Peaucellier Linkage to Get linear motion

[kkilst]

Any body has the idea of the bearing arrangement used in the 7-linkages of
Peaucellier linkage. I am making small wood-duplicating machine to use this
linkage on either sides, to get VERY smooth linear movement.
Any body interested to share with this experience ?

[dkowalcz]

I was considering making a small lathe on this principle when I first learned about linear motion linkage arrangements (there are many). However, I found that simple linear rails are far more rigid and easier to use, so I abandoned the idea.

It would be interesting to build one though, I don't doubt that if the engineering is done well that it would work. You might see if you could use one of the simpler linkages that gives very good approximate linear motion with fewer links.

[NC Cams]

There is often a difference between the math/geometry that someone was briliant enough to create and the literal application of same when it comse to to doing something that is supposedly as simple as drawing a straight line.

Unless you have preloaded, clearance free joints, it will be impossible to get perfectly smooth and linear motion.

It is easy to say "perfectly rigid pinned joint" when you're doing math, it is NOT that easy to duplicate in the day to day realities of machine tools. Anybody who's ever done suspension linkage has learned that compliance will exist in anything that has mass, cross section, length and does not have an infinite section modulus...

A preloaded linear guide will give you the most rigid repeatable and affordable approximation of a straight line. Trying to take 7 pieces of linkage and make them slop free enough to create a perfectly straight line can probably be looked at by some (myself included) as complexity in search of a need. I submit the attached as further proof of this contention:

http://mathforum.org/dynamic/java_gs...eaucellier.htm

There are easier ways to draw a straight line with mechanical devices in these days and times.

[kkilst]

Appreciate for the feedback and further information regarding Peaucellier.
At present, I have one manual wood-dupicating m/c with 4 spindles, and
using THK linear guide-way. But somehow, I personally find this contruction
is rather heavy, not sensitive, to carve wood ornamnetal type such as Bali
design, could not get the intricate parts even with small rotating tool.
Looking at LaNova-Scolpitrice design with Peaucellier linkage, and this idea
come into my mind, with 24- spindles, we still could fell the touch on the
model, I had try this old machine belong to other.
I just wonder if I could build one, start with small 4 -spindles.
My concern is the bearing construction on the joint, whether it is using
"oiless bushing" or "needle bearing"
My objective is to produce repetitive wood ornamental, master is done
with CNC router.
Anybody interested to share the experience, mostly welcome ?

[Jason Ross]

If you want ultra sensitive and really cool try http://www.newwayairbearings.com/Fam...Air%20Bushings
Air bearings need air of course and if you do not have it it is expensive.

[NC Cams]

The joints ARE the problem.

ANY clearance in a joint results in flexibility. Thus, developing a "slop free pinned and friction free joint" is the challenge. As you reduce joint preload to allow for reduced friction, you create a potential for compliance that gets magnified by most linkage motion ratio's.

Granted, the "clearance" may be at or below your threshold potential for tolerance but who's to know??? In my business, we have to hold tolerances of 0.0001". Yet, my neighbor machines parts with a +/-0.002" tolearance. His "perfectly adequate" is my "junk".

For most linkage, needle bearings may provide low friction but they also have clearance - thus slop. How much??? Can't say if it will or will not meet your satisfaction but pile the slop of several on top of each other and the parts might be pretty sloppy. BTW, our cam grinder uses a MASTER to copy grind the lobes so we are effectively doing the same thing with our process 8-16 ties per cam and linkage slop gave us fits for EXACTLY the same reason(s).

If I were making a linkage that required stiff yet flexible pinned joints, I'd be inclined to use Heim joints (spherical rod ends). Use the ones with teflon coated ball pivots which preloads them to eliminate slop. Friction is higher but low/no slop potetial.

What amazes me is he complexity that some folks will go to in order to move from one X-Y coordinate to another. In our case, we use a CNC machining center with good ball screws and precise linear guides. We can generate round parts by trepanning around the form to within 0.0001" or so - within the accuracy of some jig grinders.

And why this (a VMC) won't work to repeatedly copy wood forms or whatever with ease is something I don't understand. However, if it is a simple repetitive motion, the complex linkages that don't need a CNC control may be quite adequate.

IF that's the case, you really need to talk to bearing/linkage people as that is where your challenge lies. I'd start by contacting Heim for a catalog of their famous Heim joints (aka spherical rod ends). Alignaball used to supply them and perhaps there are others...

[NC Cams]

Caution: Jason's post on air bearings keeps redirecting you to some outfit trying to sell travel tickets.

BTW: air bearings aren't that simple. You need clean, dry air that is free of oil and water. With the clearances used in air spindles, it takes a bit more than a simple water trap to make one work. The ones we used needed chillers and filters and holding tanks etc and then they were a PITA to keep contamination out of, especially in a hot humid climate....

Moreover, air bearings and many others don't work that well under heavily loaded, oscillatory motion. This is why most bearing handbooks indicated to "contact factory for applications involving oscillatory motion"....

[dkowalcz]

Wow, 24 spindles on Peaucellier linkages!!! Someone once said there's a fine line between genius and insanity.... I have to respect your ambition.

Oil-less bronze and needles are out of the question. You probably want to use preloaded ball bearings if you're going to actually do this, for tightness as well as low friction since feel is important. The preload comes from the assembly style and shaft/housing fits, I'd buy ABEC-3 so as to be reasonably tight without going broke, and you want two per joint no matter what, unless you buy double-row bearings.

24 units x 6 joints x 2 bearings is 288 bearings minimum (depends on the specifics of the joints though). Probably you want to take your local bearing supplier out for dinner and drinks while you're at it.

I guess I don't understand why it would have to be a linear linkage for duplication - any such machine I've heard about used a parallelogram linkage for this, still a hassle to build, but way less complicated. But you would know your own process best.

[NC Cams]

CAUTION:

The ABEC class has no relavancy WHATSOEVER on the "preload" or lack thereof with respect to a bearing. Thus ABEC 1 or 3 clas would NOT have any better or worse preload.

Runout is a bit better with 3's vs 1's but that AGAIN doesn't pertain to preload

Preload comes from an intentional axial load that takes the radial clearance out of the bearing.

Double row bearings (IE 5203's or the like) are NOT preloaded. They have been cheated up to get preload by pressing them into housing and onto shaft BUT this is unpredictable and unreliable.

It is VERY VERY difficult if not IMPOSSIBLE to generate a low friction, rigid joint which ultimately is critical to a slop free complex linkage device like you're looking to make.

Math professors and the brilliant people who concieve the geometry can readily "prove" the linkages will do wonderous things think that slop free rigid joint assumptions enable you assume.

In reality, the creation of a slop free, rigid loint becomes the fly in the ointment that makes the linkages theoretically sound but functionally impossible to create realistically....

Then again, what's slop free to you might be sloppy loose to me... You have to build and decide for yourself but be advised what to expect....

[kkilst]

How to inset picture on this thread ???
I want to insert .jpg Peaucellier wood-duplicating machine.

[dkowalcz]

Here's a picture of the machine kkilst is building a copy of, with the Peaucellier section clearly visible. He requested that I post the picture for him on the board.

I can't tell what exactly they're doing about the joints though, not quite enough detail. In the Y shaped node where three pieces come together, two of them are clearly clevised. I wouldn't think that assembly could get away with simple pins and sleeves though - there are bulky dust caps on the joints which suggest perhaps mounted bearings are being used, so it could be that only the innermost member is free running on a sleeve bearing (makes sense, it has the widest contact surface since it's solid). I'm not sure it'd be possible to assemble it otherwise.

[zhagoox]

I wouldn't use the cell in a lathe but would use it to grind a perfectly flat guide way. For us who doesn't have a surface grinder it seems really tempting.