PDA

View Full Version : What is close tolerance work?



C. Tate
10-19-2001, 03:45 PM
Here is a question that is of no value other than to show the differences in the way we think about machining.

What do you consider close tolerance work.

I was talking to a man at the Mazak facility about machining and some of the things he and I had done and seen. I told him about a part we had been making in my shop that had a bore tolerance of +.0004 -.0000. I was quite sure he would think this tolerance was damn tight I sure did. He worked in some type of R&D shop for Ford in Detroit and told how they worked with tolerances measured in microns and I am sure he new that is real damn tight.

Anyway that is where the question was spawned?

SGW
10-19-2001, 05:29 PM
The man I consider my mentor, Dan Fay, used to say that people in home shops frequently didn't give themselves credit for the tolerances they achieved in fitting parts. They might not be fitted to a *specific* dimension, but the one-off fit of, say, a piston to a bore in a home shop can be within a couple ten-thousandths. As long as you control both parts and you don't have to achieve "interchangable parts" tolerance, which implies specific dimensions, it's amazing what you can achieve in your basement.

As far as making something to a specific dimension goes, I think your +0.0004, -0.0000 is about as close as I'd want to worry about on a regular basis. Measured accuracy to under half a thou starts to get problematic.

Another story: years ago I took a night class at the local voc-ed school. The instructor was a machinist at MIT's Draper Labs. He regularly worked to a tenth or less, and said when an occasional job came along with a tolerance of a thousanth, it felt like he almost could do it with a hammer and cold chisel. It all depends on what you get used to....

Gizmo
10-19-2001, 08:05 PM
My favorite quote about work tolerance:

A house framer works to the closest 1/8".
A cabinet maker works to the closest 1/32".
A boat builder works to the closest boat.

Nuttin' to do with nuttin', but I love the opportunity to say it...

halfnut
10-19-2001, 11:44 PM
Apples and oranges. You have to know what you are comparing.

Your .0004 inches is 10 microns, see you are working in microns also and didn't know it.

A micron is .001 of a Millimeter I believe.

A young fellow I know whom worked in a bearing company used to spout micron tollerances to me, I'm sorry my brain functions to the English system of measurement, and I would have to translate inch measurements to him.

.0004 tolerance is close for common work, try to get much finer and you will have to start using air gaging and temperature control-compernsation, even at .0004 you have to watch the temperatures.

Work with a fellow whom worked for a hydralic pump manufacturer, he says, most tolerance I ever had was 2 tenths, but he was running a hone and a lap, Sheffield air gaging, optical flats, surface profilometers and a inspection department at his disposal. This same fellow couldn't use a set of telescope gages and mics to a thous.

.0004 is close for ordinary methods, think of it this way, plus or minus .0002, dern, closer than you thought, why thats only 5 microns.

I just love these metric people.

BrianH
10-20-2001, 12:04 AM
.....and I worked with the Brit engineer that called everthing metric "mills".....
(for "millimeters", I guess....)

I think the point behind Gizmos quote is that it is all relative to the type of work you do. The shop I work in has been building inspection gages, but environment-wise, we're probably running the ragged edge in being qualified to inspect our work. For a gage lab it would be a different story.

Anybody can hit close numbers....there's a trick (technique) to everything. It's simply a matter of figuring it out (or being taught by one who knows).




[This message has been edited by BrianH (edited 10-19-2001).]

Thrud
10-20-2001, 12:22 AM
FYI Guys,

Metric system "rules" in science at least. Makes unit conversion so simple your dog could do it.

Hardinge HLV lathes have a spindle runout of .oooo25" or .ooo635mm or .635 microns or 6,350 Angstroms (about the wavelength of HeNe Laser light).

Starrett/Webber Chrome Carbide Master (grade 0.5) blocks have a tolerance of +-.000001" (1 millionth) or .0254 Microns or 254 Angstroms.

It is good to know craftsmanship it not dead yet...

Dave

Also 1" = 1000 mil


[This message has been edited by Thrud (edited 10-19-2001).]

toff
10-20-2001, 12:51 AM
Hi,Me again.
I agree that what you need is what to work toward. More time is lost due to over specified work than can be measured. With computer drawings going to CAM no tolerances are specified for a prototype shop. We have to make drawings for the manual machines. The draftsman isn't a machinist and every dimension is in "tenths" (four decimal places) on the drawing! We in the shop have no access to the overall drawing and do our best but, if we knew where the part was going we could save hours by dropping back to +or- a few in most places and putting the accuracy where it is required.
Lautard tells the story in a "Bedside Reader" of a fellow who got a drawing of what looked to be an indexing plate of brass. He did it right proud!! Then found it was to be a cover for a floor drain..Need I say more!
toff

bdarin
10-20-2001, 01:22 AM
To answer the original question, I try to work to +/- 0.0000. Since this is unacheivable, +/- .001 or .002 usually works just fine for most parts. For precision work, I try to get as close as possible, but +/- .0005 is acceptable error and is probably the best I can do. Was talking to a tool and die maker once, told him I can get by being off a thou or two and he said that that much slop in his work would be junk. So I guess it's all relative.

Ron LaDow
10-20-2001, 09:13 PM
I think Toff's got a point.
Given time, most any machinist can hit a couple of tenths. But does this widget need that kind of fit?
I've wasted time on fitments that really should have been looser, but that mandrel sitting in the main saddles right over there better be within a couple of tenths.

snorman
10-20-2001, 11:26 PM
I think most toolmakers talk a better game than they play. At least in my experience. Yep, I used to do that stuff too. Metal stamping die work mostly. Work quality varied enormously depending on who did it. Rarely was anything held to .0001. Plus or minus a half was the basic tolerance I guess.

The factory shut down so I had to take a job as a machinist. The closest work I do is with mating parts like a bearing that presses on a shaft. Sometimes they'll tell you: "Just make it a peck fit." I don't think they realize how close that is!

Steve

Thrud
10-20-2001, 11:41 PM
I just want to point out many members of this BBS are new to machining and still struggle with the basics - often making mistakes - and getting frustrated over it. Regardless of one's skill level, we should always endevour to do our best work and at the same time to push our own limits to new heights. A thou may be acceptable, but if you can do better in the same time or less then do it.

Dave

farmwrench
10-21-2001, 08:14 PM
I have a "free" lawn mower. My friend said his parent's lawnmower wouldn't start and the Sears shop declared it DOA. When told the symptoms I said to bring it over and I wuld fix it for the cost of parts. The starter bendix was blown and a few bucks later it was good as new. When I called them to come pick it up they said they had bought a new one and to keep it till it died or they needed it! Five years latter... I blew out a spindle for the blade, sheered the shaft when I clobered a chunk of wood that hid in the grass. I don't rember the dimensions but before I drove the bearing off the shaft I made a new shaft after much sweat on my claped out lathe to -.0005" and the dam bearing wouldn't fit. When I drove the bearing off (big fight) I found the blasted thing must have had the shaft shrunk with dry Ice and the bearing too hot to touch to make it fit. I dont rember the exact diference between bore and shaft but I then was P.O.ed and just took .002" passes till it was a snug slide. Long and short is that that spindle has mowed a lot of lawn and all the rocks my 140hp snowblower can move. Now I don't get bent about something that will work with what would be "unaceptable" as long as my final product does what I want it to and only I know what the numbers are. I liked the drain cap story!

gizmoid_52
10-22-2001, 12:25 AM
I totally agree with mr. Farmwrench. I work
for a hydraulic company and have went round and round with those who make and sell our
proprietary valve, when they came to us they
claimed total interchangability of parts, we
believe it can be done, but still haven't seen it done. they are machining the parts and once the parts are finished, they grind them to tolerance. They claimed they could
keep tolerances to 50 millionths, I haven't
seen it. They use a air bore gage to measure the bores, and then use electronic mikes to measure the spool. In the print I called out
a tolerance of a diametrical clearance of
.0002-.0005 between spool and bore.
What finally happened was their quality control man quit, he kept sending us parts
we sent back due to testing for hydraulic by pass, with all the big money they spent on
fancy grinding equipment and measuring machinery, I was out measuring them with a
12" cast iron gage I made that was tapered
.001 in its length. I simply made a mark on
on the gage where it stopped in the valve
bore and measured it with a digital mike
then turned the valve over and did the same
on the other end of the valve. then I did the
spool with the same digital mike. they flipped. he quit, and the point is big business makes statements they often can't live with. Fancy equipment is not infallable
nor are the people using it. Yes home shop
machinists can out do these guys because if they are asked, they aren't interested in what they do as much as you are, they are someplace else other than paying attention
to machine work.
Harry Pope, I have heard, said don't blame
the machine for your poor work, blame the
man. I add to this if your machine doesn't do
what you want it to, what's wrong?, your a machinist fix it!