Dimensional Shrinkage of the Maslow Frame

I have been thinking about the frame design, and one thing I thought of is the tendency of wood to shrink and expand with changing humidity. As humidity decreases, wood loses water and shrinks. Wood shrinks anywhere between 15% to .1% from its wet dimensions. It loses more width than length. That can be substantial when considering a goal of 0.5 mm accuracy over an 8 foot span. Parts cut in the winter could come out very warped if the Maslow was calibrated in the summer. Plywood shrinks 1% wet-to-dry, according to an online source, whereas solid wood only shrinks .1% in its structural direction. Pressure treated lumber is sopping wet when purchased, and shrinks substantially.

The structure should be designed to minimize how much the distance between the motors shrinks, to maximize the dimensional stability of this distance. This could be accomplished in many ways. First, solid lumber could be placed directly between the motors, which leverages the most dimensionally stable wood selection (vs. plywood). Another option is to use a steel channel placed directly between the motors. A third option is to use a steel cable stretched between the two motor arms; if the distance shrinks, the cable will sag and it will be visible. If wood is used, there could be some guidelines about how to ensure its moisture has stabilized prior to building and calibrating the Maslow.

Thanks

1 Like

This was one of my concerns when I was thinking about building my frame. @Bar had updated the frame a little while ago to use more solid lumber for that stability along it’s length. Originally, the frame had been made mostly from plywood and there were reports of the ply bowing and twisting in high humidity.

This is a top-beam frame, which we have seen the greatest successes with. @Dlang has suggested the best material for the job is LVL, which is a special type of composite structural lumber that has incredible dimensional stability. I have worked with the stuff building houses and they are great! They’re available at most lumber yards and the big home improvement places usually at 10 ft lengths.

Solid wood might be a good option as well, but it still could twist or bow and that would significantly affect your distance between motors.

YES! and this is the route I’ve taken. Compared to wood, steel has much better dimensional stability. It does flex more but as long as you get a good structural section like channel, pipe, tube, or unistrut, that won’t be a problem at the forces we’re dealing with here. I welded a steel tube frame myself, and the machine is very stable.

The easier steel frame to build is the Unistrut frame, and you may want to look into it. It’s a little more expensive than a wood frame. The thread goes into pretty good detail about it.

I like this idea, and think it could be used well with a wooden top beam frame. I think having a visual indicator that something’s off is a great idea.

If you haven’t already found it, Musings on frame design and Frame design options are good discussions on frame construction.

3 Likes

This is one of the reasons for switching to a top beam frame design.

Also, there is the option of using metal unistrut (which expands/contracts
almost as much due to temperature changes)

see the topic here ‘musings on frame design’ and the github ticket

I purchaseda LVL, and the lumberyards sell them by the foot, up to 80’ long, so
they are a good choice if you want a wood top beam.

unistrut is the other good option here.

price and weight for unistrut vs LVL are about the same.

you don’t need the entire frame to be unistrut/LVL, only the top beam

1 Like

Thanks for the responses. One thing to note is the distinction between dimensional stability and rigidity. An extremely rigid frame can still be subject to dimensional instability of wood due to humidity fluctuations.

very true but the only thing that really matters in a top beam design is the
length ofthe beam between the motors. Everything else could shift pretty
significantly and it woudln’t matter as far as accuracy.

For a wood beam, this is movement along the length of the beam, where wood is
the most stable.

If the beam (or the rest of the frame) warps, it can shift (or rotate) the
coordinate system, but all the distances are still going to be accurate. This
could matter if you were trying to use the edge of the plywood as part of your
design, but would not affect the ccuracy of any parts not including the edge
of the plyood.

Want to watch out for twist and warp in the beam as well. Humidity can cause these in low-grade lumber.

Twist is a problem because it can cause problems with the chain, but warp
wouldn’t be a big problem unless it was bad enough to cause errors in the
distance between motors

With a warped board, you can wind up with the motors at different distances above the workarea. Warp is often asymmetrical.

That is true, but since the maslow tolerates a fair bit of misalignment of the
motors from the sled (remember, Bar has never adjusted his motor distances),
unless the board warps enough to move the motors an inch or so, it won’t matter.

And if the top beam is adjustable, you can adjust this out.

if one end of the board goes up or down significantly, it will affect where the
machine cuts, but as long as you aren’t too near the edge of the workpiece (and
don’t need to align your work with the edge of the workpiece), that won’t matter
much.

dlang, I agree. Many of the dimensions of the Maslow don’t need to be extremely precise. The most sensitive dimension is the distance between the motors.

blurfl, I hadn’t thought about twisting. That does seem like a likely failure mode.

Regarding the comment about steel shrinking or expanding with temperature. This is actually a desired property, because the chains shrink or expand in the same way. Cuts could be warped if the distance between the motors didn’t fluctuate similar to the length of the chains. It seems like steel is preferred, because its dimensions would parallel the shrinking and expanding of the chains.

Structural lumber is an option if built correctly, to avoid warping. This is accomplished by laminating multiple small pieces together. The result would be something similar to lvl, except the laminations would likely be thicker. We would have to ask ourselves why not just buy lvl in the first place.

Although no numbers have been calculated, particularly in comparison to the accuracy goals of the Maslow, I have drawn this conclusion. In terms of dimensional stability, steel (because its tendency to shrink and expand is the same as the steel chains) > lvl (because it shrinks and expands .1% in its structural direction) > plywood (because it shrinks and expands 1% in both structural directions).

All the numbers stated (e.g. 1% expansion/shrinkage for plywood) so far have been maximums pulled from the internet. For example, using a 0.1% expansion and contraction over a 10 foot span results in 1012.001=~1/8 inch, which is likely too high for the accuracy goals of the Maslow. It makes sense to actually do a calculation of how significantly the lumber expands and contracts in realistic conditions. This means finding out how much the water content changes spanning a reasonable ambient range, find a sensitivity to changing water content in terms of %length change divided by %water content change, and determine a more accurate number.

Regarding the comment about steel shrinking or expanding with temperature.
This is actually a desired property, because the chains shrink or expand in
the same way. Cuts could be warped if the distance between the motors didn’t
fluctuate similar to the length of the chains. It seems like steel is
preferred, because its dimensions would parallel the shrinking and expanding
of the chains.

it’s not that simple since there is more chain than beam, if they all expand the
same % it will still cause errors

Structural lumber is an option if built correctly, to avoid warping. This is
accomplished by laminating multiple small pieces together. The result would
be something similar to lvl, except the laminations would likely be thicker.
We would have to ask ourselves why not just buy lvl in the first place.

you can also just create your beam with pieces of plywood glued together.

I don’t see why plywood would have more change than LVL

Imagine a triangle consisting of the two chains, and a line connecting the two motors. If all three lengths shrink by a same fraction, all of the three angles in the triangle would not change. Even though the whole cut would shrink slightly, it would not become skewed or warped.

Definitely outside my expertise, but the wood grain in all the laminations in LVL are in the same direction, whereas the wood grain in plywood is alternated. In plywood, the wood grain in adjacent laminations are perpendicular to eachother, such that half the laminations are oriented perpendicular to the direction of interest. LVL beams are more dimensionally stable because all the laminations are in the structural direction.

1 Like

umm, no, in LVL the grain is also criss-crossed.

According to Wikipedia, all laminations are in the same direction

1 Like

If that was the case I couldn’t see it being noticably stronger than plain
lumber.

I’ve got a LVL beam at home, I’ll double-check it.

1 Like

I don’t know how that works. The best I can think of is standard lumber has large imperfections, whereas imperfections in wood is broken up and distributed throughout an LVL beam, since they are made out of thin laminations. Imagine one large knot being cut into thin sections and distributed along the beam, or among multiple beams. Because a board is only as strong as its weakest part, LVL is stronger.

Also, LVL beams may be compressed, or densified.

Lamination creates an extremely rigid structure in the plane of the lamination. For example, a 2x4 is much stronger in the 4 dimension. An lvl, essentially, capitalizes on that strength and prevents sideways movement by adhering the laminations together. Also your suggestion that it has fewer imperfections and is denser are both effectively correct in that most imperfections are removed by a veneer selection process and the laminations are pressed together with the adhesive filling in much of what would be space in solid lumber of a similar dimension

This difference in longitudinal shrinkage of the solid wood and LVL is pretty similar. What is different in the LVL are the laminations which create a more uniform and square product. Because the veneers are dried before assembly, LVLs very rarely warp or twist.

I think you want to make sure you are using the long direction of lumber in every critical direction of the machine. Shrinkage is minimal in this direction.

I created a top beam frame, actually very similar to a stud wall with a doubled top plate to support and stabilize the motors.

3 Likes

@Joshua is correct. Wood is immensely strong but being a natural chaotic material does not behave consistently so you have to reduce the absolute strength a lot when calculating the size of timber need for a job. The lamination’s spread out the imperfections so allowing more the inherent strength to be depended on. Basic lumber is rated 16N/mm2, prime lumber 24, Oak is 35 but LVL is rated 45 or more (Metric strength values but proportional)

3 Likes