Computer Numeric Re-Considerations (VIII)

It’s time, I think to cut to the chase in this series. Over the past 7 posts on the topic of CNC automation, a technology which began innocuously enough with developments in such things as automated looms, windmills, and census survey tabulators, I have delved into some of the historical currents in technological development that have led to this point in time. Obviously, this is a highly complex topic and at best I have only touched upon certain points here and there. Readers wishing to examine this topic in further detail may wish to consider reading the works of far more learned scholars than I – such books as David Noble’s Forces of Production and America By Design, Lewis Mumford’s Technics and Civilization, and Jacques Ellul’s The Technological Society would be worthwhile starting points. In those books, if you’re looking to receive happy confirmation about the rightness of technology as the cure for all our civilization’s ills, you will likely be rather disappointed. As Pink Floyd put it, Welcome my son… to the machine! Of course, this message doesn’t market quite so well as the bright shiny futures otherwise laid out for us, so perhaps these works are to an extent condemned to a certain amount of obscurity.

It is quite obvious, reading David Noble, that advances in automation were taken up by industry for the principal reason of eradicating the ‘problem of labor’ and ‘soldiering’, and not because they were essential to 95% of the sort of work normally performed by machine shops. The vast majority of machine shop work occurs in 3-axis format, that is, in the x-, y- and z-axes. Conventional lathes and milling machines, run by humans, accomplish this sort of work easily and with good precision. Human skill is definitely involved. The large NC and later CNC machines developed primarily by the US Air Force and MIT were to solve problems in forming shapes of helicopter blades and jet aircraft frames, which are complex shapes not even do-able accurately with conventional 3-axis work. For those, what was needed was servomechanism-controlled motors to move the cutters with utter smoothness in the 3 basic axes, along with the capacity to work in two additional axes. Some may be unclear on how there could be any other axes besides x-y-, and z. Consider the 4th axis: the piece to be machined can also be spun, like a lathe spins a piece. This spin can happen with the piece turned in either a vertical or horizontal axial orientation – depends upon the type of machine. That’s the 4th axis – part rotation. The 5th axis is that the cutterhead itself can rotate in an orbital fashion.

Here’s a video of a CR Onsrud 5-axis machine cutting a sphere inside a cube using that orbital head movement and a moving table:

5-axis machining is what you would need for pattern-making complex helicopter blades, or certain types of staircase handrails, or some chair parts. Most machining however, doesn’t require 5 axes, and many machine tool businesses were not initially interested in the Air Force’s new toy, especially given the very high capital cost expenditure they represented. In order to spread 5-axis technology, the Air Force tilted the playing field a bit, both providing the machines to certain shops and then specifying that the machines be used for certain projects, and later by having product specification programs created so that the only machines which could use the programming were 5-axis. The Air Force at one point decided to stockpile machine tools, and a certain percentage were specified to be NC/CNC type, and this eventually encouraged the machine tool shops to jump into the arena. Like a lot of what passes for capitalism, state and big business are joined at the hip, cost inefficiencies/waste are socialized, as they say, while profits are privatized. Sounds like a crazy way to do things, but it works for some I guess.

CNC was not promoted because it was a cost effective approach to machining. The machines weren’t that. Plain and simple, management wanted control as a long-cherished goal, and management got that control eventually by technological means. That’s the story told by David Noble, and I am persuaded by the evidence, personally. But does that history change the fact that this technology exists and that I, and you, are faced with the fact that many of the things in the physical, manufactured environment are produced by robots and automation? Not a whit. It’s there, and some of the outcomes of that technology are admittedly good things, but not all.

Anyone who watches a film such as Manufactured Landscapes and sees people in China on assembly lines doing nothing all day but test aerosol nozzles, would be hard-pressed, I think, to come away from that thinking that it was a good thing that there were humans doing such jobs, and would prefer seeing such tasks automated. Take a look and judge for yourself:

Now, at present it must be so cheap to have a human do it all day that it makes the most economic sense for the owners of such factories, and perhaps soldiering isn’t such a problem in China, I don’t know. These workers are toiling away with dedication, doing work I don’t think I could do, or would ever want to sign up for, that’s for sure. One place where CNC manufacturing makes a lot of sense are such repetitive tasks as seen above, and yes, it is, when you get down to it, a luxury to use CNC (except when labor costs are higher, then the equation changes). But if you’re going to have a society built around mass production and consumption, then either you have humans doing utterly mindless work like that, or you have machines doing it and, as a result of those machines, people out of work. In North America, the choice has long been to automate and/or offshore the production, with a result that the manufacturing base here is very nearly non-existent these days. What’s left are service jobs. Curious too that the word service stems from the same root meaning – slave -which gives us the word servitude… but I digress.

Today I want to look outside the historical context and consider why a craftsperson may want to use CNC. It’s there, it’s a tool, and (while it might be purely illusory) I have some choice, it would appear, in whether I want to make use of that technology or not, regardless of how and why it came about, just like I do with cars, washing machines, toaster ovens, etc,. Craftsmen in the 1800’s would have had the same choice in respect to the advent of powered woodworking machinery, and craftsmen even further back in time would have a similar choice to make about the new-fangled hand saw, which could ignore the run of the wood’s grain, or, even earlier, the chisel now fixed into place in the configuration of a plane. Sacrilege!

What is new and radical in one era becomes a standard piece of kit in another, and later perhaps a quaint anachronism in an even later period – – as planes, saws, and chisels are for the bulk of woodworking companies these days. Possibly every period has those who are reactionary towards the new and deliberately choose not to employ it because of a belief in the virtue of ‘tradition’, or philosophical objections of some other sort, subjective and temporary as those ultimately prove to be in most cases. While it is not inevitable that technology marches on, I live within a system where technological change has become normalized and encouraged, and that change has a force of movement hard to resist, a river with a hard current to swim against. It could be other ways, but it is not – not at the moment at least.

Most woodworking machines are popular these days as they reduce what I might call grunt labor. Almost all woodworkers, I would imagine, would prefer to re-saw a piece of timber with a powered saw (bandsaw or circular saw) than to rip it with a handsaw, especially if there is a stack of such material to be dealt with. Now if I was a scabbard maker for a Japanese sword, I might choose to follow tradition and hand rip the 1 meter long (or so) board for that scabbard down into two, even if it took 50 or 100 times longer to do it than by machine. But if that scabbard was instead a stick 12 feet long, and I had to rip 10 of them in a row, I can imagine the pressure would be there to adopt some sort of machinery after a while. So, the scale of work tends to also be a factor in whether machinery is adopted or not. The more you have to do, and the larger the material, the more likely it is you will choose to use powered equipment. Some take machine use as the only way to work the material however, and that does have certain shortcomings. Does CNC woodworking equipment reduce labor in many circumstances? Sure it does. I’ll return to the labor issue from another angle shortly.

Another reason to use machinery is for precision, or at least, predictable results. Here, woodworking machinery is really no better than hand tools in achieving precision. That said, many woodworkers, due to lack of experience with hand tools, or at least with quality hand tools, struggle to achieve precision with them. I wonder sometimes how many Stanley planes have been thrown across the shop in frustration after another piece of fine wood was torn out by the blade? Looking in antique shops, I see no shortage of hand planes that are beat to shit. Somewhere along the line, someone used that tool with no idea as to what they were doing. it has long been a mistake, I think, to confuse apparent mechanical simplicity with actual simplicity. The humblest hand tool is a summation of a huge amount of knowledge in some cases.

If you buy a powered plane though, most times you can achieve decent results by plugging it in and running stock through (or under) it. Much easier, even if a power-planed surface is no match for a cleanly hand-planed one.

Let’s say I had a new guy in the shop and wanted him to cut the end of some sticks at 90˚. On the one hand I could walk him/her over to a powered chop saw, set up a stop and in 30 seconds or so familiarize him/her with how to use the machine, make sure they have the appropriate safety gear on, and away they go. They’d be producing work at a high accuracy level repetitively. On the other hand, I could show them how to check a try square, then sharpen a marking knife, then layout the cut lines, then get a cross cut handsaw and make that cut. Would I let them have my best handmade saw with the teeth nicely filed and set, and the blade evenly tensioned? Would you? No – you’d probably give them a ‘jobber’ sort of tool and you would well expect that they would damage the saw soon enough, either by muscling the cut and kinking the blade, or breaking a tooth, etc. And, odds are high that they would take a long time to make the cut and the result would not be all that high quality. We’ve all got to start somewhere. So then one would need to show them how to sharpen a plane or chisel, set up a plane perhaps, and then demonstrate how to clean off the end grain until a 90˚ surface is obtained. If that person did nothing all week but make 90˚ cuts, and I was willing to soak up the cost of that as their employer, then probably by the end of a week they would be getting decent at it. A month later, they would be executing fairly accurate work, at least as far as that task goes. One could lay out a similar choice between hand-planing a surface clean and machine planing it down with a super-surfacer.

Obviously, the vast majority of people would want to take that new worker over to the chop saw for the 30 second primer in operating the tool and put them to work, and not lead them down the other path. In the end, the worker who learns to make the accurate cut by hand really develops a skill which is something not inherent in the work using the powered chop saw. The quality of the work is fairly close between the two, though the chop saw will always be faster in that case.

One point in the hand saw’s favor in such a scenario is safety. A mistake with the chopsaw can be a quick way to lose a finger, while it would be tough for a person to do more than nick themselves with the handsaw.

To obtain highly accurate results with either hand or machine tools requires a lot of knowledge and patience/desire to learn the details. Once the machine is well set up though, the production speed is greater than by hand, but I would say that ultimately one can in fact achieve higher precision by hand, with the use of jigs (like, say, a shooting board), than by machine. By most machines at least. You see, most woodworking machines are not made with a particularly high degree of precision (something they share with cheaper lines of hand tools), and worse, most people using the machines do not know or care about setting up the machines to obtain precise results as they are not particularly concerned, it would appear, with getting such results as a part of what they do. I make these comments on the basis of my observations of shop practices in several shops I have worked in or visited over the years. People by and large who do woodwork are dependent upon their machinery, but curiously seem to have no grasp of the workings or adjustments of those machines. It’s a form of helplessness, let’s face it.

Now, do CNC machines add to that technological dependence? I think the answer to that is both yes and no. In most cases, I would suspect that the people using the CNC-based woodworking machines, due to the characteristic divisions of labor which are employed for reasons of efficiency, are in the zone of helplessness. The programmer will have a specialized form of knowledge that allows him to do his/her job, but they are not actually woodworking at all. Those handling the wood are often doing little more than preparing the wood blanks for cutting and then loading, observing the machine work, then cleaning and unloading. That’s not really woodwork either and, further, many of those workers would probably have no idea what to do if a glitch arose in the program. And if the machine broke down, I suspect that another technical specialist would have to be called in to fix the equipment.

On the other hand, there is an intriguing DIY aspect to CNC automation that accounts for a small slice of the pie. There is a certain segment of people designing and making their own CNC machines, welding up support tables, building a gantry, fitting in linear slide mechanisms, hooking up their own PC and then designing and programming. That’s a totally different kettle of fish and is in fact a far more empowered and engaged process of making than that which characterizes a lot of woodworking. How many woodworkers out there build their own machines? Yes, it is some to be sure, but only a very small minority of people. The vast majority of us are consumers of machinery and tools, not makers.

If you’re a designer/maker, then the DIY CNC has a definite appeal, and quite empowering it seems to me. I think if working with CNC is otherwise a division of labor where one person designs, another programs, another loads and unloads the machine and another packages the parts, then any one of those individual tasks are not the sort of woodworking I would personally like to do. It would suit some folks though, to be sure.

I think the issue, for me, really boils down to this point – am I using the machine to produce what my mind envisions, or am I working for the machine, only a cog in a larger mechanism of production? There’s a sizable gulf between the two places.

I should mention one thing about programming – in order for the outcome of the cutting to be good, the person doing the programming has to have a good knowledge of wood as a material and of tooling, and be able to envision a logical series of steps to effect the final product, and direct the machine to employ tool paths that are safe and give a clean result. So, in that sense, there is a definite connection between CNC programming and regular woodworking, in terms of how the job is carried out and how the cut out must respect the nature of the material.

Now, the 5-axis CNC woodworking machines are able to accomplish things which are very difficult (if not impossible) or very time consuming for a human to do, and in all cases, the 5-axis machine can accomplish a highly precise result. I can totally see the advantage in using such a machine for cutting out the complex forms of handrails which are employed in geometrical staircases. And if one has to make a series of identical parts of complex form and high precision, the CNC clearly offers advantages.

I’ve been communicating with several glass companies in recent weeks in regards to the glass panel for the current coffee table project. This glass top has a form consisting of 4 half ellipses that meet with re-entrant corners. I’ve found companies that want me to send them a paper or cardstock templete of the top and then they will hand-cut and hand grind the glass to that template. I’ve found other companies which have CNC glass cutting equipment on hand and I can send them my drawings directly. And guess which option I have more faith in, in terms of achieving a product which conforms exactly to my design? Yes, the CNC. I don’t have much faith, sad to say, that the hand-cutting will be done by people who will really care to work exactly to my template or who have the skill to work to the level of accuracy I want. That’s a rare sort of person, in my experience. I do trust the CNC to produce exactly what I want, within a thousandth or at worst a hundredth of an inch. Maybe I’m imagining this, as some sort of techno-fantasy? If the hand-cutting added a certain hand-cut nuance to the result which would be charming or desirable, that would perhaps change the situation for me, but no, I want the glass to be made exactly as I have designed it, and with an even edge finish and an even edge chamfer.

So, okay, I’ve looked at the issues of accuracy, repeatability, and skillful work in terms of CNC or not. Another issue, of a more personal sort, relates to the nature of the work I am engaged in from a dollars and cents perspective. Obviously, if dollars and cents were my main concern in life, then woodworking would not have been a wise career choice. That said, if the dollars and cents do not roll in with at least some frequency or reliability, then the shop doors get closed, creditors might be owed, machines might need to be sold off, depression might ensue, and personal relationships can suffer, etc. If you have a trust fund, woodworking might make perfect sense, but furniture making on a small scale is tough sledding sometimes. Well, almost all the time actually.

In the kind of work I do, as some may have noticed, I tend to employ rather time-consuming methods to achieve the end results. I use these methods because I believe they are the highest quality way to do things. This means that in order to make a living, and I mean that in the middle class sense of the term, the work isn’t what you might call ‘inexpensive’. That means that the consumer base for my work is out at the truly thin end of the wedge, if you know what I mean. Japanese architectural work, the real stuff, starts somewhere north of $500/square foot. I’ve worked on projects where the price was around $1000/square foot, and there are definitely levels above that in terms of quality and intricacy. The number of people in the US who can afford such prices are fairly tiny – perhaps in the 1% zone. And out of all those few million individuals, the number who have an interest in Japanese architecture is probably again on the order of 1~2%. Still, that is a potential market perhaps of 2500~3000 individuals, as a bit of a ‘guesstimate’. Doesn’t sound too bad, but boy it is hard to crack into that group and get the steady work that I would like to have. Since I moved to the US three years ago, while I have had some near misses in terms of projects, I still have not gained a single client for such work. And I’ve tried hard to make those contacts. So, I make furniture in the meantime, but the furniture equation still adds up to seeking clients at the thin end of the wedge, and it is still hard to find those clients as well. It is definitely true that marketing is an area requiring a lot of effort and perseverance and it can take time for things to pay off.

Obviously, it would make sense to find a way to bring the cost of the product down a bit so that I could appeal to a thicker bit of that wedge and have a dramatically larger pool of potential clients to draw from. I would also like to make the work more affordable for more people. The only ways I know of to make the price lower, while still remaining economically viable is to lower the quality/technical sophistication so that the work is faster to do, and/or to do that work faster somehow, and/or to employ cheaper materials masquerading as higher quality ones. Many shops will choose to find cheaper ways – using dowels or biscuits instead of joinery, veneered particle board or plywood instead of solid wood, etc., etc.. Make the material inherently predictable, then jig up and set up so that identical pieces can be predictably cranked out in volume, preferably, once the shop gets a bit more work on hand, by relatively un-skilled people receiving lower wages. We all know the routine.

And if the choice is not to hire unskilled people and instead to do the work yourself, to not get big, then for most the work becomes, from all appearances, rather dull. I have yet to meet someone doing the MDF-plywood substrate, glued veneer, biscuit-joined carcase, orbital sand ’til you drop and spray it with pre-cat lacquer routine on a regular basis, making boxes as it get called, who finds the work remotely interesting, who goes home at night psyched to come back the next day, who finds a delight and engagement in the challenge of the work, and who tells me with a glint in their eye about the next project they have lined up to do. It could happen I suppose.

“But, it pays the bills“.

“Hey, I’ve got a job“

Fair enough. Maybe for many the concept that work should have any fun to it at all is counter-intuitive. Maybe for many the social conditioning has taken hold that making stuff, blue collar work, the trades, is primarily a matter of physical exertion, not intellect, not the heart, not one’s passion. Too bad, because I think they’re quite wrong about that, at least in terms of what can be done.

For philosophical reasons, I choose not to go down those avenues with the parade of the boxes. So, I am left with finding ways to do the work faster without lowering the quality if I am to get the price down. My own physical speed of work can increase only so much, a few percent here and there, and in fact there’s obviously a point where the likelihood of injury would increase, or the odds of messing up a piece, as well if I’m flying along too fast. The only possibility really is to make the same exact things over and over again so the jigs are done, the bugs worked out, the material bought in quantity, and then one can crank away to heart’s content.

People suggest to me all the time that I should develop a product line, something I make in runs of a dozen, or fifty, or a hundred. Sounds good, but first there is the delicate and not to be overlooked matter of generating the sales for those massive product runs. Perhaps one day when I land a job to furnish a hotel or some pie-in-the-sky thing like that, the opportunity will be there to jig up to make 100 identical tables, but until that time, the closest situation will be making a set of chairs for a dining table, where the same piece, more or less, gets made 4, 6, 8, or 10 times over.

Another matter too is following the train of thought out a little further, of the factory approach, and see where it leads. Once I’m cranking it out, then I’m limited by machinery and employees, and once I go down that path of getting more of each I am entering into competition with Thos. Moser; get any bigger and I’ll have to open my own Furniture Barn outlet and be wondering how to take Ikea down. By that point there will be a board of directors who will probably fire me as CEO since I would be in the way of getting the profit margin to the place where investors are likely to be more attracted to the proposition. That’s not a path I want to walk, and that’s not an arms race I want to engage in.

If you’re making one-of-a-kind pieces, some would suggest that one day such pieces will be collectible. And how does the volume production of a piece affect that? Usually in a downwards manner. So I think the craftsman needs to consider all angles in terms of whether to up the volume.

Possibly, too, a client will come along who will see a piece I have made and want the exact same thing made for themselves. It hasn’t happened yet. One of the things I offer that the factory approach generally cannot is truly bespoke, one-of-a-kind pieces made for each client. I’d like to keep offering that type of product, but somewhere between the one-of-a-kind and limited production there has to be a sweet spot that would work, right?

Let’s look at the employee side of things. Right now, one of the limits on the amount I can get done is the fact that there’s only me doing it. But when I consider an employee, it’s not so appetizing a proposition actually. I know for a fact that pretty much anyone I would hire is going to make all the same mistakes I would make, plus a bunch I wouldn’t (because I’ve already made them and have learned my lesson – hopefully). Those mistakes will cost me money. I don’t meet too many folks who are as much of a perfectionist about woodwork as I am, and thus I expect most anyone I would hire would produce things at a lower quality standard than I might like. I would anticipate friction over such matters in fact. If the person is a beginner, they are unlikely to have much in the way of tools – am I to supply them? Then there’s the long learning curve with sharpening and plane set up… And if they are already skilled and tooled up, why would they want to work for me? They could have their own woodworking business.

So, in delineating some of the issues, I have come to see the CNC router for woodworking as a means of accomplishing accurate, limited batch production in the woodworking shop. One interesting point about CNC operated machines as compared to regular woodworking machines is that once the set ups are complete, the machine can be largely left alone to do the work, freeing up the craftsman to focus on other aspects of the project at the time the CNC is doing the cutting work. You can leave the machine working overnight in some cases. That’s a definite plus.

I’m already doing an extensive amount of CAD drawing, and am very familiar with the correct sort of tools paths and cut sequences one would want for various woodworking cut out tasks, so the steps in getting to the point of becoming savvy with the programming end of things are not too many in number in my estimation. I’ve already started playing around a bit with Rhino 3D, one of the more popular design software systems.

And when I think about the nature of woodwork, at least as I do it, what it boils down to really is intellectual property. That’s what design is, intellectual property. Is the fact that I can run a stick of wood through a bandsaw or planer, or chop a clean mortise some sort of unique thing to me as a worker? No. Is knowing how to go about a job from start to finish unique to me? No. Is knowing how and when to sharpen a tool, knowing what a cleanly planed surface is supposed to look like, knowing what a good fit of a dovetail consists of – are any of those unique to me? No, not at all. But the sum package of design savvy, knowledge of joinery, knowledge of historical underpinnings, knowledge of and sensitivity to materials, knowledge of hand tool and machine tool use, ability to market the work and interact with the client to produce something they really want – – it’s that package that comprises my intellectual property, and is unique to what I do and make.

Another point: when it comes down to it, few if any clients are going to care in the slightest whether I used a bandsaw to rough out the curve or an adze, or whether the dovetails are hand cut or come out of a router jig, whether the joint is held together with a Domino or a locking Japanese hako dome miter joint. Those things might matter to me a great deal, but for most clients, they care about the beauty and apparent quality of the finished product, and I think for many clients just knowing that the maker is a fanatic is all they need to know. Just as if I were to buy a Ferrari, say – I might have no idea or have little care about the metallurgical techniques behind the titanium connecting rods in the engine, but because it IS Ferrari, I know they will do their utmost. And that’s the main thing. U-huh, branding.

In some cases, the rough cutting of a thing has only distant relation to the finished product, Why on earth would it matter if the part was cut one way or another? Consider the table frame piece for the coffee table I am making right now. It’s a semi-elliptical section with a moulded edge profile and a rebate on the top surface. To make this part I might typically:

1. Cut the blank to rough size
2. joint and plane the blank close to dimension
3. make a template of the elliptical plan
4. trace the template outline onto the stock and rough it out with the bandsaw or jigsaw
5. joint and plane again, taking the stock to dimension
6. make a fixturing jig to hold the stock and process the cut in several stages using on-hand and/or custom shaper/router tooling to achieve the desired edge profile and top rebate

Or, I could do this:

1. Cut the blank to rough size
2. joint and plane the blank close to dimension
3. trace the template outline onto the stock and rough it out with the bandsaw or jigsaw
4. joint and plane again, taking the stock to dimension
5. fix stock in CNC router. Router executes the profiling cuts, using stock and /or custom tooling

In the end, I will have much the same product in my hand, a piece of wood that still requires joinery to be cut on the ends, and the surfaces planed and scraped, and finished. Did it really matter at all that the cutting of the shape and profile was achieved with a hand-guided fixturing/pattern jig on the shaper and/or router table or that the work was fixed in place on a table and a computer controlled cutter achieved those same cuts? I don’t think so. In both cases, I designed the piece and decided how I would effect the cut out, and in both cases my hands are going to be involved in getting to the finish line. In both cases a spinning cutter, a router/shaper, did the work.

In fact, cutting an elliptical form on the router or shaper involves a certain amount of risk at the ends of the cut where the curve of the ellipse swings sharply around and across the run of the grain – an easy place to have some blow-out or worse. In this case the CNC router would actually be a safer way to do the same task.

Of course, I’m not in a financial position to buy a large CNC machine right now. The 5-axis Onsrud machine featured in the first video above costs more than $100,000, maybe more than $200,000. So, if I am going to explore CNC a little further, I will need to buy an old 3-axis machine (still more than $20,000) or make my own machine. The DIY thing appeals to me a lot.

First though, I have decided that a good way to get my feet wet and see how satisfactory a product could be achieved, is by having some of the coffee table parts I designed cut by CNC router by a company in NY state which offers CNC manufacturing services. At this point they are several weeks behind schedule, but I hope to have some frame pieces in my hands by the beginning of next week or so. I’ll post up about the results on that thread and have more reflections about it at that point. Fingers are crossed. For now, I’d like to draw this thread to a close – thanks for hanging in there and I trust that the fairly long entry today hasn’t been overly tiresome.

Thanks for coming by the Carpentry Way. Comments always welcome.