Member Login

Remember Me
Forgot Password?
Join Us!

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!

Join Eng-Tips
*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.
Jobs from Indeed

Link To This Forum!

Partner Button
Add Stickiness To Your Site By Linking To This Professionally Managed Technical Forum.
Just copy and paste the
code below into your site.

Torque for NPT connections

Torque for NPT connections

I have done many searches looking for recommended torque by size and material for threaded pipe connections.  Almost all sources give a "number of turns" method.  Does anyone know why torques are not used?

RE: Torque for NPT connections

I need information about multilayer composite pipes.

thank you.

Farhad Mohkami

RE: Torque for NPT connections

NPT Tapered threads seal (in theory) by interference.  Since there are several variables that enter into determining coefficient of friction (material, pipe dope, Teflon tape) and by extension torque, it is not practical to use torque.  If you used torque one of the following would occur periodically:
1 you would thread a pipe into an oversized fitting or an undersized pipe into a fitting too far and still not get a good seal because the threadform was all wrong where interefence is generated.
2 you would thread a pipe into an undersized fitting or an oversized pipe into a fitting and not get sufficient engagement for sound mechanical connection.

Using engagement, these problems are avoided.

RE: Torque for NPT connections

I am looking for similar information on NPT torque values. I would be grateful if you could post any of the more useful information sources you have found.

RE: Torque for NPT connections

I'm curious as to why the question of torque on threaded pipe joints would arise - any particular problem that's popping up? Leaks, perhaps?

RE: Torque for NPT connections

Leaks are one problem.  Another is dissimilar materials, i.e. metal into plastic, or plastic into metal.

If it is metal into plastic and the metal is high end or the spec or oversized or the plastic is undersized then the method of number of turns can cause a real problem with the plastic portion cracking.

RE: Torque for NPT connections

To a large extent, screwed piping is an example art meeting science. Most screwed piping is tightened until it "feels right", and the fitting is pointed in the desired direction (if that's applicable). The metal/plastic piping interface can be tricky at the best of times. It's often a thin line between stopping a leak, and cracking plastic, on screwed joints, or with flanges.

When your pipe is being threaded, is anyone checking that the dies are set correctly? This is particularly important on the adjustable heads, that will do a range of pipe sizes. It's not all that often that the dies get set exactly on one of the increments in order to cut threads properly. Even the drop-head style have a small level of adjustment, and the instructions are on the back of the package that the replacement cutters come in. You should be able to screw a fitting on 3 to 4 turns, to hand tight on clean, dry threads - no tape or dope for this check. Once the dies are set, they're normally OK until the size is changed. I've never actually seen anyone count the turns after the threads are taped or doped. The fitting is screwed on as far as comfortable by hand, then wrenched until it feels right, using an appropriately sized wrench.

RE: Torque for NPT connections

Spot on, TBP. I don't know that I've seen anyone count number of turns either before or after taping. But an experienced pipe fitter will immediately know if engagement is "good", so I guess there is a subconcious level of thread counting.

As pointed out by ghillis, torquing of screwed fittings is not feasible due to the variables involved, but when you think about it, tightening screwed piping until it "feels right" is definitely torque related. Try quantifying the experience factor though......"art meeting science" - beautiful.

RE: Torque for NPT connections

Having installed quite a few kilometres of screw jointed pipe, without thinking about the "art" (I just wanted no leaks) I will chuck in my two bobs worth.
Torque by itself is meaningless. Its the progressive rise in torque you are feeling for. Screw it in until it starts to seat. Then up the force a little by yanking. If each yank gives less movement (or more jolt), you probably have a sound joint. If  the movement stops suddenly, you have probably bottomed out the pipe end and will have a leaker. If the movement doen not decrease with each yank, then you probably have a plastic valve or joint, and have stuffed it by yanking too hard. If you are lucky, it is just a bent thread, but you should have picked this up because it will not seat in the first place.
Yes, I guess there is some art involved.

RE: Torque for NPT connections

So far so good - art & science and all.  smckennz, I think that's a good description of that dynamic 'feel' - that's probably how you would program the automated pipefitter to feel it.
But how the heck do the pipe fitting designers make sure their parts are good - send 50 samples to 50 experienced plumbartists with a "how's that feel" questionaire?  There's got to be some sort of acceptable-torque-limits table SOMEwhere, doesn't there?
(I guess I'm asking about material strength.  The thread shape descriptions are pretty thorough.)

RE: Torque for NPT connections

I just had to get in this discussion.
I’ve been designing threads since the 70’s and been trying to figure a reliable torque formula for threads for the same amount of years. My conclusion so far is, not even PhD’s can solve for torque, to be reliable for the every day assembler. The reason is that there too many unknowns to equate. Every element feature such as thread height, taper, finish, lead fluctuations, yield of material, thread compound, etc. effects torque drastically. Not only that, the friction factor changes per turn and partial turn during makeup. You can get in a ball park with torque, but sooner or later it will burn you because the machining tolerances are too wide to predict what situation the two members are seeing as far as stress and strain. API has equations for torque and I believe you will find that the NPT and line pipe are the basic same threads. However most people only use these torque values as a starting point, and with a grain of salt.  Also, with experience of assembling and testing API connections and premium threads the only way to predict torque, is to have full control of the manufacturing tolerances, and because of this,  one can predict the situation and mechanical “chain of events” that occurs during assembly. With that said, I believe the real issue here is leak resistance. It is far easier to directly relate and calculate leak resistance to turns (translated to interference), rather than torque. Leak resistance is directly related to the Hoop stresses in the male and female connectors. The theory is when you assemble the male and female connections together, the female results in hoop tension values and the male is in hoop compression values due to interference. This is at the assembled state without any other stresses applied. The connection will be leak tight until the male connection reaches hoop zero due to internal pressure. Hoop zero is when the male member expands back to its machined state or you can say when it lost all of its compressive diametrical interference. Any additional internal pressure will cause a leak if the pipe body did not fail first. When assembling the connection, one should know the maximum amount of turns to make sure the connection does not yield,both male in Hoop Compression and the Female in Hoop Tension. Remember that any other stresses applied to the connection besides internal pressure, such as bending, axial tension, and/or axial compression is additive to the stresses and will approach yield and possible premature failure.
A little point of advice:
A metal female connection assembled with a plastic male will have more sealing integrity than a plastic female connection assembled with metal male connection with the same conditions. The female expansion governs the integrity.

RE: Torque for NPT connections

The plumbartists feel is more than just torque.  His basis of tightening involves torque and the rate of change in the torque.  You can feel when deformation begins when the change in rotation angle and the change in torque is no longer linear.  Experience tells the plumber when to stop to ensure a positive seal without damaging the fittings.  So knowing torque alone is not enough and neither the turn count alone is enough to determine when a joint is made properly.  I have no specific reference stating that the above is accurate; however, these are the things that I use to determine how tight I should makeup a screwed joint when I do not have a torque spec.  These factors come from my experience, and my experience is that of a car mechanic turned engineer.  If I were to make an engineering study using all of the variables, the results could not be translated into an easy to use guide.

RE: Torque for NPT connections

Mr. CRG is correct in a unique way. I never thought of entering feel in the equation, but here is a possible way of putting it on paper. The oil field uses special power tongs to assemble connections that records, via computer, Torque vs. turns, torque vs. time, and torque vs. position. The major Oil Companies have required the use of this equipment because they can get a better “feel” of the assembly (2D graph). Also these graphs are documentation for archiving in case of a mishap. As CFG stated the torque is not linear, it looks more like a stress strain curve with a rubber band in the tensile frame. Also he is very correct by stopping the torque within this curve, because he is staying within the elastic limit. Maybe the plumbers can train and calibrate their “green plumbers” arms with a Torque/Turn unit. This way they are converting their feel into a visual aid.

RE: Torque for NPT connections

We use NPTF for steel and NPT for other metals and plastics.  Our process is to use turns past finger tight for correct sealing without excessive deformation.  This is based on best practices within the valve components industries.

Torque flat out won't work for many of the reasons mentioned earlier.  If you ever look at the section in the Machinery's Handbook, you will see nominal engagement for pipe threads, both handtight and wrenched.  We do not follow these guidelines verbatim, however we do go by turns past finger tight to assure consistent sealing without excessive deformation.

RE: Torque for NPT connections

There is one application where torque on the joints is a requirement.  When a submersible pump is installed in a well, the normal direction the motor turns will untwist the threads in the pump column, if it untwists far enough, the pump and motor fall off and plummet to the bottom of the well, standard rules require a torque of 10ft-lbs per hp to ensure this does not happen.


RE: Torque for NPT connections

Hi Group,

Good info all the way down!  Although I'm aware that this is a mechanical forum, I wanted to throw in a not-so-commonly-known requirement for electrical devices using rigid or intermediate conduit in hazardous locations.  An ISA standard (I believe) calls for a minimum of five full turns at every joint or connection, in order to maintain a minimum gap length to cool hot gases that may be escaping under pressure from within.

Have a happy!

Old Dave

RE: Torque for NPT connections

The torque to which NPT threads must be tightened to obtain a leak-free joint is determined at least in part by the nature of the sealant used.  With some sealants, making the joint too tight will lead to an increased chance of leakage, as counterintuitive as that may seem.  The mechanics of the compression of these sealants in a joint during assembly are very complex indeed, so forget about systematizing this unless both the male and female parts are machined to close tolerances.  If you're dealing with pipe and 150# fittings, give up now.  Select the right sealant for the duty and then make sure you have a well-trained and experienced pipe fitter, or expect leaks.

Another consideration is the tendency of the sealant to act as a lubricant to resist galling during fit-up.  This is particularly important with materials which gall readily like stainless steels.

Building screwed pipe to seal reliably is definitely the intersection of art/technique and science.  But if you get it right, NPT threads are very reliable and extremely useful.

RE: Torque for NPT connections

In thread design of Oil Country Tubulars, which the NPT was the starting point or Grand Father, thread compound is used to plug the leak path, not to guarentee a seal. There are two leak paths located at the roots and crests. Also the thread compounds are lubricants to help lower the friction factor and help increase the gall resistance of the male and female members. Interfence is required to insure the connection is together "RADIALLY" when the connection expands due to internal pressure. If the female expands more than the interference (The female separating radially from the male), it "will" leak. Regardless of the thread compound.

RE: Torque for NPT connections

Hey guys whats up,

Just a thought, couldn't you use the info available for standard steel bolt thread torque’s by averaging the torque values between two different sizes?  For example, a 1" dia. bolt might have a maximum tightening torque of 75 ft.-lbs., and a 1 1/4" dia. bolt a maximum tightening torque of 100 ft.-lbs..  The average of the two would be half of 175 ft.-lbs. or 87 ft.-lbs..  This might be approximately right for a 1" dia. NPT pipe, but pipe is hollow.  Therefore a deration factor would be needed to compensate for the lack of strength available from the pipe as a hollow tube verses the solid bar geometry of the bolt.  I think a good approximation for a deration factor might be to use the ratio of the cross sectional area of the material of the pipe divided by the cross sectional area of the material of the bolt.  This deration factor, less than one, could then be multiplied times the average value of the bolt torque’s to arrive at an approximate torque for the tapered pipe.  My guess is around 10-25 ft.-lbs. torque for a 1" NPT. pipe thread.  This is the torque range of a bolt that is about 1/4" in diameter, which is about 1/2 the amount of steel material that the 1" NPT has, but the pipe geometry is a cylinder.  To measure this torque I think you would have to make an adapter that would place the head of the torque wrench as close to the pipe being tightened as is possible.  Maybe a square or oblong hole in a pipe wrench, if the square nose of the torque wrench won't fit into the jaws of the pipe wrench while being used.

Hope this helps.


RE: Torque for NPT connections

A good resource for this is:

Search for: NPT

I would suggest a turns past hand tighten approach based on this documentation. The reason for this is four fold: the material being tapped, the tolerances on the tapped hole, if a lubricant is being used, and the possiblity of overtightening. I would suggest that you have someone do the turns past finger tightening method and record the torque needed to accomplish this. This would give you a good initial starting point.

RE: Torque for NPT connections

The torque and number-of-turns philosophies need very much to be tempered with the requirement that elbows and tees need to wind-up pointed in a specific direction. By the time you factor that very important element in, and allow for manufacturer's tolerances on fittings & nipples, and field threading of pipe, you're pretty much back where you started - screwed piping is as much art as science.

RE: Torque for NPT connections

Generally speaking, what fails if you over-torque, no wait, I mean over-tighten, no wait, that's the same as torque, I mean over-turn, no wait, I don't mean tip over, I mean over-rotate (whew!) the joint?  If a plumber gets to the point he artfully feels is just right, then has to rotate around to point the elbow northeast, is he going to hurt something?
There's some orientation 'allowance' designed into pipe threads, but how different is that allowance in different materials and material combinations: steel, stainless, brass, copper, aluminum, PVC?  

RE: Torque for NPT connections

Far and away, steel pipe & cast/malleable iron or steel fittings are the most commonly threaded. There's some stainless. The tricky one is PVC. It's a fine line indeed between having a "weeper" and splitting the female fitting. If dissimilar materials are involved, always try to get a male PVC thread screwing into a female metal fitting or flange. PVC is usually glued, anyway. Most copper joints are soldered, grooved, crimped, etc. However, there are a LOT of really crappy threaded copper adapters on the market these days. The only brass threads usually seen are female, on things like valves, etc.

Good screwed fittings will give the feel of gradually tightening. With some practice, you can usually tell if you can get another full turn on them, or "this is it". There's some 150 malleable fittings on the market that don't always give you that "gradual" sensation. They just GRAB. Because it doesn't A/ stop in the right orientation, or B/ "feel right", they usually get removed. Quite often, they've galled, and you wind up tearing the threads up so badly you have to start over. That is very frustrating, and we avoid 150 malleable fittings whenever possible.

Also remember that "plumbers" and "steamfitters" are two separate trades. A LOT of piping problems (not just with screwed connections) are the direct result of having plumbers install & maintain steam, condensate, compressed air or process piping. Not nearly all piping is plumbing. I have great respect for a lot of plumbers, but I'd rather work on a 500 PSIG superheated steam system than plunge a toilet :)

RE: Torque for NPT connections

What a great topic and such interesting responses!

Tapered pipe fittings have many issues - even though modern life would be unimaginable without them. If you have the option to redesign your parts - another option is to look at BSPP (British Standard Pipe Parallel or ISO G) threads which are parallel - like a bolt, and rely on a sealing washer under the head of the fitting. This has much the approach as torquing a  standard fastener.

The Swagelok catalog has good information on this topic, FESTO have a huge range of ISO G threaded fasteners which rely on gaskets to seal. (no plugs-just used these parts before)

Another interesting method of joining pipe fitting is sanitary flange type arrangements.

These approaches tend to significantly reduce the forces on mating parts - but the sanitary flanges are not suitable for high pressures obviously.

RE: Torque for NPT connections

use a bigger wrench. :oP


RE: Torque for NPT connections

I have been using 1/8" NPT and 1/4" NPT Parker and Swagelok stainless pipe fittings for years, but when I went to modify a natural gas line, which was 1/2" steel NPT, I overtightened it and the male part snapped after sitting a couple of days.  

I never got the feeling the torque was increasing so I kept tightening it.  With the 3/4" and 1" pipes I did not overtighten it because there was a definite torque change.  Next time I would follow the directions in the Parker catalog.

RE: Torque for NPT connections

I went all down this thread and did not see formal mention of one thing that is probably obvious to most, but might be helpful saying to some.  In my opinion it is necessary to "start" the threading properly and in the correct alignment (if everything is perfect in the machining, I guess this would be with the male thread item aligned with the axis of the tapped hole, and that would normally be perpendicular to the tapped main pipe or fitting etc. axis).  Once this is done, the threading should of course start very easily and torque gradually increase with tightening as others have explained.
I guess it is conceivable direct taps in curved as opposed to flat/boss surfaces may increase the necessity for care in initial alignment, and (discounting issues like leaks or splits in plastic pipes) I wouldn't be surprised  cross-threaded members are perhaps more likely to cause leaks than variable torque.   

RE: Torque for NPT connections

Cross threading is a major issue as you say. And the stabbing and preliminary assembly is very important for the succesful end result. Normally an experianced assembler has experianced this before and knows that he has to throw away those parts. Also after several times he cross threads a connection, he will know that feeling and remedy the situation. If he don't, and wish it away, as soon as the pressure is applied, that connection will be telling him to fix the problem. My personal belief is over-torque is the biggest problem. Over-torque can produce a failure weeks,months,or years later. Then all that happens is a lot of finger pointing.

RE: Torque for NPT connections

As a couple of people further up the thread have said, in the upstream oil industry premiuim threads (ie non API threads) are always made up to a particular torque level (or more accurately, a particular torque range) using a power tong and a JAM unit that records the torque as the joint is made up.  These premuim connections are gas tight and do not require dope for the seal, unlike API buttress or round threads, but either seal on end of the pin into the box (typical for most VAM type connections) or on a shoulder half way along the thread (typical of Hydril connections).

A connection is made up and then accepted or rejected depending upon the shape and value of it's torque- turn curve: usually a straight line as the joint is made up, and then a curve as the torque goes into the joint.  If the curve is the wrong shape, or fails to reach a certain value or starts too soon, the connection is broken out and redone, or rejected.  Misalignement, cross threading etc are all easily identified as the graph looks wrong- the connection takes torque too early for example.

The type of dope, with a known friction factor, is closely specified, and increasingly is being pre-applied at the storage yard before the joints arrive at the rig to save time and also to ensure optimum dope application.

RE: Torque for NPT connections

Torque is required for sealing in downhole premium connections only to prevent connection from backing off downhole.    The torque is designed into the connection.   You can make low torque test plugs for premium connections to prevent the wear and tear and time required to torque equipment during pressure testing.

As for the NPT, ANSI/ASME B.120.1 says for a pressure-tight joint, the connection must be made-up wrench tight with a sealant.   Wrench tight actually has a dimension in the ANSI spec for hand-tight and wrench-tight. From this, the guidelines for making up finger tight and then tightening an additional 3 turns for sizes up to 2" are derived.  For larger sizes it drops to 2 turns past handtight.

If you don't use sealant, a leakpath will exist where the sealant would normally be.  There is only one though, not two like was mentioned un;less I have lost my mind.   

RE: Torque for NPT connections

You may want to look into the leak paths again.
A thread has a root and crest. Each are in a leak path zone. Normally the thread flanks are engaged metal to metal, however the roots and crests are in clearance. That means two leak paths. I concur with your torque theory in a way, if the threads do nothing but hold the connection together and a have separate "metal to metal seal" like premium oil field connections, then the "metal to metal seal" seals off the pressure. But NPT's are supposed to seal and hold the connection together like API 8-Round, Buttress, and Line Pipe (NPT). In order to seal, you have to have radial interference. That is where the turns come in. In order to have turns you must have torque. That is where the torque comes in.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members!

Close Box

Join Eng-Tips® Today!

Join your peers on the Internet's largest technical engineering professional community.
It's easy to join and it's free.

Here's Why Members Love Eng-Tips Forums:

Register now while it's still free!

Already a member? Close this window and log in.

Join Us             Close