Hi, my first post. I'm a professional C programmer for 30 years, mostly embedded for the last half of that. Although I can read a schematic I can't actually design my own electronics. The simplicity of the Arduino has encouraged me to learn.
Just as in this discussion, my first project is a chicken coop door opener, and I have questions about what kind of motor can hold up the weight of the door without power.
(I would have added to the existing thread, but for some reason it has been made read only - apparantly on the basis of time passed, even though the info is still current).
I have had two prototypes working. MarkI used a 3V motor and does work, but is slow and struggles to raise a 1kg test weight - whereas I was hoping that it would handle up to 3kg without struggling. So, I got myself a beefier 12V DC motor with about 4x the torque rating. The new motor happily lifts the 1kg test weight, but to my surprise it was unable to hold it up when unpowered.
The proposal in the older thread was to use a counterweight - the problem with that idea is that the door needs to weigh a reasonable amount to be sure it won't stick on the way down e.g. caused by frost. I think 1kg is about right for this.
I've got a vague idea that it's the gearbox type that's important. E.g. expensive low friction and/or epicyclic gearboxes will have this problem. Ironically, cheap and noisy gears will not. Comments?
Mind you, this does give me an idea for design change: only use the motor to raise the door, let gravity close the door for free! I'd need some way to lock the door in the open position (solenoid? - will that hold without power?).
DC motors have torque proportional to current, no current, no torque. (Well, not
quite the whole story, there is both static and dynamic friction too). In fact most
designs of electric motor are like this, its the current that produces the force one
way of another.
The normal way to exhibit a torque when powered down is to use either
a worm-gear drive or an electromagnetic braking system that clamps tight
via a spring when not energized.
Most worm gears have enough static friction to prevent back-driving from the load
completely.
There are other mechanical linkages that are self-locking, for instance a crank+piston
will fail to back-drive at top-dead-centre and bottom-dead-centre.
You could also add click-stops or detents to the system so that full motor torque
is needed to break-free from them.
[ I should add that a large reduction ratio gear such as 50:1 or more will likely have enough
friction to resist back-driving in practice ]
I have no previous experience of working with motors in an embedded project. For my mkI prototype I used the same motor as one of the commercial units I'm aware of. This was the 3V motor, and naturally in my experiments it worked the way I expected, tho surprisingly weak (vendors claims to support up to 3kg door seems... optimistic). I'm afraid I'm guilty of assuming that any similar specced (for torque) motor would work.
As you say, that 3V motor had a 500:1 gear ratio, and from the awful cheap clockwork noise I assume it has stamped steel gears - lots of friction, so no doubt that's what holds the door up. From what you say, this motor too would have a definite static friction limit, it just happens to be a lot higher than the 12V motor I bought.
Out of interest, this is the 12V motor I'm referring to. You'll notice the listing is light on details, e.g. no gear ratio. OTOH it was cheap and I thought it would be worth a punt. I had been looking at a much more expensive motor which claimed to have a much more efficient planetary gearbox - I assume "efficient" means low friction losses so I'm glad I didn't go for that!
Do you know of a small DC geared motor with a built in brake? Ideally a whole motor family so I can find one with ideal voltage, current draw etc.
Hmm. Just when I think I've got a handle on it, I make a weird discovery. If I raise the test weight up to any height, then cut the power - but before the motor goes into reverse I take the weight of the test load in my hand for a couple of seconds, then release it... the door doesn't fall! Completely repeatable. What is going on there? Is this just a borderline effect I wonder?
If you use an indirect drive, such as a pulley reduction whose load end is a crank arm, all you need to do is place a stop just past the max stroke point and allow gravity to hold it against the stop, when you close the door, the motor lifts the crank away from the stop and the door closes.
I've been lurking around a month, so I'm already aware of some of those discussions. However my goal is to learn electronics by doing, so copying someone else would defeat the purpose.
DonMilne:
I've been lurking around a month, so I'm already aware of some of those discussions. However my goal is to learn electronics by doing, so copying someone else would defeat the purpose.
At some point or other you will come to the startling and humiliating conclusion of "It has been done before!" You have to be willing to copy some things in order to learn the most important part, the "Why it works".
It's quite possibly true that it's all been invented before, but it simply isn't a fun project if I copy someone else, and I'm doing this for fun. And no, there is no room for a sideways opening door, or for a complex mechanism above the door. The coop is small, with limited room around the vertically moving door. A small self contained motor box is what it has to be.
However, I think we're drifting away from the question asked. I explained what I was doing so that you'd understand the application, however I'm not really looking for design tips (*). The question itself was quite narrow: what governs what weight of door a motor can hold without power?
And to answer my own question, it seems to depend on how easy it is to wind the gearbox in reverse from the outside, and worm gearboxes seem to be the best (hardest) for this. Unfortunately I've not yet found any worm gearboxes (with DC motors) which are cheap and have reasonable power consumption. Still looking.
(*) However, I was interested in the suggestion of some kind of lever that stops the door from falling back. I can think of several that might be engaged when the door is being raised, but I can't think how they would be released (in a single motor design) to the let the door fall again. Is there a discussion of this somewhere? Preferably with pictures?
Economy-----Used automotive winshield wiper motor, its 12V worm drive has a crank arm already won't drift under normal loading but can be overriden manually.
DonMilne:
Hmm. Just when I think I've got a handle on it, I make a weird discovery. If I raise the test weight up to any height, then cut the power - but before the motor goes into reverse I take the weight of the test load in my hand for a couple of seconds, then release it... the door doesn't fall! Completely repeatable. What is going on there? Is this just a borderline effect I wonder?
Solid friction isn't linear, it has hysteresis, also vibration can lower effective friction thresholds, if you've got borderline behaviour
you need at least twice the friction to be reliable perhaps.
Unfortunately I've not yet found any worm gearboxes (with DC motors) which are cheap and have reasonable power consumption. Still looking.
Time for you to define "cheap" and the power limitations. Worm gear windshield wiper motors and stand alone worm gear motors seem to start at $15 surplus and on ebay. A black and decker 6v cordless screwdriver can be had at walmart for $10 or less.
zoomkat:
Time for you to define "cheap" and the power limitations. Worm gear windshield wiper motors and stand alone worm gear motors seem to start at $15 surplus and on ebay. A black and decker 6v cordless screwdriver can be had at walmart for $10 or less.
I deliberately didn't define "cheap" because an amount in UK £s would be meaningless to many readers. Lets just say, I would hesitate before buying a motor for 10 times the cost of the Arduino. (I'm involved in embedded development at work as the firmware guy - I have an ingrained habit of avoiding proposing components whose cost would make the device unsellable, even though I've no current plans to commercialise this project).
As it happens I've looked at some cordless drills yesterday. From the look of them they're all axial - motor in the nose part, battery in the grip - no worm drives that I can see. I also can't think why any of them would need an ability to restrict reverse rotation when unpowered. And the cheapest decent ones are around twice the price of bare motors I can buy from a hobbyist motors site: e.g. B&D 6V cordless, £21 from Amazon. Cordless angle grinders look like they might use worm gears, but they're typically more expensive.
As for the power consumption: just in general terms it's going to be a battery operated device so I can't go with a motor that soaks up amps in normal operation. The motor in my mkII prototype sucks a constant 130mA at 12V when raising a 1kg load by a couple of feet. Of course it sucks nothing at all when the load falls back due to gravity! It would be great if I could keep those low numbers since my goal is to have the device run for up to 2 years on 4 AA batteries. The other consideration is that I'm using an SN754410 h-bridge chip, which allows 1A per channel (2A if I link channels - but I wanted to keep the second channel in reserve for a possible solenoid).
I hadn't considered windshield motors, so I'll look into them.
MarkT:
Solid friction isn't linear, it has hysteresis.
Yes, I could see that, once I had reason to think about it. I'm beginning to see a difference between what I'll call static friction and normal operating (dynamic) friction. I have a feeling that the former is what is sometimes called "sticktion". And obviously it would be higher than dynamic friction forces.
As I mentioned earlier, I have no previous experience of working with motors in an embedded project, but I'm learning more about them every day - that's the real fun with these projects.
If you have trouble with motors, try hydraulics. Not the kind of hydraulics that a bulldozer uses. Try connecting a three gallon bucket to the door over a pulley ( walmart, 80 cents ). Use a patio water feature pump ( ebay, $11 ), to pump water from a stationary bucket into the one hanging from the pulley. When enough water is in the bucket, the door goes up.
I used this scheme for a similar requirement a few years ago. Avoids the problem with gearboxes and mountings which are often unobtainable and much more problematical than electronics and software.
I also can't think why any of them would need an ability to restrict reverse rotation when unpowered.
Reverse rotation is only an issue because for what ever reason you want a coop door that weighs 3kg. You should rethink/redesign your door to be lighter.
zoomkat:
Reverse rotation is only an issue because for what ever reason you want a coop door that weighs 3kg. You should rethink/redesign your door to be lighter.
1kg (~2lb) is intended to be the typical weight. It needs to be that heavy to work reliably, because gravity is the only thing that ever allows the weight to be lowered. 3kg is just a decent engineering margin. The actual wooden door on my current coop weighs 1.1kg (when dry). Slide friction on top of that (varying in the wet and in frost, and when dirty). My margin may admittedly be a tad high, but I don't think so.
However, I don't believe the gist of your assertion is correct. Reverse rotation is an issue because the motor gearbox isn't designed to hold weights. Period. The actual weight at which the problem shows may vary from one motor to the next. If I can't rely on the next motor (with the same model number) working the same way then it isn't a real solution.