Topic Title

Hi all.

I have a question...

We have somebody here in the office that insists that he is drawing to scale ... everything to clean numbers. However, I am finding just a bit too often that when I use a floor plan, for example, in which he has primarily done all of the work, in order for me to do building sections, for example, there are these minuscule errors in the line work of his files.

Just for kicks, we turned on DataCad's accuracy setting to 256 to look at some dimension strings ... first one of his files, then one of mine. His had several goofy dimensions (X'-X 253/256", etc.), mine did not.

Just in case it makes a difference, he uses Direction-Distance, I use Relative-Cartesian.

Any Clues about what the problem could be?

Howdy Steve,

Well it could be several things. First I would guess is the difference in input methods. Does he input using the mouse or by hitting the spacebar? Curves in a drawing can really through things off. Also there have been some snapping anomalies (especially with xrefs) in the last few versions, which have made some of my drawings a little difficult to work with when set at 256 accuracy. 253/256 was a common size I was getting while trying to fix a drawing when the high accuracy was turned on. They do seem better using the last few beta versions for 11.08 (which should be out soon). It's really hard to say exactly what he could be doing different. I would suggest finding out what accuracy level you and your co-workers would like to use and set that in your base drawings. For most 32nds or 64ths is fine, but when working with small details, a higher accuracy is handy. Grids are handy too. I have started using grids just for snapping purposes because the snapping was wrong here and there. Setting the snap grid to 1/16" seems to be helping, but I still hand type all of my measurements instead of using the mouse. Just ideas.

I wouldn't get on your co-worker too much, it's most likely something funky Datacad is doing if you were to zoom in on the the cursor while drawing. There are probably a few things he could do to help, like zooming in a bit more while snapping, to make sure Datacad is actualy snapping to what you want. I opened a few old drawings from ver 9 & 10 with the new accuracy, they are 99.9% free of those 256 fractions. Only drawings worked on since probably 11.05 seem to carry a few of these weird fractions. Snapping is on the list of improvments so I'd hang in there for a bit and hopefully this will be a buggy of the past soon.

For what it's worth,
Josh

Maybe both of you should be drawing in the same relative-cartisian mode. I find that using the didtance-direction too much depends on where you point of origon is. I have the same problem in my office. I insist that he use relative -cartisin and that from time to time check that the tangents have not changed by a keystrock misspunch.

Steve -

I have found the Dir/Dist causes miniscule errors in every direction. It is what I prefer for inputing, but I have adopted the Rel/Polar (and sometimes the Rel/Cart) setting so that I don't have the errors creeping into my drawings. It is frustrating (which you have found) to have these itty-bitty errors show up.

Hmm. We played with this for a bit this morning. It does seem that there may just be some sort of glitch in DataCad ... maybe even a version-specific glitch. We opened a handful of other files that we had both done, then changed the accuracy setting to 256 and looked at the associative dimension numbers ... some files had a few errors, some had no goofy numbers at all.

I wonder if some of the problems using dir/dist don't come from thinking that ortho is locked on when it isn't.

I wonder if some of the problems using dir/dist don't come from thinking that ortho is locked on when it isn't.

Hi, Dick -

Not with this guy - I can't stand to have ortho off. Bugs the ba-jeebers out of me.

Hello,

I have noticed that sometimes if you only pull the cursor out a short distance, the angle comes out as something other than ortho -- even with ortho on...it seems to be when you have the cursor "away" from the ortho, so even though the angle shown on the screen is ortho, you end up with something oddball. I use Dir/Dis all the time, but this can be frustrating at times.

The solution seems to be to drag the cursor farther away and to try to keep it close to the ortho angle, even when you don't think that you have to.

... first one of his files, then one of mine. His had several goofy dimensions (X'-X 253/256", etc.), mine did not.

Have you certified that some of the drawing files weren't brought from DC 10 or a previous version ?

Thanks,
Miguel

Hi Miguel.

In trying to figure out what was going on, we opened several files ... some that had been converted from V10, and some that had only been worked on in V11. There didn't seem to be any consistency on what files the goofy numbers would show up in. Even some of the converted ones from previous versions came up perfectly clean, while some that had only been worked on in V11 were goofy.

Steve,

We deal here with the 'metric' units. Recently Dave has addressed a problem when we used 'millimeter' scale type, with floating point, where we had rounding errors on the last digits.

Now it is all right on RC 05 from 11.08.00.

So, my guess is that in the next update you will get it working all fine.

Thanks,
Miguel