tolerance analysis to ISO 8015 & 2768 part 1 (again!)

[KENAT]

This is a follow on to thread1103-196260.

I sent a message out to the engineer copied to the manufacturing engineering manager and my boss saying I was still concerned about interchangeability but that if that Engineering department and Production didn't think it was worth the effort of verifying/ correcting it then it could be excused check.

No one seems to just want to say 'excuse it from check' but at the same time the engineer involved is now telling anyone who'll listen that I'm wrong, I don't know what I'm doing and that I haven't read the correspondence he had with the vendor. (He may have a point on the first 2 but I did read the correspondence;-).)

I've looked at ISO 2768-1 in detail and find nothing to suggest that 2 holes, independently located from a surface by their own separate dimensions (in one plane) with tolerances from iso2768 are in any way linked to each other such that the distance between them also has to be within the tolerance from the iso.

For instance (same example as before) I have part of a hole pattern, two threaded holes in line. The first is 36.66mm from the 0 ordinate. The second is 103.34mm from the 0 ordinate. Nominal spacing therefore is 66.68mm. From the extract of the ISO I found both 103.44 & 36.66 are +-.15mm. Therefore for interfacing purposes I assumed that the spacing is effectively 66.68 +- .3 (.012”). The vendor says this is wrong and that the spacing is 66.68+-.15.

I’ve also looked at ISO 8015, which is referenced by 2768, and find nothing to change my opinion.

The only ISO referenced on the drawing is 8015, it doesn’t invoke any others.

Another example from the drawing, they have a stepped female hole (like a large c’bore hole). They specify it as R35 & R42 from iso 2768 this gives them +-.15. I’m trying to do the calculation in B6 of ASME Y14.5M-1994 so need to convert to diameters. To do so I’m multiplying both the nominal radius and the tolerance by 2 to get 70+-.3 and 84+-.3 does this sound correct?

Given the addition of ISO 8015 does this change anyone’s opinion?

If anyone thinks the vendor is correct could they let me know which ISO gives this information, and if possible give me the paragraph number.

I am going to contact the vendor but it has got so political that I wanted to back it up with a post here.


KENAT, probably the least qualified checker you'll ever meet...

 

[KENAT]

Forgot to say, if someone knows of a reference that says the vendor is definitely wrong then that would be even more appreciated as it would mean I was right;-).

KENAT, probably the least qualified checker you'll ever meet...

 

[KENAT]

So I've either had a stroke of genius or I've been looking at this too long.

ISO 2768-1 basically gives a table (

http://www.ramo.se/iso_2768_english.htm

) and the tolerance on a dimension varies based on tolerance class and size of feature/dimension.

What our vendor seems to be saying is that if you calculate a dimension from those given on the drawing then the tolerance on that resulting dimension also comes from the table, not from a tolerance stack.

It's a bit like the information you get from moulding/casting/forging shops where they say "for dimensions from 0 to 1 the tolerance is +- X, for 1 to 5 it's Y and for 5+ it's Z".

So, for example, based on the 2768 fine grade then any dimensions that are nominally/theoretically between 30 and 120 mm will always have a tolerance to that dimension of +-.15mm.

What I haven't yet got my head around is the implications of this on tolerance of mating parts, specifically matching hole patterns or coaxial diameters (as per ASME Y124.5 B6).

I'm thinking of taking a copy of their print, re-dimensioning it to match our dimensioning scheme and then applying the tolerances from 2768 to those dimensions, then do the tolerance analysis on our part. (Can't make ours match theirs as our part is basically circular datum off a central hole while there's is square datum off 2 edges with a stepped hole in it to match our part)

Does this sound reasonable or has the stress gone to my head?



KENAT, probably the least qualified checker you'll ever meet...

 

[powerhound]

KENAT,

What is the purpose of creating a dimension between the two holes? Why can't each hole remain located WRT the datum they are called out to instead of to each other? While you are correct in that if one hole were located at one end of the tolerance and the other located at the other end, they would be .3 from nominal spacing but you can't leave one hole located WRT to a datum and then change the other to be located from the hole instead of the same datum and expect to be able to add a +/- .3 tolerance and achieve the same result.

Powerhound, GDTP-T
Production Supervisor
Inventor 2008
Mastercam X2
Smartcam 11.1
SSG, U.S. Army
Taji, Iraq OIF II

 

[KENAT]

Powerhound it's maybe a bad example but it's the one the vendor and engineer have been using.

I am not creating a dimension between the holes nor am I looking at re-dimensioning the vendors part in any way, I am looking at the mating part we make.

In the case of the holes I am just trying to understand how the hole position varies so I can put the appropriate clearance hole and positional tolerance on our mating part. Can I just convert the +-.15 to diameter and use standard positional calcs or does the fact that the gap between the holes is supposedly also +-.15 mean I use a smaller figure and if so what, +-.75?

If that example is so bad as to be confusion consider my second example on the radii. They are essentially saying I'd calculate the nominal diameter by doubling the radius and then look up the table again for the tolerance on this calculated diameter.

To clarify what I stated in my second post about not being able to exactly match their tolerancing scheme. Their part is essentially square. They use 2 edges as their 0 datum for ordinate dimensioning, effectively for the purpose of simple tolerance analysis I'm treating them as datums. They essentially have a stepped female hole which our part sits in (the hole actually has a flat on one edge).

On our part I'm using one of the diameters that sits in their hole as the datum (with another datum on the straight chord for orientation).


KENAT, probably the least qualified checker you'll ever meet...