How close is close enough?

[pete]

That big box of tricks Harold is one of the main reasons I'm so much in favor of those old machining books. Many seem to think forums and Youtube are an easy way to learn. They help a lot, but I think there more of a good addition to those books if your not learning on the job and doing this as a career. None of them can teach you everything, but they can make what's needed for that hands on much more understandable. I don't really think were getting very far off the original intent of this thread and whateg0's question was so I hope he doesn't mind the side topic.

That practice even when not needed does pay off. I could use the exact same cutting tools on my quite accurate 5" swing lathe and on my 10" swing and both had very different cutting characteristics and what the preset cut depth would be against what they actualy cut. .005" accuracy didn't show much change other than the maximum depth of cut each lathe would take. But at 10ths they sure did. My equipment is really lightweight against something like a Holbrook or DS & G lathe so mine will take a much more reliable "close enough" cut using HSS over carbide simply because it's sharper. Or at least I find it a lot easier to hit what I want with the HSS.

Maybe to get this back to being OT to whateg0's original question was about how close is close enough then something like a machine tools handwheel probably doesn't demand that much as long as it runs true to the eye and feel and doesn't wobble. If I was fitting a pair of precision bearings that were spaced widely apart or even close I'd be shooting for 10ths on concentricity. I've seen a few posts on other forums much like whateg0's asked about and what accuracy is really needed in a home shop. I've read people advising that anything more than a couple of thou was wasted in a home shop and that high accuracy just isn't needed. Really? There's people out there building experimental competitive flash steam and radio controlled racing engines both of which turn at extremely high rpms. It can be real common for the metal itself to fail at the speeds there trying. A couple of 10ths with some parts isn't acceptable at all for them. I think it all depends on what your building, the parts requirements and it's end use.

[Harold_V]

After 47 years as a mechanical design engineer, I have seen more than a few drawings that had insane tolerances.

Heh! Shades of my years of experience at Sperry Utah. There was an engineer (named Nero) on the missile project who was in charge of the flight electronics chassis. The castings, which reminded one of an auto V8 engine block, made of magnesium, had a multitude of tapped holes around the two tops, upon which covers were affixed with screws. The hole location for the holes were held to ± .005", but the holes in the covers had a clearance of .03" for the screws, so missing the dimension by even five thou, the parts would still function perfectly well. I witnessed those machined castings scrapped for as little error of screw locations as a thou out of tolerance. He was junk yard dog tough on dimensions, and showed no flexibility, even when parts were submitted to MRB (material review board, for those who don't know). That was a hard thing for me to assimilate, and is part of the reason I am so anal about tolerances, as I now understand that if you hope to have your work accepted by a third party, it must be right.

In my world, close enough was what worked at the most economical price.
Now, it is whatever I feel like at the time.

I recall conversations with those above me, and how such expensive parts were sacrificed for what appeared to be no good reason. Their response was that it was not a good gamble to use a $10,000 casting when it offered the remote opportunity of failure, losing a million dollar missile. If you've ever worked in a defense facility, it doesn't take long to learn that cost is rarely an object of concern. I also recall my supervisor telling me that he didn't care how long it took me to make any given part, but he wanted the part to be good. Emphasis was on quality, not quantity or cost.

In regards to choices made for those who enjoy machining at home, not for industry, I say do what pleases you. If you fit threads by using a nut as a gauge, so what? If it serves your intended purpose, go with what works. If, however, your work is expected to be compatible with the work of others, it's best to use accepted practice and work within reasonable parameters. Without that, there can be no mass production, nor can others be relied upon to make parts from prints. Bottom line---do as the print demands. If no thread relief is shown, don't make one.

H

[earlgo]

Harold, you and I worked in much different industries. Mine was commercial where the purchasing agents were dunned if they couldn't beat a supplier out of a few cents per part. Some purchasing agents ignored fastener specs because the supplier in a remote part of the world could give them the "same" bolt for 1/2 the price. (This savings was negated by replacing bolts that lost their heads inside a distillation column and had to be replaced after a VERY costly shutdown.)
How close is close enough is entirely dependent on the situation at hand, as you have pointed out many times. But in some situations, too close is more expensive.
MBA folks all quote the formula: Profits = sales - expenses. Engineering and manufacturing are expenses, and if they can be eliminated the profits would be optimal. The never ending battle...

--earlgo
(Gee, but I miss the .... oh never mind.)

[Mr Ron]

After 47 years as a mechanical design engineer, I have seen more than a few drawings that had insane tolerances.

--earlgo

This is also a bit OT, but when I was working in the shipbuilding business, Lockheed was winding down on aerospace projects and taking a swipe at shipbuilding. They had aeronautical engineers on their payroll they intended to put to work designing ships for the U.S. navy. Their first big design challenge was to build a bridge style monorail system that would span several decks, elevators and position for transfer to other ships, similarly equipped. Well the aeronautical engineers were accustomed to working to 3 and 4 decimal places. They applied this to the ship design by trying to maintain a center-to-center distance between bridge crane rails with a tolerance of ±.003" for it's entire length of travel, which was something like 600 feet. Due to the nature of ships, this was impossible to maintain resulting in much binding and general overall failure. Here was a case where precision was not needed. Those engineers have now switched their thinking from aerospace to shipbuilding standards. I just thought you might find this interesting.

[earlgo]

Mr. Ron, I also made sure my interns visited actual working machine shops, not the company shops, but actual shops that made parts for profit. I asked the owners to give us the $5.00 tour and explain some of the issues. Hopefully the interns remembered the earful on tolerances.
Your aerospace guys take the prize on insanity. +/_.003 over 600ft?!
Hopefully the upcoming engineers have some common sense.
--earlgo

[Mr Ron]

After 47 years as a mechanical design engineer, I have seen more than a few drawings that had insane tolerances.

Heh! Shades of my years of experience at Sperry Utah. There was an engineer (named Nero) on the missile project who was in charge of the flight electronics chassis. The castings, which reminded one of an auto V8 engine block, made of magnesium, had a multitude of tapped holes around the two tops, upon which covers were affixed with screws. The hole location for the holes were held to ± .005", but the holes in the covers had a clearance of .03" for the screws, so missing the dimension by even five thou, the parts would still function perfectly well. I witnessed those machined castings scrapped for as little error of screw locations as a thou out of tolerance. He was junk yard dog tough on dimensions, and showed no flexibility, even when parts were submitted to MRB (material review board, for those who don't know). That was a hard thing for me to assimilate, and is part of the reason I am so anal about tolerances, as I now understand that if you hope to have your work accepted by a third party, it must be right.

In my world, close enough was what worked at the most economical price.
Now, it is whatever I feel like at the time.

I recall conversations with those above me, and how such expensive parts were sacrificed for what appeared to be no good reason. Their response was that it was not a good gamble to use a $10,000 casting when it offered the remote opportunity of failure, losing a million dollar missile. If you've ever worked in a defense facility, it doesn't take long to learn that cost is rarely an object of concern. I also recall my supervisor telling me that he didn't care how long it took me to make any given part, but he wanted the part to be good. Emphasis was on quality, not quantity or cost.

In regards to choices made for those who enjoy machining at home, not for industry, I say do what pleases you. If you fit threads by using a nut as a gauge, so what? If it serves your intended purpose, go with what works. If, however, your work is expected to be compatible with the work of others, it's best to use accepted practice and work within reasonable parameters. Without that, there can be no mass production, nor can others be relied upon to make parts from prints. Bottom line---do as the print demands. If no thread relief is shown, don't make one.

H

Hi Harold, I worked in the shipbuilding industry for 50+ years and the concerns expressed in the aero-space industry were shared in the shipbuilding industry. The main concern was whether something that was made and put into service on a Navy ship, could have a catastrophic effect should it fail. This could mean loss of life, loss of a system or even the loss of the ship. Functions on a ship are broken down to systems and those systems are placed into classes as to degree of failure. They range from failure that is non-catastrophic to "failure is not an option". That raises the cost tremendously. I'm sure the same concerns and execution are present in all transportation systems.

[Harold_V]

Thanks for your comments, Mr. Ron.
While it may be difficult for folks to discern what makes a difference, I offer you, as an example of what works, and what doesn't, the Falkland Island crisis of several years ago. Weapons employed had, as I recall, a mixed success, with weapons made by France and other quality conscious countries out-performing weapons from other countries. There was a clear advantage.

It's true---high quality (tight tolerance precision accompanied by good craftsmanship) costs more, but anything that doesn't perform the purpose for which it is made is never a bargain, regardless of the price. Better one spends a little more and assures proper performance, assuming cost makes a difference. That doesn't apply to overpriced items, in particular, articles of clothing with high fashion labels, which often are identical to the same articles purchased from inexpensive stores, with a different label attached.

H

[John Hasler]

Harold writes:
> It's true---high quality (tight tolerance precision accompanied by good craftsmanship) costs more, but anything that doesn't perform the purpose for
> which it is made is never a bargain, regardless of the price. Better one spends a little more and assures proper performance, assuming cost makes a
> difference.

However boring a clearance hole for a 1/2" bolt to a tenth instead of simply drilling it with the correct bit does not improve performance one iota and wastes resources.

[Mr Ron]

Thanks for your comments, Mr. Ron.
While it may be difficult for folks to discern what makes a difference, I offer you, as an example of what works, and what doesn't, the Falkland Island crisis of several years ago. Weapons employed had, as I recall, a mixed success, with weapons made by France and other quality conscious countries out performing weapons from other countries. There was a clear advantage.

It's true---high quality (tight tolerance precision accompanied by good craftsmanship) costs more, but anything that doesn't perform the purpose for which it is made is never a bargain, regardless of the price. Better one spends a little more and assures proper performance, assuming cost makes a difference. That doesn't apply to overpriced items, in particular, articles of clothing with high fashion labels, which often are identical to the same articles purchased from inexpensive stores, with a different label attached.

H

Here is another precision example which was supposed to be true. Rolls Royce had a contract to use the GM hydromatic transmission in their RR. Instead of using the transmission as is, Rolls went and tore down the transmission and rebuilt it using closer tolerances. When it wouldn't shift properly, they went to GM and explained the problem. GM told them the tolerances used in the design was correct and by refining the tolerances, they introduced problems. RR went back to the transmission as built by GM and no problems. I'm not sure this was true, but it makes a good story.

[jcarmon]

First off, I'm still pretty new, so precision seems to be relative to who your talking to. I started making muzzle brakes and gas blocks because they didn't have to be nearly as precise as I thought, and If I screw up, Its only time. I can smelt my own aluminum again. But as Ive gotten more experienced, Ive learned my eye, and my hand are more accurate than my calipers. If my eye's and hand's are pleased with the finish, my caliper is going to be fine with the results as well. Its like backing a trailer, its all in your setup.

I guess I really just like shinny things!

[Harold_V]

However boring a clearance hole for a 1/2" bolt to a tenth instead of simply drilling it with the correct bit does not improve performance one iota and wastes resources.

Which is why I don't make such recommendations. As Kenny Rogers says, in his song, The Gambler, you have to know when to hold'em, and know when to fold'em. Some folks may not have that type of logic.

I've worked, in my many years in the shop, with a multitude of "machinists", a word I'll use loosely. Some of them would NEVER do quality work, although I'm sure they might by accident. They simply don't have the proper mindset to approach critical work. For that reason, I've always recommended one work to tight tolerances, even when they don't matter. It improves one's ability to do so tremendously, so when called upon to do so, it's not an overwhelming task. And so far as speed and production are concerned, if one learns to do things properly, speed comes, like it or not. Working properly simply leads to faster production.

If anyone here has ever run more than one item, and it involved more than one setup, they most likely would understand the value of holding dimensions. By doing so, alternate reference points can be utilized to promote easier and faster setups. Disregard dimensions and the probability of that working in an acceptable fashion is drastically diminished. That quality, in and of itself, is good reason to learn to work to dimensions, using tolerance as a saving grace, not a license to run blindly in the hopes that the work will be acceptable, in spite of using all tolerance allowed, from one end to the other. I'd fire a guy who thought like that. Yeah, I like to see things done properly, and well.

H

[Gary Armitstead]

I have been watching this thread for some time now. Great comments by all. John Hasler's comment about the clearance hole for a 1/2 bolt caught my eye .

I was a die sinker in the forging industry here in Southern California from 1965 until my "retirement" in 2003. In 1990, we hit a recession that just about brought the industry to it's knees and it still hasn't fully recovered. At that time many die shops were closing or cutting their die sinker staff by almost two-thirds. Many of us had to find jobs in similar industries (machinists). What a "rude awakening" that was! Die sinkers had always made very good "money" during their careers. But to now go out in the real world of machining make less than half our normal salary (we were considered "specialists" in what we did). I was asked to come to a fairly large CNC shop (mills, lathes and EDM work). This was my first experience with prints using ISO 9000, GD&T. Geometric Dimensioning and Tolerancing. Bottom line was this new system (at least for me) brings into context exactly what this thread asks, How close is close enough? In the "old days", there would be a feature dimension of say, .500 +.003 -.003 and that was it. With Geometric dimensions and tolerancing, there is NO guesswork as to how accurate that dimension was to be. The old way was always open to "discussion", pretty much like this thread.

Take a look at this PDF file for a great explanation of this system. I know Harold knows about this, I'm sure. Not used by home machine hobbyists, but it might help.
https://synthetica.eng.uci.edu/mechanic ... GDandT.pdf