Parlez-vous engineering?

Parlez-vous engineering?

An article in the Guardian newspaper today drew my attention to a rather stunning online archive of drawings and sketches from the engineers who designed and built Britain’s railway network in its 19th and early 20th century heydey. I heartily recommend you have a browse here.

Whilst that collection is certainly a wonderful resource, I have a rather ambivalent relationship to engineering drawings. Like butterflies pinned into collection books, they are a fixed, dead representation of what once was an idea fluttering merrily through an engineer’s mind. Throughout the initial phases of concept and the first wobbly-lined and crossed-out sketches, to 3D CAD models spinning and rotating on the monitor, through to prototypes, parts take on a life of their own. They grow from chrysalis to caterpillar, evolve before your eyes and then – they are ossified, sectioned, labelled and numbered like any other sample. Their drawings sit there gathering dust on the desk, yellowing in folders or exploded into ones and zeros on a disk somewhere (most likely more than one disk…). Your drawing library is a museum. Drawings don’t do anything.  Right?

Well, drawings are the lingua franca of engineering, a translation. And like any language, they are an inexact representation of what was intended. But, like language, they are irreplaceable. Like language, they can be simultaneously fascinating, bewildering, unremittingly dull and, as you can see in the railway archive, wonderfully beautiful.

A good drawing is laden with meaning. It is the drawing that communicates what a part is supposed to be, how it is to be made, out of what materials, to which tolerances, how it is to be measured and to which performance or approval criteria. Not only that, drawings can describe a part’s relationship to other components, and they describe the status of a part – is it a concept, a prototype or is it a fully approved production item? So maybe they are less museum piece and more like baroque paintings in an art gallery, then, alive with meaning and symbolism.

Like sketches, CAD or FEA models, drawings are relatively easy to create. But they require the patience and clarity of thought of a monk to really get right. Drawings are so full of information and so key to the production of real things that getting things wrong results in great upheaval in correcting those mistakes. The slightest functional change in a drawing sets off a chain reaction of work for lots of people. Revision levels are updated. Bills of materials, ordering systems, pricing systems need to be updated; the lot. And it gets worse if you need to inform customers. Therefore, change is resisted. Therefore, getting drawings right first time is crucially important.

But, preparing drawings is still very much a chore and feels like paperwork. The question is: do engineers still feel as if they are engineering when they are ploughing through the change request and drawing management systems just because someone needed to change the length of a screw by a millimetre? I certainly don’t – in fact, I despise such paperwork with a passion. But, seen dispassionately, with the clarity of that self-same monk who created the perfect drawings in the first place, it’s still engineering, like it or not.

What do you think about drawings? Do they play a large part in your work, or have they somehow become history where you are? Is 2D redundant in this 3D printing world? Drop your thoughts in the comments below and let’s talk about *yawn* drawings…!


Thanks to SomJuan for the blueprint photo


I’m going to respond with these analogies:

Having a silent film win in an age of 7.1 surround sound is ridiculous and antiquated.

It’s like having to wait to rewind the tape when you’re done TiVo’ing your show for nostalgia reasons.

I think the 2D is now very redundant in a 3D world. I will acknowledge that I can visualize 3D objects very well (but can’t see out of 3D glasses, weird, I know) so that helps. And it helps for manufacturing. But for blueprints, it’s a waste of time.

Generating detail and assembly drawings are my opportunity to design the manufacturing process, as well as the device. I don’t particularly care whether this information is captured in 2D drawings, or embedded in a 3D CAD model, but mechanical design is more than creating geometry. Coming up with an interlocking set of geometric shapes is like drawing the appropriate free-body diagram; the work of analyzing and documenting one’s design still needs to be done. An abstract geometric “design”— consisting of exactly round cylinders, perfectly square mounting surfaces, and precisely meshing gears—is a good start. But it must be nurtured into a real-world entity.

For instance, calling out the proper tolerance on a bearing bore prevents costly rework. Most CAD models capture only geometric information, and won’t look any different based on the tolerance stack up. However, this detail is of vital importance when the machine is being assembled. If not recorded in a manner that can be easily accessed by a machinist, both money and time may be lost.

Specifying the outer diameter tolerance on a 1.000 dia roller at +.000 / -.030″ can save money, as smaller stock can be used. Will anyone know to buy 1″ round, instead of 1-1/4″ round? Not if it’s not detailed in the drawings.

And what if there is a mistake in the geometric model? I’ve run into cases where the CAD model shows a 1″ long screw thread, but I intended a 1-1/2″ engagement. These discrepancies would not be found if I didn’t go through the “chore” of checking the parts lists against my assembly drawings. Additionally, taking the time to accurately reference all of the part numbers saves time in purchasing, and reduces the chance of improper components being ordered by mistake.

If I want to save time in the machine shop, I specify a counter-bored hole as “DR & C’BORE 1/4-20 SHCS,” rather than using the full ANSI callout. The hole looks the same on the CAD model regardless of my callout, but I guarantee that every machinist I work with has a 1/4-20 counter-bore tool at the ready. Do I really care about the clearance diameter for the socket head? Not usually.

So I see the process of detailing drawings as being my best shot at ensuring that the device I designed will be properly manufactured, assembled, installed, and maintained. Of course, I’m working on small volume custom machinery, and don’t have to interact with a huge bureaucracy. YMMV.

Nice to see the online archive of designs – however, what will the young engineers of 2100 have to look at? Everything is now locked into proprietary CAD formats, which in all likelihood will not be supported that far into the future. Not to mention the fact that the obsession companies have these days for protecting their IP will mean them destroying drawings rather than let them obsolete and pass into the public domain. Perhaps by 2100 we will know more about the engineers of 1912 than 2012. Although I suppose some drawings will survive where there is a regulatory requirement to lodge them with some authority.

Good point, Ian. What will our legacy be? It all depends on the quality of those bits and bytes after a hundred years of backups, copying, deleting, archiving… I recall some studies showing how jpeg photos degrade at a not dissimilar rate to paper.

But then, paper remains compatible and openable after those hundreds of years. Who knows if we’ll even be able to open pdf files in the future?

@ Gears: I would dearly love to be able to go along with those analogies, but I’d respond by saying that The Artist brings across the magic of film better than Transformers (!) I’m with @Jeff on this one – where does the 3D model capture all of that meta-data that drawings provide?

Tolerancing is the big one for me – our friends in manufacturing need to know what leeway they have when making the individual piece within the system; quality often need to know the tolerance bands to be able to calculate capability (Cpk and the like) so they can intervene if the process is drifting (tools wearing and the like).

I’ve yet to see a CAD system that clearly defines the tolerances; once you start getting into GD&T methods, it gets veery complicated.

So, for me, drawings are a necessary pain to make sure we get what we want, as often as possible!

well :).i think still to draw some details or in place that we need to know to draw .don’t forget this is knowledge “drawing”.and it has links with neuroplaticity.

Comments are closed.