Is there a standard for what precision dimensions are set to for drawings?
I am not interested in what your firm's standards are or what you "usually" do. I want to know if there is a published article or a section about this in a textbook.
Thank you.
tduds
Dec 10, 20 1:19 pm
Doubt it.
As the saying goes: the best thing about standards is there are so many to choose from!
tduds
Dec 10, 20 2:13 pm
"I am not interested in what your firm's standards are or what you "usually" do."
Sorry, this is what this thread is going to be about. Already starting.
SneakyPete
Dec 10, 20 1:28 pm
It also depends mightily on whether it's ground-up or renovation / addition.
Chad Miller
Dec 10, 20 1:30 pm
A standard - no.
In practice though your drawings and dimensions should be as precise as the program allows. This is typically around 1/256".
Please keep in mind I'm not saying to draw and dimension things that precise (that's a whole other discussion) but you should be to recognize those measurements in your models / drawings.
proto
Dec 10, 20 2:10 pm
I use one level better than I care about.
so, drawings are usually set to 1/32" & I draw as if 1/16" matters [but try to use increments larger than 1/4"]; but only dimension to 1/8" as smallest increment. [which usually only shows up on divided dims, like stair risers]
i suppose I could just set to 1/256", but, at some point, I just don't care to see those numbers because I have a hard time understanding what that means for the materials I'm working with. Even 1/64" means almost nothing to me unless I'm trying to machine something.
Typically, circles and angles [that aren't ortho] are the only times the fractions make any sense. Brick coursing sometimes...
SneakyPete
Dec 10, 20 2:11 pm
of course.
proto
Dec 10, 20 2:12 pm
it makes sense to me in my head :)
tduds
Dec 10, 20 2:16 pm
If it's possible to set precision below 1/256" in Revit, I haven't found the setting yet. What I *have* noticed is that when importing DWGs that were set at a low precision, Revit will introduce little /256 errors here and there and my neurotic brain just can't ignore them.
SneakyPete
Dec 10, 20 2:21 pm
You'd explode at the places I have worked. Great design, world class, but the idea of precision in drawings is not there. You're drawing new construction. Make at least SOMETHING work out to round numbers.
Almosthip
Dec 10, 20 5:15 pm
Can definitely set your precision in revit.
SneakyPete
Dec 11, 20 2:22 am
That's precision for dimension rounding. You can still move things into 256ths territory. Also grab the dim string and it'll give you the real dim in the properties palette.
apscoradiales
Dec 10, 20 2:14 pm
In Imperial system, no need to dimension anything less than a 1/8".
In "metric" system (as used in Europe, for example), no need to dimension anything less than 0.5cm
In SI system (as used in Canada, for example), no need to dimension anything less than a millimetre.
There are exceptions to that; medical equipment in hospitals and or laboratories come to mind.
How accurate your drawing is in terms of units you set it to depends on the system you are using.
Keep in mind that you shouldn't have silly dims on your drawing, so be creative where you dimension to. Building buildings is not like making Swiss watches. There are acceptable tolerances for different materials during construction. Check your standards - if unsure which standards apply, look it up in your Specifications.
SneakyPete
Dec 10, 20 2:22 pm
Also get to know your tolerances. Just because it's AESS doesn't mean you can assume it's as tight as a gnat's asshole.
Chad Miller
Dec 10, 20 2:33 pm
There are quite a few exceptions but we don't design or dimension to less than 1/2" - That's about as accurate as you can get when building in the field.
caramelhighrise
Dec 17, 20 2:05 am
I'm going to start changing our models to 1/4 or 1/2" tolerance and see if anyone notices. Too many people moving my damn walls 7/256" of an inch on print day!
atelier nobody
Dec 10, 20 2:40 pm
Not in the US National CAD Standard nor the US National BIM Standard. Also not in the USACE Standards, IIRC. I have seen it a few times in some specific agencies' standards, but certainly nothing industry-wide.
I have very strong personal opinions on the matter, and have included it in every firm's CAD standards that I have been involved in developing.
Non Sequitur
Dec 10, 20 2:56 pm
I also have exceptionally strong opinions on the matter... to the point where I will advocate crucifixion to those who slightly chose to stray.
apscoradiales
Dec 10, 20 3:14 pm
One of the issues that crops-up in Canada when it comes to accuracy and dimensioning is that we are supposed to draw in the SI system (millimetres), and yet we use materials that are made in Imperial units due to the fact that they either come from US or are meant for sale to US. Concrete block seems to be the exception where it is actually made in hard millimetres.
Naturally, the metric system will never catch-on in US, so we are stuck...
tduds
Dec 10, 20 3:16 pm
Millimeters are too precise for architecture.
Non Sequitur
Dec 10, 20 3:19 pm
what do you mean, 15.9mm is a totally normal dimension to draw to... It's a constant minute item in our office... 1/2" is drawn at 13mm, 5/8" at 16mm, etc. No decimal mm. There is always someone who says "it's more precise so it's better" No, it ain't better... and I better not catch you using a dimstyle with a rounding factor.
tduds
Dec 10, 20 3:23 pm
My carpentry instructor in high school loved to say "We're not building a piano here" and I like to use that in response to the "it's more precise so it's better" folks.
apscoradiales
Dec 10, 20 3:25 pm
Not sure about the USA, but NRC published a manual back in the early 1970's called Manual on Metric Building Drawing Practice. You can google it, and download a copy. It spells out the rights and wrongs when it comes to many things related to construction and methods. That's the guide we used up here when Canada switched to the metric system (it's actually NOT metric, it called the SI system which not even most of Europe uses). https://nrc-publications.canad...
Chad Miller
Dec 10, 20 3:27 pm
about 45 years old . . . .
Non Sequitur
Dec 10, 20 3:29 pm
great resource Aps.
Chad Miller
Dec 10, 20 4:12 pm
NS - do you still use that 45 year old document in Canada?
Non Sequitur
Dec 10, 20 7:12 pm
No chad, we don’t but there is plenty of stuff available through the NRC.
apscoradiales
Dec 10, 20 3:29 pm
Non Sequitur,
you can draw 13mm or 12.7mm - it does not matter. just be careful how you dimension - don't have silly numbers.
Non Sequitur
Dec 10, 20 3:36 pm
will need to disagree here... you can't tell me you've never a floor plan get royally fucked because someone thought it was cute to offset the wall thickness to x/10 of a mm (because accurate material thickness, dawg) and it screwed up the grid lines because they then "snapped" to it. Now gids are off by 0.3mm and everything is crooked. Just me?
apscoradiales
Dec 10, 20 3:57 pm
Not sure if I'm following your train-of-thought. You set the grids to nice round numbers (hopefully) set the middle or one face of the wall onto the grid line, then dimension one string to the grids, the other string from the grid to the centre of the wall, and bob's your uncle. No?
Non Sequitur
Dec 10, 20 4:04 pm
Sorry Aps, should have added that this is in a larger group setting where one person draws to 0.Xmm and others to the accepted round number. On a large enough floor plate, this difference adds up and you don't want to check every line "just in case" when trying to align things... and yes, I've seen grids set to decimal points too. I don't think the blood from that event will ever come out of the carpet.
An interesting thing I quickly noticed in Revit is that the software prevents you from snapping lines/families to fractions of mm.
apscoradiales
Dec 10, 20 4:18 pm
Dunno how Revit works, but in acad we would always xref grids rather than having them live in the drawing - one of the reasons was that nobody can then move them or fuck around with them.
Also dimension all gwb walls to the centre of studs (think about sequence - studs go up first), and masonry to face. Wall Schedules take care of all the other info about walls.
Non Sequitur
Dec 10, 20 4:26 pm
I'm aware of sequencing, but not everyone does. We normally have grids as xref but not always. Never understood why some can't be bothered to follow basic directions.
apscoradiales
Dec 10, 20 4:58 pm
"I'm aware of sequencing, but not everyone does" LOL! They think gwb goes up all by itself? And these are the same people who spent five years in "School of Architecture", huh? Send them to the site, and let them talk to some of the older guys there...lol!!!!! I actually don't blame the poor bastards - office culture dictates some go to site, some stay in the "studio".
SneakyPete
Dec 10, 20 5:10 pm
dim to CL of studs if you will, but not in places where CLR finishes are paramount
Almosthip
Dec 10, 20 5:19 pm
Just how to set a tape measure to the center of the stud? Dim from out of foundation to stud face to stud face.....
SneakyPete
Dec 10, 20 5:40 pm
Or stop worrying about how to save the contractor money and worry about what we're hired for: design.
SneakyPete
Dec 10, 20 5:45 pm
I've asked a framer about what they prefer. They said consistency and accuracy. Where the dimension ended didn't matter provided it was consistent in the set and called out clearly where it wasn't. Missing dimensions were a much bigger hassle than someone dimensioning to face of finish, as they were used to it. They had to actually THINK about what they would prefer, that's how little thought they had given it.
gibbost
Dec 10, 20 6:00 pm
The amount of hours spent in our profession discussing how to dimension to studs/walls is laughable. I like to imagine Imhotep and his buddies having the same talk--screaming about which trade lays out where the stone goes.
SneakyPete
Dec 10, 20 6:05 pm
We argue about how to save other people money, they cash the check and don't even buy us a beer.
apscoradiales
Dec 10, 20 3:32 pm
"...tduds
Millimeters are too precise for architecture..."
Agree, but what are you going to do if that's the standard?
You can draw stuff in decimals of millimetres too, but just be careful how you dimension
tduds
Dec 10, 20 3:59 pm
Round to nearest 5/10? Half-centimeter seems like enough precision for all but the fanciest finish carpentry.
apscoradiales
Dec 10, 20 4:25 pm
Yeah, 0.5cm is good - that's what's used in "metric" countries. BTW, they don't use fractions, they use decimals instead. So, 5/10...wtf is that? ;)...!
randomised
Dec 10, 20 8:32 pm
We would call 0.5cm the 5mm that they are and never cut a mm in dimensioning, not sure if that’s written down anywhere, just common sense.
gibbost
Dec 10, 20 3:33 pm
Well, since this thread has already devolved into what the OP asked it not to . . . I'll add a quick story.
As an intern, I was having this same debate about precision on drawings. The grey hair in the room grabbed a tape measure and stretched the tape out a bit. Said, 'this is what they use in the field. the tiny marks are 1/16th . . . and most guys on site are too old to even see them. 1/8th is all you need to care about--because that's all they will care about.'
Since then, that's been my rule. As far as drafting, whatever the defaults or office templates offer in the software is what I typically end up with.
b3tadine[sutures]
Dec 10, 20 5:40 pm
I've tried, I've pounded my head on tables, and I still don't understand the need to have anything but 1/8" tolerance for most drawings, 1/16" yeah, for millwork, and details at 3"-6" = 1'-0", and even there, it's generally non-critical, as we're drawing for legibility sake. I can't be the only one that freaks the fuck when they see 29/256"
Chad Miller
Dec 10, 20 6:27 pm
The simple solution to this is draw it to the dimension you want it and you won't get those weir fractions of an inch dimensions.
Your drawing tolerance should be 1:1, don't draw a things at 11' - 3 123/256" and have your dimension round it to 11' - 3 1/2"
SneakyPete
Dec 10, 20 6:35 pm
It's not likely to end up with 256" fractions, but EQ'ing across an odd number can get frustrating quick if you look at the number. Chad's got it right, know what you want, and don't let the result just happen
.
apscoradiales
Dec 10, 20 3:40 pm
gibbost, I can't speak for USA, but this is THE standard in Canada.
Actually that's not the standard in Canada. It's a standard in Canada that some architects use.
proto
Dec 10, 20 4:01 pm
i worked on gsa projects in the past & i liked working in metric for those (tho double dipping & showing both imperial & metric was asinine)
curtkram
Dec 10, 20 7:16 pm
if you have a 45 degree angle, you better include the square root of 2 on your dimension string
Wilma Buttfit
Dec 10, 20 8:17 pm
I inherited a set of drawings that's accurate 14" +/-.
There is a standard in a textbook I've seen but can't recall right now. It isn't one standard but several, depending on what it is, foundation vs cabinets, etc. I'll think for a bit.
atelier nobody
Dec 10, 20 9:26 pm
People seem to be confusing accuracy settings in the software with users actually drawing accurately. The whole purpose of using a high accuracy setting in the software is specifically so I can't miss it if one of the drafters has been sloppy - if I query a line in the CAD file and it comes back with a x/256" length, or an angle comes back as 89.99999999°, that's how I know it's drawn wrong and needs to be corrected. If I didn't have the settings at the max, then querying that 127/256" & 89.9° might return 1/2" and 90° and leave me scratching my head about why there's a bust at the other end of that 250' line.
midlander
Dec 11, 20 3:11 am
to be fair, errors like 89.9999 etc are usually floating point number errors due to the way autocad (or any software) processes numbers. basically it starts to happen when geometry gets located very far from the drawing origin 0,0,0 - like when a plan in millimeters gets stuck into a file set up on a site plan located kilometers away from the survey origin. i mostly blame software creators for expecting users to know and workaround these issues
so tldr: computers can't actually process decimal numbers with 100% accuracy. 1.20 - 1.00 = 0.199999999999999996
SneakyPete
Dec 11, 20 1:43 pm
Having multiple types of dimension styles in a software with different precisions only works if EVERYONE is meticulous and follows the rules. Otherwise it's a complete clusterfuck.
apscoradiales
Dec 10, 20 9:30 pm
SneakyPete,
long time ago when I went to school, my teacher and future boss (British architect) taught us to dimension to centre of studs, be they wood or steel, and to the face of masonry. I don't think he invented the idea, since I have seen some very old drawings where they actually did that. He was also of the opinion that windows (in houses) should be dimensioned to the centre, and then the window schedule would address the rough (stud) as well as masonry openings. I have done window dims that way sometimes, but sometimes I would dim rough openings...depends on whether I had coffee early that day or not...;)...!
I kind prefer it over dimensioning to the face of gwb. Sometimes gwb partitions can get funky in number of layers of drywall, and their thicknesses (particularly in fire rated or acoustical walls), so as to save the drywaller some calculations, grief and possible errors, I have found out it was best for all to dimension to the centre of studs.
But, yes, consistency is one of the keys.
atelier nobody
Dec 10, 20 10:03 pm
Exterior Walls: Face of Masonry, Face of Stud, or Face of Sheathing (Outside faces, of course).
Interior Partitions: Face or Centerline of Studs, Face or Centerline of Masonry - Use centerlines for projects that will be leased to multiple tenants (BOMA Guidelines), otherwise use either faces or centerlines, just do it consistently. If dimensioning to faces always dimension to the same face - ie always left or always right for walls that are vertical in the drawing view; always top or always bottom for horizontal walls (except the exterior wall at the end of a string, which should always be to outside face).
apscoradiales
Dec 10, 20 9:49 pm
atelier nobody,
"...if I query a line in the CAD file and it comes back with a x/256" length, or an angle comes back as 89.99999999°, that's how I know it's drawn wrong and needs to be corrected...."
Yes, very true. I'll add another thing that used to drive be up the wall in acad - people turn off the ortho! You'd think two lines would be perfectly parallel to each other , but NO! Dim at one end would be one value, but at the other end would be something else. fuuuuuck!
Non Sequitur
Dec 10, 20 9:52 pm
add a fraction of a mm on the z axis and things spontaneously combust.
atelier nobody
Dec 10, 20 10:06 pm
NS - No lie. In AutoCAD, probably the single thing that will make me fly into a rage the quickest is the dreaded "Lines are noncoplanar" error.
apscoradiales
Dec 10, 20 10:08 pm
Yes, the infamous Z axis...then you wonder why you cannot snap onto the line or dimension it. 'cause it's in the bloody air, that's why!
SneakyPete
Dec 11, 20 2:24 am
There's always that asshole who scales everything by 99.99% before delivering just to try and ensure they can't (insert thing they do with cad files here)...
randomised
Dec 11, 20 2:48 am
Thanks for that tip about scaling!
Non Sequitur
Dec 11, 20 9:22 am
I had someone in the office do just that because "it did not look big enough". It was a wood truss section. changed it arbitrarily (eyeball scaling) to aprox 4x6 members... lost a whole day chasing similar errors.
Wood Guy
Dec 12, 20 9:26 am
Non, that's...wow. Yesterday I checked in on a builder framing a small addition I designed and he showed me that the window center scaled about about 2' from the face of the adjacent wall. I pointed out my dimension like that shows an exact dimension to center of window. He said he wasn't sure where that dimension like was pulled from. I pointed out the note in 1/4" font that says "dimensions to face of framing." He's a good guy and relatively new to the business, but I had to leave shaking my head. Sorry, a little off topic, but figured you'd appreciate it.
BulgarBlogger
Dec 10, 20 10:02 pm
Ok- no publication or instruction in a textbook about this. Got it. Students pay $100k+ for college academics to not give proper direction. Got it.
atelier nobody
Dec 10, 20 10:11 pm
Most architecture programs (at least here in the US), as well as the NAAB requirements, were developed back when it could be assumed that entering freshmen had learned drafting in high school; they have never taught drafting and still don't (hand, CAD, or otherwise), despite the fact that drafting classes in HS have gotten rarer and rarer.
SneakyPete
Dec 11, 20 2:25 am
Textbooks about opinions wouldn't be very useful.
midlander
Dec 11, 20 3:24 am
this doesn't actually seem like a very important issue. what problems has your work encountered due to the lack of theoretical understanding of drawing precision. i'd say i've wasted more time and money not knowing how to use my printer than dealing with dimensional precision in drawings.
b3tadine[sutures]
Dec 11, 20 9:35 am
I'll buy a textbook on opinions, as long as mine is right.
BulgarBlogger
Dec 11, 20 6:40 am
Here’s my take: In the absence of widely-accepted conventions, why not dimension to the nearest 1/8? Masonry joints are 3/8, gypsum is 5/8, actual CMU dimensions are like 7-5/8”. You will always end up with some fraction of 1/8 most of the time. So if the contractor wants to take some liberty and be less precise or more precise then so be it, but the 1/8” appears to be the best accuracy that makes sense. 1/2” (or even 1”) seem to be excessively liberal precision’s. Trying to anticipate how accurate (or inaccurate) a contractor will is just asking for problems, not just from a constructability point of view, but also from a code/accessibility point of view.
midlander
Dec 11, 20 9:05 am
sounds good let's do it!
Non Sequitur
Dec 11, 20 9:20 am
3.175mm is an odd dimension. Just make the important dimensions end in either 0mm or 5mm and you're good to go.
b3tadine[sutures]
Dec 11, 20 9:21 am
Generally, I agree with this, however now that we're in Revit, and things are what they are, nominal is not really a thing anymore. I'm not against to using 1/16th, but would still, like you push for 1/8th.
Chad Miller
Dec 14, 20 10:34 am
BB - Not a bad idea in theory however we have to keep in mind that accuracy in the field isn't that precise. For example, marking the layout of studs in say a 6,000 sf space on a concrete slab to an 1/8" accuracy would be an issue. The chalk lines and pencils / pens used to make the marks are around 1/6" wide. Even when using laser tools for marking you'll run into this issue.
Just something to consider.
Wood Guy
Dec 17, 20 9:40 am
1/8" seems reasonable for framing. Finish carpentry, at least for the higher-end residential where I work, needs to be to 1/16 or better for paint-grade, or 1/32 for stain- or clear-finish. Cabinetmaking should be to 1/128; at least that's what the best cabinetmakers I know work to. I'm usually good with 1/64 for cabinets I make.
apscoradiales
Dec 11, 20 9:34 am
A very significant issue I had with accuracy of drawings has to do with the fact that DD drawings were often drawn in feet-and-inches then scaled by 25.4 into metric. DD package is usually done by young kids under control of an old geezer architect who cannot wrap his or her head around metric system (or even CAD), while the WD have to be done in metric for permit purposes.
So, one ends ups with a thoroughly screwed-up set of drawings in terms of accuracy that you cannot fix short of redrawing everything from scratch. However, one never has enough time to do that.
This applies to Canada, not US.
Non Sequitur
Dec 11, 20 9:37 am
Aps, I had this happen a while back but it was the interiors department that took our metric building drawings and scaled down to imperial... then rescaled back to metric. Cold shivers remain from those dark days. BTW, this was for a 17-storey 400million gov new build office building, so not really a budgeting issue, just production staff incompetence.
apscoradiales
Dec 11, 20 11:56 am
Yeah, because all tables and chairs are made in feet-and-inches, unless they come from Europe. Being the government, I'm surprised they didn't want some fancy Italian furniture...! It's only taxpayers money!
BulgarBlogger
Dec 11, 20 10:01 am
why would anyone scale drawings? A 1-3/4” thick door is 44.45mm.
So when you receive drawings in metric, understand the design intent, but don’t scale...
Inform something doesn’t work, inform the designer; don’t draw elements that are not standard sizes.
Non Sequitur
Dec 11, 20 10:11 am
saying 45mm is perfectly acceptable.
randomised
Dec 11, 20 3:51 pm
When dimensioning something of 44.45 it would show as 44mm in round numbers though...oh boy!
apscoradiales
Dec 11, 20 11:54 am
BulgarBlogger.
Well, the reason why, like I said, is that many times Design Development drawings are drawn in feet-and-inches, and the Working Drawings need to be done in metric. So, you have to scale them by 25.4 . I've never had the time to re-draw everything from scratch...hence stupid looking dims.
Never mind the door frame - how 'bout the door? Is it 914.4mm or 915mm or 900mm. If you are having brand new doors made, you can have them any size, but if you are buying what's already available, chances are they're 3 feet which is 914.4. One shouldn't dimension doors anyways - set them 100mm away from adjacent wall, and that's it. Codes sometimes contribute to the mess by rounding off stuff to the old Imperial system. Should the guard be at 1070mm height or 1100 or 1050 or 1067...?
square.
Dec 11, 20 12:45 pm
can't wait until some combination of automation/ai/computation makes this entire conversation moot. it's probably closer than you think.
tduds
Dec 11, 20 1:38 pm
Can't come soon enough.
randomised
Dec 11, 20 1:02 pm
And then you hand over your drawing with the perfectly drawn measurements to the contractor...
x-jla
Dec 11, 20 5:21 pm
“Do not scale drawings”
x-jla
Dec 11, 20 5:23 pm
Maybe should add to that “Do not scale anything until we approve your tape!”
Wood Guy
Dec 12, 20 9:23 am
LOL. Hopefully it's obvious to others that the one in the middle has been re-rivetted, or else the holes would be slotted as usual. But in case it's not obvious...
Everyday Architect
Dec 12, 20 10:42 am
I thought it was an obvious photoshop.
midlander
Dec 12, 20 12:06 pm
i thought it was just a VE tape measure
apscoradiales
Dec 11, 20 1:31 pm
LOL!
One cheap Chinese-made, one American (metric not really their strength), and one Canadian - too much American influence.
Chad Miller
Dec 14, 20 12:47 pm
Or, they are all orientated differently in the z axis so they don't align.
Is there a standard for what precision dimensions are set to for drawings?
I am not interested in what your firm's standards are or what you "usually" do. I want to know if there is a published article or a section about this in a textbook.
Thank you.
Doubt it.
As the saying goes: the best thing about standards is there are so many to choose from!
"I am not interested in what your firm's standards are or what you "usually" do."
Sorry, this is what this thread is going to be about. Already starting.
It also depends mightily on whether it's ground-up or renovation / addition.
A standard - no.
In practice though your drawings and dimensions should be as precise as the program allows. This is typically around 1/256".
Please keep in mind I'm not saying to draw and dimension things that precise (that's a whole other discussion) but you should be to recognize those measurements in your models / drawings.
I use one level better than I care about.
so, drawings are usually set to 1/32" & I draw as if 1/16" matters [but try to use increments larger than 1/4"]; but only dimension to 1/8" as smallest increment. [which usually only shows up on divided dims, like stair risers]
i suppose I could just set to 1/256", but, at some point, I just don't care to see those numbers because I have a hard time understanding what that means for the materials I'm working with. Even 1/64" means almost nothing to me unless I'm trying to machine something.
Typically, circles and angles [that aren't ortho] are the only times the fractions make any sense. Brick coursing sometimes...
of course.
it makes sense to me in my head :)
If it's possible to set precision below 1/256" in Revit, I haven't found the setting yet. What I *have* noticed is that when importing DWGs that were set at a low precision, Revit will introduce little /256 errors here and there and my neurotic brain just can't ignore them.
You'd explode at the places I have worked. Great design, world class, but the idea of precision in drawings is not there. You're drawing new construction. Make at least SOMETHING work out to round numbers.
Can definitely set your precision in revit.
That's precision for dimension rounding. You can still move things into 256ths territory. Also grab the dim string and it'll give you the real dim in the properties palette.
In Imperial system, no need to dimension anything less than a 1/8".
In "metric" system (as used in Europe, for example), no need to dimension anything less than 0.5cm
In SI system (as used in Canada, for example), no need to dimension anything less than a millimetre.
There are exceptions to that; medical equipment in hospitals and or laboratories come to mind.
How accurate your drawing is in terms of units you set it to depends on the system you are using.
Keep in mind that you shouldn't have silly dims on your drawing, so be creative where you dimension to. Building buildings is not like making Swiss watches. There are acceptable tolerances for different materials during construction. Check your standards - if unsure which standards apply, look it up in your Specifications.
Also get to know your tolerances. Just because it's AESS doesn't mean you can assume it's as tight as a gnat's asshole.
There are quite a few exceptions but we don't design or dimension to less than 1/2" - That's about as accurate as you can get when building in the field.
I'm going to start changing our models to 1/4 or 1/2" tolerance and see if anyone notices. Too many people moving my damn walls 7/256" of an inch on print day!
Not in the US National CAD Standard nor the US National BIM Standard. Also not in the USACE Standards, IIRC. I have seen it a few times in some specific agencies' standards, but certainly nothing industry-wide.
I have very strong personal opinions on the matter, and have included it in every firm's CAD standards that I have been involved in developing.
I also have exceptionally strong opinions on the matter... to the point where I will advocate crucifixion to those who slightly chose to stray.
One of the issues that crops-up in Canada when it comes to accuracy and dimensioning is that we are supposed to draw in the SI system (millimetres), and yet we use materials that are made in Imperial units due to the fact that they either come from US or are meant for sale to US. Concrete block seems to be the exception where it is actually made in hard millimetres.
Naturally, the metric system will never catch-on in US, so we are stuck...
Millimeters are too precise for architecture.
what do you mean, 15.9mm is a totally normal dimension to draw to... It's a constant minute item in our office... 1/2" is drawn at 13mm, 5/8" at 16mm, etc. No decimal mm. There is always someone who says "it's more precise so it's better" No, it ain't better... and I better not catch you using a dimstyle with a rounding factor.
My carpentry instructor in high school loved to say "We're not building a piano here" and I like to use that in response to the "it's more precise so it's better" folks.
Not sure about the USA, but NRC published a manual back in the early 1970's called Manual on Metric Building Drawing Practice. You can google it, and download a copy. It spells out the rights and wrongs when it comes to many things related to construction and methods. That's the guide we used up here when Canada switched to the metric system (it's actually NOT metric, it called the SI system which not even most of Europe uses). https://nrc-publications.canad...
about 45 years old . . . .
great resource Aps.
NS - do you still use that 45 year old document in Canada?
No chad, we don’t but there is plenty of stuff available through the NRC.
Non Sequitur,
you can draw 13mm or 12.7mm - it does not matter. just be careful how you dimension - don't have silly numbers.
will need to disagree here... you can't tell me you've never a floor plan get royally fucked because someone thought it was cute to offset the wall thickness to x/10 of a mm (because accurate material thickness, dawg) and it screwed up the grid lines because they then "snapped" to it. Now gids are off by 0.3mm and everything is crooked. Just me?
Not sure if I'm following your train-of-thought. You set the grids to nice round numbers (hopefully) set the middle or one face of the wall onto the grid line, then dimension one string to the grids, the other string from the grid to the centre of the wall, and bob's your uncle. No?
Sorry Aps, should have added that this is in a larger group setting where one person draws to 0.Xmm and others to the accepted round number. On a large enough floor plate, this difference adds up and you don't want to check every line "just in case" when trying to align things... and yes, I've seen grids set to decimal points too. I don't think the blood from that event will ever come out of the carpet.
An interesting thing I quickly noticed in Revit is that the software prevents you from snapping lines/families to fractions of mm.
Dunno how Revit works, but in acad we would always xref grids rather than having them live in the drawing - one of the reasons was that nobody can then move them or fuck around with them.
Also dimension all gwb walls to the centre of studs (think about sequence - studs go up first), and masonry to face. Wall Schedules take care of all the other info about walls.
I'm aware of sequencing, but not everyone does. We normally have grids as xref but not always. Never understood why some can't be bothered to follow basic directions.
"I'm aware of sequencing, but not everyone does" LOL! They think gwb goes up all by itself? And these are the same people who spent five years in "School of Architecture", huh? Send them to the site, and let them talk to some of the older guys there...lol!!!!! I actually don't blame the poor bastards - office culture dictates some go to site, some stay in the "studio".
dim to CL of studs if you will, but not in places where CLR finishes are paramount
Just how to set a tape measure to the center of the stud? Dim from out of foundation to stud face to stud face.....
Or stop worrying about how to save the contractor money and worry about what we're hired for: design.
I've asked a framer about what they prefer. They said consistency and accuracy. Where the dimension ended didn't matter provided it was consistent in the set and called out clearly where it wasn't. Missing dimensions were a much bigger hassle than someone dimensioning to face of finish, as they were used to it. They had to actually THINK about what they would prefer, that's how little thought they had given it.
The amount of hours spent in our profession discussing how to dimension to studs/walls is laughable. I like to imagine Imhotep and his buddies having the same talk--screaming about which trade lays out where the stone goes.
We argue about how to save other people money, they cash the check and don't even buy us a beer.
"...tduds
Millimeters are too precise for architecture..."
Agree, but what are you going to do if that's the standard?
You can draw stuff in decimals of millimetres too, but just be careful how you dimension
Round to nearest 5/10? Half-centimeter seems like enough precision for all but the fanciest finish carpentry.
Yeah, 0.5cm is good - that's what's used in "metric" countries. BTW, they don't use fractions, they use decimals instead. So, 5/10...wtf is that? ;)...!
We would call 0.5cm the 5mm that they are and never cut a mm in dimensioning, not sure if that’s written down anywhere, just common sense.
Well, since this thread has already devolved into what the OP asked it not to . . . I'll add a quick story.
As an intern, I was having this same debate about precision on drawings. The grey hair in the room grabbed a tape measure and stretched the tape out a bit. Said, 'this is what they use in the field. the tiny marks are 1/16th . . . and most guys on site are too old to even see them. 1/8th is all you need to care about--because that's all they will care about.'
Since then, that's been my rule. As far as drafting, whatever the defaults or office templates offer in the software is what I typically end up with.
I've tried, I've pounded my head on tables, and I still don't understand the need to have anything but 1/8" tolerance for most drawings, 1/16" yeah, for millwork, and details at 3"-6" = 1'-0", and even there, it's generally non-critical, as we're drawing for legibility sake. I can't be the only one that freaks the fuck when they see 29/256"
The simple solution to this is draw it to the dimension you want it and you won't get those weir fractions of an inch dimensions.
Your drawing tolerance should be 1:1, don't draw a things at 11' - 3 123/256" and have your dimension round it to 11' - 3 1/2"
It's not likely to end up with 256" fractions, but EQ'ing across an odd number can get frustrating quick if you look at the number. Chad's got it right, know what you want, and don't let the result just happen .
gibbost, I can't speak for USA, but this is THE standard in Canada.
https://nrc-publications.canad...
so, the thread did not "devolve"...
Actually that's not the standard in Canada. It's a standard in Canada that some architects use.
i worked on gsa projects in the past & i liked working in metric for those (tho double dipping & showing both imperial & metric was asinine)
if you have a 45 degree angle, you better include the square root of 2 on your dimension string
I inherited a set of drawings that's accurate 14" +/-.
There is a standard in a textbook I've seen but can't recall right now. It isn't one standard but several, depending on what it is, foundation vs cabinets, etc. I'll think for a bit.
People seem to be confusing accuracy settings in the software with users actually drawing accurately. The whole purpose of using a high accuracy setting in the software is specifically so I can't miss it if one of the drafters has been sloppy - if I query a line in the CAD file and it comes back with a x/256" length, or an angle comes back as 89.99999999°, that's how I know it's drawn wrong and needs to be corrected. If I didn't have the settings at the max, then querying that 127/256" & 89.9° might return 1/2" and 90° and leave me scratching my head about why there's a bust at the other end of that 250' line.
to be fair, errors like 89.9999 etc are usually floating point number errors due to the way autocad (or any software) processes numbers. basically it starts to happen when geometry gets located very far from the drawing origin 0,0,0 - like when a plan in millimeters gets stuck into a file set up on a site plan located kilometers away from the survey origin. i mostly blame software creators for expecting users to know and workaround these issues
https://www.geeksforgeeks.org/floating-point-error-in-python/
so tldr: computers can't actually process decimal numbers with 100% accuracy. 1.20 - 1.00 = 0.199999999999999996
Having multiple types of dimension styles in a software with different precisions only works if EVERYONE is meticulous and follows the rules. Otherwise it's a complete clusterfuck.
SneakyPete,
long time ago when I went to school, my teacher and future boss (British architect) taught us to dimension to centre of studs, be they wood or steel, and to the face of masonry. I don't think he invented the idea, since I have seen some very old drawings where they actually did that. He was also of the opinion that windows (in houses) should be dimensioned to the centre, and then the window schedule would address the rough (stud) as well as masonry openings. I have done window dims that way sometimes, but sometimes I would dim rough openings...depends on whether I had coffee early that day or not...;)...!
I kind prefer it over dimensioning to the face of gwb. Sometimes gwb partitions can get funky in number of layers of drywall, and their thicknesses (particularly in fire rated or acoustical walls), so as to save the drywaller some calculations, grief and possible errors, I have found out it was best for all to dimension to the centre of studs.
But, yes, consistency is one of the keys.
Exterior Walls: Face of Masonry, Face of Stud, or Face of Sheathing (Outside faces, of course).
Interior Partitions: Face or Centerline of Studs, Face or Centerline of Masonry - Use centerlines for projects that will be leased to multiple tenants (BOMA Guidelines), otherwise use either faces or centerlines, just do it consistently. If dimensioning to faces always dimension to the same face - ie always left or always right for walls that are vertical in the drawing view; always top or always bottom for horizontal walls (except the exterior wall at the end of a string, which should always be to outside face).
atelier nobody,
"...if I query a line in the CAD file and it comes back with a x/256" length, or an angle comes back as 89.99999999°, that's how I know it's drawn wrong and needs to be corrected...."
Yes, very true. I'll add another thing that used to drive be up the wall in acad - people turn off the ortho! You'd think two lines would be perfectly parallel to each other , but NO! Dim at one end would be one value, but at the other end would be something else. fuuuuuck!
add a fraction of a mm on the z axis and things spontaneously combust.
NS - No lie. In AutoCAD, probably the single thing that will make me fly into a rage the quickest is the dreaded "Lines are noncoplanar" error.
Yes, the infamous Z axis...then you wonder why you cannot snap onto the line or dimension it. 'cause it's in the bloody air, that's why!
There's always that asshole who scales everything by 99.99% before delivering just to try and ensure they can't (insert thing they do with cad files here)...
Thanks for that tip about scaling!
I had someone in the office do just that because "it did not look big enough". It was a wood truss section. changed it arbitrarily (eyeball scaling) to aprox 4x6 members... lost a whole day chasing similar errors.
Non, that's...wow. Yesterday I checked in on a builder framing a small addition I designed and he showed me that the window center scaled about about 2' from the face of the adjacent wall. I pointed out my dimension like that shows an exact dimension to center of window. He said he wasn't sure where that dimension like was pulled from. I pointed out the note in 1/4" font that says "dimensions to face of framing." He's a good guy and relatively new to the business, but I had to leave shaking my head. Sorry, a little off topic, but figured you'd appreciate it.
Ok- no publication or instruction in a textbook about this. Got it. Students pay $100k+ for college academics to not give proper direction. Got it.
Most architecture programs (at least here in the US), as well as the NAAB requirements, were developed back when it could be assumed that entering freshmen had learned drafting in high school; they have never taught drafting and still don't (hand, CAD, or otherwise), despite the fact that drafting classes in HS have gotten rarer and rarer.
Textbooks about opinions wouldn't be very useful.
this doesn't actually seem like a very important issue. what problems has your work encountered due to the lack of theoretical understanding of drawing precision. i'd say i've wasted more time and money not knowing how to use my printer than dealing with dimensional precision in drawings.
I'll buy a textbook on opinions, as long as mine is right.
Here’s my take: In the absence of widely-accepted conventions, why not dimension to the nearest 1/8? Masonry joints are 3/8, gypsum is 5/8, actual CMU dimensions are like 7-5/8”. You will always end up with some fraction of 1/8 most of the time. So if the contractor wants to take some liberty and be less precise or more precise then so be it, but the 1/8” appears to be the best accuracy that makes sense. 1/2” (or even 1”) seem to be excessively liberal precision’s. Trying to anticipate how accurate (or inaccurate) a contractor will is just asking for problems, not just from a constructability point of view, but also from a code/accessibility point of view.
sounds good let's do it!
3.175mm is an odd dimension. Just make the important dimensions end in either 0mm or 5mm and you're good to go.
Generally, I agree with this, however now that we're in Revit, and things are what they are, nominal is not really a thing anymore. I'm not against to using 1/16th, but would still, like you push for 1/8th.
BB - Not a bad idea in theory however we have to keep in mind that accuracy in the field isn't that precise. For example, marking the layout of studs in say a 6,000 sf space on a concrete slab to an 1/8" accuracy would be an issue. The chalk lines and pencils / pens used to make the marks are around 1/6" wide. Even when using laser tools for marking you'll run into this issue.
Just something to consider.
1/8" seems reasonable for framing. Finish carpentry, at least for the higher-end residential where I work, needs to be to 1/16 or better for paint-grade, or 1/32 for stain- or clear-finish. Cabinetmaking should be to 1/128; at least that's what the best cabinetmakers I know work to. I'm usually good with 1/64 for cabinets I make.
A very significant issue I had with accuracy of drawings has to do with the fact that DD drawings were often drawn in feet-and-inches then scaled by 25.4 into metric. DD package is usually done by young kids under control of an old geezer architect who cannot wrap his or her head around metric system (or even CAD), while the WD have to be done in metric for permit purposes.
So, one ends ups with a thoroughly screwed-up set of drawings in terms of accuracy that you cannot fix short of redrawing everything from scratch. However, one never has enough time to do that.
This applies to Canada, not US.
Aps, I had this happen a while back but it was the interiors department that took our metric building drawings and scaled down to imperial... then rescaled back to metric. Cold shivers remain from those dark days. BTW, this was for a 17-storey 400million gov new build office building, so not really a budgeting issue, just production staff incompetence.
Yeah, because all tables and chairs are made in feet-and-inches, unless they come from Europe. Being the government, I'm surprised they didn't want some fancy Italian furniture...! It's only taxpayers money!
why would anyone scale drawings? A 1-3/4” thick door is 44.45mm.
So when you receive drawings in metric, understand the design intent, but don’t scale...
Inform something doesn’t work, inform the designer; don’t draw elements that are not standard sizes.
saying 45mm is perfectly acceptable.
When dimensioning something of 44.45 it would show as 44mm in round numbers though...oh boy!
BulgarBlogger.
Well, the reason why, like I said, is that many times Design Development drawings are drawn in feet-and-inches, and the Working Drawings need to be done in metric. So, you have to scale them by 25.4 . I've never had the time to re-draw everything from scratch...hence stupid looking dims.
Never mind the door frame - how 'bout the door? Is it 914.4mm or 915mm or 900mm. If you are having brand new doors made, you can have them any size, but if you are buying what's already available, chances are they're 3 feet which is 914.4. One shouldn't dimension doors anyways - set them 100mm away from adjacent wall, and that's it. Codes sometimes contribute to the mess by rounding off stuff to the old Imperial system. Should the guard be at 1070mm height or 1100 or 1050 or 1067...?
can't wait until some combination of automation/ai/computation makes this entire conversation moot. it's probably closer than you think.
Can't come soon enough.
And then you hand over your drawing with the perfectly drawn measurements to the contractor...
“Do not scale drawings”
Maybe should add to that “Do not scale anything until we approve your tape!”
LOL. Hopefully it's obvious to others that the one in the middle has been re-rivetted, or else the holes would be slotted as usual. But in case it's not obvious...
I thought it was an obvious photoshop.
i thought it was just a VE tape measure
LOL!
One cheap Chinese-made, one American (metric not really their strength), and one Canadian - too much American influence.
Or, they are all orientated differently in the z axis so they don't align.
Also a wide angle lenses and possibly photoshop.