Howdy,

I'm toying with a project idea, and while I have the mechanical elements pretty well under control I could use a little help getting oriented on the control system. Here's the deal:

I'd like to build a vertical sliding door to open my living room to my patio. By vertical I mean that it will guillotine up and down. It won't roll up on itself and it won't slide left to right. It would just shift up and hang in front of the second story of the house, completely out of the way. You don't see this very often.

The moving element would be a glass-and-steel panel that's about 15' wide x 9' tall, which a few minutes in SolidWorks tells me would weigh something in the neighborhood of 1700#. I would counterweight that with a lead slug of equal weight, connected by heavy duty roller chain run over idler cogs. So the moving mass of the door system is something in the neighborhood of 3400#.

If it were a lighter system I would plan to just operate it "manually" like a double-hung window, but with that much inertia it doesn't matter how carefully balanced it is I'll need to motorize it to start and stop it without breaking something (either the operator's back or the window panels).

Again, because of the inertia, it doesn't seem like I could safely just trigger either end with a reed switch or something and just stop the motor when it hits it. I feel some kind of position-aware, PID controlled system that decelerates gradually to a stop at calibrated Open and Closed locations would be in order.

I'm not sure if a controller that solves this problem exists "off the shelf" but if it does I don't know what key words to use to find it. If I have to engineer something myself I'm thinking that I'd use an DC brake motor to provide locking stop positions, and then perhaps some kind of linear position sensor together with a servo controller or something. But I really don't know. The extent of my work with position-aware motor systems is stepper and servo CNC routers, and I'm not sure how directly applicable that would be.

Anyone have a good starting point for me on how they'd go about controlling a door like this?

Thanks!
-Ben

 

Depending on what you deem to be acceptable, you might consider using a product called Acme threaded rod, and the corresponding nut:
http://www.dependableacme.com/class/
The use of this method transforms the design from a problem of linear motion to rotational motion. ... Select a motor that you can control the angular velocity and direction. ... PWM, stepper motor, or something else. ... Also a smaller power requireent.
This product is designed for relative motion between the rod and the nut.

 

Acme threaded rod,

That's a good idea. Like a trash compactor. Lots of weight capacity. Lots of mechanical advantage. The Acme approach is naturally slow and easy to control at a stop position, but it requires a really long screw like the jack screw in an airplane's tail. Consider the inverse: a worm gear turning a geared pulley. The worm turns at a fairly high speed so it also has the ease of stop control, but you can do that in the size of a shoe box like the gear segment you wind up to open your car window.

 

Another idea that would probably be cheaper and easier (and more robust) would be using a reel-like mechanism. For instance, you could use an old automotive winch and build a simple control circuit for it.

 

I also like the acme thread lead screw for the lifting. As to a linear position sensor? Commonly called a "string pot" I have used these units running out to 80" on a spring compression test. They can be very accurate and the linked versions should cover the range you want. Let the pot drive a differential circuitand reduce drive as door reaches the setpoint, open or closed.

Ron

 

I also like the acme thread lead screw for the lifting

They're sturdy, reliable, long-lasting, and also excruciatingly slow... you just can't spin an acme screw fast enough for a door opening mechanism to be practical... but that's my opinion

 

an old automotive winch

Yep. Already built. Just needs limit switches and a power supply, which brings the idea that you can probably buy them in 120 VAC.

ps, I disagree with the slow complaint. A whole minute to raise something that big wouldn't bother me.

 

They're sturdy, reliable, long-lasting, and also excruciatingly slow... you just can't spin an acme screw fast enough for a door opening mechanism to be practical... but that's my opinion

Damn, I should have thought that through, all my applications were slow, real slow. Good point there and a my bad. Garage door system might work.

Ron

 

What about using a Parker Hydraulic motor? It may cost more than the winch but may be easier to control and add safety features.

 

Damn, I should have thought that through, all my applications were slow, real slow. Good point there and a my bad. Garage door system might work.

Ron

I just occurred to me that an acme screw might not be such a bad idea after all. If it were coupled with a bar-lever mechanism, maybe it could be designed in such a way that the door would start closing real quick, and then slow down at the end of its traveling distance. Something similar to the way half-scissor jacks work.

​

Then again, maybe said mechanism would take too much space... I'm just brainstorming here.

 

Thanks for all the ideas!

I'd like to redirect a bit though. My concern is less about the actual power source / mechanical system and more about how to control it. What I don't want is for a 1700# door that's mostly glass to slam to a hard stop at either the top or the bottom. So whether it's an acme screw or a winch motor or anything else, I'm going to need a way of accelerating it from a stop in a non-jarring way, and then decelerating it to a soft stop in a precise location. It's the control system not the mechanical system I need help with.

On my small CNC router I just use reed switches as limits and when they're triggered the stepper stops (it runs on ACME lead screws). It's a small enough machine that this works okay. Try bringing a 1700# panel to a dead stop from any reasonable velocity and you'll break something. You need to slow it down gradually and have it come to a gentle end at a bottom or top location.

Thanks,
-Ben

 

On my small CNC router I just use reed switches as limits

It's good that you have experience in what to expect from a limit switch. (We do this with chicken coop doors often enough.)
The super-precise way to do this is with a stepper motor. Deadly accurate. You can actually count the steps between "open" and "closed".
Also deadly expensive.
https://www.grainger.com/category/s...s&suggestConfigId=6&nls_boost_words=noContent

Personally, I'm an advocate of doing it slowly. Maybe a whole minute for travel time. That's 1.8 inches per second.

Brakes and clutches are outrageously expensive
https://www.grainger.com/category/b...nd-clutches/power-transmission/ecatalog/N-aps

but jaw coupling hubs with rubber spiders aren't so bad
https://www.grainger.com/category/j...nd-adapters/power-transmission/ecatalog/N-i86
https://www.grainger.com/category/j...nd-adapters/power-transmission/ecatalog/N-i87

 

Thanks for all the ideas!

I'd like to redirect a bit though. My concern is less about the actual power source / mechanical system and more about how to control it. What I don't want is for a 1700# door that's mostly glass to slam to a hard stop at either the top or the bottom. So whether it's an acme screw or a winch motor or anything else, I'm going to need a way of accelerating it from a stop in a non-jarring way, and then decelerating it to a soft stop in a precise location. It's the control system not the mechanical system I need help with.

On my small CNC router I just use reed switches as limits and when they're triggered the stepper stops (it runs on ACME lead screws). It's a small enough machine that this works okay. Try bringing a 1700# panel to a dead stop from any reasonable velocity and you'll break something. You need to slow it down gradually and have it come to a gentle end at a bottom or top location.

Thanks,
-Ben

An up/down door seems like a good idea until you think about safety. Garage doors have detectors both for line-of-sight (IR sensors) and weight/resistance in case there is an obsticle not "seen" by the IR sensor that prevents the door from reaching the desired end point. The door must sense the obsticle before there is enough force to hurt a child that might be stuck under the door. For a garage door with spring counter balance, the resistance is fairly easy to detect since the weight of a modern sheet metal door is not so much.

With your 3400 pounds of hardware, I doubt you'll be able to stop the door with less than 50 pounds of unexpected resistance. Remember that the house will be sold eventually and it is much better to do it safely or not do it at all.

 

Oh, safety is absolutely a consideration here and one I'm taking quite seriously. There would be sensors in the system to stop based on IR. The force-feedback stop is trickier to handle though given how infrequently these would be used I think you could get away with omitting them. These doors would likely be opened only for parties and _really_ nice days. They won't be opened and closed 5-10 times a day like a garage door and I think I'd wire it so you had to actively hold the button while the door was in motion to ensure that someone was paying attention.

Anyway, safety goes way beyond just the opening/closing mechanism. There'll need to be a one way mechanism to prevent the door from crashing down in the event that the counterweight chain breaks as well. That mechanism would only be disengaged when you were actively lowering the door.

 

The super-precise way to do this is with a stepper motor. Deadly accurate. You can actually count the steps between "open" and "closed".

My concern with a stepper is that you run into "lost steps" and then end up not knowing where you are. Happens all the time with routers. Less likely with a system like this, but if you ran into an obstruction it would happen. Then you need a zeroing routing, which is a lot to expect of someone that just wants to open a door. A closed loop system (more servo-like) feels more appropriate in this application, no?

 

I think I'd wire it so you had to actively hold the button

That simplifies things. "Stand there watching it" and a limit switch should do the trick.
The upper safety could be a pawl device and you could add another button to fire a solenoid to release the pawl.
Or maybe design it so the pawl defeater is automatically powered when the upper limit switch is OK and the window receives a, "down" command from the control box.

In this case, I would install two limit switches at the top and demand that both of them agree with each other to allow movement.

Right now, I'm thinking about a chain operated hoist where you have to pull the chain a lot of times to move the load a little bit. If there is any tendency to lower by itself, the chain hoist has a lot of resistance to change caused by the load.

 

Try bringing a 1700# panel to a dead stop from any reasonable velocity and you'll break something. You need to slow it down gradually and have it come to a gentle end at a bottom or top location.

When I actually worked for a living we used acme thread lead screws to run up and down the control rods of reactors which in total exceed 2,000 lbs. Using a gear motor with a lead screw would work but would likely be slow. As to stopping and starting, they stopped literally on a dime (within 0.25"). Large industrial garage doors can weigh upwards of 2,000 lbs. How much does the weight matter if you plan a counter weight? I would likely just run with a garage door type system as used on an industrial type door. If you would like use a two speed motor or variable speed with a null detector circuit. Use a string pot for accurate position sensing. Really it is a matter of what you want to spend.

Ron

 

Good stuff. So I think the string pot is one of the key items I'm taking away from this. That would give me an absolute position indicator that's very unlikely to get confused (like a stepper that can lose steps) and that I can use in conjunction with a microcontroller and some programming to accelerate / decelerate the motion and set stop positions. Hold an "Up" button and it accelerates upward to a max velocity and stays there until it's within the upper deceleration window at which point it decelerates to a stop. Hold a "Down" button does the same thing the other way.

Ron, weight matters only in terms of it's implications for acceleration / deceleration. F=MA. When M is 1500 Kilos, if you want A to be reasonable you need a fair bit of F. That said, 20# of force can accelerate 3500# at about 0.06 M/S^2, which actually gets it to about 2fps after 10 seconds. That's not terrible. (And yes, I'm mixing up my SI/Imperial units. )

You could probably even operate it manually if you were careful to start slowing it down early enough to prevent a crash at the opposite end and used a reasonable shock absorber system at either end as a safety.

Actually, as I do some more math on that maybe manual is a better option? If you use 20# of force to accelerate to the half way point and then 20# to decelerate through the second half it would take about 12-13 seconds to open or close the door if it has 9' of travel. That's not bad at all. The shock absorber system would need to be able to safely decelerate it if someone did his or her worst and really got it moving and let it crash into the stops, but that might be an easier problem than doing this electronically. Hmmmmm....

 

You could probably even operate it manually if you were careful to start slowing it down early enough to prevent a crash at the opposite end and used a reasonable shock absorber system at either end as a safety.

That brings to mind does a system like this need a manual back up in case of power failure? I never gave that any thought but most garage doors allow for a manual backup in case of mechanical or power failure.

A good string pot can be a very accurate position indicator but they are not a cheap, I have used Unimeasure Manufacture which were real nice. Yeah, as to load we would measure the motor acceleration and deceleration under various loads so with a mountain of glass that needs considered.

Projects like this always come down to a budget. There is all sorts of incredible cool stuff out there including some sweet vari speed DC gear motors. YoYo pots like I linked to and other stuff.

Something else that comes to mind was 5 ton multi speed hoist systems we used with a very tight controlled speed. Been several years since I retired and I can't recall who made the base units we modified but they used a VFD for motor speed and a geared motor. I guess the possibilities are endless. Any manual backup would be a heck of a hand crank system. We had a giant diesel generator.

Ron

 

How about using two limit type switches at each end of the travel to indicate position?
The first would slow the mechanism near each end (say 3-6 inches) and the second would stop it at the end.