i have a solar lighting question ( heliostat )

[dacflyer]

i been trying to find some info on heliostat lighting.
(directing reflected light into the north side of my house, using a large mirror)
i been just for effect placing a mirror in the front yard. and it does make a big improvement.
anyway i planned to build a dual axis mount.. but my problem is, i cannot find out where the sun sensors are supposed to go.. do they go on the wall that the light will be shining on, or do they mount on the heliostat itself ?
i plan to get a controller from "RED ROCK ENERGY"
a few years ago i bought a solar tracker for a modified satellite dish with panels on it. the tracker was excellent.
but now i want to build me a heliostat. to direct light into the north side of my house ( living room )

can anyone provide any more info on this? i searched Utube for over 2 hrs. and i only seen demonstrations. or kits or complete units for sale.

[Externet]

Hi dacflyer.
I do the same to decrease the mold growth in my north side wall.

Sensor goes anywhere watching the sun target. But it is another way to think, a different story from solar panels. The sensor tracks the sun reflection, but the moving mirror will be at an angle and moves at half the angle the sun moves !

For a solar panel, when the sun advances -say 10 degrees-, the panel advances 10 degrees.
For a heliostat, when the sun advances -say 10 degrees-, the mirror advances 5 degrees.

Remember that you may have to track the reflection, not the sun. As putting the sensor by your window. When it sees no more mirror light, advances the mirror until it shines at the window again.

The angle depends on the location the mirror is and the location you want to illuminate. Yes, it is a mind twister. Just walk outside with a mirror on hand and keep it aimed to a fixed house window Then imagine the sun path.

But all the above does not work when cloudy

Years ago, I built a small one:
http://s588.beta.photobucket.com/user/I ... ort=3&o=29
http://s588.beta.photobucket.com/user/I ... ort=3&o=30
http://s588.beta.photobucket.com/user/I ... ort=3&o=31
http://s588.beta.photobucket.com/user/I ... ort=3&o=32
On a very different principle: Solar operated atomic clock controlled; no matter if cloudy or not, kept synchronyzed track of sun by celestial time. Did not use an optical follower for the sun position.

There is not much on the web, but this out of my favorites folder should please you somewhat :
http://www.iwilltry.org/b/projects/buil ... -lighting/

Miguel

[dacflyer]

thanks for the info.. your right, there does not seem to be a whole lot of info on them.
from what i have seen there is a way to put sensors on the heliostat itself..but how do you train it to track a certain spot is the problem that i cannot understand.
way back when i built my solar tracker, the sensors tracked the sun itself.
i used a old satellite ( C-Band ) removed the dish. and built a rack on it for holding 5- 100W panels
it tracked the sun perfectly. i have since sold it. but it was still a neat device.
i may get the urge to build another one to sell..

[Externet]

... i have seen there is a way to put sensors on the heliostat itself..but how do you train it to track a certain spot is the problem that i cannot understand...

A highly illuminated area on the wall can be recognized by a sensor(s) mounted there.
For the sensor to be at the mirrors side, will have to aim at the mirror, to track the reflection.
A sensor on the mirror side or at the target wall (or window) cannot discern if the sun moved or a cloud moved in between. If it is a cloud, advancing the mirror loses aiming.
Same as explained; cloudy days are not trackable by optical means. A clockwork mechanism would. A dedicated compfuser is another option if you like such complexity.

To discern if it is a cloud, would need more than one sensors to quantisize the amount of sunlight: if low because of a cloud, should disable tracking.

I suggest an almost better way. A convex or many convex fixed mirrors.

[dacflyer]

i think i got it now... for the way i want to do it,, your opinion on the sensors needing to face the mirror.
then it could be mounted on the mirror, then the sensors think they are looking at the sun, when it is actually a reflection...

the sensors i had on my solar tracker were very good. on a cloudy day the tracker would just ever so slowly
maybe 2-4 inches an hr. just search and hone in on the brightest spot in the sky. and on a really cloudy day it did quite well. when and if the sun came out..it would zero in on it in just a few min.
the kit i ordered was from RED ROCK ENERGY.. i thought i'd save a few bucks and build the kit..
you have to assemble it in a certain order for all the parts to fit. it is very Compaq.
i said next time, i'll just spend the few bucks and build the ready made unit..

[Externet]

... could be mounted on the mirror, then the sensors think they are looking at the sun, when it is actually a reflection...

Nope. If mounted on mirror, sensor would move too.
That is fine for solar panels, not for a heliostat.

Let me see if I find an example...

---->http://www.youtube.com/watch?v=RSjefWkZAzI ----->Which is the same as mounting the sensor far on the house wall.

By the way, my suggestion was a convex mirror; pay through the nose or do it yourself :

----->http://www.cleardomesolar.com/solareflexpanels.html

Which do you prefer ?

Miguel

[CeaSaR]

Problem I see is that you can't just track the sun as it's position changes during the year
meaning that the reflected light will wander from the intended target...

And, using reflections back to the mirror is a major alignment deal...

I suggest K.I.S.S. ...

Place 4 sensors on the building just outside the illuminated area, in a compass rose configuration.
When the light from the mirror hits a sensor, it drives the mirror mount in the opposite direction.
When no light from the mirror hits it, the motor stops. That will track the sun no matter what and
clouds will not affect it. You could even use the sensors to detect if enough ambient light is available
to signal daytime. If not enough light (ie night), it could return to a preset starting position.

My 2 cents,
CeaSaR

[dacflyer]

i still plan to build the 2 axis motorized version.
i just need to study it more..

[jwax]

You only need equatorial tracking, (sidereal drive) with seasonal tweeks to the azimuth angle.

[dacflyer]

ya, i know i could tweak it, but i'd rather have a set it and forget it operation.

[CeaSaR]

What's wrong with my analog approach, other than wire runs? Cat5
would work with a small bit of power on the house end. I guess KISS
isn't so KISS.

CeaSaR

[dacflyer]

well i was hoping for a self contained unit.. i was not really wanting to run wires across the yard..

[richfloe]

I had thought about building something similar for a different application. A mirror that would always reflect the sunlight to a specific point. Never built it (yet) but this is how I envisioned it working.

Build a stand for the mirror that can pivot up/down and right/left. Next to it, build a stand for the sun tracking sensors (described later) and link the two together so that the mirror moves exactly half the distance that the sun tracker does (gears or linkages). Adjust the linkages so that when the sun tracker mount is pointed at your desired target that the sun tracker is also pointed at your target.

Two motors are connected to either the sun tracker or the mirror mount, wouldn't matter which, that move up/down and right/left.

On the sun tracker, mount two plates that each have a pair of light sensors. On one plate they will be positioned one above the other, on the other plate they will be side by side. Place a divider between the pairs of sensors. When the sun tracker is pointed directly at the sun, all of the light sensors will be illuminated. If the sun moves, one of the sensors will be shaded. If you detect a significant difference in illumination on a pair of sensors, run the motor to bring the sun tracker back into alignment.

Provide limit switches to prevent the system from trying to track all the way down to the ground or beyond the area the sun could go. If the light level is low, disable the drive motors.

After an extended period of darkness (night time) or based on a timer that triggers in the middle of the night, run the motors back to the sunrise starting position so the tracker is close to where the sun will be. Especially in the summer, the sunset position may be so far from sunrise position that it might not be able to respond to sunrise and wont start tracking until well into the day.

Hope that made sense.

Also, mounting a solar panel to the sun tracker will give you power to run the system and since it tracks the sun, you will get max power from the panel.

Rich

[CeaSaR]

well i was hoping for a self contained unit.. i was not really wanting to run wires across the yard..

Unless the mirror mount will be run from batteries/solar, then you be running some
sort of wired power to the motors.

If not, how about wireless? What's the range of XBEE? You could use a couple of
small uP's, probably only need 4-6 IO lines to implement all functions absolutely
needed. Biggest power draw would be between the radio and the motors.

If not uP's, how about tone control? Encoder at the house, decoder at the mirror.
Motors only run when tone is present, and that would be very brief each time. Each
sensor would have it's own tone, so only 4 tones would control the movement.

Just throwing it out there for consideration.

CeaSaR

[haklesup]

What about just programming a lookup table into a microcontroller. you should be able to use your GPS coordinates with an app to determine the sun location any time of the year. you can either pre-tabulate that into memory and all you need to do is know the date and time and point in the right location. That would minimize hardware and maximize software.

see www.suncalc.net for inspiration or search on sun position calculator for more.

If you want to get fancy, you could probably adapt a low cost 7" tablet computer with GPS and either bluetooth or wifi to communicate the positions to the local controller although all you really need is time and date. I see multiple sun tracker apps in the play store, one may output useful data you can send to a position controller.