Issue with specifying a dimension - How to?

Fal Grunt · Nov 3, 2020

I'm not an engineer, and throughout my time in the trade, very rarely MADE prints. There is one unlucky soul here on the forum that has to navigate his way around my terrible blueprints and I have one that I am sending that frankly I am not sure how to specify some specific features.

My background is in Tool & Die, and as dumb as this may sound, the shops I worked at, rarely worked to tolerances. We had a dimension, we made it to that dimension. At the one big shop I worked at, they had all sorts of tolerancing, but you did your damndest to NAIL the nominal dimension, because chances were, if you used the tolerance, the parts wouldn't work together.

I have a part that is being made for me. It is a threaded shaft. The threads are an oddball bastard. I designed both parts, and made them as simple as possible. Once the shaft is made, I will make the "nut" that threads on. There are two diameters, along with the thread, that mate to the nut. The two diameters keep the nut concentric and shoulder the nut flush to the shaft.

I don't really care about the dimensions within reason. What I CARE about, is consistency. One of the diameters is dimensioned at .299", plus or minus .001". Not exactly close tolerance, and enough clearance that the nut would wobble on one end of the spectrum and be tight on the other. Specifying plus or minus .0002" makes it a stupid expensive part.

Now, the question at hand.

Short of tightening the tolerance, how do you specify that consistency is key?

For reference also, we aren't talking about 10,000 parts. We are talking about small quantities. 30, 50, 100 piece jobs.

mhajicek · Nov 3, 2020

By tightening the tolerance. Period. You need to be able to send that print to any vendor at any point in the future and get good, interchangeable parts, otherwise it's a bad print.

You say you don't work to a tolerance, but just nail the dimension; that tells me you need better measuring equipment. Everything is made to a tolerance, it's just a matter of how big.

And if parts made using their tolerance don't work, the tolerances were specified wrong.

Mike1974 · Nov 3, 2020

Would it be doable to match them in pairs or groups? Years back we kept getting a no quote on two parts. I asked to see the print. 1 part was something like +.0002" / -0.0, the other -.0002" / +0.0

Not terrible in todays world, but for a few pieces, either more trouble than it's worth, or the parts were going to cost a small fortune. I convinced the engineer to leave one part marked up at say +/-.001", and mark the other one slip/press (one of those) to detail xx. Problem solved, got what we need for manufacturing, got a fair price, someone (hopefully) made a bit of profit from the job.

edit: I see you are making one of the mating details? Maybe have a print note "+/-.002 OK, but group into .0005" increments - ie .299 +/-.0003 <> .300 +/-.0003 etc "

Mtndew · Nov 3, 2020

Fal Grunt said:
Short of tightening the tolerance, how do you specify that consistency is key?

One way to specify it, is to add a note to the dimension saying something like: "All parts alike within .001 for this feature"

gustafson · Nov 3, 2020

mhajicek said:
By tightening the tolerance. Period./QUOTE]

No

no no no no no no

No

IF one has a 1 inch pin one wants to fit in a 1 inch hole, it is not engineering to make both parts pus or minus nothing

If there is no other reason, on dimensions the pin .994 + - .005 and the hole 1.005 + -.005
IF you need a tighter clearance, as the max there could be .021, then tighten the tolerance. Like +.005 minus 0 and -.005 plus 0

Tightening tolerances tighter than what you need is first year engineer stuff

mhajicek · Nov 3, 2020

gustafson said:
No

...
IF you need a tighter clearance, as the max there could be .021, then tighten the tolerance. Like +.005 minus 0 and -.005 plus 0

Tightening tolerances tighter than what you need is first year engineer stuff

You said "No", and then talked about specifying a tighter tolerance.

I agree that one should never specify a tolerance tighter than needed, that adds unnecessary cost. The key is to specify the RIGHT tolerance for the application. Modern manufacturing is based on interchangeable parts. If you use custom fitted parts you're just kicking the can down the road; someone sometime down the road, will need to have a replacement part made. They'll send your print to a supplier, get the part, make sure it meets the print, and then wonder why it doesn't fit.

gustafson · Nov 3, 2020

mhajicek said:
You said "No", and then talked about specifying a tighter tolerance.

I agree that one should never specify a tolerance tighter than needed, that adds unnecessary cost. The key is to specify the RIGHT tolerance for the application. Modern manufacturing is based on interchangeable parts. If you use custom fitted parts you're just kicking the can down the road; someone sometime down the road, will need to have a replacement part made. They'll send your print to a supplier, get the part, make sure it meets the print, and then wonder why it doesn't fit.

I said you can tighten the tolerance, not as the answer to a problem
proper tolerancing is the key

Explain how my tolerancing would not fit?

yeah, it could result in a 21 thou clearance which may not be what you want but then you tolerance it for the fit you want, not tighten it an pretend that will do something

I can pick up a set of drawings tolernced to 20 millionths and make them not fit and yet be in tolerance

tighter tolerance is not the answer

CarbideBob · Nov 3, 2020

Fal Grunt said:
...... We are talking about small quantities. 30, 50, 100 piece jobs.

Confused a bit. You are making the mating piece to fit the supplied parts?
Once multiple POs over time are issued does one batch need to match another? If that true may as well tighten the tolerance.
If a one time never repeat order I can see your want but never to be made again?

Agree that a note on the quote, print and PO may do what you seem to be after but even all three notes leave a possible problem floating out there particularly if not in real good contact with the supplier.
.001 is too loose, .0002 is too expensive. What is needed in "consistency"? How was that number determined?
Is .0003 "batch consistency" all that much different than .0003 tolerance?
If one can hold a range on a run it is not that much more to move where the zero is running. If you can not make that move then you can not control that range. (admit some exceptions here)
Bob

mhajicek · Nov 3, 2020

gustafson said:
yeah, it could result in a 21 thou clearance which may not be what you want but then you tolerance it for the fit you want, not tighten it an pretend that will do something

If that 21 thou clearance is not acceptable, the answer is to tighten the tolerance. How do you get a guaranteed closer fit on interchangeable parts without tightening the tolerance?

Mike1974 · Nov 3, 2020

Depends alot too on end use of part(s). Our parts are made such they are an unit, nothing interchangeable between assemblies. Yes the unit needs to fit as a replacement, but not one half or one piece. If Fal can get assemblies and not worrying about supplying replacement parts, I don't see an issue with making them in groups for "best fit".

Fal Grunt · Nov 3, 2020

mhajicek said:
By tightening the tolerance. Period. You need to be able to send that print to any vendor at any point in the future and get good, interchangeable parts, otherwise it's a bad print.

If I need to make these parts again in the next 30 years I would be really surprised. I suspect I will cross that bridge when I get there. While I agree with you, tightening the tolerance to where I could send it to anyone and get good parts would likely make the part too expensive. If it is too expensive, then that means I won't sell any of them, which I guess solves the problem of needing to send them somewhere to be made in the future.

mhajicek said:
You say you don't work to a tolerance, but just nail the dimension; that tells me you need better measuring equipment. Everything is made to a tolerance, it's just a matter of how big.

Sure, but then I don't care what you have for measuring equipment, we can always add a zero? We can keep adding zero's until NIST doesn't have accurate enough measuring instruments. What would the point in that be?

mhajicek said:
And if parts made using their tolerance don't work, the tolerances were specified wrong.

ummm yes?

Mike1974 said:
Would it be doable to match them in pairs or groups? Years back we kept getting a no quote on two parts. I asked to see the print. 1 part was something like +.0002" / -0.0, the other -.0002" / +0.0

Not terrible in todays world, but for a few pieces, either more trouble than it's worth, or the parts were going to cost a small fortune. I convinced the engineer to leave one part marked up at say +/-.001", and mark the other one slip/press (one of those) to detail xx. Problem solved, got what we need for manufacturing, got a fair price, someone (hopefully) made a bit of profit from the job.

edit: I see you are making one of the mating details? Maybe have a print note "+/-.002 OK, but group into .0005" increments - ie .299 +/-.0003 <> .300 +/-.0003 etc "

That is an interesting idea.

Mtndew said:
One way to specify it, is to add a note to the dimension saying something like: "All parts alike within .001 for this feature"

Also an interesting idea I had not thought of.

CarbideBob said:
If a one time never repeat order I can see your want but never to be made again?

Agree that a note on the quote, print and PO may do what you seem to be after but even all three notes leave a possible problem floating out there particularly if not in real good contact with the supplier.
.001 is too loose, .0002 is too expensive. What is needed in "consistency"? How was that number determined?
Is .0003 "batch consistency" all that much different than .0003 tolerance?
If one can hold a range on a run it is not that much more to move where the zero is running. If you can not make that move then you can not control that range. (admit some exceptions here)
Bob

This is likely a one time order, and likely never to be made again.

To expand on what you are saying... let's use your .0003" as an example. You said exactly what I am thinking, I am just not sure how to make it on the print.

Currently the print is marked .299 plus or minus .001. Using your .0003" as an example, I really don't care if the parts are .299 plus or minus .0003" and I really don't care if the part is .2997 plus or minus .0003" and I really don't care if the part is .2983 plus or minus .0003".

I can make the mating part accordingly. What I struggle with is if the parts start at .2980 and migrates up to .3000 through the run. I don't want to be individually running each part, offsetting accordingly. Or needing to arrange them from smallest to largest, etc.

I think I have great communication with the supplier, he has been helpful, offered suggestions, told me I'm an idiot (in much nicer verbiage), said no, said yes, disagreed, agreed, and so far I *think* we have worked well together.

I also think he does good work and I am HONESTLY not worried about the parts, but after thinking about this print for awhile I really couldn't come up with a better way of marking the print. Plus or minus .001 is too much...

mhajicek · Nov 3, 2020

I don't see why it would cost any different. If I can make a run of parts all match within +/-.0003" I can make them in whatever specific +/-.0003" span you like.

Guest · Nov 3, 2020

The only issue I could see with tightening the tolerance is if the diameter's length was like a little over 4.5" long and was made from stock that is .0015" out of round ran through a guide bushing that wasn't brand new. Definitely would not want to try to hold +/-.0003 under those circumstances. The guide bushing issue is due to the fact they are now made to order from Southwick & Meister on a 5-6 week lead time.

mneuro · Nov 3, 2020

Can you make a "GO" and "NOGO" version of the mating part. Then you dont need to specify a tolerance at all on the print.

gustafson · Nov 3, 2020

mhajicek said:
If that 21 thou clearance is not acceptable, the answer is to tighten the tolerance. How do you get a guaranteed closer fit on interchangeable parts without tightening the tolerance?

You tolerance to what you need, you do not first increase the tolerance

again, rookie engineering

How many drawings have you seen with a close tolerance only to realize a bolt goes through it?

mmurray70 · Nov 3, 2020

If you cant accept a 0.002 range you need to tolerance tighter, no way around it. You pay for the size of the range, not the specific size the range is.

Did a job for a customer the other day with a +/-0.0015 tolerance on OD. Customer returned with the part and said it wont fit into mating part which was warped from welding. Asked me to take 1/16th off it lol. Big drop in tolerance there.

mhajicek · Nov 3, 2020

gustafson said:
You tolerance to what you need

Exactly. And when you need a tighter tolerance you need a tighter tolerance. If you want a closer fit, that's a tighter tolerance. If you want a guaranteed clearance, but not too much clearance, that's a tighter tolerance. It's not that difficult a concept.

Guest · Nov 3, 2020

One must consider if they want the K-mart blue light special or want to pay Nordstrom prices.

TeachMePlease · Nov 4, 2020

gustafson said:
I can pick up a set of drawings tolernced to 20 millionths and make them not fit and yet be in tolerance

Not if the designer knew what he was doing.

AARONT · Nov 4, 2020

If you give me a print and it's toleranced .299" +/-.001" but say they all need to be the same within .0005" to each other. I see it as .299" +/-.00025" and it's going to be an expensive part. There's no way around it. If you want them all to be within the same size range, then spec it as such.

Issue with specifying a dimension - How to?

Titanium

Diamond

Diamond

Diamond

Diamond

Diamond

Diamond

Diamond

Diamond

Diamond

Titanium

Diamond

Guest

Guest

Cast Iron

Diamond

Stainless

Diamond

Guest

Guest

Diamond

Stainless

Similar threads