Scaling a drawing isn't done very much, but it is done.

For instance, it's the middle of the night, the fabricator or machinist can't contact you, the part is due in the morning, and you left off a dimension. No way is he not going to deliver _something_.

As a matter of course, every drawing should say DO NOT SCALE somewhere, typically in the small print in the title block, or on a border, just to give you a tiny bit of ammunition in the occasional finger pointing battle.

There's another reason. Even a good diazo contact print can be off by a couple of percent from the original, just because of curvature and thickness on the exposure drum and shrinkage from the developer. Never mind that photocopiers aren't always in perfect calbration, even at 1:1, or what transmission as a .pdf file can do.

In your case, since you won't always be delivering your product on standard size sheets, it might be a good idea, call it a courtesy, to include a scale bar similar to what you find in the legend of a map, say along a border, so a desperate fabricator can cut out the little bar and use it directly on your print, no matter what size the print arrives at.

<tangent>
One time I was sitting next to a very experienced drafter in an AutoCAD14 shop. I was tasked with converting a couple hundred of his 2D drawings into 3D models.

No problem, I thought; just copy/paste a view or two onto the correct plane at the correct place, extrude, and do a few Boolean operations.

It was a friggin' disaster; I had to do every damn model from scratch.

I asked the old timer why his dimensions didn't exactly match the corresponding lines. He said he gave up on associative dimensioning years before, because it didn't work for him. All of his dimensions were 'overridden' and manually edited.

His lines either failed to meet, or crossed each other at intersections. Always by .003" or so.

The guy had run AutoCAD for years with 'near' as a running osnap.

So DO NOT SCALE applied to his work even when using AutoCAD's internal measuring tools. He even drew rectangles whose nonadjacent edges were not parallel....
</tangent>

Mike Halloran
Pembroke Pines, FL, USA

An addition to Mike's point about diazo prints... not only are they off, they are off by different proportions (L vs H). So a vertical and a horizontal scale bar would be needed.

"Know the rules well, so you can break them effectively."
-Dalai Lama XIV

ok thanks for the input guys. It always seemed a little old school as I said. But nowadays we just send STEP files so I dont even think that we need to include the scale bar. I just wanted to get a second opinion. Im thinking that even if I did include a scale bar the machinist would be hesitant to use it because of printing errors as you already mentioned...

That's why you put an 'L' or cross hairs or something so they can scale in both dimensions separately as I long ago learnt to do when necessary.

How much time or effort do you save by not putting a scale?

yea fair enough, but when you have a sheet with several detail views all with different scales it gets annoying if you start creating scales in both dimensions for each different scale... I understand where you are coming from.

Not sure we're talking the same thing but...

Drawing has some default scale specified in the title block somewhere.

Individual detail views etc. have scale specified if they differ from the default sheet scale - every CAD package I've worked on will do this at least semi automatically (i.e. you click a button to say 'display scale' or similar and this is often set to default on for detail views).

You then just have one global 'graticule' - the cross hairs or 'L' scale I mentioned that is AT SHEET SCALE - i.e. if the the drawing were printed at true 1:1 the graticule would be 1:1. User has to apply the 2:1 or whatever to this graticule as required for individual views.

Scaling, as in placing your machinist's scale on the drawing to derive an approximate measurement, is NEVER a good practice. Any machinist that has been through decent training will not do it. With CAD nowadays, it's easy enough to print off a correctly scaled DXF if one needs to make a template for some reason.

It is better to have enough ideas for some of them to be wrong, than to be always right by having no ideas at all.

Quote (MikeHalloran)

...

I asked the old timer why his dimensions didn't exactly match the corresponding lines. He said he gave up on associative dimensioning years before, because it didn't work for him. All of his dimensions were 'overridden' and manually edited.

...

I have encountered this trick as well, and I was not happy. This should be an immediate firing offense. Our shops have requested DXF files from us for years. Now, they are asking for STEP files. These are to scale, with the exception of your bilateral and limit tolerances.

--
JHG

I've never used a drawing to scale when machining. But when I worked in sheet metal forming we did it all the time to check periphery and contour with mylars.

Back in the late 1800's drawing didn't have any dimension and fabricators HAD to scale off blue prints to be able to make the "part", but the problem is obvious, they didn't get the accuracy that was required and the part became a one-off assembly ...working but not inter-changeable. Then in the early 1900's standards were devolved and dimension now required, so fabrication of parts was uniforms from fabricator to fabricator. But the machinists still had the tendency to "scale from the drawing", plus as Mr. Halloran has indicated ...IF it's after mid-night and the parts are due that day, and that last dimension is missing or "messed up" someway, what other means is available? When I first started in this profession some 40 years ago, doing machine drawings, I can't tell you how many times I had to add the note ..."Do NOT scale from drawing"! Wish I had a nickel for everyone ...not sure I'd be rich, but I'd be well off!

Eredhel,
You used the term "Mylar" ,very often a Mylar was a dimensionless plot, full size, of a component. There was no scaling.
There is a standard and a set of tolerances for using " Dimensionless prints/plots ", but it is not considered scaling.
B.E.

You are judged not by what you know, but by what you can do.

Scaling directly from mylar does not pose the issues that scaling from vellum or paper does as it is a much more stable material.

"Know the rules well, so you can break them effectively."
-Dalai Lama XIV

If the drawing invokes Y14.5 then scaling is prohibited per Fundamental rule 1.4(b) anyway.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech



SSG, U.S. Army
Taji, Iraq OIF II

@powerhound,

Correct me if I'm wrong but the topic at hand is one of the very exceptions to that Fundamental rule cited in 1.4(b). "Undimensioned drawings"

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5

I don't see where there is any mention that the drawings are undimensioned.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech



SSG, U.S. Army
Taji, Iraq OIF II

That may've been my assumption. I don't know why you would ever scale a feature from the print if there was already a dimension that sizes it anywhere. The only case that necessitates 'scaling' manually is if there is no way to ascertain the size from the given information (from any view) at which point I figure it's an 'undimensioned' feature.

Unless that ASME definition is saying that the -entire- drawing must be undimensioned for it to allow scaling. Which, in retrospect may be the intent. That it's only allowed for controlled masters and the like.

_________________________________________
NX8.0, Solidworks 2014, AutoCAD, Enovia V5

A practice that I used to see a lot in my early days of my career was scaling of missing dimensions or acquiring dimensions that were more easily accessible from a vector other than shown on the print. I haven't seen this done in at least 20 years. My opinion is that this was included in the standard to specifically disallow that. I've never seen a purposely undimensioned print.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech



SSG, U.S. Army
Taji, Iraq OIF II

I've had to scale from incomplete interface drawings lot's of times when unable to get more detail.

It happens, it's bad practice, but then it's bad practice not to give out meaningful interface detail to customers trying to use your kit but it happens all too often.