What is the Optimal Tolerance for Close Fit in Millimeters?

[jarvezs]

Need best "mm" for close fit

I am for the first time building an object for fabrication online with "e-machineshop" free CAD software and odering service.

I would like to know what is the minimum desirable space between two objects to allow for the closest (tight) fit with the least amount of friction. I just want everything to fall in place. I have no experience in physics. I am an economist but, what can I say, I am searching for a new hobby.

Help me someone, please. I would prefer if you give answer in millimeters "mm". (Oh! I tried googling my problem but found nothing.)

Thanks

 

[inha]

Friction isn't something that happens on a mm distance scale so my answer would be zero. The objects would have to be in touch for there to be friction. If you want a tight fit with min. friction you should rather look at the material properities, lubrication, fabricate the pieces to fit well and not use excessive force in clamping them together.

 

[Pengwuino]

Why not use lubricant?

 

[rcgldr]

Motorcycle engines have some pretty tight piston to cylinder clearances, just a few thousandth's of an inch. You need some clearance for temperature and lubrication. I'll have to dig up some old motorcycle magazines to get the exact clearance values.

Regarding the lowest amount of friction, air tracks have almost no friction in the direction of travel.

 

[FredGarvin]

You're interested in limits and fits. Anyone who does design work would be able to help you out with this so it's not a physics problem per say.

Since you are interested in the metric system, take a look here:
http://www.roymech.co.uk/Useful_Tables/ISO_Tolerances/ISO_LIMITS.htm

This will give you a brief overview of standard fits. Each class has a particular description with how you want the assembly to go together. My suggestion is that you would probably want something like H7/h6 which is:

H7/h6 Locational-Clearance fit:
Provides snug fit for locating stationary parts, but can be freely assembled and disassembled.

If you describe the parts and their functions, we can make sure that this is indeed the fit you need. Tolerancing plays a large role in your fits, so watch your stack ups. Dependig on your actual fit, you may also consider specifying a surface finish requirement.

 

[brewnog]

Bigging up Fred's post here. It depends heavily on your application. Big, crude rotating machines like rock crushers generally have pretty loose fits, precise things like machine tools have much tighter fits.

Let us know your application, expected service life, operating conditions and the like and we should be onto something. Alternatively just open any design handbook.

 

[Mech_Engineer]

I would like to know what is the minimum desirable space between two objects to allow for the closest (tight) fit with the least amount of friction. I just want everything to fall in place. I have no experience in physics. I am an economist but, what can I say, I am searching for a new hobby.

Unfortunately, this subject is much deeper than you are hoping. You need to take into acount maximum easily achieved tolerances, and the fact that price goes up exponentially with tolerance. Additionally, you need to decide what you are trying to achieve and what is needed to achieve it in terms of backlash, assembly, repeatability, manufacturing, etc.

Just wanting "everything to fall into place" is not a very good requirement because everything could theoretically fit together with grossly huge tolerancing.

We COULD just throw out numbers, but the fact is we need to know more about what you are trying to have made, otherwise they mean nothing...

 

[brewnog]

Unfortunately, this subject is much deeper than you are hoping. You need to take into acount maximum easily achieved tolerances, and the fact that price goes up exponentially with tolerance. Additionally, you need to decide what you are trying to achieve and what is needed to achieve it in terms of backlash, assembly, repeatability, manufacturing, etc.

Setting tolerances aside, the question of limits and fits (for a specific application) is a very well documented one, and easy to answer.

 

[Averagesupernova]

I'll just mention that oil clearances in the crankshaft bearings in an engine are around .5 to 1 thousandth of an inch. (Do the conversion yourself.) As clearances get tighter and tighter, other specs need to have tighter controls. Thermal expansion, out of roundness, runout, all these things need to improve when the oil clearance gets closer. So as someone mentioned before, the friction doesn't occur until the parts actually contact each other. So the type of lubrication is often what determines how close parts are machined to fit. And the type of lubrication is often determined by other properties of the machine.

 

[FredGarvin]

You have to be careful when dealing with bearings, which are a science unto themselves. Clearances around the outer race have many different implications other than traditional fits. For example, a larger clearance may be required because of the need for squeeze film dampening. Unless you're looking at something that is really low speed with low relative loading, there are probably more dynamics reasons that drive bearing fits than anything else.

 

[jasc15]

Unfortunately, this subject is much deeper than you are hoping. You need to take into acount maximum easily achieved tolerances, and the fact that price goes up exponentially with tolerance. Additionally, you need to decide what you are trying to achieve and what is needed to achieve it in terms of backlash, assembly, repeatability, manufacturing, etc.

Just wanting "everything to fall into place" is not a very good requirement because everything could theoretically fit together with grossly huge tolerancing.

We COULD just throw out numbers, but the fact is we need to know more about what you are trying to have made, otherwise they mean nothing...

He is right, you have to account for tolerancing in the machining, which is specified by you the designer, just don't give the machinist a headache with unrealistic tolerances. you are looking for a clearance fit, rather than transitional or interference fits. here is a link that gives tolerance ranges for these types of fits, depending on your application. this is simplified using the shaft hole example. http://www.engineersedge.com/general_tolerances.htm

 

[jarvezs]

You guys are the best. You have pointed me in the right direction for sure "Tolerances, limits and fits". I agree with FredGarvin that a H7/h6fit might be useful "Snug Fit". I will however be using a combination of clearances to get a variety of desired fits. I located a very good table at the following link, you could have a look at it. "ISO Metric Tolerance Chart" >>http://www.ex.ac.uk/~sritchie/hydro/datasheets/D780Ref.pdf<< Mech_Engineer, you couldn't be more right that the subject is much deeper than I thought. Thanks all for your comments. The emachineshop seems to be up to the task as the number of digits allowed behind the decimal point seems to be more than sufficient.

Oh! I like the idea of the emachineshop so I feel compelled to do a bit of advertising. Simply its an online free cad software (very easy to use) for custom design and ordering of parts given a wide range of materials to choose from and also a wide range of machines. It seems quite simple to me. There is a short video tutorial that gets you started instantly. You can get a price quote with shipping charge by just pressing a button. When you are finished. Just order with your credit card and wait for it to arrive. Check it out at http://www.emachineshop.com/. Its really cool.

 

[jarvezs]

Oh! One other thing. If you are interested in Building Circuit Boards, they have a sister company that allows you to do it online. I have no interest in this field but I won't put it pass a couple of you guys. A link is on their webpage but you can also go directly to http://www.pad2pad.com/index.html .

 

[FredGarvin]

There are a few companies out there that do PCBs for you. We used to use a different company quite a bit at a different job I had in NY.

Good links.

 

[jarvezs]

Thanks, I've been still researching the Tolerance matter. Much of you might already know but just to get it out there. Recall I posted a table. There is also a calculator that seems to do a very good job. All you need to do is enter your Basic / Nominal Diameter and choose the type of fit and it will give you the limits based on the ISO Metric Standard. Check it out at http://www.infosystemspro.com/isofits.htm . I am not sure I want to spend an almost $40 dollars on it though given that much of the information is already in the table. However, it seems to be a bit more percise as it accommodates entry of decimal numbers for the Basic Diameters.