The Firing Line Forums

Go Back   The Firing Line Forums > Hogan's Alley > Handguns: The Revolver Forum

Thread Tools Search this Thread
Old March 16, 2017, 12:34 PM   #51
Heavy Metal 1
Senior Member
Join Date: January 17, 2012
Posts: 231
Why can't the CNC machine self-adjust for tooling wear? It seems to me there could be a feedback system (laser?)that "tells" the cutter to go 'x' amount deeper.
The blood runs free, the rain turns red, give me the wine, you keep the bread.
Heavy Metal 1 is offline  
Old March 16, 2017, 04:20 PM   #52
Driftwood Johnson
Senior Member
Join Date: January 3, 2014
Location: Land of the Pilgrims
Posts: 1,750
For those of us not in the industry (and/or not in Q.C.), what is "tolerance stacking" and how does it apply to S&W revolvers?

Tolerance stacking is a term used in all industries that manufacturer physical products. It is not related to the firearms industry alone.

Many people think CNC (Computer Numerical Control) equipment magically makes perfect parts. In another life I was a CNC operator, CNC programmer, and draftsman. I may be able to clear up some misunderstandings.

The first thing you have to understand is that no process, and no machinist, can make a perfect part, no matter how hard they try. When I say perfect, I mean a part that perfectly matches the dimensions.

I am going to only discuss old fashioned plus and minus tolerancing, I am not going to even venture near geometric dimensioning and tolerancing, which is what most companies are using today.

Here is a very simple part drawing. It is a simple block, ¼” thick, 1 ½” long and 1” wide, with a 5/8” hole drilled through it. The center of the hole is 1/2" from one side of the part and 1/2" from one end.

The numbers with the + - symbols are the tolerances. Tolerances dictate how much variation from the perfect dimensions I will allow for my part. If I gave this drawing to a machinist with just the basic dimensions he will ask me what the tolerances are. The reason for this is that the more closely the actual part comes to the perfect dimensions on the drawing, the more expensive it is to make.

What this drawing is saying is that I will accept a part that is between 1.48” to 1.52” long, .98” to 1.2” wide, and .23 and .27 thick. In addition, the hole must be .620 to .630 in diameter, and it must be located .49 to .51 from one side and .49 to .51 from one end.

These are very, very loose tolerances. Part length, width, and thickness have a tolerance window of .040 each. That’s more than 1/32”. The position of the hole has a tolerance window of .020 in X and Y and the diameter of the hole has a tolerance window of .010. Fred Flintstone could make this part blindfolded and every part would pass incoming inspection. So no money would have been wasted making bad parts that did not meet the dimensions and tolerances called out in the drawing.

But tolerances do not come out of the thin air. They usually are related to the dimensions of mating parts.

Just for the sake of argument, let’s say that a 5/8” diameter rod is going to go into that hole. If I dimension the diameter of the rod the same way, chances are it will not fit. If I tell the lathe operator to make me a rod that is .625 +/-.005 in diameter, he can certainly make them and I will have to accept every rod he makes from .620” to .630” in diameter. And I have accepted every plate with a hole from .620 to .630 in diameter.

A rod at the high end of the tolerance, .630, will not fit into a hole at the low end of the tolerance, .620 in diameter. So I will pay for a lot of parts that do not fit to each other, unless I want to pay my assemblers to mix and match until they find parts that fit each other. Not the way to go in mass production.

If I want to guarantee that all my rods will fit into all my holes, I will have to change the tolerances to something like .625 +.005/-.000 for the hole diameter and .624 +.000/-.004 for my rod diameter. So then I will accept holes between .625 and .630 in diameter and rods between .620 and .624 in diameter. There will be something between .001 and .010 of slop for the fit of the rod in the hole. This may or may not be acceptable, depending on the design concept of the parts. I may have to tighten the tolerances further, depending on the design concept of the assembly.

The rub is, I have just made my parts more expensive. My machinist will have to be more careful not spending his time making parts with holes too small. Any he makes with holes too small my incoming inspector can reject. The same is true of the rods. Any that come in too big will be rejected. So costs have gone up to make the parts, and the machine shop passes those costs on to me. And that is just the fit of a rod to a hole. We have not mentioned the .020 positional tolerance in both directions for the location of the hole. Depending on the design concept of the final assembly, we may need to tighten the tolerances of more dimensions.

That is what tolerance stackup is. We used to call it doing a tolerance study. You run the numbers to make sure the parts will fit together in the final assembly and function the way they are supposed to. Our part is incredibly simple. In the real world of manufacturing, parts are much more complex and there are usually many more dimensions that are related to the dimensions of mating parts. So lots of dimensions have to be carefully examined and toleranced. And usually, the more precise the parts are, the more they cost.

When I was in the design end of things, we routinely toleranced all our dimensions at +/-.005. That was pretty routine and any machine shop could meet those requirements. Of course, we were making fairly small parts, usually less than 1 foot long. Tolerances have to relate to the real world the parts live in. You are not going to tolerance the window openings in a commercial airliner at +/-.005.

What does all this have to do with CNC? Most CNC machines today are very precise. They can do +/-.002 without batting an eye (a sheet of paper is about .004 thick). But what does that really mean? A manual machinist who was turning the wheels on a Bridgeport took into account the backlash in the system. Meaning how much slop there was in the mechanism so the tables would move when he started turning the cranks (huge over simplification here).

CNC equipment does not defy the laws of physics. There will be a certain amount of slop for the X,Y tables and the quill in even the most expensive modern equipment. But most will use some sort of electronic feedback loop from the motors to allow for that. In other words, if the program tells the table to move 1.001", the feed motor will stop moving at 1.001. Tool wear is another thing. Over the life of an endmill, as it wears, the cut it makes will become narrower. I’m not sure how the most modern equipment allows for that, but there will be a process.

But the bottom line is, CNC equipment is not magical.

A great deal of manufacturing today is done directly from 3D CAD (Computer Aided Design) directly to the CNC machine. An engineer creates a 3D model of his part in his computer, and sends the file to the CNC programmer. The programmer still has to take that information and create the CNC program, it does not happen automatically. He will decide which operations are done first and which operations are done after that, he will choose tool diameters for certain cuts, he will determine spindle speeds and feed rates, and a bunch of other things. If he screws up any of these parts of the program, he may break a tool or he may turn out crap. And somewhere in his notes, or in the design, the engineer will have built his tolerances in.

Then there is the matter of fixturing. Our simple part could be made from one simple set up in a vice. A more complicated part may require several setups to complete it. Meaning you may have to pop it out of the vice and reposition it to cut features on another surface. Repositioning a part invites inaccuracy to creep into the part. Parts that are made over and over again will have specific fixtures made up for them, to hold the parts in various positions in the machine. Then, before the program is started the machine will be zeroed in X,Y, and Z to the new position of the part. The most modern Machining Centers may be able to pick up a part from one fixture and reposition it on another fixture, without the assistance of an operator, rezeroing before continuing, I don’t really know.

I had a conversation with an engineer from S&W a number of years ago mentioning the amount of precision their CNC equipment was capable of. I really don’t remember the numbers, but they were very precise, much better than +/-.002. It was down to a few ten thousandths of an inch.

But the machines are still not perfect. Parts still have to be designed and toleranced correctly or they will not fit or function as they are supposed to.

There are several reasons why CNC adds value to the production process. One operator (not as skilled as a good machinist) can operate several machines simultaneously. There is no human error that creeps into the parts, such as the machinist accidentally turning the cranks too far, the machine is programmed to move the tables the correct amount, and allow for any slop in the tables. And CNC equipment can generate complex curves, which could only be done on old milling machines by following patterns.

But machine precision still enters into the equation, no matter what, there will still be a tiny amount of variation from part to part. Whether or not the amount of variation is significant is another matter.

And parts still have to be designed correctly, or it will be garbage in, garbage out.

P.S. A few years ago I had a job as a technician in a start up that was making some pretty interesting electro-mechanical devices. One of my tasks was to build prototypes. Every time an engineering change was made to the parts, I would build up a few assemblies to make sure everything fit. I can't tell you the number of times I went to the design engineers to tell them what did not fit where. It all boiled down to manufacturing tolerances related to the types of materials the parts were made from.

Last edited by Driftwood Johnson; March 16, 2017 at 11:23 PM.
Driftwood Johnson is offline  
Old March 16, 2017, 07:23 PM   #53
Senior Member
Join Date: November 10, 2014
Posts: 607
I have quite a few S&W revolvers, all but one are older models. I have a M63
that is not pinned. I only have adjustable sight models. I have owned 100s of
S&Ws. I am not interested in any new production models. The quality control
has come down hill, just like all the other gun makers. Still it is hard to believe
multiple guns would have made it out to market with same defect.
Drm50 is offline  
Old March 19, 2017, 05:05 PM   #54
Join Date: October 31, 2016
Posts: 80
How do you know other than cylinder gap, lockup stiffness, cylinder wiggle, overall fit/finish that a used S&W is worth purchasing? Revolvers I know have many parts and I'd be worried someone dropped the gun or that if it came from a gun store, who knows what abuse someone put on it. I've seen some beautiful used Smiths that I have passed on just because I was afraid to buy from someone I did not know. I had a used Model 19 once I bought from Bud's and it had some flaws and I was not comfortable keeping it. I love the old pinned/recessed models, hey were hand fitted whereas today they appear to use mim parts and less attention to details. However I'd still take a modern Smith over a Taurus any day.
kyguy1 is offline  
Old March 20, 2017, 11:49 AM   #55
Senior Member
Join Date: August 19, 2004
Posts: 7,025
As mentioned in detail above, CNC machines & processes result in neither perfectly uniform nor uniformly perfect parts.

CNC IS much more precise, on the whole, in turning out to-spec consistency than a bank of skilled machinists who used to do them on machinery that demanded more operator skill as opposed to programming skill.

But, tolerances do vary, and the "stacking" term simply means an accumulation of variances from part to part in a given assembly (here, a gun).

As examples of extremes at both ends, if ALL of the parts in a particular gun are manufactured/machined/CNC'd/cast/MIM'd to exact specs & assembled together, that gun would come closest to the older individual hand-fitting that produced guns prior to the CNC & sintering/MIM process introduction.
Minimal to practically zero tolerance stacking.

On the other end, if ALL of those parts were individually outside specs, either under or over, as the result of cutting head wear, mold wear, or whatever, the combined sum of those variances adds up (or stacks up) to result in a gun that may be loose as a goose (relatively speaking) in general, or may have this part here not interacting ideally with that part there, because four other parts surrounding them being off-spec don't allow them to mate correctly.
Major tolerance stacking.

Most guns are somewhere in between, and there's no way to have every gun made by any given company always come out in the first category using CNC & casting technology without hand fitting.

One of the contact guys I dealt with at S&W (recently retired) had worked his way up through the ranks to an executive position.
He started out as an assembler/fitter, and while he didn't do the mill work, he did use the parts produced by those skilled machinists in building S&W revolvers.

That process took a part from a parts bin, and fitted it into the "puzzle" (gun) by installing it & checking to see how close it was.
If close enough, it'd require very minor finish work (a stone here, a stone there).
If not quite as close, maybe a little more filing, a little more dressing.
If too far out of spec, rejected.

Today, Smith & Wesson doesn't have skilled fitters, for the most part, in the regular production lines.
They have assemblers.

They pretty much adopted Colt's production model when Colt developed their first sintered revolvers with the MKIII Series.

Pick a part out of a bin, stick it in, if it fits & functions reasonably well, that's it.
If not, keep pulling parts out of the bin till one does fit & function.

With Colt's MKIIIs, their sintering process (and later their cast process in the MKV) spit out parts close enough to finished form to be pretty much drop-in assemblies.

The advantage to those non-fitted methods was in saving manpower hours, cutting back on salaries paid to highly skilled labor, and speeding up production.
The disadvantage is that we don't get the hand fitting that used to produce such well-built revolvers.

CNC processes can come close, but still don't achieve the same end result.
Tolerance stacking is simply built into the program nowdays, where-as in the past it was adjusted, removed, or compensated for by a very skilled pair of human hands.
DPris is offline  
Old March 20, 2017, 06:04 PM   #56
Senior Member
Join Date: July 26, 2001
Location: midwest
Posts: 2,033
Junk. Get rid of it.
agtman is offline  
Old March 20, 2017, 07:44 PM   #57
Senior Member
Join Date: February 22, 2008
Location: SW Washington state
Posts: 1,412
Good Job!

Thanks for taking the time to accurately describe tolerance stacking. I was doing design engineering in a former life, and encountered all of which you so nicely described.

Then I discovered the sales reps made lots more money......hell if I can design it and build it, selling it should be a piece of Quiche.

Actually the job of selling is a LOT harder. It is more lucrative tho.

Thanks for taking the time.
Freedom is not free
Ricklin is offline  
Old March 20, 2017, 08:20 PM   #58
Driftwood Johnson
Senior Member
Join Date: January 3, 2014
Location: Land of the Pilgrims
Posts: 1,750
Howdy Again

I worked as a designer for a small prototyping department for a few years. We had a machinist assigned to our group who made up all our parts. Every once in a while he would call me up and say, 'Hey, how important is that 3.015 dimension?' Or something like that. I would look at my assembly, and tell him whether or not I could accept the part, or whether he had to make it over again.
Driftwood Johnson is offline  
Old April 2, 2017, 12:29 AM   #59
Senior Member
Join Date: November 3, 2015
Posts: 138
I'm a little late to this party, but thank you for the informative thread.

Here's a pic of my new (purchased 2016) 4-inch 686+. It's a tack driver. Slightly odd positioning of the yoke (lower) but I don't think it affects anything.

Yes, the weapon was loaded for the picture. But don't worry- I was only drinking whiskey when I took it.
Attached Images
File Type: jpg FullSizeRender-7.jpg (243.4 KB, 40 views)
Fivesense is offline  
Old April 2, 2017, 11:10 AM   #60
Senior Member
Join Date: December 28, 2006
Posts: 3,793
However I do remember being completely surprised seeing the ball detent on a really big S&W a year or two ago. I seem to recall it was one of the really big X frames, but I cannot be sure.

This will require more investigation.
My X-Frames have the ball detent as well as my P.C. 629 Magnum Hunter.

As for the new Smiths, I just bought a new 4" 686 a few weeks back. Walked past the revolver case at my LGS and saw it there with a $499 price tag on it. Thought it might be used and in great shape. Clerk told me it was new, just received from their sister store in the next town. I asked to look at the gun and if the price was correct. He told me yes that was the correct price. I already have a prelock, 4" 686-4. I also have a 6" 686 -5 and a 5" 686-8. I didn't need another 686....but for $499? Must be something wrong I thought and looked it over closely. Barrel was perfectly straight, nice cylinder gap, crane/frame fit was great, crown good for a new Smith, rifling looked good and action nice and smooth for a new gun. Another guy who saw me ask for the gun was waiting for me to put it down after seeing the price tag, but I didn't before I told the clerk I'd take it. Wasn't till checkout they discovered they had priced the gun wrong. But by then, I had already paid for the BG check and transfer(this shop makes you pay for this up front first) and the gun had been successfully transferred. They did a lot of grumbling(even politely asked me to at least pay their cost of the gun) but in the end, decided to honor their mistake. Since I did not need the gun, even at that price, if there had been any issues, I would have walked away. I actually looked for an excuse not to buy it....but couldn't find one, not at that price. In the long run, it's been put away and will be a graduation present for my youngest when he graduates from college this spring.

CNC, MIM is not cheap. All they do is eliminate the "hand-fitting" because the precision of the parts means it is now unnecessary. While there is a amount of "tolerance stacking" in them, there also was when skilled assemblers hand fit those same parts. CNC and MIM do not have "Mondays or Fridays" like humans do. They do not have good days and bad, nor do they come in sick or hung over. Nor do they get tired or bored at the end of the day. I worked construction for 45 years. When trusses first came out, when air guns became the norm, folks cried that the craftsmanship in home building was gone. No, the craftsmen were still there, they just used different tools and different products to produce the same product. While these new systems made it easier for unskilled folks to assemble a product, it was no different than when those unskilled folks used rafters and framing hammers. In many cases, today's homes are built better, than those old "craftsman" style homes built back in the 30s. I see the same with firearms. While processes have been changed and certain materials, because of availability and cost have changed, there are still quality firearms being produced. While nostalgia is big right now from us Baby-Boomers and it's nice to own things built the "old-fashioned" way, most of it is just that, nostalgia. My -5 686 outshoots my -4 686 and the -8 is the most accurate of all three. The new 686 has not been shot yet, but I'm sure if there is a problem, the lifetime warranty will fix the gun for as long as my young son lives(yes, it was registered to him to S&W as the original owner). Only difference I see in the new 686 vs the older -4 is the Hillary hole, the hammer and trigger, and the finish is not quite as smooth on the new gun, but that could be from years of polishing from me. Oh, and the new rubber grips. The barrel aligns better on the newest 4'' 686 as compared to the -4(which has the rear sight far to the right). Looking at the newest 686 in the safe, I wish all the new Smiths were like it.
buck460XVR is offline  
Old April 2, 2017, 11:27 AM   #61
Senior Member
Join Date: November 26, 2016
Location: Minnesota
Posts: 866
A few days ago I bought a S&W Model 29-10 Classic.

Because of this thread, I spent much more time inspecting the gun than I normally do. The new gun is 100%.

My X-frame is also 100%.

So there are at least two newish guns that defy the trend(????)

disseminator is online now  
Old April 3, 2017, 06:44 PM   #62
Senior Member
Join Date: December 14, 2004
Location: Northern Indiana
Posts: 6,117
No photos, no proof you must be a liar?
So I have attached some photos.
Note the Crown photos, the scratches are a lot worse than they show in the photo and 3 of them went to the land and into the barrel.

A year ago I ordered a 69 and the following report I made at the time. I have a heard of N frame S&W and know the quality they are capable of. This thing was more Taurus like.
S&W did fix the multiple issues but a gun like this to arrive in this fashion was very disheartening. S&W took care of the issues but these were issues that a company with S&W's history I should not have to put up with.

The rest of the thread below is at the following.

Well it showed up last night and this is the letter that I just wrote to S&W.

Smith and Wesson
I ordered and received last night a new M69 with the markings I have added below including the markings on the blue box.
A little story about myself, I am a collector of N frame S&W’s, my prime collection are of older Model 24’s and 29. These I have multiple of including 2 Lew Horton 24’s. 25’s 27’s and a 1955 Target in 44 special. I know a quality gun when I handle one.
This 69 is not one of them. The following are my concerns.
1. The grips are a poor fit along the back of the grip. They don’t align, and to me they seem to be manufactured poorly where the back sides go together are very thin. When griping the gun I can feel the two edges roll under my palm. This gives a very poor feel.
2. Single action is very nice but too heavy. I can live with it but I would guess 5 lb. But the smoothness is inconstant. Sometimes you can feel a little creep sometimes not.
3. Here is the worse problem. This is by far the worse double action I have ever handled. The first quarter of the pull feels like dragging bricks through gravel and I was shocked at the weight. I have no scales but I have no guns that come close. This is a very poor trigger for a S&W.
4. The cylinder release is a joke. Two bricks sliding together.
5. This one surprised me coming from S&W. I know the guns have to be fired so you can retain a fired case but can’t you at least clean the gun. Only two cylinders have been fired from the look of the front of the cylinder with one at 12 and the other two chambers around. The ring I don’t mind but the barrel has so much copper fowling that I was just simply surprised. You must be using very cheap jacketed bullets or something.
6. The barrel shroud. There is a flat spot that runs along the left side of the shroud at 7 o’clock. Yes a flat spot. There are also discolorations on the side. A scratch at the shroud to frame aria and the barrel crown has a small scratch that runs to the edge of the rifling. This is small but it’s also the worse location for a nick or scratch.
7. One final thing. When I got home with it I dry fired it. This is where I became very unhappy with the double action. In less than 20 dry firings the cylinder has a ring.
Attached Images
File Type: jpg junk Grip.jpg (144.8 KB, 31 views)
File Type: jpg origional crown scratches poor casting..jpg (151.8 KB, 31 views)
File Type: jpg Copper factory.jpg (63.8 KB, 30 views)
It was a sad day when I discovered my universal remote control did not in fact control the universe.

Did you hear about the latest study.....5 out of 6 liberals say that Russian Roulette is safe.
Ozzieman is offline  

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

All times are GMT -5. The time now is 01:28 PM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2018, vBulletin Solutions, Inc.
This site and contents, including all posts, Copyright © 1998-2018 S.W.A.T. Magazine
Copyright Complaints: Please direct DMCA Takedown Notices to the registered agent:
Contact Us
Page generated in 0.10668 seconds with 11 queries