Gunsmithing FYI: Tolerances vs Clearances

[gnochi]

In the interest of reading fewer posts by people who don’t know what they’re talking about, I’m going to lay down the law on one of the most misunderstood concepts in the precision rifle world.

Specifically, tolerances and clearances.

99% of the time, when someone who isn’t an engineer (so, me, @308pirate, and a few other people) says “tolerances” on this website - for example, complaining about loose tolerances - they are actually talking about “clearances”. They also have no idea what they’re talking about because loose clearances are not inherently a bad thing. Too-tight clearances are inherently a bad thing, so if you have to pick a direction to err...

Nominal clearance is how much gap there is supposed to be between two components. If the bolt rattles around like it’s dancing the Macarena, the actual clearances are loose. If a gnat’s testicle can cause the action to lock up, the actual clearances are tight. Clearances allow you to coat your actions with insert-coating-here, operate in dusty/muddy/icy environments, and can make it slightly easier to bind the action if you’re running the bolt weird. Clearance is critical to proper function, and the range of possible clearances for a product is something that is very much engineered for a specific purpose.

Tolerances are how much the size of something varies from part to part. Tight tolerances have less variation, loose tolerances have more variation. If every individual piece is hand fit, tolerances don’t matter because you start with an interference and remove material as needed. If you’re in a mass production environment, you need to make sure an arbitrary bolt can work with an arbitrary action - and if the combination of nominal clearance and achievable tolerance can result in an interference, you get to deal with horrifically expensive rework, possibly of an entire production run.

The problem is, cutting tolerances in half increases manufacturing cost by an order of magnitude. As such, when Remington or someone is making a bunch of cheap rifles, they increase clearances by an astronomical amount to make sure the ancient machines and tools that they never budgeted to maintain, and therefore are incapable of holding reasonable tolerances, deliver parts that can technically fit together. And hey, for someone who thinks minute of deer at 50 yards is all anyone needs, and who thinks cherry-picked 3-shot groups are enough to prove accuracy, that works just fine. That describes pretty much everyone who bought Remington rifles in the last couple decades, and the manufacturing is cheap, so the profit potential is phenomenal, and so only a conglomerate of assholes intentionally trying to suck out every penny of value into their own pockets could drive company into bankruptcy.

On the opposite end of the spectrum, you have companies like Defiance, with tight tolerances that allow them to have really tight clearances that are legendary for locking up in dusty environments - unless you tell them you’re going to be operating in dust, or cerakoting, or whatnot, and they are perfectly happy to loosen clearances up for you while keeping tight tolerances. They know essentially exactly how much clearance they’re going to have between the bolt and the action body, and know what the clearances need to be for pretty much every use case to feel perfect, but since the default buyer tends to only care about how tight the bolt is in the action body that’s what they default to.

Then you get to some of the other cases who hold very tight tolerances, so they can have variable clearances for different parts of the bolt cycle. Borden and ARC actions both have looser clearances when the bolt is traveling, and use various methods to remove the wiggle when the bolt is locked up.

Anyway, in summary:

• Nominal clearance: how much gap there is supposed to be
• Actual clearances: how much gap there actually is, which should be within tolerance bounds of the nominal clearance
• Tight clearance: less ability to handle dust/etc
• Loose clearance: people complain
• Tolerance: how much the gap is allowed to vary from part to part
• Tight tolerance: really expensive, but allows good control of clearances, and prefits, and interchangeability, and...
• Loose tolerance: really cheap

 

[gnochi]

Note that I didn’t go into “nominal” vs “actual”, because while it’s technically important, for lay purposes it’s not necessary and just adds confusion.

Edit: it was bugging me, so I made a few edits. If you want to learn everything about parts fitting together, read up on ASME Y14.5 Geometric Dimensions and Tolerances. You’ll quickly learn that holes are egg shaped, nothing is flat, nothing is parallel, everything is crooked, and we can engineer around it anyway.

 

[iceng]

There is a mum joke in this, but I won't say it cos I'll get banned.

TL;DR tolerance is how close something is to the pre-determined desired size. "Drill this hole at 10.00mm +0.02mm, -0.01mm". Your tolerance is 0.03mm.

"Make a pin to fit in that hole, we want 0.1mm clearance per side so it rattles down easily with grease and won't touch the sides.". You have the same tolerance on the pin diameter, relative to the hole size.

 

[alamo5000]

Clearance= there is supposed to be a .25 inch gap between X and Y

Tolerance= the part is supposed to be .375 inches thick, but we can live with .374 or .376

Tolerance stacking= when multiple parts of a machine are all within 'tolerances' but when you put those specific parts together it can cause a problem. Part A might be .001 off, Part B might be .001 off, and Part C might be .001 off so when you add it up it winds up being .003 off, which together is out of tolerances

Some guns (like an AK) are designed to have a lot more clearances than other types of guns but a lot of the time they have more clearances and a wider range of tolerances both.

"Slop" is more along the lines of tolerance stacking or just plain old crappy machine work (or whatever process is needed to manufacture the unit). A Taurus pistol is going to have a hell of a lot more slop in it than a Bul Armory race gun will.

 

[iceng]

Customer "my Glock rattles when I shake it".

Retail person "that a feature".

Designed clearance, held to tight tolerances

 

[Mr.BR]

In defense of Remington and co ...

Clearance and/or tolerance is not down to shoddy machines or tools, rifle actions are far from rocket science they also do not need tight clearances and/or tolerances to function or even to function well ,cost/benefit to the average user is where its at . As long as millions of folks want 400-600$ rifles that is what you have to settle for. Making 50$ bolts and receivers is an art in itself ,in every process seconds spent on one or other machining operation on each part, costs lots of $ down the bottom line, making cheap components like 30$ hammer forged barrels takes a huge investment compared to one required to make cut rifled barrels and if done well like you see in Sako,Tikka etc these Hammer forged barrels can shoot well enough.

Unlike 'custom' actions that are generally machined from a billet, practically all serial production parts are machined from castings or forgings to reduce machining time and the problem is not machining precision but warping inherent to heat treatment processes, but it's a conscious choice to avoid any unnecessary post heat treat machining in large scale serial production. Note that a 400-600$ MSRP rifle needs to cost roughly 100-125$ in raw manufacturing costs.

In the grand scheme of things action is also not the primary driver of rifle precision, these are more along the lines of the bullet,barrel,bedding, the shoddiest Mosin-Nagant action you can find will run 0.5moa with these three elements being of decent quality.

That doesn't mean 'custom' actions (although most are hardly custom as they are made in thousands each year)are a waste of money and they cost $$ for a reason.
I myself tend to fuck up my action designs at the stage where they need to be cheaper and easier to make. Either messing up with fancy tool steels, not quite as fancy as that used for the bolt and bolting insert in Theis's action but enough to add considerable cost making an action that will survive me by a couple of centuries or in case of my current project fucking up the design i terms of machine time on a one-piece receiver/chassis that requires considerable time machining on 5 axis machines again ballooning the cost. Gains from using fancy tool steel or wasting machining time will likely newer really show on a target ,but making just another uninspired Remington clone like most is not an option either.

 

[gnochi]

the problem is not machining precision but warping inherent to heat treatment processes

So I agree with most of your post. Potential warping (which is inconsistent for a multitude of reasons, not all of which Remington can control) is another one factor that needs to be accounted for in the “how much clearance do I need in the raceway” engineering efforts.

Let’s look at headspace, though. Depending on the cartridge somewhere between +/- 0.0015in to +/- 0.003in is acceptable. This is essentially a stack of three tolerances:

• Tolerance from the bolt face to the back of the bolt lug
• Tolerance from the action lug front to the action front
• Tolerance from the barrel shoulder to the headspace-diameter datum in the chamber

Because of warping and such there’s also a small amount of adjustment for relative angle, etc., but even splitting that tolerance into thirds is easily achievable at massive scale, and since it’s only one tolerance that needs to be tight on each part (and can be done in a single operation from a single fixture) it’s a small increase in overall cost - literally, maintain your machines and tools made since the late 1800s, and they can hit +/- 0.0005 on a nominal dimension less than an inch. And instead, they let the bolt+action part of headspace float at like +/- 0.005 and cherry-pick barrels.

The floating bolt head and barrel nut are one of the things Savage did very, very right, because at least they acknowledged that tolerance stackup exists and should be accounted for in their design process.

 

[Mr.BR]

End of the day that is up to the management, some manufacturers obviously make better end products than others not necesarily due to manufacturing limitations.

 

[gnochi]

End of the day that is up to the management, some manufacturers obviously make better end products than others not necesarily due to manufacturing limitations.

Ain’t that the truth!

 

[alamo5000]

End of the day that is up to the management, some manufacturers obviously make better end products than others not necesarily due to manufacturing limitations.

It's up to management in some ways but if the guy operating the machine doesn't care or gets sloppy if they don't have redundant QC it can get bad.

I've had that happen a couple of times. Fortunately only two or three times ever but at the end of the day the employees need to take pride in what they do but sometimes they don't.

If you have ever ran a business, even a bad secretary or somebody hired to answer the phone can screw things up. The guy packing and shipping boxes can mess up someone's day as well.

 

[YammyMonkey]

So, what you're saying is words actually mean things, and precision (and accuracy) in speech is as important as precision (and accuracy) in shooting?

 

[E. Bryant]

Excellent post. Next, please explain to everyone that "QC" doesn't just mean "someone measures the parts a bunch of times".

 

[THEIS]

Hi,

Tune in next week for explanation of process control vs quality control, lol

Sincerely,
Theis

 

[gnochi]

Excellent post. Next, please explain to everyone that "QC" doesn't just mean "someone measures the parts a bunch of times".

I’m not the right person to speak on that. However...

Hi,

Tune in next week for explanation of process control vs quality control, lol

Sincerely,
Theis

Yes please!

 

[Faronth]

This thread is giving my flashbacks to my first career out of college. 11 years give or take spent in the world of QC. I let my CQE credentials lapse many years ago but there is a ton of good information in this thread.
I traveled and trained people in SPC, DOE, Gage R&R, ISO9000 in automotive, steel mnfg, precision tool mnfg, food processing, chemical mnfg, banking, railroads, and hospitals you name it. I am not saying I have seen it all but I have seen a hell of a lot. With all the stories I have been told of honest line level screw ups and management with no commitment to process control hiding bad data I often wondered why my TV did not just blow up when I got home and turned it on.
On the other hand it was always nice to find the ones who were really committed to producing the best product for their customers. They are certainly out there and worthy of our support.
As my former boss inscribed in one of his books he signed for me, "Quality and honesty are the same thing". Words to ponder.

 

[brianf]

This thread is giving my flashbacks to my first career out of college. 11 years give or take spent in the world of QC. I let my CQE credentials lapse many years ago but there is a ton of good information in this thread.
I traveled and trained people in SPC, DOE, Gage R&R, ISO9000 in automotive, steel mnfg, precision tool mnfg, food processing, chemical mnfg, banking, railroads, and hospitals you name it. I am not saying I have seen it all but I have seen a hell of a lot. With all the stories I have been told of honest line level screw ups and management with no commitment to process control hiding bad data I often wondered why my TV did not just blow up when I got home and turned it on.
On the other hand it was always nice to find the ones who were really committed to producing the best product for their customers. They are certainly out there and worthy of our support.
As my former boss inscribed in one of his books he signed for me, "Quality and honesty are the same thing". Words to ponder.

actually quality and honesty aren’t the same in manufacturing, you would hope they would be though.

Quality while manufacturing is adherence to product specifications created by the QCU/QCG.

Those specifications are a mixture of regulatory and manufacturer mandates.

the final signed off specs are then sent to the customer when marketing the product etc.

IF LEGAL, the customer can require specs that have a enormously wide range of test results or measurement criteria...making what you/I would call crappy product.

yet those products still pass all of the product specific criteria. So, it can pass with flying colors but be a piece of crap.

Farther up stream you get in to
process validation
Product validation

which cost more initially but can actually be used to reduce the QA scheduled checks on the floor during production.

 

[sirhrmechanic]

In the interest of reading fewer posts by people who don’t know what they’re talking about, I’m going to lay down the law on one of the most misunderstood concepts in the precision rifle world.

Specifically, tolerances and clearances.

99% of the time, when someone who isn’t an engineer (so, me, @308pirate, and a few other people) says “tolerances” on this website - for example, complaining about loose tolerances - they are actually talking about “clearances”. They also have no idea what they’re talking about because loose clearances are not inherently a bad thing. Too-tight clearances are inherently a bad thing, so if you have to pick a direction to err...

Nominal clearance is how much gap there is supposed to be between two components. If the bolt rattles around like it’s dancing the Macarena, the actual clearances are loose. If a gnat’s testicle can cause the action to lock up, the actual clearances are tight. Clearances allow you to coat your actions with insert-coating-here, operate in dusty/muddy/icy environments, and can make it slightly easier to bind the action if you’re running the bolt weird. Clearance is critical to proper function, and the range of possible clearances for a product is something that is very much engineered for a specific purpose.

Tolerances are how much the size of something varies from part to part. Tight tolerances have less variation, loose tolerances have more variation. If every individual piece is hand fit, tolerances don’t matter because you start with an interference and remove material as needed. If you’re in a mass production environment, you need to make sure an arbitrary bolt can work with an arbitrary action - and if the combination of nominal clearance and achievable tolerance can result in an interference, you get to deal with horrifically expensive rework, possibly of an entire production run.

The problem is, cutting tolerances in half increases manufacturing cost by an order of magnitude. As such, when Remington or someone is making a bunch of cheap rifles, they increase clearances by an astronomical amount to make sure the ancient machines and tools that they never budgeted to maintain, and therefore are incapable of holding reasonable tolerances, deliver parts that can technically fit together. And hey, for someone who thinks minute of deer at 50 yards is all anyone needs, and who thinks cherry-picked 3-shot groups are enough to prove accuracy, that works just fine. That describes pretty much everyone who bought Remington rifles in the last couple decades, and the manufacturing is cheap, so the profit potential is phenomenal, and so only a conglomerate of assholes intentionally trying to suck out every penny of value into their own pockets could drive company into bankruptcy.

On the opposite end of the spectrum, you have companies like Defiance, with tight tolerances that allow them to have really tight clearances that are legendary for locking up in dusty environments - unless you tell them you’re going to be operating in dust, or cerakoting, or whatnot, and they are perfectly happy to loosen clearances up for you while keeping tight tolerances. They know essentially exactly how much clearance they’re going to have between the bolt and the action body, and know what the clearances need to be for pretty much every use case to feel perfect, but since the default buyer tends to only care about how tight the bolt is in the action body that’s what they default to.

Then you get to some of the other cases who hold very tight tolerances, so they can have variable clearances for different parts of the bolt cycle. Borden and ARC actions both have looser clearances when the bolt is traveling, and use various methods to remove the wiggle when the bolt is locked up.

Anyway, in summary:

• Nominal clearance: how much gap there is supposed to be
• Actual clearances: how much gap there actually is, which should be within tolerance bounds of the nominal clearance
• Tight clearance: less ability to handle dust/etc
• Loose clearance: people complain
• Tolerance: how much the gap is allowed to vary from part to part
• Tight tolerance: really expensive, but allows good control of clearances, and prefits, and interchangeability, and...
• Loose tolerance: really cheap

Very well-written. I may borrow this for some car folks applications!!

Sirhr

 

[sirhrmechanic]

Couple of fun 'tolerance vs. clearance' stories...

One is that I was on the board of a specialty machine/engineering shop. They did a lot of aerospace parts, worked for NOAA, NASA, Darpa, iRobot... all of them. So some really incredible parts with very small tolerances came in. But they also did a lot more run-of-the mill work...

The 'fun' part of tolerance came in with the seats of CAD software some clients used... where the tolerance on plans was default-set in their shops to .0005" Half a thou. The engineers designing the part can change that tolerance with a check box. And the software applies that tolerance across the part... though the engineer can change specific places (like hole spacing) for tolerance, etc.

The issue was that some of the engineers/designers never changed the default tolerance. So things came in with a default tolerance of a half-thousandth. Parts that in no way needed that kind of tolerance. Battery boxes, housings... things that would have worked just fine with, an external tolerance of .002 or even .005". So the protocol was to call the company/engineer and say "Hey, you have these tolerances at half a thou and if we run them to that tolerance, it will cost X. But based on the prints, your part will work fine if this, this, this and this are done to to tolerance .003"... it will cost X/5.... 1 fifth as much." Most of the time, the engineer/designer would say... Oh, yeah... good catch. Sure let's change this, this this this... and they'd sort out the tolerances in about 10 minutes, they'd send a new print and then the parts were made. Part of the motto of "We don't always get you the part you asked for, but you always get the part you need..."

Every now and then, however, someone would be "No... I want what I want and MY print specified 1/2 thou." And that would be that. The most ludicrous example was a "sheetmetal" aluminum battery box for a prototype vehicle. That was specified to a half-thou. And the very junior engineer would not budge and it was 'his' design. So... numerous e-mails later... we kept the paper trail. Made a "sheetmetal-like" battery box that met a half thou tolerance... (It WAS beautiful.... machined out of a billet! It was a perfect battery box for the Space Shuttle. But it wasn't. It was for an MRAP-type truck. They could have bought one from Summit racing for $35. I think it ended up being a $25,000 battery box. Which we could have made for $350. Or bought from Summit for $35. If the engineer had not gotten his ego up. The company blew a cork. We showed the paper trail, including several notes from us saying "here is what we recommend for this application... This does not need to be toleranced this way.... How about a... You sure you don't want to go off the shelf?" And the company paid the bill after much consternation. We never worked with that engineer again, though. Always wondered if they fired him.

The second is a great military history story... which is about the famous Rolls-Royce Merlin engine. Developed in the late 1930's, it powered a huge number of British and American aircraft... even things like PT boats... tanks... It was a war-winning technology. In 1940, with Nazi invasion possible, Rolls-Royce wanted to duplicate the production capability in safe North America. Plans were sent to Canada and then down to the US where Ford and Packard were to take up production. So they took the prints and made a bunch of prototype engines and nothing... I mean nothing... fit. They looked at the prints... the parts were made to the prints. But they could not for the life of them assemble an engine. So Rolls-Royce sent a man named Alec Harvey-Bailey over to the U.S. Harvey-Bailey was THE production guy at Rolls-Royce and was a production genius within the British system. And he got to the U.S. and Packard and they showed him their print room... and it was lovely. And their machine shops... and he was duly impressed. And their assembly areas which he thought were lovely. And then he asked "Right, this all fine. Please show me your fitting department?"

I can only imagine the response... "What in the hell is a FItting Department?" I'd have loved to be a fly on the wall.

You see, the way the Americans worked, they made parts to extremely tight tolerances. And to very good clearances. So that they could be assembled in line-fashion by, essentially, unskilled assembly workers. The tolerances of American parts made them 100 percent interchangeable and able to assemble with little or no final finishing steps. They went from machine to assembly instantly. The British system made all parts oversize (or undersized on holes) and relied on rows and rows of "Fitters" to take parts from final machined to fitted assembly. Usually these were older workers, past their ability to work the heavier jobs. In wartime, it was women. And they would take a group of parts and 'fit' them establishing tolerances with reamers, stones, files and other tools until the "Mechanism" was perfect. Then the mechanism would be fitted to the next assembly (by a fitter). And the car or aero engine final assembled with individually-fitted assemblies was perfect. But would not interchange with 'other' engines or cars. If an aero engine needed rebuild, it had to go back to the factory. Which they did, running trucks and trains through the night delivering fresh engines to airfields for fitting to airframes.

The Americans simply made parts that fit. Packard re-drew and re-toleranced all the plans. They set up the machines to make exact parts that were 100 percent interchangeable. And to this day, a Packard Merlin and a Rolls-Royce Merlin are NOT interchangeable. Ask the guys at the Reno Air Races!

Anyway... it was a small story in a big war. But one that explains the difference between an AR-15 and a Purdey really well.

And one must also remember that in the British "Craft" system, a huge part of their labor force was PAYING between 20 and 100 pounds a year for the privilege of working as an apprentice and learning a valuable trade. So between ages 15 and about 22 or 23... you weren't only free labor... you paid to work there! So who needed to do silly things like 'Design for Assembly?" Use 1 bolt when 20 would do? And eliminating those jobs would have destroyed the England (and arguably European) economies. Plus, in England's resource-poor island economy, materials were expensive and labor cheap. Machine tools were expensive, so were jigs used.... hard to justify with craft production. So designing complexity while making best use of expensive materials and minimum manufacturing infrastructure was par for the course. And last, the British designed things to last 100 years. Whether a toaster or a car. And goods were for the middle-class and upper class. And that was a small population. So volume was not in the cards, quality was.

America on the other hand developed a citizen-laborer system, not a craft system with apprenticeships and journeymen and masters. Men's labor was valuable (childrens, not so much). But progressive legislation and unions and a booming job market in Industrial America ensured that even assembly workers made 'good' money (The ford $5 day). But materials were cheap and who needed to be sparing? Use a sledge-hammer, not an ice pick. Over-design. And then make things in mass with specialty tools and patterns and jigs for mass production. And design for obsolescence ensuring that your customers need a new one every three years... keep those assembly lines moving. And paying average workers enough that they can afford to buy the products they make. Great big cycle!

So... history lesson for the morning. Culture plays a role in technology just as much as... technology plays a role in technology.

And tolerance and clearances matter! Big time! As is knowing the terminology.

I highly recommend this book to anyone who 'really' wants to understand the subject: https://www.amazon.com/American-System-Mass-Production-1800-1932/dp/080183158X

Sorry for the Amazon link... I'd normally find another source, but they seem to have put everyone else out of business.

Cheers,

Sirhr

 

[THEIS]

Hi,

Brief tolerance example picture....

You see the specifications with allowed +/-

For example my below bolt guide/bolt stop area in the receiver frame does not need to be held to 1thou across each and every receiver, 2thou will suffice

The "clearances" are built in either way.

Sincerely,
Theis

 

[gnochi]

Hi,

Brief tolerance example picture....

You see the specifications with allowed +/-

For example my below bolt guide/bolt stop area in the receiver frame does not need to be held to 1thou across each and every receiver, 2thou will suffice

The "clearances" are built in either way.

View attachment 7569081

Sincerely,
Theis

I’m guessing that the other zero point is the front face of the action? Ordinate dimensions just make life so nice. I’m surprised you’re doing toleranced dimensions instead of GD&T though.

For the uninitiated: toleranced dimensions are concerned with size, GD&T/Y14.5 is concerned with fitment and takes into account warp/nonperpendicularities/etc. In this particular case a feature of size is perfectly appropriate, but I would have done it as follows:

EDIT: Image attached

• A/B/C square flags are “Datums”, which tell us how to fixture the part for measuring. They should also be features that we use when connecting the part to other major things!
• Dimensions with boxes are nominal values for where things should be.
• Since the B datum (recoil lug), not the front face, is how we mount the surface, I tossed in a quarter-inch offset. This isn’t great practice but if we care about distances from the front face it’s acceptable to have an offset zero.
• The parallelogram and upside down T are “flatness” and “perpendicularity” respectively. I should have used parallelism instead of flatness, since flatness plus datum is obsolete, but I also used pen. The alternative is to use the semicircle “surface profile” which gives us imaginary boundaries that we shouldn’t go past. The boundaries are offset from the nominal surface by half of the value, and when we’re fixturing the part in the measurement or gauged fixture we do so aligning with Datums in the orders indicated in the boxes.
• I sketched in the section view so I could clarity that what we actually care about for fitting the bolt stop is the profile of the surfaces in question, but only where it’s cut out (ie between the top and bottom edges).

This reminds me: “A drawing clarifies the worst acceptable part in relation to a platonic ideal.”

 

[THEIS]

Hi,

I don't deal with warping and such because everything is machined post heat treatment....everything!!

Sincerely,
Theis

 

[gnochi]

Hi,

I don't deal with warping and such because everything is machined post heat treatment....everything!!

Sincerely,
Theis

Don’t go measuring microns now, your blood pressure will skyrocket...

But yeah, GD&T acknowledges the reality that even with perfectly stress relieved material, and the best equipment in the world, the ways aren’t perfectly aligned and fixturing and tooling aren’t perfectly rigid. As such, let’s focus on getting everything to fit together, because a parallelogram can be in spec for features of size, but defeat the purpose of wanting a square.

Old school machinists tend to hate working with GD&T because it’s harder to deal with on manual machines than “go this far, ish”.

 

[THEIS]

Hi,

You past my level, lol...this is one of those things that definitely fall into the category of "I have people for that", lol....
I know where the power button is at and where the assembly areas are at, lolololol....everything in the middle; I leave to smarter guys than myself in that field.

Sincerely,
Theis

 

[gnochi]

Couple of fun 'tolerance vs. clearance' stories...

One is that I was on the board of a specialty machine/engineering shop. They did a lot of aerospace parts, worked for NOAA, NASA, Darpa, iRobot... all of them. So some really incredible parts with very small tolerances came in. But they also did a lot more run-of-the mill work...

The 'fun' part of tolerance came in with the seats of CAD software some clients used... where the tolerance on plans was default-set in their shops to .0005" Half a thou. The engineers designing the part can change that tolerance with a check box. And the software applies that tolerance across the part... though the engineer can change specific places (like hole spacing) for tolerance, etc.

The issue was that some of the engineers/designers never changed the default tolerance. So things came in with a default tolerance of a half-thousandth. Parts that in no way needed that kind of tolerance. Battery boxes, housings... things that would have worked just fine with, an external tolerance of .002 or even .005". So the protocol was to call the company/engineer and say "Hey, you have these tolerances at half a thou and if we run them to that tolerance, it will cost X. But based on the prints, your part will work fine if this, this, this and this are done to to tolerance .003"... it will cost X/5.... 1 fifth as much." Most of the time, the engineer/designer would say... Oh, yeah... good catch. Sure let's change this, this this this... and they'd sort out the tolerances in about 10 minutes, they'd send a new print and then the parts were made. Part of the motto of "We don't always get you the part you asked for, but you always get the part you need..."

Every now and then, however, someone would be "No... I want what I want and MY print specified 1/2 thou." And that would be that. The most ludicrous example was a "sheetmetal" aluminum battery box for a prototype vehicle. That was specified to a half-thou. And the very junior engineer would not budge and it was 'his' design. So... numerous e-mails later... we kept the paper trail. Made a "sheetmetal-like" battery box that met a half thou tolerance... (It WAS beautiful.... machined out of a billet! It was a perfect battery box for the Space Shuttle. But it wasn't. It was for an MRAP-type truck. They could have bought one from Summit racing for $35. I think it ended up being a $25,000 battery box. Which we could have made for $350. Or bought from Summit for $35. If the engineer had not gotten his ego up. The company blew a cork. We showed the paper trail, including several notes from us saying "here is what we recommend for this application... This does not need to be toleranced this way.... How about a... You sure you don't want to go off the shelf?" And the company paid the bill after much consternation. We never worked with that engineer again, though. Always wondered if they fired him.

The second is a great military history story... which is about the famous Rolls-Royce Merlin engine. Developed in the late 1930's, it powered a huge number of British and American aircraft... even things like PT boats... tanks... It was a war-winning technology. In 1940, with Nazi invasion possible, Rolls-Royce wanted to duplicate the production capability in safe North America. Plans were sent to Canada and then down to the US where Ford and Packard were to take up production. So they took the prints and made a bunch of prototype engines and nothing... I mean nothing... fit. They looked at the prints... the parts were made to the prints. But they could not for the life of them assemble an engine. So Rolls-Royce sent a man named Alec Harvey-Bailey over to the U.S. Harvey-Bailey was THE production guy at Rolls-Royce and was a production genius within the British system. And he got to the U.S. and Packard and they showed him their print room... and it was lovely. And their machine shops... and he was duly impressed. And their assembly areas which he thought were lovely. And then he asked "Right, this all fine. Please show me your fitting department?"

I can only imagine the response... "What in the hell is a FItting Department?" I'd have loved to be a fly on the wall.

You see, the way the Americans worked, they made parts to extremely tight tolerances. And to very good clearances. So that they could be assembled in line-fashion by, essentially, unskilled assembly workers. The tolerances of American parts made them 100 percent interchangeable and able to assemble with little or no final finishing steps. They went from machine to assembly instantly. The British system made all parts oversize (or undersized on holes) and relied on rows and rows of "Fitters" to take parts from final machined to fitted assembly. Usually these were older workers, past their ability to work the heavier jobs. In wartime, it was women. And they would take a group of parts and 'fit' them establishing tolerances with reamers, stones, files and other tools until the "Mechanism" was perfect. Then the mechanism would be fitted to the next assembly (by a fitter). And the car or aero engine final assembled with individually-fitted assemblies was perfect. But would not interchange with 'other' engines or cars. If an aero engine needed rebuild, it had to go back to the factory. Which they did, running trucks and trains through the night delivering fresh engines to airfields for fitting to airframes.

The Americans simply made parts that fit. Packard re-drew and re-toleranced all the plans. They set up the machines to make exact parts that were 100 percent interchangeable. And to this day, a Packard Merlin and a Rolls-Royce Merlin are NOT interchangeable. Ask the guys at the Reno Air Races!

Anyway... it was a small story in a big war. But one that explains the difference between an AR-15 and a Purdey really well.

And one must also remember that in the British "Craft" system, a huge part of their labor force was PAYING between 20 and 100 pounds a year for the privilege of working as an apprentice and learning a valuable trade. So between ages 15 and about 22 or 23... you weren't only free labor... you paid to work there! So who needed to do silly things like 'Design for Assembly?" Use 1 bolt when 20 would do? And eliminating those jobs would have destroyed the England (and arguably European) economies. Plus, in England's resource-poor island economy, materials were expensive and labor cheap. Machine tools were expensive, so were jigs used.... hard to justify with craft production. So designing complexity while making best use of expensive materials and minimum manufacturing infrastructure was par for the course. And last, the British designed things to last 100 years. Whether a toaster or a car. And goods were for the middle-class and upper class. And that was a small population. So volume was not in the cards, quality was.

America on the other hand developed a citizen-laborer system, not a craft system with apprenticeships and journeymen and masters. Men's labor was valuable (childrens, not so much). But progressive legislation and unions and a booming job market in Industrial America ensured that even assembly workers made 'good' money (The ford $5 day). But materials were cheap and who needed to be sparing? Use a sledge-hammer, not an ice pick. Over-design. And then make things in mass with specialty tools and patterns and jigs for mass production. And design for obsolescence ensuring that your customers need a new one every three years... keep those assembly lines moving. And paying average workers enough that they can afford to buy the products they make. Great big cycle!

So... history lesson for the morning. Culture plays a role in technology just as much as... technology plays a role in technology.

And tolerance and clearances matter! Big time! As is knowing the terminology.

I highly recommend this book to anyone who 'really' wants to understand the subject: https://www.amazon.com/American-System-Mass-Production-1800-1932/dp/080183158X

Sorry for the Amazon link... I'd normally find another source, but they seem to have put everyone else out of business.

Cheers,

Sirhr

Thanks for the reading material!

I’m actually dealing with something similar right now with my job; the structures engineer has an aerospace background and loves match-drilled holes for mounting brackets. I work with 1 ton batteries that flex the frame until they’re fully bolted in, and want to be able to put bolts through holes that have moved by a fraction of an inch.

 

[Mr.BR]

Hi,

I don't deal with warping and such because everything is machined post heat treatment....everything!!

Sincerely,
Theis

Machining post-heat treatment does not prevent warping! High end materials can and do minimize it.

Machining major operations pre and the minor post-heat threat is the norm and seems to be used in your bolt as well judging from the picts.The reciever that is not realy load bearing as you are using a bolting collar can be machined from a Q&T barstock of 1.7225/ 42CrMo4 , but that does not mean it will not warp. Warp is just not that consequential as clearance between bolts and receivers are large enough and mayor warping hotspot like around bolt handle raceway (when present),the tangs in receivers tend to open up away from the centerline so do not cut into the clearance between the bolt and reciever.

*The Hoplite arms action concept is quite interesting,by the way any specfific reason for an ''non guided'' firing pin tip and firing pin stop directly on the cocking piece? Is that driven by the requirement for higher pressures & a more resilient short firing pin tip?

 

[Ludasmith]

Machining post-heat treatment does not prevent warping!

A good engineer knows when certain things matter and when they don't and leaves the things out that don't. Anything that doesn't matter is waste and waste adds no value.

They seem to be making parts just fine, no point in going deeper for no reason.

 

[THEIS]

Machining post-heat treatment does not prevent warping! High end materials can and do minimize it.

Machining major operations pre and the minor post-heat threat is the norm and seems to be used in your bolt as well judging from the picts.The reciever that is not realy load bearing as you are using a bolting collar can be machined from a Q&T barstock of 1.7225/ , but that does not mean it will not warp. Warp is just not that consequential as clearance between bolts and receivers are large enough and mayor warping hotspot like around bolt handle raceway in receivers tend to open up from the centerline so do not cut into the clearance between the bolt and reciever.

Hi,

Only on the internet can you find a guy to tell you at what stage you machine parts, that A..he has neve seen, B..he is in completely different country than the machining.... lololol

I know for a FACT, since I was there, lol...nothing of the bolt is machined pre heat treat.

Speaking of that bolt you referring to....Have you ever seen Aermet 100 being machined, lol?

Sincerely,
Theis

 

[Mr.BR]

Hi,

Only on the internet can you find a guy to tell you at what stage you machine parts, that A..he has neve seen, B..he is in completely different country than the machining.... lololol

I know for a FACT, since I was there, lol...nothing of the bolt is machined pre heat treat.

Speaking of that bolt you referring to....Have you ever seen Aermet 100 being machined, lol?

Sincerely,
Theis

On the picts you posted in Hoplite Arms thread with bolts and parts in different stages of finish it looks like the bolts are turned and machined pre heat treat and it seems no a closing cam , extractor cut and bolt handle cut pre heatreat. Unless that are some preproduction parts in the pictures.

Aermet 100 is a fancy material yes

Hoplite Arms

Hi, So this is one of the great benefits of having a Mastercam guru running the programming and machining... He can make improvements on the fly; that not only look better cosmetically but improves production efficiency at the same time. Seconds lead to minutes, minutes lead to hours and...

www.snipershide.com

 

[alamo5000]

Hi,

I don't deal with warping and such because everything is machined post heat treatment....everything!!

Sincerely,
Theis

I have a question for you since you seem to know a lot more than us average folks.

Anybody else that knows feel free to answer as well.

I am far from an expert but are "heat treating" and "stress relieving" two different things?

They both use heat but I am trying to wrap my head around the difference in processing.

 

[B-P-UU]

Next week: Runout vs. Concentricity

This is another area where keyboard commandos are, well... wrong.

 

[2aBaC̶a̶]

Pull my finger...

 

[lariat]

Couple of fun 'tolerance vs. clearance' stories...

One is that I was on the board of a specialty machine/engineering shop. They did a lot of aerospace parts, worked for NOAA, NASA, Darpa, iRobot... all of them. So some really incredible parts with very small tolerances came in. But they also did a lot more run-of-the mill work...

The 'fun' part of tolerance came in with the seats of CAD software some clients used... where the tolerance on plans was default-set in their shops to .0005" Half a thou. The engineers designing the part can change that tolerance with a check box. And the software applies that tolerance across the part... though the engineer can change specific places (like hole spacing) for tolerance, etc.

The issue was that some of the engineers/designers never changed the default tolerance. So things came in with a default tolerance of a half-thousandth. Parts that in no way needed that kind of tolerance. Battery boxes, housings... things that would have worked just fine with, an external tolerance of .002 or even .005". So the protocol was to call the company/engineer and say "Hey, you have these tolerances at half a thou and if we run them to that tolerance, it will cost X. But based on the prints, your part will work fine if this, this, this and this are done to to tolerance .003"... it will cost X/5.... 1 fifth as much." Most of the time, the engineer/designer would say... Oh, yeah... good catch. Sure let's change this, this this this... and they'd sort out the tolerances in about 10 minutes, they'd send a new print and then the parts were made. Part of the motto of "We don't always get you the part you asked for, but you always get the part you need..."

Every now and then, however, someone would be "No... I want what I want and MY print specified 1/2 thou." And that would be that. The most ludicrous example was a "sheetmetal" aluminum battery box for a prototype vehicle. That was specified to a half-thou. And the very junior engineer would not budge and it was 'his' design. So... numerous e-mails later... we kept the paper trail. Made a "sheetmetal-like" battery box that met a half thou tolerance... (It WAS beautiful.... machined out of a billet! It was a perfect battery box for the Space Shuttle. But it wasn't. It was for an MRAP-type truck. They could have bought one from Summit racing for $35. I think it ended up being a $25,000 battery box. Which we could have made for $350. Or bought from Summit for $35. If the engineer had not gotten his ego up. The company blew a cork. We showed the paper trail, including several notes from us saying "here is what we recommend for this application... This does not need to be toleranced this way.... How about a... You sure you don't want to go off the shelf?" And the company paid the bill after much consternation. We never worked with that engineer again, though. Always wondered if they fired him.

The second is a great military history story... which is about the famous Rolls-Royce Merlin engine. Developed in the late 1930's, it powered a huge number of British and American aircraft... even things like PT boats... tanks... It was a war-winning technology. In 1940, with Nazi invasion possible, Rolls-Royce wanted to duplicate the production capability in safe North America. Plans were sent to Canada and then down to the US where Ford and Packard were to take up production. So they took the prints and made a bunch of prototype engines and nothing... I mean nothing... fit. They looked at the prints... the parts were made to the prints. But they could not for the life of them assemble an engine. So Rolls-Royce sent a man named Alec Harvey-Bailey over to the U.S. Harvey-Bailey was THE production guy at Rolls-Royce and was a production genius within the British system. And he got to the U.S. and Packard and they showed him their print room... and it was lovely. And their machine shops... and he was duly impressed. And their assembly areas which he thought were lovely. And then he asked "Right, this all fine. Please show me your fitting department?"

I can only imagine the response... "What in the hell is a FItting Department?" I'd have loved to be a fly on the wall.

You see, the way the Americans worked, they made parts to extremely tight tolerances. And to very good clearances. So that they could be assembled in line-fashion by, essentially, unskilled assembly workers. The tolerances of American parts made them 100 percent interchangeable and able to assemble with little or no final finishing steps. They went from machine to assembly instantly. The British system made all parts oversize (or undersized on holes) and relied on rows and rows of "Fitters" to take parts from final machined to fitted assembly. Usually these were older workers, past their ability to work the heavier jobs. In wartime, it was women. And they would take a group of parts and 'fit' them establishing tolerances with reamers, stones, files and other tools until the "Mechanism" was perfect. Then the mechanism would be fitted to the next assembly (by a fitter). And the car or aero engine final assembled with individually-fitted assemblies was perfect. But would not interchange with 'other' engines or cars. If an aero engine needed rebuild, it had to go back to the factory. Which they did, running trucks and trains through the night delivering fresh engines to airfields for fitting to airframes.

The Americans simply made parts that fit. Packard re-drew and re-toleranced all the plans. They set up the machines to make exact parts that were 100 percent interchangeable. And to this day, a Packard Merlin and a Rolls-Royce Merlin are NOT interchangeable. Ask the guys at the Reno Air Races!

Anyway... it was a small story in a big war. But one that explains the difference between an AR-15 and a Purdey really well.

And one must also remember that in the British "Craft" system, a huge part of their labor force was PAYING between 20 and 100 pounds a year for the privilege of working as an apprentice and learning a valuable trade. So between ages 15 and about 22 or 23... you weren't only free labor... you paid to work there! So who needed to do silly things like 'Design for Assembly?" Use 1 bolt when 20 would do? And eliminating those jobs would have destroyed the England (and arguably European) economies. Plus, in England's resource-poor island economy, materials were expensive and labor cheap. Machine tools were expensive, so were jigs used.... hard to justify with craft production. So designing complexity while making best use of expensive materials and minimum manufacturing infrastructure was par for the course. And last, the British designed things to last 100 years. Whether a toaster or a car. And goods were for the middle-class and upper class. And that was a small population. So volume was not in the cards, quality was.

America on the other hand developed a citizen-laborer system, not a craft system with apprenticeships and journeymen and masters. Men's labor was valuable (childrens, not so much). But progressive legislation and unions and a booming job market in Industrial America ensured that even assembly workers made 'good' money (The ford $5 day). But materials were cheap and who needed to be sparing? Use a sledge-hammer, not an ice pick. Over-design. And then make things in mass with specialty tools and patterns and jigs for mass production. And design for obsolescence ensuring that your customers need a new one every three years... keep those assembly lines moving. And paying average workers enough that they can afford to buy the products they make. Great big cycle!

So... history lesson for the morning. Culture plays a role in technology just as much as... technology plays a role in technology.

And tolerance and clearances matter! Big time! As is knowing the terminology.

I highly recommend this book to anyone who 'really' wants to understand the subject: https://www.amazon.com/American-System-Mass-Production-1800-1932/dp/080183158X

Sorry for the Amazon link... I'd normally find another source, but they seem to have put everyone else out of business.

Cheers,

Sirhr

How’s this: I got drawings into the shop with a true position for holes of ZERO tolerance off the datum. Did I kick that back? You bet your ass I did. Even for a component that was meant for space and re-entry that was a bit much.

ETA: fucking iPhone doesn’t know what a datum is. Had to change it from “dating”.

 

[ghostrider272]

Alamo, heat treating and stress relieving are processes that are used for different results. Both can use heat to complete. For example, I can heat treat a material to increase the strenght. I can use heat to stress relieve to remove internal stresses of a material. These are simplified examples. Others may provide more details.

 

[...]

I have a question for you since you seem to know a lot more than us average folks.

Anybody else that knows feel free to answer as well.

I am far from an expert but are "heat treating" and "stress relieving" two different things?

They both use heat but I am trying to wrap my head around the difference in processing.

Heat treating is an all encompassing term......which can include stress releaving.

Typically when people mention "heat treat" ,they are talking about hardening or tempering

"Hardening" is heating the steel to a temp just below it's melting temperature (usually) and then quenching it in water or oil.....this is what we see them do on forged in fire, and it takes a soft steel and makes it really hard....but brittle

"Tempering" is heating that now hard and brittle steel up to a lower temperature (around 300-600 degrees depending on what level of temper)....that removes some of the hardness and also some of the brittleness

"Stress releaving" does just what it sounds like......through machining, forging, or otherwise working the steel, material "work hardens".....think of this like compacting snow, it eventually gets more compacted and harder and harder the more its pressed together....stress releaving heats up the steel to just below it's melting point, and then allows the steel to cool slowly....this causes the grain of the steel to revert to it's pre-work hardened state, softening the steel, and removing any stresses