Go/NoGo gauge tolerance

[dogbural]

Hi,

I have a spacer with ID 8.1 (+0.2/0).
If I would like to make a Go/NoGo Gauge, what should be their diameters?

If we make a Go gauge being machined at 8.10, is this the correct Go gauge? Or should it be below 8.10, say 8.09 instead?

 

[IRstuff]

I don't understand; if your design tolerance is +0.2/0, then isn't 8.3 acceptable per design? Why would you design to tolerances that aren't acceptable (GO)?

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[SWComposites]

the tolerance is 8.1 + 0.2 / -0.0, so the acceptable dimensions are 8.1 - 8.3.

You need two go/no-go gauges - one at 8.31 and one at 8.09. If either fits in the ID, the part is no good.

 

[3DDave]

The 8.09 should always fit (GO) unless the hole is too small. 8.31 should never fit (NO-GO) unless the hole is too large.

The question of how much too small or how much too large Go/no-go gauge sizing is usually on the 10% rule - make it 10% of the tolerance range and to the inside. So larger than 8.1 and smaller than 8.3. They can be closer to the limits if you want to spend more money on the precision of the gauge making. For diameters, there are pre-made sets with very small increments available.

 

[3DDave]

See

https://www.qualitymag.com/articles/92337-the-infamous-no-go-gage

 

[Belanger]

3DDave, your go gage of 8.09 would accept a bad part (undersized).
Dogbural, think of it this way -- the ideal go gage would be 8.1 and the ideal no-go gage would be 8.3. But since gages are real-world items, they need tolerance (usually a total of 10%, or 5% to each gage; see ASME Y14.43 or the article that 3DDave linked to). And the tolerance on the gages should be applied in such a way that a borderline good part may be rejected but a borderline bad part never accepted. Thus, the go gage should be 8.11.

Regarding the no-go gage, I think a point that everyone is missing (including the article) is that a no-go gage should be able to get deeper in the hole (think of a telescoping gage), rather than merely butting up to the entry or exit face of the hole. So unless it's sheet metal or some other thin piece, the common image of a two-side gage (red and green) isn't ideal. A hole might be the correct size at the entry point, but then barrel open to a huge diameter in the middle of the depth; that won't be discovered by the go/no-go gaging that's often depicted.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[3DDave]

No wait - the 8.09 IS outside and it would accept parts but, guess what - that's an acceptable practice, one of several, in the making of gauges.

I even had this conversation with Meadows about the fact that applying 10% of the defining tolerance for a datum feature was wrong because it would allow features that depended on those datum features and were out of spec to be accepted and would reject parts that are in spec. Meadows said it wasn't a problem.

10% is a fun amount - it exists only because we use the decimal system. It's entirely arbitrary, but it is also for the convenience that moving a decimal point is easier than any other calculation.

However real manufacturing equipment is not discretized on making decadal errors. Why not 8.213%? Because it is for making the gauge makers and inspectors job simpler, not to make acceptance of the parts better.

The tolerance should be based on how closely to the limit the designer has accounted for. Unfortunately if they have done what machine shops love and used "all the tolerance" then that amount allowable would be zero.

It ends up being a sneak play to hide from engineers the true results being generated out of the factory to have rule of thumb discounts on the tolerance limits.

 

[Belanger]

The 10% business (or 5%) is arbitrary, but no -- regardless of your percentage -- you shouldn't put that tolerance on the negative side of the 8.1 limit. That's why I disagreed with your 8.09, because you're opening the door to bad parts being sent to the customer. It's best to put any gage tolerance on the side where it adds material to the gage.
So the main purpose of my post was to remind the OP about two issues: 1) allocating that gage tolerance so that it's not bilateral, but rather so that it prevents a bad part from being delivered to the customer; and 2) the fallacy of having a no-go gage be a plug style.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[IRstuff]

So the main purpose of my post was to remind the OP about two issues: 1) allocating that gage tolerance so that it's not bilateral, but rather so that it prevents a bad part from being delivered to the customer; and 2) the fallacy of having a no-go gage be a plug style.

The issue is that the gage is not perfect, so machining a gage at 8.0 requires some tolerance, which means that the plus end of the tolerance would result in a gage at 8.0 + epsilon, thereby rejecting a part that's exactly at 8.0; so how much good material do you want to throw away vs. dealing with fit issue further down the line?

The apocryphal story of Ford and Mazda transmissions comes to mind. Ford used to mix and match engines to transmissions, since they tended to have a lot of parts that would be at the edges of the tolerance range, while Mazda's production philosophy was to build everything to the middle of the range. If the OP's manufacturing process is statistically well centered, the 8.0 ± epsilon might be OK, since there's little chance of rejecting in-spec parts, but if the process tends to bang into the edges of the tolerance, then there may be an unacceptably high rejection rate.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[LittleInch]

I think the real difficulty here is that this is an ID so it needs two gueage which one is a no go if it DOESN'T go in the hole ( the 8.10mm one) and one which is no go if it DOES fit in the hole (the 8.3mm one), but mostly that one shouldn't fit in. Easy to get that wrong.

It's much harder to get something to fit inside something as a guage than the external one IMO. You need to get the angle just right and also not damage good parts by nearly getting the bigger one inside.

These are also very small fittings and small tolerances.
Can't you just measure it with a feeler guage?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.