Go Down

Topic: Garage door, getting the position (Read 4938 times) previous topic - next topic

bld

#30
Jun 13, 2010, 10:22 pm Last Edit: Jun 13, 2010, 10:25 pm by bld Reason: 1
I don't think that is a big problem. What you see there is what is gathered over a 5 year period, and I haven't cleaned it a single time. The owner before me was 96 years old, and I am pretty sure he didn't either.

The thing with all the webs on are also only the end that is going down on the ground, there are only a tiny one in the area where the wheel are, and not even close to it.

I think it is the best shot I got, no matter where I look, there are something in the way for my plans. Also tried to look at where I was planning on adding the rotary encoder, and I can't do that either, because it got a protective cap there, and if I remove it there can get dirt down into it, plus it will loose half of its support keeping it in place. So I am starting to run out of options :( All that damn safety, impossible to work with...
captain-slow.dk | non contagious!

zoomkat

With 196cm travel and 5cm accuracy desired, you would need ~78 sensors. Sounds expensive and complicated. I suggest you duplicate the door spring/pully system internal to the door, externally, so a single pot can be used. Using a long skinny screen door spring (or even some thin bungi cord), some wood and nails, a length of string, and a couple of other incidentals, one could construct a gizmo that would stretch the small spring with a slider pot attached similar to the big door spring. I like to keep gizmo experiments less than $10.  
Google forum search: Use Google Advanced Search and use Http://forum.arduino.cc/index in the "site or domain:" box.

bld

The problem is just to find space to actually do that. If you look at the picture, you can see the metal wire going in the box where the spring is in the end. That box is going a bit down, and the wheels are then running inside there. So there are no room for anything on the outside.

Tried with some string before, and it almost got squeezed into the rails where the wheels are. Would then have to drill a hole in top of the box so it can go out there, the next problem is then to connect it to the door, because there are no room there because of the outer guiding rail..

I could then just connect it to the other end, but I am going too look at it at least twice daily, plus it is very open there, so it would be very easy to see, and I would like to make it close to invisible. Think I will go to bed and see if I get some ideas during the night.
captain-slow.dk | non contagious!

zoomkat

Keep the string clear of and external to any mechanical areas of the door. You could connect the end of the string to the top of the door, and the mini pully/spring/pot setup on the back wall of the car port. As the door closes the string would be pulled along the ceiling. As the door opens the string would be retracted by the spring/pully setup. Note the operation of multi sheave block/tackle setups in the below link. Long pulls of the rope result in a small amount of travel in the block. For the door, the string end would be connected to the door top, and the spring to be streched would be connected to the block. As the actual load is only sufficient force to retract the string, it probably can be lightly built. That being said, I think in the end you will really only find door closed/not closed of any real value.

http://en.wikipedia.org/wiki/Block_and_tackle
Google forum search: Use Google Advanced Search and use Http://forum.arduino.cc/index in the "site or domain:" box.

bld

Quote
That being said, I think in the end you will really only find door closed/not closed of any real value.


Yeah, but that value isn't even for me, it is for the arduino. I am planning on adding a lot of things to it. I want it both to measure rain, and if it starts to rain, it should then close the door enough so it wont rain into the garage, but still allow air to pass under. Also measure humidity, so if it gets too high, then open the port a bit, if it doesn't help, then open it some more, but still without allowing it to rain in.

Going to look in some shops tomorrow for stuff that can be used, and how much it costs.
captain-slow.dk | non contagious!

mrmeval

The more pulses an encoder puts out the more precise your measurement. Most encoders put out A B and Z. A and B tell you what direction and where. Z tells you when you're at a known fixed location.

It could take your motor 10 revolutions to go from open to close. So you multiply 10*1024 to get the pulses you'd need to count. For human readout divide that by the length in mm or cm to get a conversion factor. The Z output is a convenience you could use limit switches but if the encoder is working and you're counting all the pulses you will only need those if the arduino or encoder fail so they should be wired as cut offs.
If it was designed by man it can be repaired by man.

zoomkat

If one wanted to count chain links, a limit switch like below could be used. If counting is used, then calibration routines would be needed upon startup after power loss and such.

http://www.radioshack.com/product/index.jsp?productId=2049719
Google forum search: Use Google Advanced Search and use Http://forum.arduino.cc/index in the "site or domain:" box.

zoomkat

For proof of concept at only $2, I think the switch would be something easy to try. One can tweek the switch to minimize the "beat up" factor. Low tech solutions often have associated increases in low tech issues (noise, debounce, etc). An optical gizmo that looks thru the chain links generating on/off as the chain moves might be a cleaner solution. The door mechanism apparently is well sealed to limit access which limits options.
Google forum search: Use Google Advanced Search and use Http://forum.arduino.cc/index in the "site or domain:" box.

GordonEndersby



I was pondering this the other night after reading some of the thread.

How about a largish potentiometer on its side with a weighted pendulum attached to the shaft.
If this was mounted in a box and attached to the door,in the right orientation, as the door moved the pendulum would turn the shaft of the pot and the resistence would change.
it would be safe from dust and dirt and you could map the resistence to the door position.

Gordon

cr0sh

Quote
How about a largish potentiometer on its side with a weighted pendulum attached to the shaft.
If this was mounted in a box and attached to the door,in the right orientation, as the door moved the pendulum would turn the shaft of the pot and the resistence would change.
it would be safe from dust and dirt and you could map the resistence to the door position.


Something like this could work, but you would need one box per door section if you wanted to be able to stop the door at any position (rather than just up and down) - plus the math to integrate all of the potentiometers might get hairy. You might be able to get away with three such boxes - one on the top section, one on a middle section, and one on the bottom section.

As far as potentiometers are concerned, you would want to use weights on the shaft to effect a pendulum response, but you wouldn't need a long lever pendulum, provided that the shaft of the potentiometer turned freely enough. This might mean some shopping and sampling of potentiometers, though, unless you can physically handle them. A largish knob with a blob of formed lead at one spot (maybe a fishing weight) could be used as the "pendulum".

Or - you could use an optical solution consisting of an LDR or similar light detector plus a clear disc with a radial gradient mounted on a shaft to detect its position (hmm, maybe a weighted alluminum disc with a slit that starts narrow and increases in width to change the amount of light seen?).

Its an interesting idea...

:)
I will not respond to Arduino help PM's from random forum users; if you have such a question, start a new topic thread.

tkbyd

#40
Jun 20, 2010, 11:26 pm Last Edit: Jun 20, 2010, 11:26 pm by tkbyd Reason: 1
The You-Tube video is a big help. Lots of possiblities!

Put a magnet on the little black thing that travels along the track.

Mount a 2x4 alongside the track. On it, at strategic points, mount reed switches. Very easy to move them to "fine tune" whatever "it can stop here" points you set up.

bld

#41
Jun 21, 2010, 07:29 am Last Edit: Jun 21, 2010, 07:48 am by bld Reason: 1
Getting a lot of interesting ideas! Going to be hard to choose the right one to try ;D

Quote
Put a magnet on the little black thing that travels along the track.

As mentioned earlier, already got one there.

Quote
Mount a 2x4 alongside the track. On it, at strategic points, mount reed switches. Very easy to move them to "fine tune" whatever "it can stop here" points you set up.

True, that is already how the end stops work, but that is not what I need.
captain-slow.dk | non contagious!

tkbyd

#42
Jun 21, 2010, 10:43 am Last Edit: Jun 21, 2010, 10:43 am by tkbyd Reason: 1
For the mechanics of a "reel" which winds out and winds back, to which you can add a "rev counter", your could hack into a fancy dog-walking lead. Or a carpenter's measuring tape. For the latter, feed it through a pinch roller to get a rotating element to "watch"

If you want to be REALLY wild (!) and get the absolute position sensor you have always said you want:

At the back of the garage, mount a disc with it's plane parallel to one of the "sides" (ie. not back) of the garage. It needs to spin on a shaft. From the shaft to the "thing" travelling along the door track, a string. And there needs to be a spring to "wind" the shaft "up", as the door opens. With the garage door closed, the string would be pulled out/ unwound from the shaft. The diameter of the shaft could be done approximately at first, and then fine tuned so that from "door shut" to "door open" the disc rotates almost 360 degrees. Whew! Almost there!

You then put a Grey code... http://en.wikipedia.org/wiki/Grey_code (Skim down to the "rotaty encoder" figure)... on the face of the disc, "look" at it with sensors... and you're done!

daveg360

I know it's hardly in the spirit of this thread but.... do you really need to be able to set the position of the door to a 1% degree of accuracy?  I can't imagine why you'd need anything more than fully closed, fully open and perhaps a mid point?  this could be achieved very easily with limit switches.
If your system involves lethal voltages/life critical/flamable elements - you probably shouldn't need to ask.
The Arduino != PC.

bld

#44
Jun 21, 2010, 05:39 pm Last Edit: Jun 21, 2010, 06:12 pm by bld Reason: 1
Quote
I know it's hardly in the spirit of this thread but.... do you really need to be able to set the position of the door to a 1% degree of accuracy?  I can't imagine why you'd need anything more than fully closed, fully open and perhaps a mid point?  this could be achieved very easily with limit switches.

Of curse, I could even do it without using an arduino or anything else. But that is not what I am making. I am not only making it this way because it is something that is that necessary, it is because I want to learn more about it at the same time.

Also, the controller in the current motor is slowly dieing, no idea why, but the range of the remotes are getting shorter and shorter, right now it is down to about 5m so it can just exactly reach inside if i drive up to the door and stops close to it. When I got it the range was closer to 20m. It is not the batteries or antenna inside either remote or controller. So when it dies completely, the arduino is going to be connected directly to the relays driving the motor, along with a current sensor to check if it draws too much (collision). Then when that time comes, I would also like to be able to make it with variable speed, so it starts and stops gently.

I could just time it, but the thing is that the speed is very different, depending on the weather. In summer with some rain, it opens and closes fast, because the rubber seals on the outside easily slides over the moving door. In dry and cold it is moving slower, I think the difference is around 5 seconds.

So maybe it is not needed right now to know where the door is, but for future projects it would be very handy to make it right first time. And at the same time learn some more about it too.

I am considering the dogleash/measuring tape, that would be a pretty good solution I think! A measuring tape would have the right tension needed for a rotary encoder, and at the same time be small enough to not be easily noticed.


Update

Just found one!

Old one, but it seems to be perfect for the job!

The roll inside is held in place by the outer ring...





This is close to be the PERFECT ratio! For the multi turn potmeter to travel from one end to the other, it takes 234 cm, and the door moves 195 cm from end to end.

All this took was 2cm of solid grounding wire, bend the ends and drilled two holes and made a two small cuts into the inner circle on the drum. Then just scratched the end of the pot and soldered it directly on...
captain-slow.dk | non contagious!

Go Up