I am working on a project that needs some shallow 1/8- 27 npt holes. There
is no way around it there just isn't the space so I am going to use a short
projection tap. I will be tapping using a verticle mill with the work held
in a vise. Anyone got some advice? Why do npt taps need so much lead? This
has bugged me for some years.
I think that that "so much" lead you refer to is because the thread is
tapered (1 in 16) so that when the tap is driven into a straight bore
hole it has to remove more metal that a straight thread tap would.
You might want to experiment with putting a slight taper (about 1 in 16)
in the tap drill sized holes to make it easier for the tap to do its thing.
What Jeff said. Pipe threads rely on a 100% thread in order to seal
properly. The tap drill must be smaller than the minor diameter at the
tap's full depth so you end up with the required 100% thread.
When tapping with a short projection tap, you must either use a standard
tap, followed by the short projection, or ream the proper taper before
To tap your holes, use a center in your drill chuck to follow the tap as you
turn it with an adjustable wrench. A very nice center can be made from the
shank of a 3/8" end mill, assuming you have either a toolpost grinder,
cutter grinder, or a precision grinder at your disposal. It's important
that the center be *on center* so you tap runs vertically. Once you have
it, you'll keep it for just such use. I have three in my toolbox. Keep
the tap well lubricated, and back it up every 1/4 turn to break the chips.
My common practice is to remove the tap at about one turn before full depth
and clean out the hole and lubricate well once again, then take the last
turn or so as required. That way the finished thread is nice and clean.
Stop tapping when you have 5 threads above the finished surface showing on
the tap. That will put you in the ball park for a proper thread
diameter/depth, assuming you don't have a plug gage at your disposal.
Not really Harold. The common NPT pipe threads rely on some kind of
"pipe dope" sealant to properly seal. Permit me to pontificate:
Take a look at the thread form of NPT threads in Machinery's Handbook
and you'll see that the crests of the male and female threads don't come
to a sharp point, but are flattened. The roots of the threads are closer
to a sharp point, so the crests don't fully fill the roots. This is done
on purpose to make sure that the flanks of the threads will be sure to
contact each other to support axial loads on the joints.
That flattening of the crests of the threads leaves a small trapezoidal
shaped path spiraling along the full length of properly torqued engaged
threads, sometimes referred to as a "spiral leak path". It's that path
which has to be filled up with some kind of "pipe dope" to keep it from
leaking, and the choice of what dope to use, paste, teflon tape or
sometimes string varies with the pipe's contents and the installer's
preference. I still prefer using Rectorseal or a paste type dope over
teflon tape for water service.
It's not always easy to make sure the teflon squishes into where it
needs go to completely fill the spiral leak path. And, backing off a
joint a bit after tightening and then retightening it can displace the
teflon so the joint won't seal. That generally doesn't happen to me with
paste type dopes while they are still "soft".
NPTF threads are what's called "Dryseal" threads. The last letter in
that appreviation, the "F", stands for "Fuel".
When the NPTF thread form was originally developed it wasn't easy to
find pipe dopes which would withstand fuels, so a thread which would
seal without needing pipe dope was designed. NPTF threads have a sharper
crest on the male threads and are designed to create a metal to metal
crush seal along the spiraled thread path. They are often found on
fittings made of brass, where the mallability of that metal makes it
easier for the crush seal to work.
A few lines swiped from Machinery's Handbook describing NPTF threads are
say it better than I can:
American National Standard Dryseal Pipe Threads for Pressure-Tight Joints.
Dryseal pipe threads are based on the USA (American) pipe thread;
however, they differ in that they are designed to seal pressure-tight
joints without the necessity of using sealing compounds. To accomplish
this, some modification of thread form and greater accuracy in
manufacture is required. The roots of both the external and internal
threads are truncated slightly more than the crests, i.e., roots have
wider flats than the crests so that metal-to-metal contact occurs at the
crests and roots coincident with, or prior to, flank contact. Thus, as
the threads are assembled by wrenching, the roots of the threads crush
the sharper crests of the mating threads. This sealing action at both
major and minor diameters tends to prevent spiral leakage and makes the
joints pressure-tight without the necessity of using sealing compound.
All that's well and good if the male and female threads are both
manufactured to specs, something that doesn't always happen,
particularly with some of the imported junk being sold in the big box
stores and elsewhere nowadays.
I have a spring loaded "tap following" center I treasure for just that
use so I don't wish for an "extra hand" to work the drill press or lathe
tailstock feed to keep the point in the tap's center hole. Looks sort of
like Fig. 1 on this page, I think it cost me less than five bucks from a
cheapo tool place.
Harold and Jeff,
Thanks for the insights.
The Short projection tap came in but wasn't short enough for the
application, I had only .350" to drill, I took the tap to a grinding house,
they ground .150" off and restored a 15 degree 2 thread lead and it did the
I was able to thru a few test holes to get the drilled depth down to .300" .
That is about the shallow limit I expect for a 1/8-27 npt tapped hole that
seals, the testing confirm or dispute this. Do you guys have some idea how
it will shake out?
Assembly is done with some sort of anerobic sealing pipe dope, testing is
helium leak blah blah blah.
Harold's suggestion that the hole be tapped with a regular projection tap
before going to the short projection tap was a confirmation of my initial
thoughts, thanks for that.
I used a spring loaded tap follower to do the regular projection tap, the
short projection tap didn't have a centering pilot on the end so I just
followed the lead of the previous.
I agree that the spring loaded is sweet for small taps but a solid center is
the only way to go when you are using a tap that you don't have a tap wrench
big enough to drive it and must go to a wrench. It happens to me in the
1/4 npt size, I hung a 10 pound bag of shot on the quill lever using a
solid center and that worked.
Jeff, are nptf threads designed to be taken apart and put back together?
Thank you both, when I see posts from you guys I read closely.
Snip lots of good stuff-------
The point is that tapered pipe threads *are* a 100% thread, they do not
provide for minor diameters on tapped holes that have clearance. By drilling
the proper (undersized) hole, the tap established the diameter, and yields a
100% thread. The length of the tap insures that enough of the hole is
removed to generate the full thread. In other words, they are not 75%
threads like the typical unified straight thread. The could permit the
sealant of choice to be easily displaced by internal pressure. It is very
desirable for the clearance, if it exists, to be minimal to prevent that
If you'll check figure VII.2, page 6 of the Screw-Thread Standards for
Federal Services, 1957, Handbook H28 --Part II, you'll see that there is a
minor allowance for maximum clearance, but the optimum is no clearance, and
a it provides for a fit with a slight interference. That for a standard
NPT. It clearly states that the threads, when *properly wrench tightened*,
with sealant, form a pressure tight assembly.
I have no quarrel with your statement about the "spiral leak path", but it
is a result of mechanics, not a desired affect. Given the ability to
produce perfect thread forms easily and reliably, it wouldn't be there. And
I fully agree, the intention is for the flanks of the threads to have
intimate contact, but that is with a wrench tightened assembly, not a hand
tightened one. Even with an interference fit on the root and crest of the
threads, that isn't a problem, due to deformation. Again, the H28
handbook addresses that.
The problem is in generating and preserving a sharp V thread form, so both
the major and minor diameters are truncated as an offset. There may, or may
not, be clearance when two pieces come together.
I, too, have a spring loaded center. Custom made, in fact. Heat treated
and precision ground to perfection, with perfect clearance on the
telescoping portion, which was precision honed. No slop. That's why I
use the solid one I recommended. The pressure of tapping with pipe taps
renders spring loaded centers useless, especially when the tap is driven by
an adjustable wrench. It's not uncommon for the tap to lean right out of
the center, which really screws up your thread. Don't use yours for pipe
Hard to say. Did you try fitting anything in the hole after it was tapped?
I'd be concerned that you may have tapped the hole oversized enough that a
fitting could bottom out before tightening due to the removal of the end of
the tap. Beyond that, sounds like you did good.
Yep! I was very careful to advise the use of a solid center for that very
Agreed, no quarrel from me. I was referring to "what is" not "what
But I'd hate to think of the results if plumbers all over the world
decided to boycott using any kind of pipe dope.
I haven't, but I don't doubt that what you say would happen if I tried
that. I went off half cocked again by mentioning it on this
thread...Nothing new for me.
Most of the things I fiddle around with are "small stuff" and I doubt if
I've ever used my spring center on anything over a 5/16" tap.
That is good, because it (the spiral leak path) is there every time.
I do not know, maybe the standard NPT threads "in principle"
do not show this spiral leak, but in practice they all do. This
makes me think the "principle" being relied in is somehow
incorrect for the application.
One can visualize the spiral leak path dramatically by overpressuring
a threaded NPT fitting that has had some kind of sealant used during
assembly - SWAK, teflon tape, pipe dope, etc.
The Swagelock folks did this as a demonstration (hyrdrostatic test)
to show how their fittings would not fail before either the tubing
itself ruptured or the thread sealant was ejected from the
spiral leak path. That's right, a spiral of teflon thread sealant
is squirted out of the NPT male/female joint when the overpressure
You can create mass spec helium leak tight NPT joints without that much
trouble. I do it regularly with plain 'ol teflon tape.
1) be sure the threads are in good condition and clean of any
previous application of sealant.
2) place two and exactly two wraps of teflon tape around the male
fitting. Be sure to have it wound the correct direction, ie, so it
tightens upon assembly. (hold the tape in your hand and screw the
fitting into it)
3) leave the first thread on the male fitting untaped.
4) work the tape into the male threads with your fingers before
5) do not over tighten.
Sounds like it would be pretty easy to create a leak-proof joint,
considering the most you're dealing with is atmospheric pressure.
Have you worked with any high pressure hydraulic fittings? I built a
filter press that used a 10,000 PSI manual pump. It's nothing short of
amazing how a perfectly tight and good joint can seep a tiny bit of oil. I
never did get it to quit. The joint would stay perfectly dry for hours,
but wet after a day of sitting under pressure.
It's a pretty sensitive technique.
Consider your hydraulic example. How much oil does it take
to make your fitting 'damp' after a day?
1 cc would be way too much. How about one cubic mm? That
would work out to 10e-3 cc of oil.
Say that oil works its way out over about one day. In seconds
that's: 24 x 60 x 60 which is about 9e4 seconds.
So the leak rate could be said to be about: 10e-8 cc/second.
The leak testers I'm talking about measure down to about 10e-10cc/sec
or thereabouts. It would blow the meter right off the scale!
It would find a leak that would take 100 days for you to see, in