frogfot
Hazard to Others
Posts: 212
Registered: 30-11-2002
Location: Sweden
Member Is Offline
Mood: happy
|
|
Linear motor
Maby I'm reinventing the wheel but I got this idea of a linear motor. Long time I had an urge to make a mechanical arm, but the cost of
hydraulic/pneumatic parts was to great and the linear motors based on gears was even more expensive. Both types of devices are nearly impossible to
make at home.
Now comes the idea, using a DC motor to rotate a threaded rod, and the linear movement will be done by a nut positioned on the threaded rod, kinda
like in the pic below:
Well, I guess this idea is not new, I've seen such driving system as a GC syringe injector, but searching google revealed nothing... Why is this type
of setup not common?
Anyway, sofar, I see only advantages:
It's cheap as hell and easy to construct. A DC motor from anywhere and a threaded rod from a hardware store would cost below 5US$.. Ofcaurse there
would be a problem to find a nut..
Since the threaded rod works as a "gear-down" (eng?), it would be easy to make a powerfull linear motor. Just use enough big motor.
Decent speed is also an advantage. If proper motor is chosen, say 2800rpm and thick enough threaded axis.
Very easy to control speed/power, just use pulse width modulation (PWM) as in any other DC motor application.
Very easy to set up for feed-back. Attaching a disc on the axis with black/white strip and reading it with simple phototransistor/photodiode setup one
can easily detect 1/10 of a revolution.
Precision would be very big. With above mentioned speed and feed-back control you could make steps of less than 1 mm.
The lenght of the movement is only limited by the manufacturer of threaded rods! My store carries 1 m long threaded rods, which is more than enough
for most purposes.
Is there any disadvantages? Friction?
One disadvantage is that precision is limited by the nut's gap (it's increased with nut's diameter)
On my to-do list I have a mechanical arm with several of these linear motors.. It will have PWM all over it. The motors will be controlled with PWM
(to make soft start/stop) and the feed-back phototransistor/photodiode will be also on PWM (to ignore the daylight when the phototransistor will read
the feed-back black/wight strip). Whole shit will be controlled by AVR.
The only problem is that I don't know what the arm will do 
Sofar, I can think of it doing some preset movements..
|
|
|
evil_lurker
National Hazard
Posts: 767
Registered: 12-3-2005
Location: United States of Elbonia
Member Is Offline
Mood: On the wagon again.
|
|
First off, I'd like to say you have an awesome website and I'd love to see some more synthesis on there.
But, in this case your putting the cart before the horse. You need to know what you want your arm to be able to do before you start designing it.
[Edited on 24-6-2006 by evil_lurker]
|
|
|
frogfot
Hazard to Others
Posts: 212
Registered: 30-11-2002
Location: Sweden
Member Is Offline
Mood: happy
|
|
Well, you have a point there, but I had an initial idea and it was to stear the arm with a joystick. The microcontroller would convert x,y,z
coordinates (obtained from joystick) into arms different angles. This shouldn't be too complicated math, at least to write in C language..
Anyway, the arm will just be a place to test several things that are totally new for me. The software that will drive the arm is of secondary
importance 
|
|
|
YT2095
International Hazard
Posts: 1091
Registered: 31-5-2003
Location: Just left of Europe and down a bit.
Member Is Offline
Mood: within Nominal Parameters
|
|
that`s quite an old design and has been used for an equaly long time, basicly because it works so well, I`ve used this design myself to great effect,
but be very carefull with it, it can exert Massive amounts of pressure more than capable of taking a finger off!.
car jacks employ the same principal and a 70 year old granny can effectively lift a car 
[Edited on 24-6-2006 by YT2095]
\"In a world full of wonders mankind has managed to invent boredom\" - Death
Twinkies don\'t have a shelf life. They have a half-life! -Caine (a friend of mine)
|
|
|
enhzflep
Hazard to Others
Posts: 217
Registered: 9-4-2006
Member Is Offline
Mood: No Mood
|
|
IIRC this is also the mechanism used to controll the elevator control surfaces of various aircraft. Though presumably not an issue here, several
air-crashes have been attributed to the failure of this very mechanism an account of insufficient lubrication and/or the part being employed for too
great a period, leading to mechanical failure of the thread on either the screw or the nut.
I forget the flight number, but IIRC it was a TAA flight that crashed after a spectacular spiral from 10,000m (somewhere between 10 and 20 years ago)
|
|
|
chromium
Hazard to Others
Posts: 284
Registered: 27-6-2005
Member Is Offline
Mood: reactive
|
|
| Quote: | Originally posted by frogfot
Is there any disadvantages? Friction?
|
I used this principle for film scanner type device to move module with linear CCD sensor and lens. I hoped very high precision but this was not the
case. I did not manage to connect threaded rod exactly enough to the end of stepper motors shaft. It seemed easy but i always got some excentricity so
that module was moved back and forth turing each rotation and lot of precision was lost. Movement of nut was not uniform too (as thread was not of
highest quality and frictional forces were changing all the time).
So i droped this idea and rebuilt device with more complex but always relialble belt drive.
My conclusion was that this can be used for high precision systems only if manufactured with very high precision.
|
|
|
franklyn
International Hazard
Posts: 3026
Registered: 30-5-2006
Location: Da Big Apple
Member Is Offline
Mood: No Mood
|
|
It is always best not to try to reinvent the wheel unless there
is a great need to do so and existing technology is wanting. The
definitive tome for devices as this is the mechanical drives
reference issue of Machine Design journal which describes in
depth all commercial types of linear actuators , as they are called.
.
|
|
|
hinz
Hazard to Others
Posts: 200
Registered: 29-10-2004
Member Is Offline
Mood: No Mood
|
|
Why don't you use a stepper motor, if you use microprocessors it's easy to controll these. Then you don't need any photodiode, you can just drive the
nut aggainst an end-switch, collect the position at which the switch get's contact and add X mm per step-revulution of the motor. Since you have
controll about each step-revolution of the motor you just have to count the revolutions to get the position.
Nice stepper motors can be found in printers and scanners, they use the motors to shif the printing cartridge or the CCD/CMOS photochip of the
scanner.
|
|
|
frogfot
Hazard to Others
Posts: 212
Registered: 30-11-2002
Location: Sweden
Member Is Offline
Mood: happy
|
|
| Quote: | Originally posted by YT2095
but be very carefull with it, it can exert Massive amounts of pressure more than capable of taking a finger off!.
|
`
mm, I like what I hear!! 
Chromium, in my case, such very precise movements are not that necessary. The nut on threaded rod that I bought from hardware store can be moved less
than 1 mm without turning..
Franklyn, I knew that this wasn't any new idea but there was a problem to find anything on this on google... thanks for telling me how these type of
motors are called and where to look for info on them
Hinz, the main reason I've chosen DC motor was speed. Some time ago I've used steppers with homemade controllers (found on google), they couldn't
rotate fast, they just stalled after increasing frequency of the driving pulse.. Speed is needed when so many revolutions are needed to be made..
Using feed-back with DC motors seems like an ideal solution. I'm comparing this with a human arm: when arm needs to be moved to grab something the
driving signal moves the hand and the eyes follow the movement.
It's sometimes nice to feed-back a stepper motor too.. think if it takes too big load and misses couple of steps..
About the end-switch sensor, I'll have those. In the beginning of prog the arm will move to activate all the switches and thereby resetting all angles
to a known values. After, the angles will be followed by counting 1/2 of the axis revolution.
Oh, whole yesterday I was writing a schematic test prog for the arm. It's done for the most. The things it will do is, to move the arm to several
predefined positions (that is, move to A then to B then to C... etc). The movements will have soft start/stop. Having this last option I just had to
use two AVR's, otherwise it gets too complicated. AVR A will say required PWM level to the AVR B and further, AVR A will take care of the feed-back
and come up with motor direction. The AVR B will basically generate PWM pulses.
Anyway, not it's time to aquire some bearings and metal to actually start doing something. Maby I'll post some pics on the progress of this project.
I'll be able to play with it mostly on saturdays and sundays.. so this will take time..
Wondering if mods think it's ok to have this thread in miscellaneous section..? It doesn't have to do about chemistry..
EDIT:
If, lubrication is important than maby it's a good idea to somehow enclose the threaded rod from the outside so that no crap will get into
threading...
[Edited on 25-6-2006 by frogfot]
[Edited on 25-6-2006 by frogfot]
|
|
|
YT2095
International Hazard
Posts: 1091
Registered: 31-5-2003
Location: Just left of Europe and down a bit.
Member Is Offline
Mood: within Nominal Parameters
|
|
the one we did involved spot welding the end of a motorbike brake cable to the hexagonal nut, and then a plat on oppsing faces of the same nut, in
effect making a Runner using the same 2 8mm bar stock we used to support the hall thread with.
it was a stand-alone device at that point, we decided to use the brake calippers off a mountain bike as the "Claw" for the robot.
we had no motor as yet, so my mate sugested we used my Drill (Mains and 750Watts), we did.
we used a glass milk bottle as the item to grab.
needless to say it Exploded!
now I don`t know how many of you think BAH Glass bottle EASY! what is YT talking about!??
well YOU try it!
bare hands even!
\"In a world full of wonders mankind has managed to invent boredom\" - Death
Twinkies don\'t have a shelf life. They have a half-life! -Caine (a friend of mine)
|
|
|
Twospoons
International Hazard
Posts: 1281
Registered: 26-7-2004
Location: Middle Earth
Member Is Offline
Mood: A trace of hope...
|
|
Have a look at the leadscrew on any lathe or mill. You will notice that the thread has a square profile - this is for strength IIRC. Backlash can be
reduced by using two travelling nuts set up so they can't rotate, with a strong spring between them. This is at the expense of friction. Friction
can be reduced using a fancy recirculating ballbearing thingy, instead of a nut, designed just for this job. Small precision leadscrews can be
extracted from CD rom and floppy drives.
You can even buy the whole motor/screw assembly from RS Components ($$$!)
Helicopter: "helico" -> spiral, "pter" -> with wings
|
|
|
![[*]](images/xpblue/default_icon.gif){kind=icon}
