Register to reply

Build a Linear Motor?

by h2oski1326
Tags: build, linear, motor
Share this thread:
h2oski1326
#1
Feb1-08, 01:09 PM
P: 82
Background:

I am a first year PhD student and I just started graduate school this January. My background is in ME so this is a bit out of my realm.

My current project requires the use of a tubular linear motor, possibly similar to the one shown here:

http://www.copleycontrols.com/motion...ologies/motor/

I want to use a device like this (moving magnet only) to control a gas compression device. The trouble is I need to control the dimensions of the device as much as possible (i.e. it needs to fit within a specific space envelope). From what I can tell, this rules out all the off the shelf products I could find.


Specifications (approx):


The device needs to be capable of the following:

Peak Force: ~50lbf (222 N)
Const. Force: ~20lbf (89 N)
Accel.: ~15-20G

The device it will control (the piston) will have an OD of roughly 0.59 in (1.5 cm), and a stroke of about 0.39 in (1 cm). It would have an operating frequency of about 50Hz and be spring return (i.e. the motor alone does not have to handle keeping the magnet within operating space).

Question:

Does anyone have any experience building a device like this from scratch? If so, how difficult is it, would you recommend it for a research project? What are the major difficulties I should look out for?

I am trying to decide if it would be worth the effort to research the function of these devices so I can build my own, or, approach an outside vendor and attempt to get one custom made.

Again, I am an ME by education so the design of this device is a bit outside of my traditional range. Thanks in advance.
Phys.Org News Partner Engineering news on Phys.org
An innovative system anticipates driver fatigue in the vehicle to prevent accidents
Squink personal factory aims to make circuit prototyping easy
Catching grease to cut grill pollution
berkeman
#2
Feb1-08, 02:48 PM
Mentor
berkeman's Avatar
P: 40,639
Those are pretty high force and acceleration numbers for a linear electric actuator/motor, IMO. To get that force, hydraulic actuation would seem a better choice. But to get that acceleration..... not sure.

For electric actuation, the force and the acceleration would seem to be competing specs -- it takes a big magnet to get much force, but a big magnet is harder to accelerate.

Your dimensions and spring return are almost consistent with a solenoid design, but I'm still not sure about the force and acceleration numbers. Can you say more about what you are trying to do? Maybe there is some alternate way of accomplishing what you need to do.
Q_Goest
#3
Feb2-08, 08:39 PM
Sci Advisor
HW Helper
PF Gold
Q_Goest's Avatar
P: 2,903
Hi h2oski,
Have you looked through any of the patents yet for similar devices? Thereís quite a few. Hereís just a sample:
Multistage compressor: 6,015,270
Recip pump: 6,203,288
Liquid metal piston: 6,183,206

Iím familiar with one such device. It worked well, but in the end it failed to become marketable because it was simply too expensive. The competition, a conventional pump using a crankshaft and gears, proved to be much less expensive, and the driver (an electric motor) put out much more power. The linear device didnít have any significant advantage and the cost was much higher.

Linear motors are sexy, but from my experience (which is considerable) they donít confer any inherent advantage. Theyíre relatively large and inefficient, primarily because they have to stop and change directions on every stroke unlike a motor which is simply rotating at high speed in one direction continuously. They arenít smaller or lighter or anything like that. If the force produced by the magnetic field isnít acting on something thatís moving fairly quickly, the power output is minimal. Linear motors have a niche, but itís not in the compressor or pump field.

You asked if itís difficult to build from scratch. Iíve designed reciprocating pumps and compressors for the past 12 years and have yet to find anyone coming out of college that could even start to design something like this and Iíve met a lot of guys that have tried. I guess that means its very challenging. I wouldnít recommend it quite honestly. ďThe devil is in the details.Ē The hardest part is simply all the detail work. Seals, valve design, dynamic and flow analysis all have to come together. Designing pumps and compressors requires a whole host of analysis such as stress, dynamics, heat transfer, thermodynamics, fluid flow, manufacturing and assembly, testing, knowing the codes and standards, etc.... Everything you need to know about pumps and compressors you have to learn from scratch. The knowledge of how to design all the little bits and pieces of a good pump or compressor are not things you can simply read in a book. What works best is often kept proprietary.

If youíre really that interested in designing your own compressor or pump using a linear motor, find a mentor who can give you some guidance and explain what works and what doesnít and why. Purchase a linear motor and forget about making your own. Theyíre no big deal anyway. I think youíll have your hands full trying to figure out how to make the rest of the machine.

Sorry if that seems kinda harsh, but Iíve seen too many good engineers try to bite off elephants. I never understood the need for experience as a college grad Ė I honestly thought my education was fairly comprehensive and as such, the design of any device simply required the application of basic engineering principals. Now I have a half dozen patents in pumps and compressors and realize I still have way too much to learn.

h2oski1326
#4
Feb3-08, 11:12 AM
P: 82
Build a Linear Motor?

Thanks for the replies!

For some clarity, here is some more background on the project. I am developing a miniature linear compressor for a small scale refrigeration system. A linear compressor is desirable because of the compact form that it can take (compared with rotary compressors with similar pressure and flows). This makes a compact design, with a single moving part. Also, these devices (tubular linear motors) use gas bearings, which is also desired because of maintenance issues.

Quote Quote by berkeman
For electric actuation, the force and the acceleration would seem to be competing specs -- it takes a big magnet to get much force, but a big magnet is harder to accelerate.
I realize that the forces and accelerations are high. However, I know this technology is capable of these types of numbers (or at least close enough to work with) just based on what I have found on the net and speaking with a few other engineers. Also, these force/accel estimates are rough and strictly 'back of the envelope' type analysis, so they are likely high estimates. I will need to do something more refined before I begin my build.

Quote Quote by Q_Goest
You asked if itís difficult to build from scratch. Iíve designed reciprocating pumps and compressors for the past 12 years and have yet to find anyone coming out of college that could even start to design something like this and Iíve met a lot of guys that have tried. I guess that means its very challenging. I wouldnít recommend it quite honestly. ďThe devil is in the details.Ē The hardest part is simply all the detail work. Seals, valve design, dynamic and flow analysis all have to come together. Designing pumps and compressors requires a whole host of analysis such as stress, dynamics, heat transfer, thermodynamics, fluid flow, manufacturing and assembly, testing, knowing the codes and standards, etc.... Everything you need to know about pumps and compressors you have to learn from scratch. The knowledge of how to design all the little bits and pieces of a good pump or compressor are not things you can simply read in a book. What works best is often kept proprietary.

If youíre really that interested in designing your own compressor or pump using a linear motor, find a mentor who can give you some guidance and explain what works and what doesnít and why. Purchase a linear motor and forget about making your own. Theyíre no big deal anyway. I think youíll have your hands full trying to figure out how to make the rest of the machine.
I certainly appreciate your candor, and I am sure that this won't be an easy task, which is why it has a two year schedule for modeling, design/prototype, and optimization. My adviser is a 'compressor guy' as are many of the other fellow graduate students in my group. If I knew exactly what I was doing it wouldn't be research!

I am leaning toward having someone build the motor for us but I thought I would try and get some other opinions first. I am planning on leaning on the expertise of my group in addition to some good practices found in the literature but the motor itself was something I was just not at all sure about.

Q_Goest have you had experience building motors or are you a mechanical guy?

If anyone has any other thoughts I would certainly like to hear them. Thanks!
LowEParticle
#5
Feb3-08, 03:45 PM
P: 11
Quote Quote by h2oski1326 View Post
Background:
The device needs to be capable of the following:

Peak Force: ~50lbf (222 N)
Const. Force: ~20lbf (89 N)
Accel.: ~15-20G

The device it will control (the piston) will have an OD of roughly 0.59 in (1.5 cm), and a stroke of about 0.39 in (1 cm). It would have an operating frequency of about 50Hz and be spring return

Does anyone have any experience building a device like this from scratch? If so, how difficult is it, would you recommend it for a research project? What are the major difficulties I should look out for?
I've designed a linear motor somewhat similar to what you need, although capable of much less power. Here are some thoughts to try and help:

1. Your acceleration requirements are commercially achievable, but your power output requirements are very high.

2. Your stroke length (1cm) is quite short, which is certainly good. It opens the possibility of designing a tubular slider motor's 1st cousin: a voice coil motor. There are four basic pernutations:
A) The core is the prime mover in a moving coil design
B) The core is the prime mover in a moving magnet design
C) The outter case is the prime mover in a moving coil design
D) The outter case is the prime mover in a moving magnet design.

3. Your spec for the piston OD is 1.5cm. This is fine as long as you can achieve it by simply attaching a 1.5cm diameter rod to the end of the linear motor's actuator. If, however, you intend to use this value as a limit to the outter dimensions of the entire motor, or even the core in a (B) design above, you are essentially hammered from the get-go. You will be unable to generate the current density or the magnetic flux density to achieve the forces you need. I would go for a larger motor case, with a small piston attached to the end of the prime mover.

4. Tubular linear motors (including voice coil motors) are largely axis-symmetric problems. That means that you can get a free 2-dimensional magneto-static solver, such as Dr. David Meeker's FEMM program, and without actually building any prototypes get a feel for what it might take to design your motor.

5. The hard-core part of the design for any linear motor is the electromagnetics. You need to maybe read some emag, play with a magneto-static solver, and perhaps make friends with an electrical engineer or a physicist . Your problem is essentially an applications one, so I would tend to go towards the EE, but that's just my take on it. It may be useful for you to know that of the typical disciplines within electrical engineering, "motors" is one.

You don't actually state the requirements for the dimensions of the motor case, but from the stroke spec and the piston OD spec I'm worried that the motor case needs to be very small. Can you tell us what the case dimension bounds need to be?

Dave
dlgoff
#6
Feb3-08, 06:25 PM
Sci Advisor
PF Gold
dlgoff's Avatar
P: 2,700
I've been following this thread with great interest. The idea of making a miniture compressor is cool. But like berkeman mentioned, there might be another way.

I don't know if this technology will work for your application, but you might want to look into Piezoelectric Motors. Here's a couple of links that show some Compact Linear Motors.

http://www.physikinstrumente.de/prod...otor_index.htm

http://www.sensorsmag.com/sensors/Te.../detail/361287

Regards

EDIT: I looked at some of the specs. from the devices in the first link and the N-214 / N-215 NEXLINEģ OEM linear drives has these:
* 20 mm Travel Range
* 0.03 nm Resolution Open-Loop and 5 nm Closed-Loop
* To 400 N Push / Pull Force and 600 N Holding Force
* Self Locking when Power down, No Heat Generation
* Non-Magnetic and Vacuum-Compatible Working Principle
h2oski1326
#7
Feb4-08, 11:49 AM
P: 82
Quote Quote by LowEParticle View Post
There are four basic pernutations:
A) The core is the prime mover in a moving coil design
B) The core is the prime mover in a moving magnet design
C) The outter case is the prime mover in a moving coil design
D) The outter case is the prime mover in a moving magnet design.
This is what I would like to use. With the prime mover serving as/attaching to a compressor piston. Ideally the space created inside the coils would serve as the compression chamber to save space.


Quote Quote by LowEParticle
3. Your spec for the piston OD is 1.5cm. This is fine as long as you can achieve it by simply attaching a 1.5cm diameter rod to the end of the linear motor's actuator. If, however, you intend to use this value as a limit to the outter dimensions of the entire motor, or even the core in a (B) design above, you are essentially hammered from the get-go. You will be unable to generate the current density or the magnetic flux density to achieve the forces you need. I would go for a larger motor case, with a small piston attached to the end of the prime mover.
Can I not increase the magnetic flux density by increasing the length of my pistion/coils and/or adding thicker layers of coils around the prime mover?

Quote Quote by LowEParticle
4. Tubular linear motors (including voice coil motors) are largely axis-symmetric problems. That means that you can get a free 2-dimensional magneto-static solver, such as Dr. David Meeker's FEMM program, and without actually building any prototypes get a feel for what it might take to design your motor.

5. The hard-core part of the design for any linear motor is the electromagnetics. You need to maybe read some emag, play with a magneto-static solver, and perhaps make friends with an electrical engineer or a physicist . Your problem is essentially an applications one, so I would tend to go towards the EE, but that's just my take on it. It may be useful for you to know that of the typical disciplines within electrical engineering, "motors" is one.

This is a fantastic suggestion and I have been doing some similar research. I will look into this FEMM program.


Quote Quote by LowEParticle
You don't actually state the requirements for the dimensions of the motor case, but from the stroke spec and the piston OD spec I'm worried that the motor case needs to be very small. Can you tell us what the case dimension bounds need to be?

This is open ended, as is most of the project, but ideally as small as possible.

Quote Quote by dlgoff

I don't know if this technology will work for your application, but you might want to look into Piezoelectric Motors. Here's a couple of links that show some Compact Linear Motors.
This is certainly intriguing and I hadn't thought of these yet. I have looked into these motors in the past for an unrelated project and they seemed to have some limitations. The high force model looks promising but slow. I will have to keep PI on my list of vendors to call.

Thanks all for the replies so far!


Register to reply

Related Discussions
Non-Linear Control System and Motor synchronization Mechanical Engineering 4
I have to build a simple electromagnetic motor Introductory Physics Homework 1
Linear Syncronous Motor Computing & Technology 1
Linear Synchronous Motor velocity? Electrical Engineering 0
Physicists build world's smallest motor ...Berkeley News General Physics 2