Maybe a dumb questions? Engineering vs pure mass for workbench (Paulk vs Roubo)

[Matt Lau]

Dear fellow Neanders,

This may be a dumb question, but it won't be my first.

Over the past few years, I had to be pretty much a nomad (due to hunting for work, relocating, etc).
While I longed for a Roubo or Nicholson to work off of, I simply didn't have the space or capability to get/make/carry the material in my trusty Corolla.

After 3 failed attempts to make a folding underhill workbench (first attempt: beetle infestated wood from craigslist; second attempt: beetle infested wood from BORG; third attempt: beetle infested wood from a reputable lumberyard), I finally bought a Blum workbench (which works pretty good). It has minimal racking, much less wobble than any Sjoberg, and is relatively light.

More recently, I've been messing around with a japanese workbench (not done yet) on sawhorses based on Stan's suggestion (check post on beach woodworking, beer, and tacos). I've found it to be remarkably capable, nonfussy, and sorta fun (like a lowtech version of a Paulk bench).


Now here's my question: How much is it mass vs resistence/stiffness/lateral support a key variable in workbench design?


I have a 2"x18"x70" thick maple butcherblock that's been air-drying for at least 40 years.
It's the perfect size for a workbench (I focus on mainly guitars).
Any thoughts on this?

I'm tempted to splay the legs outwards a bit like a Moravian bench, or just use a quickly screwed together 2x4 base. Ideally, I'd like to stick my shopvac + dust deputy underneath it.

[Karl Andersson]

Hi Matt,
I built a Scandinavian-style bench out of doug fir several years ago, and it's fairly light (probably under 200 lbs). It gets its lateral stability from the left-side trestle being about 3' wide (the main bench is about 18" wide), and from having softwood feet on a rough chip-board floor. I meant to put a shelf on it that I could load with heavy tools, but it really hasn't been needed. In my experience, you just have to use some extra thought about what you're doing and a lighter bench works just fine. Its primary job is workholding, you just have to take some responsibility for how you apply force and plan a little.

The only time my bench moves a little is if I get sloppy and push too hard while ripping a board held vertically in the shoulder vise. Allowing the saw to do the cutting solves that problem.

Mortising and endgrain chisel work just need stiffness, and your block sounds like it would have plenty. For planning, you might want a little more base width, but I'd only do it in one direction (i.e. plumb legs in the front, angled in the back) so you're not tripping over the legs. Maybe I'm just clumsy, but the thought of working around 4 canted legs sounds like aggravation to me. You might consider knockdown trestles instead of mortised legs - they can be extended to the rear also. If you think you'll have a problem with the bench skating away from you while planing cross-grain or sawing on the benchtop, you could put a stringer right at floor level that you can step on to help hold it still. Sharp tools and some finesse usually keep things from moving around with a lighter bench - and your butcher block sounds like it would outweigh mine by quite a few pounds.
Karl
bench01.JPG

[Jason Dean]

ETA: Karl pretty much said it all already.

The mass of the bench (and friction between the feet of the bench and floor) needs to be great enough to oppose the pushing force you are applying to your plane otherwise you are going to be chasing the bench all over your shop.

The stiffer your bench is the more energy from your mallet blow will go into chopping your joinery and the less energy will be absorbed by the flexing/wracking of the bench resulting in more efficient cleaner joinery.

[glenn bradley]

I think the "innovative technology" benches are cool. For me however, in cars, stereos, bandsaws and workbenches nothing beats raw mass and horsepower. ;-)

[Patrick Chase]

I think the "innovative technology" benches are cool. For me however, in cars, stereos, bandsaws and workbenches nothing beats raw mass and horsepower. ;-)

Mass is king if you've got a fixed location for your big, heavy bench. Clever engineering come in handy when you're space-constrained or when you need to do work on site.

I'm pretty tightly space constrained and am in the early stages of building a torsion box benchtop (with "filled cells" for dog holes and vise) and accompanying sawbenches/legs. We'll see in short order how clever my engineering is (and I might even post pictures!). Designing to accommodate expansion of the solid (poplar) fill has actually been the most interesting part so far.

[Stanley Covington]

Now here's my question: How much is it mass vs resistence/stiffness/lateral support a key variable in workbench design?

I have a 2"x18"x70" thick maple butcherblock that's been air-drying for at least 40 years.
It's the perfect size for a workbench (I focus on mainly guitars).
Any thoughts on this?

I'm tempted to splay the legs outwards a bit like a Moravian bench, or just use a quickly screwed together 2x4 base. Ideally, I'd like to stick my shopvac + dust deputy underneath it.

Matt:

A bench that is useful for all hand woodworking operations needs to be stiff.

One important aspect of this is a stiff top, one that resists impact forces of hammering, chiseling, and sawing, thereby allowing those forces to be transmitted to the wood being worked instead of making the top deflect, vibrate, and bounce. Torsion box designs can easily produce the necessary strength and resistance to deflection, but they lack resistance to vibration of a solid thick top. Most lightweight benches fall short in this criteria, even though they look good on the computer screen and in promotional photos. I think your maple butcherblock will be adequate.

Another important criteria for a useful workbench is resistance to racking forces. Planes and saws put a lot of horizontal forces into the workbench and unavoidably make it rack to some degree or another. The longer the bench's legs, the greater the racking forces a given amount of horizontal force will produce. This racking matters because every minute movement of the benchtop in response to these forces equals wasted energy and potential decreases in accuracy. Accuracy suffers most when racking forces induce eccentric movement (rotation in the horizontal plane) of the benchtop and not just back/forth/right/left movement.

So how does one minimize racking movement?
(1) Minimize the length of the legs. This is the solution the traditional Japanese bench used while sitting on the floor employs.

(2) Add shear panels. Boards or plywood rigidly secured between the legs. This is one of the engineering solutions I believe you referred to, and is relatively modern. This method is what keeps most modern 2X4 wood structure buildings from falling over during high winds and earthquakes. It is the most cost-effective method of increasing resistance to racking. Don't forget that if one is to avoid eccentric displacement/movement, the shear panels must be installed uniformly at front back and sides. This is not always convenient. In a workbench, this method limits access to the space underneath, and gets in the way of one's knees.

(3) Add diagonal bracing between legs, or between top and legs. This is another engineering solution, but is by no means modern. This method can be especially lightweight, but if one is to avoid eccentric displacement/movement, the bracing needs to be installed at front back and sides, just like shear panels. This is not always convenient, and has the same downsides as shear panels.

(4) Use moment-resistant joints between legs and top, and between legs and spreaders. This is the classical solution for workbenches. A moment-resistant joint resists racking forces because the joint itself is stiff and flexes very little. A through-tenon at a thick benchtop is a very reliable method of producing such a joint. Tenons at the end of deep spreaders between legs, either secured with glue/pins or bolts, are also a proven solution to racking forces that works very well. This method does not obstruct access to space underneath the workbench, nor does it get in the way of legs and knees. With some thought, a knock-down workbench can be made. This solution is easiest to accomplish with a thick top, and thick legs, and deep spreaders. I don't think great weight is an absolute necessity, but weight is natural consequence of making the joints effective.

BTW, the semi-dovetail tenon joint of the Roubo bench may look good, but from structural viewpoint, it is greatly inferior to a simple through-tenon.

The reason traditional workbench designs have endured is because they actually work in practice and not just in AutoCad.

Matt, you need a stiff, solid top. With some planning, and if you can accept some compromises, you can make legs and/or a base that are lightweight and effective. The screwed-together 2X4 base will work, but storage underneath will be difficult. I have made several of these over the years, but they were only temporary.

Splayed legs are nice, and can reduce the effects of racking forces to a degree, but are more difficult to make with precision. They will not compensate for a bouncy top.

If you are finally settled down, and are tired of the temporary halfway solutions, I recommend you develop some performance goals for your workbench, choose a classic design that is most likely to achieve those goals, adapt it to your needs (lots of planning and drawing are helpful, I promise), and do it right instead of over and over again.

Stan

[Curt Putnam]

Stan, I am of the opinion (based on zero research) that the semi-dovetail/tenon joint in a Roubo bench evolved because, with that arrangement, the leg vise bears on the leg and not the benchtop. The tenon portion serves to keep the bench from wiggling when planing and doing other such work. I suspect that the folks of old did not do intricate bench joinery for looks. JMO

[Patrick Chase]

A bench that is useful for all hand woodworking operations needs to be stiff.

One important aspect of this is a stiff top, one that resists impact forces of hammering, chiseling, and sawing, thereby allowing those forces to be transmitted to the wood being worked instead of making the top deflect, vibrate, and bounce. Torsion box designs can easily produce the necessary strength and resistance to deflection, but they lack resistance to vibration of a solid thick top.

As I said before, mass is desirable if you can dedicate the space.

With that said, if you've ever worked with honeycomb engineering materials you'll know that they can be made surprisingly damp and vibration/bounce resistant. I'm relying on a couple tricks to mitigate that:

1. I do much of my serious pounding in line with the dog holes, so I'm planning to fill the torsion box cells with solid wood (except for strain-relief slots to accommodate expansion) along that "track". That also makes the dog holes much more useful/effective.

2. In other parts of the box I'm planning to use thin (probably 1/8") diagonal cross-reinforcements in each cell, such that every point on the top surface is within 1" of an internal stringer, reducing any "trampoline effect".

I'm working on a smaller proof-of-concept box right now to test those out. You're absolutely right about racking - I'm assuming I'll need diagonal anti-rack members as you describe.

[Pat Barry]

As I said before, mass is desirable if you can dedicate the space.

With that said, if you've ever worked with honeycomb engineering materials you'll know that they can be made surprisingly damp and vibration/bounce resistant. I'm relying on a couple tricks to mitigate that:

1. I do much of my serious pounding in line with the dog holes, so I'm planning to fill the torsion box cells with solid wood (except for strain-relief slots to accommodate expansion) along that "track". That also makes the dog holes much more useful/effective.

2. In other parts of the box I'm planning to use thin (probably 1/8") diagonal cross-reinforcements in each cell, such that every point on the top surface is within 1" of an internal stringer, reducing any "trampoline effect".

I'm working on a smaller proof-of-concept box right now to test those out. You're absolutely right about racking - I'm assuming I'll need diagonal anti-rack members as you describe.

Interesting. I hope you can post a few pictures of the concept to help visualize it.

[Stew Denton]

Stanley,

I appreciate your post a great deal. I hope to build a bench someday, but it will be after I build a shop (hopefully). I have also thought a lot, as have lots of other folks here, on what sort of bench to build. Your post explains a lot as to the "why" of traditional designs, and as such will certainly have an impact on what I may eventually build.

Again, thanks for the very insightful post.

Stew

[Derek Cohen]

Matt, I have used two types of benches (not including the doors over trestles). The one I have at present is a really heavy Roubo style. It is thick and stiff, and it stays put when I work on it. It is stationed about 12" from a rear wall, but I can move it back a little (with some effort) if needed.

My previous bench was used for about 20 years and was lightly built in hardwood. It was stiff, but was capable of racking. However, it was bolted to the rear wall (just an L bracket at either end), and this caused it to be rock solid. Really solid.

The point is, light is OK as long as you can lock it down.

Regards from Perth

Derek

[Stanley Covington]

Stan, I am of the opinion (based on zero research) that the semi-dovetail/tenon joint in a Roubo bench evolved because, with that arrangement, the leg vise bears on the leg and not the benchtop. The tenon portion serves to keep the bench from wiggling when planing and doing other such work. I suspect that the folks of old did not do intricate bench joinery for looks. JMO

I understand why you feel that way, Curt.

Let's analyze the Roubo leg design using as few engineering terms as possible and without any of the mandatory diagrams.

In a workbench that employs a leg vise, a twin tenon is an excellent, time-proven way to deal with the torsion forces that the leg vise applies to leg and joint. The Roubo design has this feature. Check. Is it necessary for all four legs? No. The other 3 legs could be made with more efficient joints.

Looking a little closer at the twin tenon concept, we notice that the Roubo joint is not exactly a twin tenon, but rather a rectangular tenon combined with a dovetail tenon. Making a dovetail tenon is more work. Does replacing one of the twin tenons with a dovetail tenon add strength to the joint? No, it does not. The dovetail tenon is strictly decorative.

Notice that the dovetail tenon is not entirely housed, but is exposed at the benchtop's edge. Does this strengthen the joint? No, clearly it weakens it. Does it make the benchtop's edge, which is the most-used and abused portion of the workbench stronger? No, clearly it doesn't. Rather it weakens it.

So looking at the dovetail tenon as a part of the leg joint, we can conclude that, while decorative, it takes more time and skill to fabricate, and in return for that effort, the workbench is made weaker. Whether or not this weakness is critical or not is another matter, but I suggest you refer to Mr. Schwarz's account of breaking off a part of his workbench top adjacent this dovetail tenon.

You mentioned the benefit of having the leg exposed at the vise jaws. This looks like a positive thing at first, but a close look reveals that this detail is helpful only when the workpiece being clamped in the vise is either centered on the exposed dovetail tenon, or spans the exposed dovetail tenon and bears equally on both edges of the benchtop to the left and right of the exposed dovetail tenon. A less-than-ideal situation arises when a narrow piece of wood is clamped at the edge of the vise's jaw while bearing only on the benchtop edge, and is not bearing on the face of the exposed dovetail tenon. The resulting off-center force will tend to cause the angled faces of the dovetail tenon to spread out and break off the weakened edge of the benchtop. Clearly, the exposed dovetail tenon at the vise is not an improvement over a plain double tenon housed entirely in the benchtop, and with the vise jaws bearing on a continuous uninterrupted benchtop edge.

Another problem arises when the width/thickness of the leg/tenon shrink/expand due to water content change (humidity), and these dimensional changes do not closely match those of the benchtop. These dimensional changes may be identical, but that is unlikely. If not identical, a step will develop at the benchtop's edge within the vise's jaws. This is not good. But if the benchtop edge is left continuous instead of having an exposed mortise cut into it, no problem.

So while most people will agree that the Roubo leg joint detail with its exposed dovetail tenon is interesting and quite craftsman-like, a simple analysis shows that it is not a functional improvement over a conventional housed twin tenon joint.

Those considering using the Roubo detail will need to decide for themselves whether or not the enhanced appearance and novelty are worth the increased time, effort, risk, and possibly reduced strength and functionality inherent in the design.

Nothing at all wrong with making functional tradeoffs in the name of tradition and fashion IMO. We all do similar things all the time. For example, I wear a necktie more than I like, and while neckties are both traditional and fashionable, they are uncomfortable and have NO functional value.

Stan

[Stanley Covington]

As I said before, mass is desirable if you can dedicate the space.

With that said, if you've ever worked with honeycomb engineering materials you'll know that they can be made surprisingly damp and vibration/bounce resistant. I'm relying on a couple tricks to mitigate that:

1. I do much of my serious pounding in line with the dog holes, so I'm planning to fill the torsion box cells with solid wood (except for strain-relief slots to accommodate expansion) along that "track". That also makes the dog holes much more useful/effective.

2. In other parts of the box I'm planning to use thin (probably 1/8") diagonal cross-reinforcements in each cell, such that every point on the top surface is within 1" of an internal stringer, reducing any "trampoline effect".

I'm working on a smaller proof-of-concept box right now to test those out. You're absolutely right about racking - I'm assuming I'll need diagonal anti-rack members as you describe.

Please do a write-up for us. It would be invaluable.

Stan

[James W Glenn]

As a child I occasionally heard my father muttering about new lightweight steamrollers....

[Kees Heiden]

My bench certainly isn't heavyweight. It's made of fir/spruce in the Roubo kind of model. Top is just shy of 4" and the legs are similar. It totally suits my needs, it isn't bolted to the wall, but leans against it which helps with sawing. It sure doesn't move when planing, either it doesn't groan when chiseling.

So, I guess, it doesn't need to be a 10-tonner.