Topic: Dwg Accuracy

[PHX cadie]

There are 2 things that are bothering me, so I've decided to throw them out to experts. There may not be an answer to these questions since I can not completely explain exactly what I did but......(its also American Idol nite so why not post)

Do you keep Osnaps running? Good/bad?

There are times I will draw a boundary, usually plines, (no I'm not in the habit of c for closing, that would be one fix), but I always have Osnaps running (not nearest, apparent intersection or extension). Yet if I try to hatch, "boundary not found", sure enough after zooming in forever on each intersection, there will be one that is .0001 open. How does this happen with running Osnap?

Another nuisance is offset.
Currently in a Mechanical position I will offset from the centerline, 4" pipe, offset 2 each side. Yet the boss and myself have come back to a dwg and do a dist, it will be 3.9999999999.  The boss is sensible enough to know no one in the bull pit is offsetting a line 1.99999995 each side to purposely get 2 lines 3.999999 apart. It is possible that once or twice when doing a dist I did not snap Per, but I also know I have. I've become so paranoid I'll set the UCS to the centerline then offset.

Like mentioned there may not be an answer, just thought I'd put it out there for discussion

[M-dub]

We use SNAP a lot.  It definitely doesn't work for everyone, but most instrument / electrical drafting, it's a necessity.  It makes the job go faster, easier, more efficiently and it sets up the same qualities for the next person to modify that drawing, which in my case, is usually me!  Osnaps are quite valueable as well, but if you're not familiar with your F9 key, give it a try...if the drawing type / discipline allows for it to be used.

[Dinosaur]

I have seen similar problems when offsetting linework for streets WHEN I have some osnaps running.  Additionally, with 2005 I found a significant number of errors with the endpoint osnap and I have always found the intersection osnap to be a bit unpredictable.

[Bryco]

Many people don't use the magnet for snapping. If you are snapping to an endpoint of a line if you hover over the line anywhere past halfway from the midpoint to the endpt you will see the endpoint box appear, click now and your line/pline will snap/ move to the end. Anytime you see it jump you know the snap has worked, whereas zooming in to the end point doesn't seem to always work. I think it's tied into not allowing the screen enough time to refresh enough to allow the magnet to adjust. Although most people know this, they don't force themselves to hover for that annoying split second.   

[hudster]

Since AutoCAD 2006 there has been a fuzz variable in the hatch command, so you can specify a gap in the boundary and if the gap is less than the fuzz variable it'll go ahead and hatch it. Way cool.

[Bob Garner]

I'm running ACAD 2004 and get those same problems.  I always keep osnap running and sometimes the hatch says illegal boundary.  Usually it's faster to just drag and copy the lines to be hatched, retrace them, hatch that, and drag the hatch to where it should be.  I know, that's a real abomination but when pressed for time, production rules.

Also, I carefully measure stuff out and still get the 3.99997" instead of the 4" I specified.

I hate this!!

[drizzt]

I don't always keep osnaps running, just use shift right click (e for endpoint etc). unless I am doing a lot of linework that needs to snap. I have seen this problem while running osnaps and offset at the same time to.

[sinc]

Something to be wary of is tracking.

If you have tracking enabled, it is easy to accidentally miss your osnap, and get a tracking vector radiating from the osnap instead of the osnap itself.  I like going into the options and setting the tracking to "Shift to Acquire".  With this option, you have to hit the shift button to enable the tracking vector.  This helps prevent missed osnaps.

[hyposmurf]

One thing I find annoyign is when you have your OSNAPS running and it snaps to a point outside of the viewable drawing area.It just shouldnt do that but even in 2007 this can happen.

[CADaver]

One thing I find annoyign is when you have your OSNAPS running and it snaps to a point outside of the viewable drawing area.It just shouldnt do that but even in 2007 this can happen.

Why shouldn't it do that?? I use that feature all the time, especially for ROTATE Reference.  When I osnap to an ENDPoint I want that endpoint no matter where it is, onscreen or not.  Heck, half my rail curves have centers in the next county, if I zoomed out far enough to get that center onscreen, I wouldn't be able to see the arc segment to select it.

The problem with not closing is a result of multiple running osnaps, and not waiting for point aquisition (as Bryce points out).  Never use more than one or two running osnaps unless you're willing to slow down and wait for every point aquisition.  And unless you're running a real heavy-weight computer with a couple of gigs of ram, never use running osnap at all, over large multiple xrefs.

We've traced most of our fatal exception errors to running osnaps, and nearly every "lock-up" complaint stems from running osnaps over large multiple xrefs.  While running osnaps seem like a good idea at first, we've found them to cause much more trouble than they are worth. 

[sinc]

We've traced most of our fatal exception errors to running osnaps, and nearly every "lock-up" complaint stems from running osnaps over large multiple xrefs.  While running osnaps seem like a good idea at first, we've found them to cause much more trouble than they are worth. 

Are you absolutely sure of that one?  I use running OSNAPs and XREFs all the time, and have not noticed any "lock-up" problems.

I have problems with OSNAPS (both the running kind and the other kind) that plain don't work - especially the INT and APPINT.  It's pretty annoying.  But I've not noticed anything like what you're describing.

I suppose part of that might depend on what you think is a "real heavy-weight computer".  Does a 2.8GHz Pentium 4 with 1GB of RAM qualify as one of those heavyweights?

[Dinosaur]

sinc, I had just about given up on APPINT until I made this discovery . . . You MUST make the first selection where there is only one entity to select.  It absolutely will not work when there are multiple entities overlapping where you make the first selection.  This does not affect the second selection.

[Keith™]

One of the biggest problems I have encountered with offset (actually just about any geometry command) and running osnaps is that if your pickbox is set large enough to overlap any other object where your target point is located, then the target point is sometimes overridden by a running osnap point. For example, if your pickbox covers 1" at the current zoom factor and you have 2 objects 12" apart, many times I have encountered that when you would offset one of them 11 3/4" toward the other object, that the offset location may be affected by a running osnap that snaps to the object 12" away. Also, another issue with running osnaps is when you use implied direction vectors for commands.

Now, I have found many times I have offset something(or rather thought I had offset something) by simply using implied direction vectors and a distance, except that I didn't have the correct vector ... I may have thought I had 90 deg, ,but in reality I had something line 89.987 because I didn't have ortho turned on.

Believe me, I have many times swore I had used offset onlyto go back and read the command line to see I had thought offset ony to see I had entered the copy command. It is something I suspect we all do from time to time ... particularly when the result is the same we tend to overlook it.

[M-dub]

Believe me, I have many times swore I had used offset onlyto go back and read the command line to see I had thought offset ony to see I had entered the copy command. It is something I suspect we all do from time to time ... particularly when the result is the same we tend to overlook it.

I do the same thing occasionally.  I'd like to say that I've picked up the mistake every time, but I can't be sure.  It's more difficult when the result may not be exactly the same, but very close...

[CADaver]

Are you absolutely sure of that one?  I use running OSNAPs and XREFs all the time, and have not noticed any "lock-up" problems.

yep, can recreate the problem at will (R2002).  Of course we're talking files that average 10-20 megs.  And it's not as big a problem if you only have one or two osnaps running.

I suppose part of that might depend on what you think is a "real heavy-weight computer".  Does a 2.8GHz Pentium 4 with 1GB of RAM qualify as one of those heavyweights?

Yeah, for us that would be a pretty heavy hitter.  Currently our base machines are 2.0Ghz P4 w/ 512M ram.  We're beginning to roll out new hardware this quarter, 2.8MHz P"D" 2G ram.  Should roll out R2007 soon, Wooo-Hooo.