Author Topic: Mechanical Design  (Read 29888 times)

Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Mechanical Design
« on: March 16, 2011, 09:40:15 PM »
Hi Guys,

I'm trying to put together a heliostat design to use with Gabriel's Arduino program. I'm impressed with what has been done so far with simple hardware and wood structures, but I would like to come up with something a little more solid and permanent while still keeping costs as low as possible. There are a few basic design parameters that I will need to nail down before much progress can be made. I'm hoping you folks can give me a little feedback.

1. What is the minimum resolution  needed? I know I can do better than say 0.1 degree steps, and .01 is probably not that difficult, but how good is "good enough"? Is there any disadvantage to really fine resolution?

2. How about repeatability? Any slop or flex in the mechanism will affect this of course. It wouldn't be very hard to design a really tight and rigid system using precision motion components but it gets expensive in a hurry. Again how does one decide what is "good enough"?

3. How big should the mirror be? I'm thinking about using 12x12 inch mirror tiles since they are cheap and available, but what is the trade-off between multiple small machines and fewer bigger ones? Clearly there is no simple answer to that, but it's something to ponder. If the electronics and mechanism can be made cheap enough, I like the idea of smaller machines that can be deployed in numbers to meet the task at hand, as opposed to fewer big, heavy, expensive machines. Plus, starting small will give me a chance to get everything working and gain some knowledge which can then be applied to something bigger if desired.

Any thoughts/ideas? I'm looking forward to being part of this group and seeing what we all come up with.

Regards,
Jim


zininzelfdoen

  • New Member
  • *
  • Posts: 21
    • View Profile
Re: Mechanical Design
« Reply #1 on: March 18, 2011, 10:01:20 AM »
Hi Jim.

We @zininzelfdoen are working on a system that uses two linear actuators to move 3 mirrors of 0.5 m2. We will still have an interesting (experimental) job to adapt software and electronics for using sattelite actuators. Maybe we can help eachother with this.
These are the actuators I ordered, they are really cheap. Believe me, we spent ages finding them for this price. Uptill now we are working on getting the normal software to work. Later we will try to make it apt for the new design.

http://www.miracle-moon.nl/shop/satelliet/accessoires/draaibaar/jaeger-superjack-10-inch.php5

kind regards,
daan


Gabriel

  • Administrator
  • Hero Member
  • *****
  • Posts: 648
    • View Profile
Re: Mechanical Design
« Reply #2 on: March 18, 2011, 03:00:24 PM »
Hey Jim,

I'll try to answer your questions based on what I've noticed so far with my own heliostat.

1. The minimum resolution in large part depends on how far away your target is. The window I reflect the light into is about 50' away from the heliostat, and I think that 0.1 degree steps would actually probably be good enough for it. The sun moves 0.25 degrees per minute, and I have my heliostat update its position every 3 minutes. Even though it is 0.75 degrees off by the time it finally updates, it still stays on target fairly well. You can see this for yourself in the time lapse video I made on the heliostat projects homepage of this site.

So it's up to you how far you want to go with the resolution I guess. More certainly wouldn't hurt. The only disadvantage I can think of is that a higher resolution heliostat might be slow compared to a low resolution one. Not a big deal once everything is up and running since the sun itself isn't very fast, but it can be a pain when debugging if you have to wait a long time for the machine to reset itself.

2. From what I've seen with my heliostat, rigidity (or in my case the lack of it) is not too terribly detrimental to the heliostat's success. I can forcibly move mine along either the altitude or azimuth directions a degree or two, but it still ends up returning to its original position and reflecting the light to correct spot. Granted, that spot is only 50' away. Like before, it mostly depends on how far away you'd like to reflect the light. Make it better than I did obviously, but don't break the bank to do it.

3. I like the idea of several smaller machines too since they are more discrete. Something large would look odd (or more odd) if it sits in the front yard like mine does. Somewhere between 1-2 square meters might be about right?
I would make it as large as possible without being unwieldy, if that makes sense.

As far as keeping the costs down is concerned, I think it really just comes down to sitting down with some actual numbers and doing the calculations to see what size turns out to be the cheapest.
A few things that I have found out when building mine are:

Hook up wire for the stepper motors is one of the more expensive items (unless someone knows of a cheap source that I wasn't able to find), and the more heliostats you have the more you'll need of it.

Making the electronics and wiring everything together for each heliostat is very time consuming.

Aligning each heliostat and programming the targets is also very time consuming. I have ideas for making this task go faster by adding features to the Arduino program, but either way it will still take time out of the day.

Like you said, I would start small and first see if you can just get everything to work right.


Take Care!
Gabriel


Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #3 on: March 19, 2011, 06:46:36 PM »
Thanks for the input guys.
Daan, I have seen actuators like that here in the US, but never seriously considered them for my project.  A DC servo type drive like that may be more appropriate for the bigger systems, but for now I'm going to stick with steppers. I am more familiar with them and also I have some left over from my CNC projects. In fact I have a couple dozen smallish NEMA-23 motors that are too weak for anything but the smallest CNC machines. I'm thinking they should be powerful enough to handle a 1 or 2 square meter mirror depending on how I do the gearing.
Speaking of gearing, I'm thinking about using worm gears. I discovered a real simple way to make them using a spiral tap to cut the teeth. They can give extremely high ratios in a small space, and because they won't "back-drive" you can de-power the steppers when they are not moving, just as you can with screw drives.

Another "geometry" question, how big of an issue is having the mirror hinged near the glass surface? I understand there will be some non-linear "shifting" of the reflection if the hinge is at some distance from the mirror, but at what point does this become a problem? Can the software be set up to compensate?

Gabriel

  • Administrator
  • Hero Member
  • *****
  • Posts: 648
    • View Profile
Re: Mechanical Design
« Reply #4 on: March 20, 2011, 06:05:14 AM »
Any errors added due to the mirror being hinged away from the glass surface will be negligible, so I wouldn't worry about it. I've played around with this before in Sketchup using simulated reflected lines, and didn't notice an issue.


Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #5 on: March 21, 2011, 11:27:29 PM »
Well I did a little CAD doodling, and here's a concept for an AZ-EL drive using NEMA-23 steppers and worm drives. The cross shaft would be supported at the ends of the big square tube by a bushing in a plate that is not currently shown, or maybe just a solid plastic plate with a hole in it. I'm thinking steel rod or maybe even black iron pipe for the shafts, threaded rod or a sawed-off bolt for the worm screws, and brass or plastic for the pinion gears. The axis shafts would turn in plastic or bronze bushings where they pass through the square tube and end plates. Right now I'm showing a 12x12 inch mirror, but it could be scaled up to whatever the motors and gear-train could handle.
Any thoughts/comments?
Jim

Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #6 on: March 21, 2011, 11:34:15 PM »
More pics.

Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #7 on: March 21, 2011, 11:34:48 PM »
One more.

Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #8 on: March 25, 2011, 11:38:26 PM »
I have a question regarding the electronic hardware.
When I went to order my real-time-clock module I came across some comments on the Sparkfun site regarding the accuracy of the DS1307 module.
Check out the comments on the bottom of this page:
http://www.sparkfun.com/products/99
Any thoughts/comments? I don't mind spending a little more for a better clock if it will help keep things tracking better.
I assume it would take some tweaking of the Arduino program though, and my programming skills are rudimentary at best.
So, I'm thinking I'll just stick to the DS1307 and existing code, unless someone wants to update the program for a more accurate clock.
Jim

PS. An update on the gearbox design - I had the idea that using standard threaded pipe nipples and fittings from Home Depot would be a simple and cheap way to do a prototype, and would be easily reproducible by others since the stuff is so readily available.  2 trips and $50 worth of plumbing parts later I'm still not quite there. The main trouble is that the fittings and tapered threads are very non-precision so I need to figure out a way to compensate. So... I still think my general concept is workable but I'm still scratching my head over what materials to use.

Gabriel

  • Administrator
  • Hero Member
  • *****
  • Posts: 648
    • View Profile
Re: Mechanical Design
« Reply #9 on: March 26, 2011, 05:35:46 AM »
I think some of the comments there are an exaggeration or quite possible made by individuals with a faulty chip, but supposedly there will be some time loss/gain as the months go on. I was a little disappointed that Sparkfun didn't make that more clear when I first bought the chip, otherwise I might have upgraded to something more accurate myself. That said, my RTC has kept the correct time for about six months now. Worse case scenario, I just have to update the time every 6-12 months. I don't mind that since I'm always tinkering with the program anyway, but other people might not want to do that.

The same chip is available at this link. It's a little cheaper, but it needs to be soldered together. This is what I'm using now since I replaced the one from Sparkfun after I thought I fried it by accidentally wiring it wrong. Turns out I just drained the battery though. After replacing it it works again. Always nice to have a backup I guess.
http://www.adafruit.com/index.php?main_page=product_info&products_id=264&zenid=610967d039a114f29a28fe12c82de4a9


I just noticed that the same site also has another clock module which is supposedly more accurate.
http://www.adafruit.com/index.php?main_page=product_info&cPath=42&products_id=255

According to them, "The I2C interface is very straightforward and virtually identical to the register addresses of the popular DS1337 and DS1307 RTCs, which means that existing code for the Arduino, Basic Stamp, Cubloc, and other controllers should work with no modification."

So I guess that it should work with the Arduino Sun Tracking / Heliostat program without having to rewrite any of the code. If you try it and it works, let me know so that I can recommend it to people.

Good luck with your design!

Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #10 on: March 30, 2011, 12:09:39 AM »
Thanks Gabriel.
I dropped the water-pipe version for now, but may pick it up again later. It would be nice to come up with a design that can be built "tinker-toy" fashion using Home Depot parts, but for now I'm going to just buy some raw stock and do a little machining since I have the tools. Once I get something working I can focus on making something easier for those with more basic tools.
Looks like I'm going with the Adafruit clock. That's where I bought my Arduino and they were easy to deal with. Was thinking about trying their "motor shield" as well after looking at Bernard's postings, but your transistor driver is so simple and the code is proven so I'll probably stick with that. For some future machine I might want to go with a more powerful, commercially made stepper driver. To that end, have you thought about doing a "step and direction" type output to accommodate this?
Regards,
Jim

Gabriel

  • Administrator
  • Hero Member
  • *****
  • Posts: 648
    • View Profile
Re: Mechanical Design
« Reply #11 on: March 30, 2011, 06:52:01 AM »
Hey Jim,

The code is based off of the Arduino Stepper motor library (http://arduino.cc/en/Reference/Stepper?from=Tutorial.Stepper) , so it would be easy to switch to a "step and direction" driver board.
At the top of the Arduino_SunTracker tab, you'll see "Stepper Motor Setup" section of the code.

Just replace the

  //Pins used for altitude stepper motors
  Stepper altStepper(int(steps), 2, 3, 4, 5);
  //Pins used for azimuth stepper motors
  Stepper azStepper(int(steps), 6, 7, 8, 9);

with

  //Pins used for altitude stepper motors
  Stepper altStepper(int(steps), 2, 3);
  //Pins used for azimuth stepper motors
  Stepper azStepper(int(steps), 4, 5);

and it should work with most of the typical two pins per motor driver boards.


Take Care,
Gabriel Miller

Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #12 on: April 03, 2011, 11:17:47 PM »
Thanks for that info Gabriel. I'll refer back to it when I get to that stage.
Well, I finally cut some metal. See pics below. I still need to make the pinion gears, bore the end plates for the crass shaft, and add some bushings. Also need to figure out a mount. Since I already have a bunch of galvanized pipe and fittings from the first aborted attempt I think they'll get used for that.
I'm moving and starting a new job next week so might not make much progress for the next couple weeks. Will keep chipping away and ordering parts though.
Jim
« Last Edit: May 26, 2011, 11:28:13 AM by Jim McMillan »

Gabriel

  • Administrator
  • Hero Member
  • *****
  • Posts: 648
    • View Profile
Re: Mechanical Design
« Reply #13 on: April 04, 2011, 06:55:56 PM »
Wow!

That's looking really good! It looks like it should be very precise too. I can't wait to see how those worm gears come together.

Congrats on the new job!

Jim McMillan

  • Jr. Member
  • **
  • Posts: 61
    • View Profile
Re: Mechanical Design
« Reply #14 on: April 06, 2011, 12:42:09 AM »
Thanks. I think I went a little overboard using 1/4 inch wall tubing, in fact the whole thing seems like a bit of overkill. I figure I gotta start somewhere and I can always lighten it up in the next version, or keep adding mirrors until something breaks.
FYI - the ACME screws I'm using are 10 TPI and I'm shooting for 36 teeth on the pinion gears. My motors are 400 steps per turn. If I did my arithmetic right that's 0.025 degree per step.
Jim