27/64", 155/256", 10' 1/32" ... these monsters are rare, but deadly. What do you do? Chase them down and kill them? Ignore them and hope they'll go away?
When they show up in a dimension - do you override the string? Set up a new grid and start stretching everything back until it reads right? Change the precision in the dim properties?
Me: I like nothing better than spending an afternoon following one of these suckers back to its lair (layer?) and slaying it. But knowing that, back when, standard procedure was to just erase the number and write in a new one, how to justify that? And knowing that Acad is accurate down to the billionth, how do these things arise in the first place? Is it all just inappropriate reliance on the perpendicular Osnap?
You know what I'm talking about, Free. Perpendicular, but perpendicular to what? Usually perpendicular to the last random spot you happened to click on. Deadly.
Yeah, I change the tolerance, too, but ... it just feels wrong, right?
does it sometimes look like it's snapped to the end of a line sometimes, but when you zoom in really close you see that it has actually snapped at a point 1/64" away from the endpoint of the line?
Who would have guessed that "tomorrow's superior technology," capable of getting a tiny satellite to pass between Saturn's rings like a perfect billiard shot, would in fact, nearer to Earth, cause our architectural practitioners to abandon fine tolerances altogether (on the "page" at least) in favor of a merely Medieval level of precision ?
Exactly, SDR, what's a 1/128" when you're traveling between planets? Actually, a lot.
Sometimes I wish I was drawing for NASA, I'll bet no one there ever says 'dude, if you dimension to anything finer than 1/2", the contractor's just going to laugh at you.'
When dimensioning things, it's helpful to know which way the tolerances should go to avoid trouble: is bigger better in this case, or smaller ? "A perfect fit is an interference fit," as an old boss of mine (no perfectionist himself, I note) used to say . . .
Some of my drawings are loaded with these, I've been trying to figure out if it's my fault or not. I don't use perpendicular snap, but I use tracking a lot, others use perpendicular. I fix the busts as much as I can, as they tend to multiply if you don't. No one else cares.
Ya - but just enough to be dangerous. Some projects he draws up he goes into drawings to revise his designs. Its awkward. Dont get me wrong, its only 1/164, or 5/32 we are talking about on a midrise apartment building floor plan - It lays out fine in the field. But its just the principle ( no pun intended) of it. Everyone delegates - I sketch or draft a module - and tell a draftsperson to insert into whatever set. But when you own the company youcan feel free to just dive right in I guess.
1/2 our projects are on Revit now - that pretty much keeps the principle at bay.
Revit brings up some trickier issues. A few of us are licesned, have 10+ experiance and generally know what the hell is going on. I find Revit to be extremely accurate as long as you know what your building. The yunger folks are being screwed a bit because they are asked to draft plans - but revit isnt drafting. So I spend a lot of time editing models. They re still grasping with nominal sizes and relations to grid lines - Revit is a beast to teach that on. So Ive become their Construction technologies teacher they should have had in school in order to get them to not create models 5/64" off.
How about those moments when you're zooming or panning and all the sudden, just for a second, one of your straight layout lines, a line you've based everything else off of, goes jagged.
strawberry, that is my total pet peeve, I can't stand it when two lines that should be perpendicular are actually just slightly off from it, the worst is when it is a grid line, that drives me nuts. I do nearly all my drafting with the orthogonality button on and only turn it off temporarily when I specifically need to, it seems to help keep things from getting slightly skewed.
i hate when that happens, like a horizontal that is 179d.
this stuff really isn't a problem unless you're relying on the software to "tell" you that what you're designing is OK instead of knowing (or having a good idea) what the computer/software is supposed to be giving as an answer; like only knowing math by way of a calculator. garbage in=garbage out
thankfully most of the things we build are typically only built to the 1/8", for now anyway
I'm working on a building in metric right now, and the units have a lot to do with making it easier to eyeball problems before they crop up, and fix them once they do. Or failing that, it's much easier to just round it off in your head.
I was thinking about the problem (metric vs imperial/us customary) and it occurred to me that only the latter seems to have the quality of offering finer and finer gradations as one approaches the irreducible (smallest) objects. This isn't actually so, of course, but by making it easy to continually halve the previous measurement, down to a portion much smaller than a millimeter, the effect is certainly there.
But the simplicity and logic of metric measurement is irrefutable. . .
set things to an even dimension. if i'm doing an addition with an as-built or something in it, i'll take my dim string and stretch it over 1/32 or 1/16 or whatever so it reads even. i used to use the over-ride, but if someone stretches the wall, the dim will read the same and you don't want that. never send out drawings with those kind of fractions. the builders will think you're idiots.
interestingly enough, if we did change to the metric system, the modules that we build with [8", 16", 4', 8' etc.] wouldn't change. the only thing that would change is the way they are called out.
I respect the convention of drawing to scale -- and (in the old days, at least) it had the benefit of allowing the drafter to learn from the drawing. But in any event the user really needs only a diagram with dimensions -- so the numbers that are intended are the important element, no ?
SDR I think the computers are fine, it is opperator error. If you don't pay attention to your snaps as strawberrry suggested you can make mistakes. You can also make mistakes in the drafting by not keeping lines perpendicular and lengths accuate. It's important to correct mistakes early before they get propogated throughout a drawing and become difficult to correct.
The War on Error
27/64", 155/256", 10' 1/32" ... these monsters are rare, but deadly. What do you do? Chase them down and kill them? Ignore them and hope they'll go away?
When they show up in a dimension - do you override the string? Set up a new grid and start stretching everything back until it reads right? Change the precision in the dim properties?
Me: I like nothing better than spending an afternoon following one of these suckers back to its lair (layer?) and slaying it. But knowing that, back when, standard procedure was to just erase the number and write in a new one, how to justify that? And knowing that Acad is accurate down to the billionth, how do these things arise in the first place? Is it all just inappropriate reliance on the perpendicular Osnap?
yes! you understand! i've tried to explain this to others but no one gets it ..
if i don't fix them, just set the tolerance to 1/4" and usually everything is fine.
You know what I'm talking about, Free. Perpendicular, but perpendicular to what? Usually perpendicular to the last random spot you happened to click on. Deadly.
Yeah, I change the tolerance, too, but ... it just feels wrong, right?
i stretch it back to the mistake if possible, but usually just squeeze down tolerances to 1/4" as well
its even better when its the boss that did it
does it sometimes look like it's snapped to the end of a line sometimes, but when you zoom in really close you see that it has actually snapped at a point 1/64" away from the endpoint of the line?
it is insanity.
Who would have guessed that "tomorrow's superior technology," capable of getting a tiny satellite to pass between Saturn's rings like a perfect billiard shot, would in fact, nearer to Earth, cause our architectural practitioners to abandon fine tolerances altogether (on the "page" at least) in favor of a merely Medieval level of precision ?
Exactly, SDR, what's a 1/128" when you're traveling between planets? Actually, a lot.
Sometimes I wish I was drawing for NASA, I'll bet no one there ever says 'dude, if you dimension to anything finer than 1/2", the contractor's just going to laugh at you.'
i keep my tolerances at 1/16" and redraw.
it's not the software's fault per se, it's the operator (though Revit makes precision a bit tricky sometimes)
When dimensioning things, it's helpful to know which way the tolerances should go to avoid trouble: is bigger better in this case, or smaller ? "A perfect fit is an interference fit," as an old boss of mine (no perfectionist himself, I note) used to say . . .
Hey, drawing is easy. . .http://www.youtube.com/watch?v=qkkDwD52xCs
evilp, your boss knows autocad?!?
What I can't stand is when you trim to an arc and autocad creates a miniscule space between the end of your line and the arc.
how much of this - aside from the dimensional issues - has to do with the monitor/graphics card and their limitations?
my tolerance is typically set to 1/8 for everything but details 1 1/2" - 6" = 1'-0", all else nothing should matter.....
weem's issue definitely sounds like the curve display resolution. If you type 'regen' you'll see the line and curve meet.
Some of my drawings are loaded with these, I've been trying to figure out if it's my fault or not. I don't use perpendicular snap, but I use tracking a lot, others use perpendicular. I fix the busts as much as I can, as they tend to multiply if you don't. No one else cares.
"evilp, your boss knows autocad?!?"
Ya - but just enough to be dangerous. Some projects he draws up he goes into drawings to revise his designs. Its awkward. Dont get me wrong, its only 1/164, or 5/32 we are talking about on a midrise apartment building floor plan - It lays out fine in the field. But its just the principle ( no pun intended) of it. Everyone delegates - I sketch or draft a module - and tell a draftsperson to insert into whatever set. But when you own the company youcan feel free to just dive right in I guess.
1/2 our projects are on Revit now - that pretty much keeps the principle at bay.
Revit brings up some trickier issues. A few of us are licesned, have 10+ experiance and generally know what the hell is going on. I find Revit to be extremely accurate as long as you know what your building. The yunger folks are being screwed a bit because they are asked to draft plans - but revit isnt drafting. So I spend a lot of time editing models. They re still grasping with nominal sizes and relations to grid lines - Revit is a beast to teach that on. So Ive become their Construction technologies teacher they should have had in school in order to get them to not create models 5/64" off.
i never had a Construction technologies class...
my program was designed for autodidacts, with the exception of structures and mech/elec
How about those moments when you're zooming or panning and all the sudden, just for a second, one of your straight layout lines, a line you've based everything else off of, goes jagged.
In many of our drawings there is something rotated at 90.06 degrees.
strawberry, that is my total pet peeve, I can't stand it when two lines that should be perpendicular are actually just slightly off from it, the worst is when it is a grid line, that drives me nuts. I do nearly all my drafting with the orthogonality button on and only turn it off temporarily when I specifically need to, it seems to help keep things from getting slightly skewed.
i hate when that happens, like a horizontal that is 179d.
this stuff really isn't a problem unless you're relying on the software to "tell" you that what you're designing is OK instead of knowing (or having a good idea) what the computer/software is supposed to be giving as an answer; like only knowing math by way of a calculator. garbage in=garbage out
thankfully most of the things we build are typically only built to the 1/8", for now anyway
two words: go metric
that is an absurd comment, the units used have nothing to do with the degree of accuracy in measuring.
I'm working on a building in metric right now, and the units have a lot to do with making it easier to eyeball problems before they crop up, and fix them once they do. Or failing that, it's much easier to just round it off in your head.
Quick: is 127/256" closer to 7/8", or 3/4"?
um, looks like 0.496" to me. or, if you prefer 12.44mm.
Yeah it was a trick question. ;D
Well, 765 does make a point.
I was thinking about the problem (metric vs imperial/us customary) and it occurred to me that only the latter seems to have the quality of offering finer and finer gradations as one approaches the irreducible (smallest) objects. This isn't actually so, of course, but by making it easy to continually halve the previous measurement, down to a portion much smaller than a millimeter, the effect is certainly there.
But the simplicity and logic of metric measurement is irrefutable. . .
set things to an even dimension. if i'm doing an addition with an as-built or something in it, i'll take my dim string and stretch it over 1/32 or 1/16 or whatever so it reads even. i used to use the over-ride, but if someone stretches the wall, the dim will read the same and you don't want that. never send out drawings with those kind of fractions. the builders will think you're idiots.
interestingly enough, if we did change to the metric system, the modules that we build with [8", 16", 4', 8' etc.] wouldn't change. the only thing that would change is the way they are called out.
I think it is sloppy snapping, not bad fraction computation.
I respect the convention of drawing to scale -- and (in the old days, at least) it had the benefit of allowing the drafter to learn from the drawing. But in any event the user really needs only a diagram with dimensions -- so the numbers that are intended are the important element, no ?
|| || || || || || || || || || ||
|| || || || || || || || || || ||
|| || || || || || || || || || ||
|| || || || || || || || || || ||
|| || || || || || || || || || ||
|| || || || || || || || || || ||
Those are parallel walls. Your dimensions round off to 1/4". You could be off by as much as .25" x 11 = 2 3/4" = BOZO
Accumulated error. Got it. Not good. Why aren't our computers doing what they were designed to do ?
SDR I think the computers are fine, it is opperator error. If you don't pay attention to your snaps as strawberrry suggested you can make mistakes. You can also make mistakes in the drafting by not keeping lines perpendicular and lengths accuate. It's important to correct mistakes early before they get propogated throughout a drawing and become difficult to correct.
I blame overuse of the "stretch" command as the main perpetrator.
it's quick, but if you don't pay very close attention to what you are doing, drawing errors can add up pretty quickly this way.
Block this user
Are you sure you want to block this user and hide all related comments throughout the site?
Archinect
This is your first comment on Archinect. Your comment will be visible once approved.