Tools That Make Other Tools
August 19, 2009 10:27 AM   Subscribe

How did we go from basic rough hand tools to the point of having really precise tools? Specifically, how do you make a better tool from the ones you have on hand?

I came across an article on Make: Magazine about casting parts and building your own machine shop that got me thinking about how we as a species have arrived at very precise tools that are capable of creating a perfect sphere. The statement that got me thinking was:

"The order is important, because each tool requires the use of the previous machines in its construction."

How did we progress from only bare hands to these very precise machines? What fundamental laws of the universe can be harnessed to build these tools? I would expect that it required the development of measuring devices that could also become more precise.

Some of the other things that got me thinking:

The "Universal Measure" (as I found out from that scene in Stephenson's Quicksilver with Hooke and the mercury on the table), or the ideas of "straight", "square", geometric construction, linkages, etc.

Are there any good resources that speak about this? How would one go about recreating the level of precision we have now if trapped on a (technologically) desert island?
posted by toomanyplugs to Technology (17 answers total) 13 users marked this as a favorite
I know very little about this, but your post made me recall a neat series of books I heard about.

Dave Gingery wrote a series of books about making your own metalworking shop from scratch. You can see the books on Amazon. One reader says:

With this series you take an aluminum storm door, a bag of charcoal, some scrap wood and sand, and an old vacuum cleaner. With them you create a milling machine, lathe, drill press, scraper, and many fascinating deluxe accessories that actually perform (and in many ways out-perform) some of the same machines you buy at Sears.

One website gives some specs for the lathe you will build: 7" capacity metal cutting lathe accurate to 0.001", 12" between centres.

0.001" accuracy sounds pretty good when you've built the thing from scratch, casting the parts yourself through sand casting.
posted by splice at 10:38 AM on August 19, 2009 [3 favorites]

Some early stone tools, like pressure flaked micro-blades, were as sharp and most likely sharper than any knife in your kitchen.
posted by nestor_makhno at 10:46 AM on August 19, 2009

The Foundations of Mechanical Accuracy likely covers what you're interested in: The subject matter is based in large part on the "four arts" underlying the attainment of mechanical accuracy: geometry, standards of length, dividing the circle, and roundness. A fifth section covers the techniques and applications of the Universal Measuring Machine.
posted by zsazsa at 10:48 AM on August 19, 2009 [1 favorite]

Handwork can actually be pretty precise, if you work carefully, and check yourself systematically. Many amateur astronomers grind their reflecting telescope mirror blanks to 1/4 wavelength of visible light accuracy across an 8" or 10" diameter blank (and sometimes even larger), essentially by hand, using a method first perfected in the 15th century.
posted by paulsc at 10:59 AM on August 19, 2009 [1 favorite]

I had also thought of the lens grinding example.
A slight swerve, but you should read "The Endurance". about the Shackleton expedition. The ships carpenter makes all kinds of things with 3 or 4 hand tools. He's one of the great characters in this amazing drama.
Generally, most of us have no idea what humans are capable of making, because most of us use tools that require minimal or no skill to use. It's a truism that master craftsmen should be able to make things more accurate than their tools. It seems mysterious because you haven't developed (I assume) true mastery of a physical craft. Do you have 10,000 hrs to spare?
posted by Carmody'sPrize at 11:19 AM on August 19, 2009

You make the first leadscrew by hand. You now have a semi-accurate lathe. Using that lathe you make a new leadscrew that is more accurate, which you use in a new lathe that is more accurate that your first one. And so on.
posted by Rhomboid at 12:02 PM on August 19, 2009

Thanks zsazsa for that book link, I'll see if I can't scare one up to look at.

Carmody'sPrize, I am familiar enough with engineering and dabble enough in woodworking to understand this idea, though I can't attain it myself (yet).

After thinking a bit more, I think my question isn't so much about whether or not the precision can be obtained with more or less complex tools, but more how do you know?

I think I'm trying to get at more a calibration issue. How can you measure accuracy with a device that isn't already itself more accurate than what you are trying to measure? And in the grand evolution of our tool-using culture, how does that first more precise measuring tool get calibrated?

Some things I understand can be calibrated by using the world, like thermometers (ice water = 0C and boiling water = 100C), but what about length? How do you calibrate a micrometer without already having a calibrated micrometer (or similar)?
posted by toomanyplugs at 12:34 PM on August 19, 2009

My blacksmithing teacher used to say "the only tool you can't make by hand is the anvil", though, of course you could probably cast one and shape it by hand.
posted by judith at 12:35 PM on August 19, 2009

In that case, you need to read How Round is your Circle?.
posted by Rhomboid at 12:59 PM on August 19, 2009

In the specific case of measuring distance, once you have a given reference (like a piece of material that is exactly an inch) then there are various ways of precisely dividing that length into equal subparts using geometry. So now you can measure e.g. tenths of an inch accurately. You can then make a vernier scale using two such divided measures (one with 10 divisions and one with 11) and that lets you measure down to even greater precision (1/100th). So really the only question is where do you get that initial reference length, and that is just something that initially has to be arbitrary. If you were on a desert island you could just make up your own inch.
posted by Rhomboid at 1:07 PM on August 19, 2009

Seconding zsazsa. Magnificent book. Available directly from the Moore Tool Co.

It hadn't occurred to me until the book explained it: if you're on the desert island, you start with three rocks and you get a flatness reference. The genius of the book is it follows just as though you were on a desert island.
posted by jet_silver at 1:49 PM on August 19, 2009

In high school geometry, we were taught all manner of paper-folding tricks with old computer (paper) tape. If you make a fold in a piece of tape (use adding machine tape), then crease the fold against the edge of the tape, then crease the *new* fold against the other side, back and forth, you get 60 degrees rapidly. It can be proven with geometry that the error - the difference between the angle of the crease and 60 degrees - that error is halved with each successive crease. If you do two one fold, then another from the same vertex, then switch edges (so you do two folds on each side, back and forth), it will give you 72 degrees. Etc. So intelligent method design can yield great results from garbage starting points.
posted by notsnot at 1:50 PM on August 19, 2009

You would enjoy the documentary series Connections. It's 30 years old, but it's all about how one innovation unexpectedly enables others, and follows the crazy paths of these strange connections between technologies across time.

Episode One on youTube

The other episodes, and more
posted by -harlequin- at 2:49 PM on August 19, 2009 [2 favorites]

It's also important to remember how critical your materials are. If the only thing growing on this desert island is coconut trees and the ground is all sand, then it's going to be hard to make much.
posted by TheManChild2000 at 3:02 PM on August 19, 2009

Johanssen blocks. Effing amazing. My dad was an engineer and he had to make them for his apprenticeship. They are blocks of steel that are so smooth and so flat that they will adhere to each other through molecular attraction alone. And you can make them at home!

Basically, you take three fairly-smooth pieces of steel (call them A, B and C) and coat them with polishing compound. Rub A against B and note where the compound is rubbed off. That's a high point, so it gets polished down. Rub B against C and do the same. Rub C against A and do the same. Once you don't have any obvious high points left you just keep going through the A-B, B-C, C-A permutations and each block will polish its counterparts.

You use three blocks so that a high point in A doesn't make a matching low point in B: by using three blocks the high point in A makes a low point in B which makes a high point in C which will then rub against its counterpart in A, grinding both flat. When the three blocks are done you can slide them against each other and it will take a fair bit of strength to get them apart. This is because of the inter-atomic van der Waals force, and it is ever so cool.
posted by Joe in Australia at 8:49 PM on August 19, 2009

There's a good article that outlines this issue from the wiki link that Joe in Australia provided called "the Joy of High Tech." I'm including the link here for posterity.

(still, how does one measure the length of one gauge block without another, more precise gauge block?)
posted by toomanyplugs at 8:53 AM on August 20, 2009

(still, how does one measure the length of one gauge block without another, more precise gauge block?)

You can't, of course - but you don't need to. If you're genuinely starting from scratch then you have to create your own reference measurement - your best block becomes your reference and you use it to make others. It might not be a "real" inch or centimeter, but it's your inch or centimeter and it precisely measures everything you make. If you're just interested in making a home workshop then you do what every other workshop has to do on occasion, which is send your measuring tools off to be calibrated.
posted by Joe in Australia at 6:10 PM on August 20, 2009

« Older When almost perfect is too complicated   |   Help calculate the calories in my favorite NYC... Newer »
This thread is closed to new comments.