Taking measurements off hard copies

[awntie]

my engineers expect that when they use a scale on any hard copy the measurement will be accurate. experience tells me this is not the best practice but have no clear grounds for discouraging the reliance on this method. is it common in your offices?

[Wanderer]

my engineers expect that when they use a scale on any hard copy the measurement will be accurate. experience tells me this is not the best practice but have no clear grounds for discouraging the reliance on this method. is it common in your offices?

I agree with them that the measurements SHOULD be accurate. But, I still have always offered warnings with it, because I am familiar with the quality of drawings we received. I learned early on in my career that the text of the dimensions was often 'off' of the measurement in AutoCAD. I had to ASSume that the manually overridden dimensions were the true intent, but, who the heck knows.

[dkoch]

I work for an architecture/engineering firm. Our contract documents explicitly state that drawings are NOT to be scaled. If a dimension is not given and cannot be calculated from the dimensions that are given, the contractor is to ask us what the intended dimension is. In addition to the overridden dimension values that Melanie mentioned**, even a perfectly drafted layout in a computer cannot be reliably measured on the print. There can be errors in the actual printing process, and, worse, paper is not a stable medium. Even if the print is dead on at the moment it is printed, paper expands and contracts with changes in temperature and humidity. It does not take much to generate an error of several inches on a 1/8" = 1'-0" plan drawing.

** - These are definitely bad practice. Back in the bad old days of hand drafting, when the precision of the actual drafted linework would have been less than computer-drafted linework, it would have been more common to accept just changing dimension values when there are last-second changes and no time to erase ink on mylar and redraft. Edits are much easier when using a computer to do the drafting, so the "no time" excuse becomes a lot less justifiable, but I am sure it still happens. (Just not on my projects.)

[cadtag]

It _should_ be accurate, but the level of precision of a triangle scale versus the level of precision of CAD and a 1200 dpi printer is a whole different ballgame. Throwing a scale on a plot and trying to build from the interpolated distance is.... a dubious practice.

Using a scale to check a dimension on the other hand, is quite reasonable. There should be minimal difference, between the stated dimensional distance and the scale measurement. And if there's a large discrepancy, something needs to be fixed.

[Monkey With No Name]

Depends on their definition of "accurate". Using a scale rule alone has it's limitations on accuracy and has approx. 5% margin of error. Engineers I know that scale from drawings are aware of the printing limitations and perform checks if they are doing so. If they measure on a 1:50 plan something that is 1000 mm they expect it to be anywhere between 950 mm & 1050 mm. If they want something more accurate they know they need to find a spec, manufacturers' literature or something.

Surely "do not scale from this drawing" is a legacy thing from the days of overwriting dimensions as already discussed? Aren't we all now modelling and drawing at 1 to 1? If not, why not? What's the point of having a scale on the drawing if you can't use it?

[dgorsman]

There's still a point, and you've made it - there's that assumption that because this was done on a computer, it will be accurate. But the other posts make the points as to how inaccurate it can be: paper medium, plot-to-fit reduced size paper/plot-to-fit from PDF, printed line width, skill/experience of the person doing the scaling, etc. Those "in the know" won't expect something highly accurate, and will file the RFI (request for information) rather than assume if it's not on the drawing. Those who aren't need that reminder of "If it's not a number, it's not accurate".

[Ed Jobe]

If not, why not? What's the point of having a scale on the drawing if you can't use it?

It's there to provide information on the space being modeled, to give context. Its not there to give permission to scale off the dwg. In today's world, drawings convey information on what is to be built. From the perspective of the dwg originator, if you don't specify something, then you get what you get. That's why David mentioned a note that requires you to ask about missing information. Missing information just happens, mistakes happen. From the perspective of a user of a dwg, if you scale off of a dwg, you assume you know the originator's intent. Assuming so opens the door for trouble. Rather than assume, isn't it just better to double-check and ask them what their intent was?

[dkoch]

...Surely "do not scale from this drawing" is a legacy thing from the days of overwriting dimensions as already discussed? Aren't we all now modelling and drawing at 1 to 1? If not, why not? What's the point of having a scale on the drawing if you can't use it?

Yes, but I have seen plenty of 1:1 models and drawings that were not done to exactly the intended dimension, particularly when architectural units are involved. Not sure how it happens, but I have seen some drawings where it looked like someone eyeballed things rather than take the time to type in precise values. It is not much fun having to follow behind such sloppy work (where even the column center lines cannot be trusted).

[remi678731]

Double amen to that.
I routinely get dwgs from others that corners aren't 90 they are 89.89999 or something similar. Or a 10 feet is actually lists 9.89 in properties window. There really is no reason for such sloppy work. It takes just as much work to draw it wrong *** it does to do it right

[Wanderer]

Double amen to that.
I routinely get dwgs from others that corners aren't 90 they are 89.89999 or something similar. Or a 10 feet is actually lists 9.89 in properties window. There really is no reason for such sloppy work. It takes just as much work to draw it wrong *** it does to do it right

As a client, I've actually seen this done intentionally. They'll run a program on their CAD file that skews everything just slightly. Not badly enough that it would be visibly off on the printed drawings, but so much so that you have to work pretty hard to re-use the CAD file (which, btw was in our specs, so ergo quotes, so we paid for a workable cad file that was intentionally damaged for 'intellectual property' reasons... I know IP is a thing, but, seriously, you're renovating a 150 yo building and I spec'd cad files as a deliverable, so get over yourselves).