Significant figures and tolerances 8

[microbus]

thread301-176582

I dug out the above referenced thread and hope to get the groups opinion on this real world example:

I have a customer that has requested a coating thickness of 2 to 4 microns. Their drawing doesn't reference any dimension or tolerancing standard.

Actual thickness of the part is measured at 4.2 microns. I have argued to my QA department that 4.2 should be rounded to 4 as the significant digit in the specification is a whole number. In other words, they did not require 2.0 to 4.0 microns.

Should the reading of 4.2 be rounded in this case?

 

[desertfox]

Hi microbus

I disagree with your logic, in your first line you stated that your customer required a coating thickness between 2 and 4 microns and thats what it should be, your tolernce band for coating thickness is 2 to 4 microns not 1.8 to 4.2.

desertfox

 

[microbus]

Thanks for the reply.

My logic would actually allow for a thickness range of 1.50 to 4.4 on a specification of 2 to 4. 1.5 would be rounded to 2 and 4.4 would be rounded down to 4.

What I am trying to understand is whether there is a difference as far as rounding and significant digits goes for a specification of 2 to 4 versus 2.0 to 4.0? Does "2 to 4" imply "2.0000000000+ to 4.0000000000+"?

 

[desertfox]

hi microbus

No theres no rounding 2 to 4 is 2.0000000000000000000 to 4.000000000000000000000 thats it

desertfox

 

[microbus]

Got it, thanks.

 

[Compositepro]

In the real world of manufacturing your interpretation of +/- one half of the least significant digit is commonly used. Only your customer can tell you what he really meant or needs. But if the customer did not care to be very explicit about a very tight tolerance requirement then he probably does not need the tighter requirement.

The statement that 2 to 4 means 2.00000000... to 4.000000... is not correct. That would require measuring instruments that are even more precise.

 

[desertfox]

Hi Compositepro

I agree about the number of zero's after the 2 and 4 however the point was that a customer had requested a coating thickness of between 2-4 microns and what I was saying was thats what it should be between those figures or the max and min figure

desertfox

 

[Compositepro]

But, Dessertfox, if you think about it you are actually contradicting yourself. I understand your point but measurement and tolerancing are not so simplistic. Every measurement has a tolerance. Only the customer can define what the requirements are. If you are using a gauge that is only accurate to a half micron what does a reading of 4 microns mean? The issue can get very complicated. So if the customer requires greater precision he needs to be more explicit. There are whole books written on the subject and they don't provide a simple answer to the question. It depends.

 

[desertfox]

Hi Compositepro

Every dimension as a tolerance I agree, if you write on a drawing or specification that the thickness is between 2 and 4 whether it be microns, millimetres,inches then thats what it must be, anything above 2 and anything below 4 is acceptable if its 1.8 or 4.2 its not.
If your measuring something then you need to make sure the tool your using is adequate for that purpose.
I have no doubt that its a complicated issue but unless their is any mention in the specification or on the drawing about significant figures which the OP as not mentioned in any of is posts then he is wrong to assume that there is.

desertfox

 

[fsincox]

I must agree with desertfox. While I understand your point and your post specifically states that no standard is specified, there is a standard on dimensioning and tolerancing and to pretend it does not exist is detrimental to our common understanding, communication and commerce. We are currently using a standard, English as a language, to communicate; to change the definition of words arbitrarily would not facilitate that, either. Dimensional limits on a drawing have always been interpreted to be absolute. If you feel you method of measuring is not accurate enough and the .2 is a measurement error that is your business, your customer may not feel the same. If the spec is intended to be 4.2 it should be changed to say so.

 

[KENAT]

Per any relevant standard I'm aware of what Desert Fox says about 2.0000000... to 4.0000000... is correct, for example ASME Y14.5M-1994 section 2.4.

Your get out is that your customer doesn't referance a standard (double check it's not in your or there standard terms or contract though, even if not on the drawing). However, do you really want to play that game? Is it worth winning this one at the expense of maybe upsetting the customer longer term? Yes, you'd probably win in a court of law but do you want it to get anywhere like that far, where is the cost/bennefit.

Maybe you go back to the customer, ask what they want (Absolute or rounding sig figs) or maybe just ask if the 4.2 is OK on some kind of waiver etc. and change your process to meet it next time. Find someway to meet in the he middle.

Posting guidelines faq731-376

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

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

 

[microbus]

Kenat,

FYI - I did check their terms before posting and there is no reference to any standard. The issue has been resolved and the customer is happy to accept the parts at 4.2 microns. After discussions with the customer, they actually want >2 microns not to exceed 5 microns. The "2 to 4 micron" specification was a carry-over from a previous supplier's high end limitations. They are updating the specification to read "2.0 to 5.0" microns. This helps us both to have clarity.

There was no game being played here. You apparently misunderstood the motivation for my post. I was looking for technical advise about the definition and interpretation of a somewhat vague specification, not business or legal advice.

 

[KENAT]

microbus, I put what I did because you explicitly pointed out that they didn't reference a spec. Part of the reason for the development of the specs was effectively as a legal document explicitly detailing what you would accept. So in this case, the technical specification and legal ramifications are intrinsically linked.

The concept of absolute limits is a fairly fundamental one in dimensional specifications on drawings, given that you weren’t familiar with this I thought you may not have been familiar with other factors I went on to comment about.

Sorry if trying to take a broader view of the question upset you.

Posting guidelines faq731-376

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

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

 

[mfgenggear]

Glad that the issue is resolved. ^_^


All parties are correct.

The tolerance that is specified should be held.
but yet it's not a relistic tolerance that can be held.
if I remember 1 micron = .000039"
so 4 micron =.000156"

which = .000117 tolerance which is not gonna happen.

I would have the drawing change MM or Inches thick
that is reasonable. without knowing the fit, form or function is tough to say.

.0001-.0005 plate is resonable to me.

 

[ewh]

So what happens in a couple of years when another vendor is chosen to make the parts and wants to know if 5.2 microns is acceptable per the original arguement? The original problem has not been solved, only avoided.

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[mfgenggear]

ehw

the lower portion of the my previous post states
the drawing requires changed.
this would be the long term fix.

it is possible this is an old drawing & their customer is reluctent to change it. even if it's wrong.

 

[ewh]

I was refering to microbus' comment that "The issue has been resolved". The underlying issue has not been resolved, only the clarification of what the customer will accept. Changing the units doesn't change the problem. Dimensions are given as limits and significant figures shouldn't enter into it.

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[mfgenggear]

ultimately it's upto the owner of the drawing what is required. Regardless of our opinions.

beating up microbus about this is just wrong.

The drawing is not practical.
and unless we know the fit form & function.
it's not worth arguing.

 

[ewh]

It doesn't matter if the drawing is practical or not, and I'm not beating up on microbus. Fit form and function also don't enter into it. The problem is not accepting that a dimension and it's tolerances are limits that are not subject to rules concerning significant figures. This is not opinion, but the basis of what is deemed acceptable when manufacturing to a drawing. As such, it is a point worth arguing.

"Good to know you got shoes to wear when you find the floor." - [small]Robert Hunter[/small]

 

[mfgenggear]

ok then

the tolerance is 2 to 4 microns
in the inch = .000078"-.000156"
which is .000078" tolerance.

what type of plating can hold that tolerance.
my cnc grinders can hold that tolerance.
maybe if I lap them to 3 light bands flatness.

normal plating that I work with day in and day out
is normally .0005' tolerance (12 microns)

ok