Shop dimensioning and tolerance standards

[mrwolfdog]

I am the fabrication supervisor at a major paint company in the U.S.. In the Fab Shop, the welders make stairs that lead up to tanks, small footbridges that go over driveways, flood areas, hand rails, stainless 1oo gl mixing tanks for sample batches, and reverse engineered "stuff" to keep the plant going. The problem I am having is with a recently adapted policy to base our shop standards primarly on ASME Y14.5. The tolerance range of the things we make range from +- 1/16 to +- 1/4. The drafter that proposed this standard wants to dimension in decimal form, using .X as +-.1. I have looked for examples of this kind of callout all over the internet and can fine none. No matter what the tolerance, weather +- .1or +- .5, I want my guys to have the most defined target possible. If the designer wants 12.75, his only choice is .8. This seems very limiting, considering fractions have many more choices. Also, is there a Design standard or Engineering standard that says that whenever possible, to use standard numbers in the design? .1,.2,.3,.4,.6,.7,.8,.9 are not standard numbers. Standard number being a decimal equivalent of a common fraction. Maybe I am too old school but this just doesnt seem right. Any advice would be gratefully accepted.

 

[ewh]

There is no requirement to use decimal dimensions in Y14.5 AFAIK.
Fractional dimensions seem well suited to your environment. It would be far easier to continue their use (as long as item definition is not compromised) than to make those who have to use them adjust only for the reason of pleasing the designer.
At my current workplace, fractional dimensions are the norm, with decimal dimensioning being reserved for more precise machined components, with few problems.

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

 

[pmarc]

Very briefly saying, the method chosen by the drafter simply does not fit to your environment.

Y14.5 does not show application of fractional inch tolerances, but I would not say you can't use them for your purpose - as long as it is clearly stated somewhere on the drawing how to read it.

 

[SteveMartin]

There's nothing that says you can't have a 12.75 +/-0.10 dimension. It sounds like your problem is more with a lazy drafter who wants to use one default tolerance to cover all dimensions (a bad idea in my opinion). Perhaps you need to work out a more detailed default tolerance block, or even get rid of default tolerances altogether. Even if you don't get more than a single default tolerance for .X dimensions, there's nothing (in Y14.5) preventing your designers from putting other dimensions with explicit tolerances on the print. Most default tolerance blocks I've seen say something to the effect of 'unless otherwise specified'.

I don't know of a standard that says to use common fractional numbers. That could be part of a company standard referenced on your drawings. ASME Y14.5 says that decimal dimensions shall be used, with only a few exceptions (stock sizes of pipe & lumber sizes for examples). A lot of designers I've known prefer to work to common fractional sizes, but it's not mandated or preferred anywhere I know of. For the work I do it doesn't even make sense anymore (all the machines generally have digital readouts) - though your situation may make the fractions make sense due to the equipment used by your shop.

I disagree that fractions have more choices than decimals. Considering the following options for a value between 0-1: Allowing fractions down to 1/16 (a .XXXX decimal) and decimals down to .X gives 16 fractional options and 10 decimal, but opening decimal to just .XX gives 100 options and moreso for every decimal thereafter. The finest fractional measures I've seen in use (1/64ths) offers 64 options. The finest decimal measures I've seen in use (.XXXX) offers 10,000 options. If you're willing to subdivide far enough they have equivalent options, but in practice decimal has more.

 

[KENAT]

ASME Y14.5M-1994 1.6 "Decimal dimensioning shall be used on drawings except where certain commercial commodities are identified by standardized nominal designations, such as pipe and lumber sizes." (C'mon ewh - you knew that.)

However, there is no reason the decimal equivalents of fractions can't be used.

I'm not sure I fully understand the questions, why not directly tolerance all the dimensions if you can't come up with a block tolerance scheme that meets your needs?

Else come up with a scheme like the more common tol blocks I've seen but adapted for your looser tols, e.g.

.X = +-.25
.XX = +-.125
.XXX = +-.063

Remembering that trailing zero's on dimensions can be used to invoke the tighter tols. If you want a looser tol than the block tol would invoke then you have to directly tolerance the dimension.

There is some standard somewhere about using only selected numbers but it doesn't correspond to what you're calling standard numbers as I recall.

Posting guidelines faq731-376

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

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

 

[ewh]

Somewhere in the midst of my past, you're right, I knew that, just in too lazy of a mood to actually break (my own copy of) the standard out right now...
Being subject to my current period of serfdom where manufacturing drives engineering, I have come to accept working under the conditions prevalent. I long for the days of actual communication of engineering design to manufacturing rather than the other way around.
In their (weak) defense however, almost everything is fabricated from standardized structural components. I'm not in a quixotic mood to fight their many transgressions for such a short contract as this.

On second thought, I just might try to endear myself even more by pointing out this slight transgression from the standard purportedly used... nothing to lose as a short timer.

As for other ET members who may work here, I'm kidding... really! And this usually isn't considered a "soft" forum, but one that actually pertains to my assignment.;-)

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

 

[MechNorth]

Hmm. In '94, "1.5.2 U.S. Customary Linear Units. The commonly used U.S. customary linear unit used on engineering drawings is the decimal inch." which almost seems to contradict 1.6. However, 1.6 is more definitive and I would tend to go that way. On the otherhand, people regularly deviate from the Y14.5 standard to suit their corporate needs. When they do, and when Y14.5 is called out on the drawing, they need to just add a note on the drawing or in a corporate addendum to identify the deviation.

Personally, I've heard of a palette & crate manufacturer that went to GD&T and continued using fractional inches as that's what the tools and equipment were built for, and I've worked with one fabricator who does fractional & decimal inch in the same drawing without issue. Where fractional-inch dimensions are used, they typically adhere to fractional-inch tolerances as well.

As to the accuracy, chances are that the difference is at the 3rd or 4th decimal place and in weldments you're not that precise to begin with. If it's crippling fabrication, then fight back, otherwise get the new rulers & such and flip.

Jim Sykes, P.Eng, GDTP-S
Profile Services

http://www.profileservices.ca

TecEase, Inc.

http://www.tec-ease.com

 

[mrwolfdog]

Good feedback people. One thing that I did not mention is that we have a machine shop in the department. My prefered tolerance block layout is as follows:

X/X=+/- 1/8 (Band saw / Plasma cutter / welding ops.)
X.XX=+/- .02 (primaraly for use of raw stock edges)
X.XXX=+/- .005 (General machining tolerance)
X.XXXX=+/- .0005 (press fits etc.)

These are just the defaults. The actual design will dictate the actual tolerance when diferent from the default.

The drafter that wants all decimal call-outs has wanted them forever. He recently attended a GD&T class and observed the mandate for an all decimal format/standard. So, he is introducing ASTM 14.5. ASTM 14.5 is great for cars, airplanes, cell phones, flashlights and such but not for factory support.

My main issue with single place callouts (X.X) is that you loose the real intended target. If we need to modify a door that has a desired dimention of 24-1/4" the print will have to say 24.3 or tighten the tolerance to 24.25 (remember, X.XX = +/- .02)The welders can't do that.

Thanks again folks

 

[ewh]

As noted, you can still follow the standard as long as any exceptions to it are defined somewhere on the drawing. It is a mandate only as long as such exceptions are not accounted for.

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

 

[fsincox]

I have delt with just this limitation on inch dimensioning all my life and feel you pain. I have mentioned the same issue here many times, thanks for putting a new face on an old problem.
Frank

 

[MintJulep]

This is exactly why making tolerances a function of the number of significant digits used in the dimension is stupid.

Perhaps for your application you would be better of with something ISOish.

If the dimension is less than a foot tolerance is 1/32"

If the dimension is a foot to three feet tolerance is 1/8"

Anything longer than three feet tolerance is 1/4"

 

[btrueblood]

Is it really "shall be", or is it "should be" like the rest of Y14.5?

 

[KENAT]

You could have different tol blocks for fabricated V machined parts. Generally I shy away from customizing tol blocks that are usually standard because the customization often gets overlooked but where you have 2 obviously separate categories it might work. I actually had someone modify our tol block similar to what the OP's follow up post proposes for some foam parts one time. I didn't like it but ended up allowing it (this was back in my (attempt at) checking days.

Back at my place in the UK there was no standard tol block. We would select the tolerance that worked for the most dimensions on that drawing and put that single value in the tol block in the title area.

Then for every dimension that needed a different tol we would directly tolerance it.

Seemed a perfectly workable system, and I don't recall any complaints about cluttered/busy drawings because of all the direct tols. I also don't remember drawings taking lots longer, however it did teach me to think about what tolerance I'm actually specifying which is all to easy to forget with the typical US block, or even reliance on ISO 2768 etc.

As to losing the intended target with X.X - that only happens if the drafter is too lazy/stupid to directly tolerance that dimension. I often have .375 or similar that I don't want to invoke our default 3 place decimal +-.005 tolerance on. So I get crazy and directly tolerance it +-.030 or similar.

Mint, stupid is too strong, it has its use if used properly. However, it's not the be all and end all for every situation. The ISO system is equally open to abuse.

Btrue, I copied that from 14.5, no paraphrasing (though if you look close enough someone may find a typo/incorrect auto correct). Y14.100 defines 'shall' but I don't care enough to type that one up. Given all the stuff in section 2.3 about leading and following 0's I'm pretty sure they didn't want fractions.;-) Whether they were justified in this I'll lead to others to espouse.

Posting guidelines faq731-376

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

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

 

[drawoh]

mrwolfdog,

This might be a good time to switch to metric.

As per ASME Y14.5, millimeter dimensions are shown without trailing zeros. With most CAD packages, you have to manually apply tolerances to each and every dimension. This actually, is not very much work, unless your required tolerance is difficult for the shop to do, in which case, you have detected a production problem.

Your painters and welders could learn to appreciate the lack of need for decimals or fractions on millimeter dimensions.

I usually double my standard tolerances. ±0.1, ±0.2, ±0.4, ±0.8, ±1.6. This converts closely to your old fractional English tolerances, with 1.6mm being pretty close to 1/16".

JHG

 

[fsincox]

drawoh
I agree, rounding to the nearest mm (.039") for fabrication is not a blg deal to the nearest inch seems only practical when measuring land.
Frank

 

[SteveMartin]

For the record, I agree that default tolerance schemes based on number of decimals are silly. I'm saddled with one right now, but some fights can't be won.

Having said that, the only problem with having a 24.25 dimension on a print where .XX gives an overly tight tolerance should be a designer/drafter too lazy to put a direct tolerance on that dimension. I have seen a title block that claimed all tolerances were per the block - no exceptions (which meant anything directly toleranced was overdefined/unclear), but in general block tolerances are for dimensions that do not have a tolerance directly specified. Often direct tolerances are tighter than the general ones, but they can also be much looser. You could have 24.25 +/-5.00 if you really wanted.

 

[CheckerHater]

Not to sound like a troll, but if your tolerances are +-.25, the difference between 12.25 and 12.3 is NOTHING in comparison. Get over it. Welders won't detect it anyway.

 

[ptruitt]

Just get the Standard and read it. It is not too long.

Peter Truitt
Minnesota

 

[fsincox]

ptruitt,
Sorry, Which one did you mean?

 

[ptruitt]

I was just replying to mrwolfdog, fsincox. If he just reads Y14.5-(XXXX), he can decide for himself if his staff is really limited by the Standard's requirements, or not. Apparently, meeting the Standard is settled law there. If the question related to some of the more advanced GD&T concepts, then we might all add our two cents. Simply reading the Standard would be most enlightening, I would think.

Peter Truitt
Minnesota