop the surfaces. Speaking of which, what would you recommend as a good, very accurate, way of making a mirrored parabolic dish? I was going to cut some Mylar into sections, stick them together, and then place them in a sort of parabolic cradle made of card slotted together. It's kind of hard to explain fully, maybe I should do an instructable on it! If it works...

I recommend making a compound parabolic dish because it does not need to be so accurately made or pointed so perfectly at the sun all the time. Brian

Mmmm, those things are quite interesting, I only just stumbled across the idea and I´ll look into it. From what I can tell tho the don`t have as sharp a focus as standard parabolas, which I need in this application.

I've actually been doing exactly this recently, but rather than paper I used a computer. You might want to do the same.

There's an excellent cad package called Rhino3D, you can download an evaluation version here:

It lets you (among many many other things, I use it mainly for 3d animation) bounce lines off'v modeled objects. If you contact me on sugarandfat at yahoo d0t co d0t uk I'll take you through the process, it's incredibly simple.

The results I got were that as the sun moves off to, say, the right, the concentrated light very quickly (as in degrees, not tens of degrees) swings out to the left and defocuses. I was modeling it with a smaller parabola at the focus to bounce the light back down at 90°, so the difference was much larger, but I wouldn't want to have the whole thing off by more than two or three degrees. Depends on what you're doing with it.

Oh, and, I'm not sure I completely follow what you guys are discussing here, but surely finding the tangent at a point on a parabola is simple. Draw a line straight down to your point, then from the point to the focus. A line that bisects these two will be the normal, 90° perpendicular to the normal is the tangent. Right? I think that's right...

Construct a tangent to each point, then copy the angle of incidence to construct the angle of reflection.

I can easily do it for the point a in my diagram but how do I do it for point x? For rays of light coming in at the same angles. Straight down is easy but what about 30 degrees off vertical for point x? Check the elevation and plan in the second image to see what I mean. Thanks Brian


. Instead of trying to measure and reproduce the angles, use a compass to copy the angles.

Thats what I did in the picture, The compass was lined up with a tangent to the curve and the angle was copied to the other side. The problem arises because I am also measuring tangents to a 3 dimensional surface. See the second picture. I should have elevation, plan and side view so that I can represent the rays in 3 dimensions but even if I did, I have no clue how to draw a tangent to a surface that is almost the shape of half an egg shell.

I meant that I used a protractor. You could use a compas but the problem is that you would have to draw a tangent every time. with the protractor, the bottom of the protractor is the tangent! A lot easier and quicker.