rounding off to 10mm

[Walterke]

Is there a way to round off numbers to the nearest 10mm?
As in 2776mm becomes 2780mm.

Second: while searching this forum for an answer to this, I found this (from 2007):

When rounding, examine the figure following (i.e., to the right of) the figure that is to be last. This figure you are examining is the first figure to be dropped.

If it is less than 5, drop it and all the figures to the right of it.
If it is more than 5, increase by 1 the number to be rounded, that is, the preceeding figure.
If it is 5, round the number so that it will be EVEN.

The thread in question, however, does not make a reference to any standard on this. I'm interested to know if this is the same way for ISO.

Thanks.

NX 7.5
Teamcenter 8

 

[Walterke]

Damnit, this was supposed to go in the siemens NX forum. Well maybe the second question can be answered here.

Are there admins here who can move threads?

NX 7.5
Teamcenter 8

 

[MintJulep]

Just red flag your post here with an "oops" note to the admin.

Repost on the correct forum.

 

[CheckerHater]

There were no standards on rounding since 1946.
In general math, if it's 0,1,2.3.4 round down; if it's 5,6,7,8,9 round up.

 

[Walterke]

@MJ good tip, but I'll just create a new thread there since question 2 fits better here anyway.

@CH that's how I was always thought to do it, but apparently NX does (or did in 2007, I should check it out) this differently, claiming it was according to ASME standards. Apparently it's a statistical thing. Full explanation can be found here:

http://www.eng-tips.com/viewthread.cfm?qid=197969

NX 7.5
Teamcenter 8

 

[CheckerHater]

Thank you Walterke,

The “explanation” doesn’t really explain anything.
How rounding 1.504 to 1.5 is different from rounding 1.514 to 1.5?
The person who came with this explanation probably didn’t do well in school, if he is afraid of 0 being “not a real number”.
Rounding off 0 is as legitimate as rounding off 1.

BTW, have you heard of “fourth” rule:
“In rounding off numbers, if the first number dropped is 5, and all the numbers after the 5 is not zero, then the last number kept shall be increased by one. For example: 1.6501 = 1.7”

It looks like everyone makes them up as they please.
Also, the question still open: which standard? ASA Z25.1 – 1940?

 

[Walterke]

It's not about the difference between 1.504 and 1.514.
It's the difference between 1.55 and 1.65. Both would be rounded off to 1.6 according to JRB's explanation. (so one rounded up and one rounded down, even if both have the same number to round off)

NX 7.5
Teamcenter 8

 

[CheckerHater]

W,
I was talking about the “explanation” in the link you provided.
It was saying that rounding 1,2,3,4 down and 5,6,7,8,9 up creates un-even statistical results as if 0 didn’t exist or it was not “the real number”.
And, no, I am not buying into rounding to even numbers.
Also I don’t understand concept of “rounding” applied to design. You round your experimental data. When you design you make part the way you want it. There is nothing to round.

 

[KENAT]

So that's why Solid Edge does funky rounding.

Why is it they allow the actual drafting standards to be broken will nilly then decide to strictly follow some obtuse ASME standard most have never heard of.

Darn CAD companies.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[Walterke]

@CH, this kind of rounding is also used in the financial world because when you add everything up at the end, you're statistically closer to the real number (and some numbers just HAVE to be rounded at some point).
I don't really see where he says that 0 is not a real number though...
All I can find is

Remember that "rounding off" a zero does not change the value of the number being rounded off.

As for rounding in design, as soon as angles and circumferences are involved, you're pretty much bound to round something off eventually...

@KENAT, that's what I'm trying to find out

NX 7.5
Teamcenter 8

 

[CheckerHater]

Well, exactly,

"rounding off" a ONE does not change the value of the number being rounded off, "rounding off" a TWO does not change the value of the number being rounded off.

But you know what? It is pointless discussion. I'd like to see actual standard first.

 

[Walterke]

More a discussion on semantics then anything else if you ask me. So you're right, it's a pointless discussion.

And finding the standard is the goal I had when I posted this in this forum, since the people here seem to be more experienced on the subject.
Is there an ISO standard that says to round off differently?

NX 7.5
Teamcenter 8

 

[CheckerHater]

I don’t think ISO really needs it.
They do not have trailing zeros and they do not use number of decimals to indicate tolerance.
You can specify 12.3456 +/- 0.1 and be happy.

 

[dgallup]

Well, yes and no CheckerHater. You can specify that but then you need a measurement system accurate to 5 decimal places to check it so it would still be frowned upon.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[Walterke]

You have a point, but I feel like, at times, it might cause confusion when 2 people round off in a different way. (since ISO doesn't describe how you SHOULD do it, either way is correct then?)

NX 7.5
Teamcenter 8

 

[CheckerHater]

All I am saying is that dimensions and tolerances are representing fit form and function, not some lame attempt to legislate the mathematics.
And Walterke, if you are using metrics (judging from your example “2776mm becomes 2780mm”) you cannot use number of decimals to represent tolerance in ASME as well.
It's 2776 +/- 10 if you have to. (gdallup, we can have measurement system measuring to whole millimeters, can we?)

 

[Walterke]

I understand your point. The rounding off to 10mm wasn't a tolerance problem, it was just for aesthetics in a brochure (no important dimensions were harmed in the making of the brochure)

And yes, we have measurement systems that can measure whole millimeters

I'm still trying to come up with an example of how rounding off in different ways can cause problems later on but I can't seem to get anything formed properly.

NX 7.5
Teamcenter 8

 

[CheckerHater]

I think the only real problem with rounding is converting inch to metric, or back – you have to make sure you didn’t create larger tolerance, but you can always tighten you final result a little bit.

 

[KENAT]

Rounding off can cause all kinds of problems going from theoretically perfect model, to rounded of drawing because someone wants to invoke +-.01 from a 2 place decimal instead of +-.005 from a 3 place decimal etc. (your value may vary depending on what default tolerance values you're abusing).

It doesn't need to cause these problems (maybe there's the odd case but darn rare) but lazy/thoughtless/uneducated drawing producers often go in that direction.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[CheckerHater]

OK, let’s apply common garden-variety 6-th grade arithmetic to rounding.

Say, your drawing demands something like .XX+/-.03, .XXX+/-.003.

You model feature to .375. BTW, no one held gun to your head, you could make it .400.
You want it to be +/-.03, so you round it to .38, which results in limit dimensions .41/.35. (This is what machinist will check)
When you apply +-/.03 to .375 you get limits .405/.345.
Round them to same two-digit, you get .41/.35.

Same with even number.
Model .125. Round it to .13. Limits are .16/.10.
Apply +/-.03 to .125 you get limits .155/.095.
Round limits to two-digits and get .16/.10

As long as you stick to same rules you get same results. Not to mention that most of “rounding” problems are coming from designing in fractions and using number of decimals for tolerancing. Both belong in the museum.