Accuracy question

[tk369]

When I use a calculator, 1-.475=.525, every time. Yet, I'm now checking a model (someone elses, but built to exacting standards) in NX, and 1-.475 now equals .52497890521!

This has bugged me for years. Is there something in the NX code that has to be updated? This is computer-based modeling. It seems to me that a computer would/could/should be a bit more precise.

It has crossed my mind more than once that the powers-to-be at Siemens are much more concerned with changing the colors of icons and doling out new releases to capture these color changes (at sky-high rates) than they are at precision. Some of the changes have me wondering if the code writers actually USE the software, or just use their intellect to change it.

IMHO, NX is the best modeling software out there, but as a checker, I can't understand why the discrepancies in simple arithmatic can't be fixed.

ted kralovic

VisVSA, NX-6, Macbook, iPhone 3GS, Garmin 765T, Garmin Forerunner 405, Garmin eTrex Vista Hcx, among others

 

[eex23]

Really cannot understand what are you talking about.

 

[tk369]

Here is a PowerPoint detailing my concern.

My WCS is set on the upper surface, and I am interrogating the flat on the lower surface in the -Z direction. You can see the end point in yellow.

ted kralovic

VisVSA, NX-6, Macbook, iPhone 3GS, Garmin 765T, Garmin Forerunner 405, Garmin eTrex Vista Hcx, among others

 

[cowski]

How do you know for sure that NX isn't reporting the correct value?

 

[looslib]

Your model does not led to any 1-.425 calculation.
There are a nuimber of reasons why the one point is showing a different Z value than you expect.
Have you verified the plane on both of the feet?
I am guessing this is an assembly, so you have mating conditions to acount for.
NX is a very precise modeler, but it is still subject to the old adage of Garbage In, Garbage Out.


"Wildfires are dangerous, hard to control, and economically catastrophic."

Ben Loosli

 

[Walterke]

First of all, why do you complain about "the powers-to-be at Siemens" not updating their code, when you are using an outdated version of NX? (at least your sig says you're using NX 6)

Second, I have never seen this behavior in my short time (1yr) of using NX, the only similar thing I've seen is that sometimes a dimension of, for example, 10mm becomes 9,999999999(I never counted the 9's)mm for no apparent reason.

You say the model was not built by yourself. Are you sure that the person who created it did not do something wrong? What software (and version) did he use?
Did you at any point convert the file to another format (step or anything alike)?

NX 7.5
Teamcenter 8

 

[tk369]

loosib..
"Garbage in Garbage out" is uncalled for. The endpoints of each face of the 2 bottom feet are EXACTLY the same Z distance from the absolute X/Y plane. And the difference between the 2 planes (on a calculator, I checked and used floating decimal just in case) is exactly .525.

Yet, when I put my X/Y CSYS on one face or the other, and get information on the end points of the other face, I get the error.

Prime rib in. Pork chop out. It just doesn't make sense. Any thoughts?

ted kralovic

VisVSA, NX-6, Macbook, iPhone 3GS, Garmin 765T, Garmin Forerunner 405, Garmin eTrex Vista Hcx, among others

 

[capnhook]

Ha! Prime rib in, pork chop out......that's a good one. Mind if I use that one? Thanks.

Proud Member of the Reality-Based Community..

 

[cowski]

Can you post the model, or at least a dumbed down version that shows the same problem?

 

[looslib]

Yawn.......
One man's Roast Beef may be another man's Garbage.

Since you didn't create the model, have you checked ALL mating conditions of the assembly as well as all part sizes?
Is this an original NX created model/assembly, never been translated out or in?

Is your assembly and inspection process fine enough to detect the .00022 difference between what NX shows and your theoretical .525?

Started using UG2/NX in 1987.
Started using Pro/E in 2001.


"Wildfires are dangerous, hard to control, and economically catastrophic."

Ben Loosli

 

[Dave K]

I've had issues like this in the past, and I don't recall an instance where I didn't find one of the faces was either off a tiny bit, or at a fraction of an angle to the other, etc. Sometimes it can be a pain in the arse to find though.

-Dave

NX 5.0.6.3mp7