So...how big is the part modelling space actually in SW04? I can't draw a rectangle sketch that is 900 m x 900 m. Is this something to do with my own settings like units etc. or is it really that small?
- posted
20 years ago
So...how big is the part modelling space actually in SW04? I can't draw a rectangle sketch that is 900 m x 900 m. Is this something to do with my own settings like units etc. or is it really that small?
SW has a size limitation - I think for a part it is 1000m either side of origin (total 2000m) and an assmbly is 500m either side of origin (total
1000m) or maybe it's the other way around?Merry :-)
I found this in the knowledge base:
Total space is a 1000m cube centered on the origin, part or assembly. Any linear or diameter dimension must be less than or equal to 1000m (500m radius).
It seems to me that 900m should fit, unless maybe it's way off center?
I can draw a circle with 900 m diameter, but not rectangle that is 900 m x
900 m, strange.
Sounds correct, can do D900m circle, but not 900x900m rectangle
It must be a 1000m Sphere not cube that would make sense because if it could be a 1000m cube you could potentially have a line as long as 1732 and change meter line from corner to corner
Corey
Hmmm.
I was able to make a square sketch centered on the origin 1km per side. However, it didn't like it. The sketch kept going over-defined as I gradually increased the driving dimension. Usually, I had to drag the sketch a couple times to get it to realize it really could solve.
A driven dimension measuring the diagonal correctly reports 1.41km.
I extruded the 1000m square to make a 1000m cube, diagonal measurement:
1732.0508m. I was not able to make a mid-plane extrusion of 1000m, I had to use a bi-directional extrusion 500m each way.Now can you make a sketch with a line from corner to corner?
you can extrude your square kilometer by 1 micron and have the correct volume. My point is, this "limitation" makes sense because MCADs have to be precise in small dimensions. The format for floating points used in SW is the C "double" format, which guarantees 15 decimal digits of precision.This means the largest and smallest dimensions that can be added and still give a correct result must not have a ratio larger than 1/10E15. If you want to measure nanometers (1E-9 m) and still have some margin for geometric operations, you can't allow dimensions much larger than 1E3 m.
Philippe Guglielmetti -
Other modelers get around this by allowing users to pick the minimum dimension. If SW allowed me to say that my minimum dimension was one micrometer, then I could make models with sizes up to 1 km. If I want my minimum to be 1 mm, then I could go to 1000 km. The downside is that they would have to figure out how to handle parts that are done to different minimums. The other guys did it, so I suppose SW could as well.
Jerry Steiger Tripod Data Systems
Wow, amazing what you find out. I always though the environment was infinate. DOH....learn something new everyday.
I knew that there is limit, but I thought that the space is much bigger..
May I ask what it is you want to model that is so large?
That wasn't actually me...bridge maybe
Ok....so lets just say, hypothetically speaking, if you had something bigger, that had to go together, what would you do? And is there an environemnt that goes bigger? Would it make sense or even do-able to have an infiate environment?
as I explained earlier in this thread, you can't have a larger world if you want nanometric precision. Right, some software let you choose a tradeof between size and precision, SW doesn't. But you still can decide to scale your whole model by, say, a factor 1/1000, or 1/1'000'000. Then you just add a few zeros in your drawing scale...
You could scale with meters it is really simple 1CM = 1M then you could fit a model that is 100 times bigger per-say. Then if you already have modeled other parts that fit @ normal scale that need to be in an assembly you can make a 1/100 scaled config.
Is the Assembly space constrained by this same 1000 meters?
Oops ... I deleted it before getting this far ...
Many systems allow you to choose database units: inches, mm, feet, light years ... Internally, that's just a flag for units conversion and perhaps dimensioning ... the data is all stored in numbers (mostly floating point) and only that flag tells the units of measure.
You can see this in the header of IGES files & such too ..
Beyond that, most PC based systems use the MS/Intel
32 bit data format these days (internally). So that determines where the exponent and mantissa are (and how many bits are used for each) though DOUBLE might be used sometimes (64 bit floating point but the calculation times on 32 bit hardware go up fast).Anyone recall 8 or 16 bit integers? The maximum integer limits how many entities a part may contain, among other things ...
IIRC The IBM 360 series and above allowed for a wide range of floating point formats .... running CAD/CAM on s system optimized for counting pennies could produce some strange results .... DEC also used a different word size IIRC ... and exactly where in those bits (no matter how many) the decimal point is is, at times, important.
HTH and is not too much OT.
Would it make sense or even do-able to
Just talking out my butt here, but if 64-bit systems become ubiquitous, it would be possible to design a CAD system with 2^32(4.3billion) times more precision, and therefore more modeling space. This would use quad precision instead of dual precision. (4X 8-bit vs 2x 8-bit floating point precision).
Would it be worth it for a mainstream CAD system? I'm guessing probably not. I think it more likely the programmers would pack the 32 bit data into the 64 bit data path to speed up data processing. I'm really out on a limb now, as I don't understand the mechanisms they used to do this with 16 bit data when 32 bit systems became available. Furthermore, I have no idea how much performance something like SW could get from such methods. But I suspect SW would rather have speed benefits than the capacity to model the Solar System with Angstrom precision.
To create an infinite environment with the dimensional precision we have now, you would need arbitrary bit-precision. That would require special algorithms used to handle numbers larger than quad precision. As I understand it, such systems are referred to as bignum (big number) systems.
See:
you can make a circle 1000m dia. and a square 707.10678119 a side
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.