Interesting Discussion Is "rounding" acceptable when measuring parts? Tolerance Limits - What is acceptable?

D

David Bear - 2010

Is "rounding" acceptable when measuring parts? I know that I should already absolutely know the answer to this. However, I am confused about the answer. I spent many years of my life obtaining an education that said that 8.24 is acceptable is the specs are 8+/-0.2. It would be unacceptable if the spec was 8+/-0.20 (note the additional significant digit). In practice, I see parts measured at 8.201 rejected because the spec is 8+/-0.2. Have I missed something in my educational journey?:confused:
 

Jim Wynne

Leader
Admin
Re: Tolerance Limits - What is acceptable?

Is "rounding" acceptable when measuring parts? I know that I should already absolutely know the answer to this. However, I am confused about the answer. I spent many years of my life obtaining an education that said that 8.24 is acceptable is the specs are 8+/-0.2. It would be unacceptable if the spec was 8+/-0.20 (note the additional significant digit). In practice, I see parts measured at 8.201 rejected because the spec is 8+/-0.2. Have I missed something in my educational journey?:confused:

I think we've had this discussion before, and if I remember correctly, I was amazed to find that there were several people who didn't believe that 8.24 is >8.2. To answer your question, the upper limit is absolute and is not affected by the number of decimal places in the specification. When tolerances are classified by the number of decimal places in the specification, it's a matter of convenience, and does not convey permission to to round if the specification has clearly been exceeded.
 
C

Craig H.

Re: Tolerance Limits - What is acceptable?

FWIW, I try to include significant digits according to the device that will be used to measure the attribute, no more, no less.

In direct response to your question, I would think that if it is possible the customer should be consulted. And, the additional significant digit, whether it be .00 or .05, should be added to the spec.
 

Statistical Steven

Statistician
Leader
Super Moderator
Significant digits are set by the specification. If the specification is 8.0 +/- 0.2 then you have to round 8.24 to 8.2 and it is acceptable. Had the specification been 8.00 +/0.20 then 8.24 is out of specification.
 
D

David Bear - 2010

Re: Is "rounding" acceptable

There appears to be a great differance of opinion among those in this forum.
My confusion comes from all that darn mathematical training on significant digits.

My problem is that I have issues with Design Engineers. How can a formed sheet metal part have a hole to hole measurement of 120+/-0.2mm? or a foam gasket have a dimension of 60+/-0.5mm? Manufacturers will send me PPAPs in spec and then our CMM operator will measure the first part at 119.77mm. I am highly concerned that the part has a good Cpk, but the dimension could be 118mm and would be acceptable for my application. It costs a lot of money to get a part into the acceptable range.

I have not seen any cases where the Design Engineers specify how to measure the parts. If the tolerance is +/- 1mm, is it just as acceptable to use a scale as to use a CMM? I know which device I want to use and when, but it may not be what the Design Engineer had intended.

note: I used to be a Design Engineer, so I'm allowed to complain about them.
 
S

suziwann

Is "rounding" acceptable when measuring parts?

I think it depends on the part, it's application and the particular dimension.

After extensive testing on 1000s of products, I found that we were unable to 'round off' numbers where critical dimensions were concerned or we suffered failures.

With parts going to some countries the consequences of getting it wrong were 'minor,' but with other parts going to other countries the consequences could have been fatal.
 
B

Bill Ryan - 2007

Re: Is "rounding" acceptable when measuring parts?

Significant digits are set by the specification. If the specification is 8.0 +/- 0.2 then you have to round 8.24 to 8.2 and it is acceptable. Had the specification been 8.00 +/0.20 then 8.24 is out of specification.
Sorry but my "automotive world" mindset just blew a gasket (don't worry - happens all the time :rolleyes: ). Being in that mindset, I'm with Jim - 8.24 > 8.2 and is out of specification. I suppose it's possible that in another industry 8.24 is NOT out of specification and rounding is OK but I can sure see problems if there isn't a standardized method for measurement interpretation. Put simply, an 8.24 diameter shaft is going to have some problems assembling into an 8.2 diameter hole regardless of the "industry".
:2cents:
 
Last edited by a moderator:
G

Gert Sorensen

Re: Is

Is "rounding" acceptable when measuring parts? I know that I should already absolutely know the answer to this. However, I am confused about the answer. I spent many years of my life obtaining an education that said that 8.24 is acceptable is the specs are 8+/-0.2. It would be unacceptable if the spec was 8+/-0.20 (note the additional significant digit). In practice, I see parts measured at 8.201 rejected because the spec is 8+/-0.2. Have I missed something in my educational journey?:confused:

Well, if you've missed something, then so have I. Of course, rounding is acceptable. The issue is not rounding, the issue is specification. If you have an absolute demand that a part cannot exceed e.g. 8.2400 (and not 8.24) then it must be specified. Lack of specification, and lousy specification, is generally a source of problems. Design specified to 0.000 mm is - unfortunately - quite common in my field, but the specification does not necessarily match the requirements for the rest of the kit, or the design engineer "forgot" to define how to measure the specification :notme:

IMHO time spent clarifying specifications and measuring methods for these are really well spent.
 

Jim Wynne

Leader
Admin
Re: Is "rounding" acceptable when measuring parts?

Sorry but my "automotive world" mindset just blew a gasket (don't worry - happens all the time :rolleyes: ). Being in that mindset, I'm with Jim - 8.24 > 8.2 and is out of specification. I suppose it's possible that in another industry 8.24 is NOT out of specification and rounding is OK but I can sure see problems if there isn't a standardized method for measurement interpretation. Put simply, an 8.24 diameter shaft is going to have some problems assembling into an 8.2 diameter hole regardless of the "industry".
:2cents:

It's not just the automotive world, Bill. Drawing specifications mean what they say unless there is bilateral (customer/supplier) agreement to the contrary. It can't be any other way. I would like for someone--anyone--who believes that specification limits are tacitly elastic to explain to me why we also insist that measuring instruments are capable to at least one decimal place beyond the spec limit.
 
B

Bill Ryan - 2007

Re: Is

Design specified to 0.000 mm is - unfortunately - quite common in my field
So you are saying that 0.0003 mm over the specification is acceptable? Sorry but I don't buy it. I used to think along the "classical" rounding theory also, but it just isn't so in automotive manufacturing (and I would highly doubt aerospace (example) thinks rounding is OK).
 
Top Bottom