Author Topic: Hi from Nice, France  (Read 1470 times)

jlp6k

  • New Member
  • *
  • Posts: 2
    • View Profile
Hi from Nice, France
« on: November 21, 2013, 04:51:26 PM »
Hello there,

I'm Jean-Louis Paquelin, I live in Nice on the french riviera where there is plenty of sun.

I'm trying to build a cheap heliostat with the following characteristics:
  • less than 100
  • a flat mirror up to 50x50 cm
  • user friendly (e.g. with a web interface)
I've built a fully functional proof of concept with an aluminum pan-tilt head actuated by two servos. Firstly, the servos were driven by an Arduino.
The mirror was oriented by computing the sun position instead of catching it with 4 photoresistors as it is often done on small project like mine.
But I wasn't satisfied with the user interface as using the Arduino terminal isn't very user friendly. So I switched to a board more powerfull and yet ready for network connection, the raspberry pi.
The servos were then controlled with a small USB interface from Pololu. I rewrote the code in Python and added a simplistic web interface.

I was very pleased so far but I discovered that when I tried to replace the 12 cm (a CD) I was using as a mirror by a larger plastic mirror, the tilt servo started to buzz showing it had hard time to keep the mirror at the right angle.
I tried to use lighter support material with sticker mirror but I ended up concluding the tilt overhanging is a bad idea. However, I designed a pan-tilt head with counter-weigths compensating the mirror weigth, but I dropped the idea as I saw another drawback with this idea...

This year there were two tornado-like episodes with 200+ km/h wind. These winds ripped tiles off my roof and I thought that I should consider this for my heliostat. Even with a balanced pan-tilt head, the servos will surely break on high wind and may suffer from even the lightest winds.

Here is where I am.

After a pause, I restarted this project looking for other ways to build the mechanical part of my heliostat and I think that linear actuators can solve problem of wind resistance.
In a direct driven mirror (as with the servos) the wind (and even the weigth of the mirror) apply a radial force against the servo axis and may crush the gear train. In a linear actuator driven mirror, the wind force is transferred axially against the shaft mostly in a static way.

Then I (re)discovered the Cerebral Meltdown website where I read a couple of time that using linear actuators isn't such a good thing. I didn't find why yet but I'm still digging the forum.

Even if it doesn't solve the wind issues, balancing the design is a good idea. I drew a couple of sketches involving two frames (more of less like Gabriel's second prototype http://www.cerebralmeltdown.com/projects/heliostat_num1/default.htm).
I haven't built it but as it is still direct driven, the servos won't last long.

I like the servos as they provide a kind of absolute positioning but they are not well suited to power linear actuators or worm gears. And if you think about continuous rotation servos, you'll lose the absolute positioning. Using stepper motors will be a better option in this case and this seems to be proven by many heliostats on this website. But I don't want stepper motors as they are quite expensive (remember my 100 limit) and they may silently lose some steps in case of hard mechanical strain.

So I imagined a new approach which may be cheap, efficient and easy to setup.
I plan to use simple DC motors with worm gears (to get reduction and static transfer of the wind force) and I'll add an accelerometer+compass sensor on the mirror to get some feedback of its position. Then I may be able to control the motors thru a PID controller.

Before getting my hands dirty, I'll wait for the comments, suggestions or ideas of this distinguished gathering.

Thanks,

jlp


Paul L

  • Jr. Member
  • **
  • Posts: 98
    • View Profile
Re: Hi from Nice, France
« Reply #1 on: November 21, 2013, 07:41:05 PM »
Hey Jean-Louis,

    I wish I was in the French Riviera right now - it was -3C this morning in Victoria, Canada...I can almost taste the snow! :)

    The main reason I don't like the idea of using linear actuators for my heliostats is because it limits the range of motion.  I don't have my heliostats set up now, but if I remember correctly, the maximum Azimuth I could achieve was around 60 degrees either side of directly South, and for summer, that doesn't capture all of the possible solar potential in the mornings and evenings.  On the plus side, it's cheap and (fairly?) easily put together.  Worm gears would be a much better setup, but they are more expensive, and you can't buy them off the shelf...

I doubt that anything would stand up in 200km an hour winds though.  At that speed, I'd just head to the basement and thank my stars that I survived!

Anyways, I think building a heliostat for that 100 Euros, even with using steppers, is definitely doable.  There are some pretty good deals on Ebay if you keep your eyes open.  I'm pretty sure that I built mine for less than that - though I did lower my costs by buying in larger quantities and building three. 

If you do proceed with DC motors, it'd be nice if you detailed your progress here - like you say, steppers are more expensive - and I'm always interested in lowering costs.  :)

Cheers,

Paul


Gabriel

  • Administrator
  • Hero Member
  • *****
  • Posts: 650
    • View Profile
Re: Hi from Nice, France
« Reply #2 on: November 25, 2013, 05:10:40 PM »
Hey Jean,

Like Paul said, using Linear actuators limits the range of motion, which is the main reason I'm trying to steer away from them. If you use the same sort of DIY ones that I did, you also might run into the same issues with freezing rain that I did. Basically, ice would cover the threads and keep the machine from moving.

I have heard a few people mention that they were going to try and build a heliostat using servos or DC motors because the aren't as expensive. I haven't actually seen much of a price difference between the two though, provided you compare similarly sized motors.
I bought some of these awhile back, and they seem to be able to do the job fine, provided they're geared down.
A couple more links to compare prices.
Stepper motors: http://www.phidgets.com/products.php?category=23
Servo Motors: http://www.phidgets.com/products.php?category=22

I have always wanted to see someone find a way to use DC motors though, mostly because it seems like it is a lot easier to scavenge DC motors than stepper motors (although I guess even DC motors are going to become hard to find as we move further into the digital age). I had at one point considered using a potentiometer for positioning, but never got around to trying it out. I was a little worried that it might not be accurate enough, but it might be fine or possibly wouldn't even matter depending on the application.

Gabriel

jlp6k

  • New Member
  • *
  • Posts: 2
    • View Profile
Re: Hi from Nice, France
« Reply #3 on: December 22, 2013, 05:54:17 AM »
Hello,

Thank you for your comments about (not) using linear actuators.

So I'll probably use DC motors. But this makes me thinking about servomotors again. After all they are DC motors with some gears and a potentiometer to provide some feedback. I've found some cheap heavy duty servos on the web http://www.hobbyking.com/hobbyking/store/__84__189__Servos_Parts-All_Servos.html but I don't know how reliable they are. I'll probably give them a try.

Anyway, I'm not at the mechanical/electrical design yet. I've bought a Pololu 6DoF magnetometer+accelerometer board http://www.pololu.com/product/2124 to try the idea I exposed before. I try to have an absolute positionning from (magnetic) north and ground. Elevation angle is given by the accelerometer.  It's accurate and easy to handle. Azimuth is given by the magnetometer.
It was quite easy to program a toy stabilized pan-tilt device. But the azimuth value is biased by the position/tilt of the sensor in space and the magnetic sources in the vicinity (e.g. my pan-tilt head servo motors or my laptop harddrive).
There is a handfull of tilt compensation algorithms for magnetometers. They are used in quadcopters and smartphones but they seem to have a limited range of efficiency (I've read it's around a +/-40 degrees tilt). I have to work on that but alternatively I could buy a separate magnetometer and accelerometer affixing the magnetometer on the panning part and the accelerometer on the tilting part.
In order to avoid (or correct) the magnetic effects of the suroundings, I may use a lookup table of corrections mixed with an interpolation between rows of the table.

Questions, comments and advices are welcome.

Happy holidays,

Jean-Louis

Gabriel

  • Administrator
  • Hero Member
  • *****
  • Posts: 650
    • View Profile
Re: Hi from Nice, France
« Reply #4 on: December 27, 2013, 07:45:17 PM »
Hi Jean-Louis,

I have zero experience with magnetometers or accelerometers, so I can't really give any advice. I don't think I have encountered anyone else who had tried to build a heliostat using this method, so you may even be the first. The closest that I can think of is Bob's on this thread, but it still doesn't seem very similar though. http://cerebralmeltdown.com/forum/index.php?topic=349.0
It would definitely be interesting to see how well it works out.

Gabriel