Request for help with spring design

[John2004]

Hello,

I would like to ask if someone could please provide some feedback on a unique spring design problem.

Thanks
John

PROBLEM:

I have a radial plate cam, with an oscillating roller follower. The small cam is oscillated manually by hand, via a 6" long lever. At the neutral or resting position, the roller follower is in contact with the approximate center of the cam curve or profile. From the neutral position, the cam can be rotated clockwise, (by pushing the 6" activation lever down) or counter-clockwise (by pulling the activation lever up). The maximum angular displacements of the lever and cam are 16 degrees clockwise from the neutral position, and 14.5 degrees counter-clockwise from the neutral position.

When the cam is rotated clockwise from the neutral position, the force the roller follower exerts on the cam curve decreases, but when the cam is rotated counter-clockwise, the force the roller follower exerts on the cam profile increases.

At the neutral position, the roller follower is in contact with a "dwell" portion of the cam profile, and as such, the force the roller exerts on the cam profile does not create any torque on the cam at the neutral position. However, when the cam is rotated from the neutral position in either direction, the roller force then creates a torque on the cam.

When the cam rotates clockwise from the neutral position, the roller moves towards the cam rotation axis. When the cam rotates counter-clockwise from the neutral position, the roller-follower moves away from the cam rotation axis.

When the cam rotates clockwise from the neutral position the cam follower roller force goes down. When the cam rotates counter-clockwise from the neutral position, the roller force goes up.

I have springs connected to the cam in a series parallel array, i.e., 4 springs on each side of the cam, which hold the cam at, and return the cam to, it's neutral position, when the activation lever is pushed down or pulled up from the neutral position, and then released. The opposing springs are pre-stretched at the neutral position with equal forces on each side of the cam. This provides a "balanced feel" in the activation lever, which is important for the design.

My problem is, when the cam rotates clockwise from neutral 4.637 degrees, the un-balanced force of the springs needs to be about 50 pounds to overcome the force of the follower roller and return the cam to the neutral position when the activation lever is released. Since the force of the follower roller and the force of the springs are nearly balanced at this point, it is easy to push the lever down from the neutral position. The roller follwer force "helps" to push the activation lever down.

However, when pulling the lever up from the neutral position, not only does the force of the follower roller increase, but the force of the opposing return springs also increases (i.e., the un-balanced force will also be 50 pounds at 4.637 degrees counter-clockwise).

Also, the roller follower force works against the activation lever when pulling the lever up from neutral.

I need a system that provides a "balanced feel" in the activation lever at the neutral position, but allows the force to pull the lever up from neutral, to be small, like it is when pushing the lever down.

I tried using different (lower) rate springs on one side, but when I layed it out in my CAD system, and did the math, I could not get everything to work out.

It seems the opposing spring forces must be balanced at neutral, and the unbalanced force of the springs must be 50 pounds at 4.637 degrees of clockwise cam rotation, but somehow smaller when the cam is rotated counter-clockwise 4.637-degrees.

I basically want the force to pull the lever up from neutral to be as small as possible through the entire 14.5 degrees of counter clockwise rotation. It’s desirable to keep the forces necessary to move the lever as low as possible in either direction, either clockwise or counter-clockwise from neutral. The forces look good for pushing the lever down from neutral, but I would like to see lower forces for pulling the lever up from neutral.

I cannot really increase the length of the activation lever, since I am using a stock lever and want to keep using it. The space I have to work in is very small and constraining. The entire cam is also adjustable up and down by a total travel of 5 mm, which complicates things further.

The springs I am using now have about a .390" OD, a 1.8" free length, and a 115 lb/in rate. It would be difficult to increase the OD of the springs very much do to limited space.

Please let me know if anyone has any suggestions for a type of spring, spring system, or design that my work.

Thanks for your help.

John

 

[GregLocock]

Wow. Could you plot, or list, a curve of desired actuation force vs angle? I think you have described it in essence, but we've recently spent a while analysing a problem where one error in the initial descritpion invalidated most of the thread.

FWIW, on cars we use polyurethane foam cylinders as spring aids (ie they are durable and powerful and smaller than coil springs). By fine tuning their profile and the amount of preload we can get virtually any non linear characteristic we want. They can be turned on a lathe for small quantities, or injection moulded for production.



Cheers

Greg Locock

 

[israelkk]

Added to GregLococ questions. Are you looking for a spring design or a systm design? Can you can define the spring properties (forces vs deflection, life cycle etc.) and the space limitations as though as the spring type (Extension, compression, linear, non-linear etc.

http://www.webspawner.com/users/israelkk/

 

[zekeman]

After reading your thread, I think that you will have to furnish a drawing of this design including the geometry of the cam and the oscillating follower, including all of the forces or loads including inertia, if any. Also, I don't understand your need for such stiff springs to hold the cam in neutral equilibrium since you are so concerned about forces in the lever.
Again, it would be impossible to help you without further detailed calarification.

SS, consulting engineer

 

[John2004]

Hello everyone,

Thanks for your feedback.

I'm going to try to put a drawing up on a free web host, to clarify the problem. I will come back with a link to the drawing.

For Greglocock: can you tell me where I can find more information on the polyurethane foam cylinders ? Are they strictly custom made & if so, where could I have them made?

I am presently using regular extension springs with hook ends. I don't "think" opposing compression springs will work, and I don't think a torsion spring will work. As the cam is rotated, the extension springs on one side of the cam extend by about .012" per each degree of cam rotation, and the springs on the other side of the cam retract by about the same amount. Which side extends or retracts depends on the direction of cam rotation from the neutral center position. The spring hooks connect to the cam at about 0.673" from the cam rotation axis (center of spring is at this distance). All the parts are very small.

Everything fits into a space that is only about 1.75 inches thick. Since the lever is activated intermittently by hand, the life cycles dont't need to be as high as most industrial automated machinery. I think it would be considered light duty service.

I think I may need a different "spring system" but it must be cost effective, work in a compact space, and provide a "balanced feel" at the neutral position.

Thanks again,
John

 

[GregLocock]

Boge Elastofabrik (sp?) of Germany make ours but for one offs we just buy polyurethane foam and turn it on a lathe.

Cheers

Greg Locock

 

[Warpspeed]

I am having difficulty clearly picturing your requirement in my own mind. As the others have suggested a force displacement diagram might help a lot.

I doubt if spring design alone will fully answer the problem, but there are some tricks with over centre springs and bell cranks and similar kinematic linkages that can produce some wonderful non linear effects with springs.

 

[John2004]

Hello everyone,

I was finally able to put some drawings and descriptons on the internet of the spring design and/or lever force problem I posted in this thread. I think this will make the problem much more clear. I am hoping this will allow additional comments and suggestions.

Here is the link:

http://www.cdtd.50megs.com/

I would sincerely appreciate any additional feedback or suggestions on this problem. I would like to consider as many viewpoints as possible. I hope I can find a suitable solution to the problem.

Thank you.
John

 

[nbucska]

Could you use a second CAM.2, coaxial with the CAM.1
and use the spring-loaded roller arm of this CAM.2 to supply the complementing force so that the sum of the two torques would change as required?

Can you define by equation or table the position of CAM.1
and the required torque on the handle fixed to it?



<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[John2004]

Hi Nick,

Thanks for your reply.

I had thought of using a mirror cam profile located 180 degrees opposite the profile shown in the drawings, and then using a spring loaded "extra roller" to offset the load of the existing roller. I would still need a couple of opposing extension springs connected to the cam solely to provide a balanced feel in the activation lever at neutral. However, the springs could be small.

I don't think I really have the space to do it. The follower arm pivot point, and the cam, are adjustable up and down by a total distance of 5 mm, and this also makes the dual cam curve approach hard to implement.

I had also thought about using a track cam that can both push and pull the follower, and then put a spring on the follower to oppose the follower load. However, space constraints make that hard to implement to. It is also hard to mount the cam follower with this method. I would need to mount the roller on some type of shoulder screw or something, to get the cam track to fit "around the roller". The OD of the roller is only 3/16", and the shaft the roller rotates on is only 1/8" diameter. If I use a track cam, it looks hard to provide a roller with a strong mounting.

I don't need much spring force at all on the left side of the cam, just enough to pull the cam off the high dwell, which is not hardly anything since the torque required to rotate the cam when the roller is in contact with a dwell, is practically zero.

The main reason for having equal opposing forces on the cam at neutral, is just to provide a "balanced feel" in the activation lever at neutral. Otherwise, the problem would be easy to solve, I could just put springs on the right side of the cam that are completely retracted at neutral, and then put one very small spring on the left side of the cam, to pull the cam off the "high dwell" after the cam has been rotated to the extreme counter-clockwise position, and the activation lever is released. However, I think I will then lose the balanced feel in the activation lever at neutral, if I do this.

I need some way to have a balanced feel in the activation lever at neutral, but somehow have lower spring forces on the left side of the cam, when rotating the cam counter-clockwise from netural. Or, somehow gain leverage when rotating the cam counter-clockwise from neutral. But as explained on the site, all of the standard or obvious methods of gaining leverage are not really available due to design constraints.

Please let me know if you have any other suggestions.

Thanks
John

 

[nbucska]

The "mirror cam" doesn't have to be at 180 degrees- it
may be at any angle -- since it my be mounted on the same
axis as the first cam i.e. on the drawing behind or in front of
the first cam so it may even be at identical angle.

You may use an e.g. rack and pinion to convert the rotary
motion to linear and use twin linear cams, if it helps
in locating it.

What about some electronic solution?


<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[John2004]

Hi Nick,

Thanks for your message. I am afraid an electric solution would not be possible. The mechanism must be hand operated, and there is no way to provide some type of servo or electric assist.

The only real place I could put the mirror profile is right underneath the cam rotation axis, where the springs are presently connected to the cam.

The height of the cam is adjustable up and down by a total distance of 5 mm. The cam plate does not have bearings on each side. The cam plate is mounted to a hub and it has a cantilever bearing mounting, sometimes referred to in bearing catalogs as an "overhung shaft" Bearing mounting.

There is only one cam height adjustment slider, located on one side of the cam only, about 1.5 inches from the cam plate. With an extra spring loaded mirror profile, I think it will bind the cam height adjustment mechanism (dowel slider), and make adjustments difficult. I would have to put an adjustment screw on the extra spring loaded follower, and go back and forth between two adjustment screws, hoping they would not bind.

Please let me know if you have any other suggestions.

Thanks
John

 

[zekeman]

Consider a spring configuration similar to a toggle. The linear spring at neutral would pass through the center of rotation with no spring torque at neutral. If you rotate clockwise it would increase in torque, rotate counterclockwise it would reverse the torque direction, which is what you are looking for.
It would work better if you could use a pneumatic cylinder at constant pressure in lieu of the spring.

 

[John2004]

Hi Zekeman,

Thanks for your feedback.

I don't think I can use a pneumatic cylinder to assist, since both the motion & the "responsiveness" of motion between the activation lever and the roller, has to be controlled manually by have via a lever.

I am having a hard time visualizing your toggle spring idea, and how it could be implemented. Could you give me a little more feedback ? I don't think I can wrap anything around the camshaft, due to design constraints.

Your idea is interesting to me. If it can reverse the spring torque, and still provide a balanced feel in the activation lever at neutral, that would seem to do what I need, if it can be implemented in a very small space.

Thanks
John

 

[nbucska]

Hi John:
Would it not be possible to replace the two (red/violet)
spiral springs with a flat spring with a roller in its
middle ?

This is obviously a more complicated assembly -- there
must be a way to accomodate the few components needed
even if it requires some additional transmission-
shaft,rod,cable,belt,chain, etc.

I don't know where do you have space or what kind
of compromisses you can afford to make. Obviously
the best general solution for the force compensation
is the double cam. I don't know the accuracy required
etc...

By the way how do you know my name ?




<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[John2004]

Hi Nick,

Thanks for your message.

I know your name because you gave it to me when you emailed me from the site I uploaded the drawings too.

I have now determined that I could probably fit a "track cam" in the available space, that can both push and pull the roller. I could then put a spring on the follower, which opposes the follower load. I would still need opposing extension springs on the cam to provide a balanced feel in the activation lever at neutral, and to hold the cam at the centered dwell point. However, they should be able to be smaller extension springs than I am using now. I think this arrangement will make it easier to rotate the cam.

I would use the added spring on the roller to "almost" pull the roller off the center cam dwell at neutral, (off the inside profile of the track). Then, the outside track profile would move the roller when rotating the cam clockwise from neutral, an the inside track profile would move the roller when rotating the cam counter-clockwise from neutral.

However, my main concern is that the needed clearance between the track and the roller could cause problems. If the clearance is too small, the roller will jam in the track. If the clearance is to large, there may be a shock or "bump" felt in the activation lever at neutral, when the roller changes from one side of the track to the other.

Presently, the cam follower bushing is a stock item from

http://www.peerinc.com

(part # 02FB03). These bushings are really made to be pressed into housings, and the engineering department will not give an OD tolerance on the bushing, just a Min. & Max. tolerance on the housing ID.

In order to get a track cam to work, I think I will need a custom close tolerance to roller. I cannot go larger than 3/16" OD on the roller, because the minmum radius of curvature of the cam curve will then get too small.

The roller presently rotates on a standard 1/8" diameter hardened dowel pin.

Could I turn a piece of steel to close tolerance, harden it, and then use that as a bushing ? I tend to think there might be noise and excessive friction. Even at hand operated speeds with intermittent use, I wonder if wear will be a problem. I would prefer to offer a self-lubricated maintenance free bushing.

Maybe I could somehow turn the OD of the Peer bushing to close tolerance, or burnish it to size by pressing it through a die. The bushings are so small (3/16" OD X 3/16" wide)

Another thing that comes to mind is to use a "U" shaped spring at the bottom of the cam, but I am not sure how to implement it. Extension springs are appealing becasue they make everything easy to adjust.

Please let me know if you have any other thoughts.

Thanks
John

 

[John2004]

Hi Nick,

I have another idea that may be viable. I could machine a "rib" in the plate cam and put a roller on each side of the rib. This should eliminate the "binding" problems that you can get with putting a single roller in a cam track.

This will still allow the cam to both push and pull the follower, and I could then use a spring on the roller to offset the cam-follower load. I would put just enough opposing spring force on the roller to "almost" pull the roller off the center cam dwell at neutral. Then the inner rib profile would move the inner roller when rotating the cam clockwise from netural, and the outer rib profile would move the outer roller when rotating the cam conter-clockwise from neutral.

What do you think of this idea ? I will still need opposing extension springs attached to the cam to provide a balanced feel in the activation lever at neutral, but I think they can be smaller. I think it will probably make the activation lever easier to move in each direction. Do you agree ?

The wall thickness tolerance on the bushings is 0.001", so I think it looks promising (i.e., I can keep clearance between the rollers and rib small).

Please let me know what you think. If you or anyone else has feedback on this or other approaches, I would still like to consider all options, but this idea seems most promising so far.

Thanks
John

 

[John2004]

After checking into the rib cam idea, it looks like the pressure angle of the inner rib profile may get to high, since the cam has to be a little smaller to allow for the extra roller.

I am not sure If I can get away with using a very thin curved rib, but it may be possible since there is a roller on each side of the rib i.e., the rollers can possibly give support for the thin rib. If I make the rib 1/4" wide, it seems that makes the cam to small.

Please let me know if anyone has any other suggestions.

Thanks
John

 

[nbucska]

John:
As I see on the drawing there is not too much non-linearity
in the roller_angle vs handle_angle. Couldn't you just
omit the whole thing, move the output directly with the handle and trust in the human_senses -- brain -- hand --
handle feedback loop to position the output correctly ?

Why do you have to ballance the force?


<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[John2004]

Hi Nick,

Thanks for your message.

I must use a cam. I need the dwell on the cam to prevent changes in the follower load, from displacing the follower at neutral. I cannot use another non-backdriving mechanism like a worm gear, or lead screw.

The cam must "block" the follower at neutral. At neutral, the springs do not balance against the follower load, because the roller is in contact with a dwell on the cam profile at neutral, and as such, the roller load will not create any torque on the cam, or cause it to rotate.

Therefore, at neutral, the springs only oppose "each other" and this is only necessary to provide a balanced feel in the activation lever.

Thanks
John