Scaling From Unknowns

An unknown to an unknown? Did you state that correctly, or am I missing something?

Just wondering...

Nope your not missing something.

I believe you took a measurement from point A to point B on the drawing and then the same on the model but then the formula came into play.

It may have been in one of Sheperd Paine’s “How to” books. wished I hadn’t loaned his books out a few years ago.

If X is the distance between two points on the model, and Y is the distance between the same two points on the picture, and Y1 is the distance between two points that you want to locate on the model, then X1 = Y1 x X/Y. The trick is to make sure that perspective (things farther from the camera lenses are smaller) doesn't distort the results.

Geez…I shouldn’t think of things like this when I’m at work…er…working.

Robert, the only things unknown are the scale of the drawing/picture and the scale of the model.

If I take the distance between point A & B on the model and divide it by the distance between A & B (same points) on the drawing doesn’t that give me a “multiplier” to scale the rest of the drawing to the rest of the model?

Told you I was having a major brain cramp!

MG,

I guess my confusion (quite common by the way) is in figuring out what IT is that we're trying to do. Tailspinturtle has (it appears) a formula, but the question is, what good will it do? Without at least one known factor - a measurement, a distance, something - we can't do anything constructive with it except punch in hypothetical numbers and see what comes out.

If your goal is to check the accuracy of various points on the unknown model by comparing various points on the photo, the point was already made that perspective can severely hamper this process. Even photos taken at exact right angles to the subject's various profiles will have a certain amount of distortion due to the lens in the camera (fisheye effect). This can be fairly minimal, but still something to be considered. If you're using a drawing, this is obviously not a factor.

You're right, MG. This is one of those subjects that could drive someone batty. It's still interesting and I'd love to see where you're going with this.

Enjoy your modeling...

Dan,

Your method is correct (and the same as my formula in effect) as long as you take into account perspective and distortion. I used it to locate and size the roll thruster ports on the lower wing of an X-35B model from a photograph, for example.

T

Robert, I’m not really going anywhere special with this. I am just now starting to advance my model building skills beyond “out of the box” type builds.

For the project I’m working on right now (see Sci-Fi forum) I have a kit of unknown scale and a good set of drawings of unknown scale. I build a little bit of everything and plan to use this to improve all my builds.

MG,

So you're looking to add details to the model from items on the drawing? Or are you just looking for a way to check the accuracy of the model based on the drawings?

Let me know if either of these questions is correct.

Fun stuff...

In the case of my Sci-Fi project it is to add omitted details and to correct flaws.

MG,

Now it becomes clear and the answer is relatively simple.

First, you have to determine which parts of the model are accurate to begin with. This is needed in order to create your baseline.

Select two, three, even five parts of the model that you consider to be accurate "as is". Measure each part and record these values X1, X2, etc. Now measure each corresponing part on your drawing and record these values as Y1, Y2, etc.

Now divide each X value by it's corresponding Y value. For example, if part X1 on the model is 1/2" and the same part on the drawing is 3/4" (Y1), then you would divide 0.5 by 0.75 and get 0.667. This number will always be less than zero if the model is smaller than the drawing and greater than zero if the model is bigger than the drawing.

Repeat thes process for several parts that you suspect are completely accurate on the model. If all of your results are within a few thousandths of an inch, than you know your on the right track. If one of the results is completely different than the others, throw it out.

Let's say your results for five parts are:

0.478
0.481
0.462
0.471
0.470

They all look pretty close, so that's good. Now you just compute the mean - in this case 0.472 - and this is your multiplier.

All measurements from the drawing are multiplied by this number to get the size that the model part should be. Pretty simple overall.

Note that this will actually be somewhat more accurate if you work with millimeters instead of inches.

Hope this wasn't too confusing.

Enjoy your modeling...