View Full Version : 'Custom' Tool question: Measuring Manifold nut inside angles?

T. Cotten
07-25-2017, 12:34 PM

I encounter a lot of different manifold nuts of a lot of different models, and it seems now some reproductions vary significantly from OEM on the inside angle.

It is easily accommodated if it can be accurately measured.

This is posted in "Custom Tools" because I probably must make the instrument, but I haven't a clue what it would look like.

Any ideas?

Thanks in advance as always,


T. Cotten
07-26-2017, 01:11 PM
So then,....

How would a "pro" measure it?

Geometry isn't working for me.


07-26-2017, 11:40 PM
With a CMM.

07-27-2017, 07:18 AM
I would get a rough idea of the angle using a protractor, then I would mount it to an angle block. Put the angle block on a sine plate and adjust it until the angle was parallel with the surface plate. Or cut one in half and use an optical comparator.

T. Cotten
07-27-2017, 09:38 AM
Thanks Mark!

What's a CMM?
(And is it affordable?)

And Big Lake Bob!

What kind of protractor, or angle block fits inside of a manifold nut?
(Some are only 1 5/16" ID.)

My customers aren't anxious to cut up their vintage hardware, and cutting up new ones is pretty expensive, particularly if you never know what angle the next batch will be.

The problem arose when a fellow's mass-produced PEEK seals failed in modern nuts.
(I have never had an issue with my machined seals, but then I don't use any of the same parameters.)

So I cut some round stock to 60, 61, 62, etc., inked the bevels, and rubbed them into various nuts, OEM and reproduction.
Two degrees is significant, and apparently enough to make a seal fail.


07-27-2017, 11:04 AM
You could make a casting of the angle, and lip of the manifold nut with Bondo. With careful casting of just a section of the inside of the nut, you might have enough of a sample to measure the angle. I've had this same problem with trying to determine an angle inside a gland nut, or packing nut. . . . Tough nut:)

T. Cotten
07-27-2017, 12:05 PM
I considered sealing wax, Eric!

But the bevel is actually wider than the ID of the threads. I would have to break it out, and once again, we do not want to damage the hardware.

So then I thought if I could put a dial indicator in my toolpost set at 90 to the nut in the chuck, I could measure the geometry accurately enough, but no.
One arm of the indicator is fixed, and cannot follow the bevel.

Perhaps I'll hold it with a spring.


07-27-2017, 12:39 PM
Tom, this is the method I was talking about to roughly check the angle, and NO, the protractor is not set at the manifold angle. Just quick pictures to show there is an available means. I like your lathe set up to check the angle, but why not use a standard dial indicator and run it back and forth using the compound while adjusting the angle on the cross slide until the indicator reads '0' all the way across the angle. Looks like a quick easy means to me, great simple idea.

And may I suggest that if you use your lathe setup as you pictured, turn the nut 180*. It will be easier to use an indicator as I said if you are not trying to clear the length of the threads. Basically go from the opposite side.

T. Cotten
07-27-2017, 01:22 PM
Thanks Bob!

My protractor lacks a tiny adjustable cross-slide rule like that. I was about ready to cut it up.

I understand using the lathe postholder to read the angle, but you lost me with the 180 flip. Working in the open side of the nut seems easiest to me!


07-27-2017, 02:21 PM
It is easily accommodated if it can be accurately measured.Depending on what you consider "accurate," a set of angle plates might be all you need:


They're the right size to insert in that nut and if the internal angle of the nut is the same as the plate the bottom surface of the plate will be parallel to the outside flat on the nut. Up to 5-deg. it only takes one plate, and beyond that two plates held together would give 1-deg. resolution.

Of course, there are more accurate ways of making this measurement, some of which have been mentioned by others, but none are nearly as quick as using these angle plates. Again, whether these plates are accurate enough for what you need depends on what accuracy you need.

T. Cotten
07-27-2017, 03:41 PM
Golly BoschZEV!

Just holding a straightedge against the narrow bevel is quite a trick, and squaring it up with my protractor doesn't reproduce.
And locking the nut down on the mill way robbed me of feel altogether.

So I would have to be schooled as to how to stack and apply the plates, unless they are magnetic!
This is one more reason to ask some local mentors, especially if they might have some plates..

At this point I would be happy with readings reproduceable to the degree, although I suspect a half a degree can make a difference in the compression zone on the seal.


07-27-2017, 04:55 PM
So I would have to be schooled as to stack and apply the plates, unless they are magnetic!
At this point I would be happy with readings reproduceable to the degree, although I suspect a half a degree can make a difference in the compression zone on the seal.To test this I went out to the garage and, after looking for something to test it on, decided on the L00 taper on a lathe chuck. I have my mill leveled, and the door jam should be pretty close to plumb, so I held the plates against the taper with my finger and sighted between the edge and that of the jamb. Eight degrees was too little and 9-deg. (i.e. a 4-deg. and a 5-deg. plate held against each other and against the taper by a finger) looked perfect. I then looked up what the taper really is and found it to be 8-deg 17' 50" (i.e. ~8-1/3 deg).

Since it is easy to see the difference of 1-deg. this way, and since the door jamb easily could be off by 2/3-deg., my conclusion is this technique is good to an absolute accuracy of at least 1 deg. Maybe it's good to 1/2-deg. but I'd hesitate to claim that unless I tried it with a rod in the background adjusted to be perfectly vertical.

As the (modified) saying goes: fast, cheap, accurate -- pick any two. This one gives you fast and cheap. Although, 1-deg. accuracy isn't all that bad (unless your application requires better than that).

07-27-2017, 05:19 PM
I found this site that sells neodymium magnets, and I was thinking; if you could find one narrow enough to fit in the nut, and stick to the angle, you could then stick the base of that magnet to a piece of ground steel to establish your sine. You could then use basic trig, or angle plates, or a protractor to find the angle. As I said, Tom, I've had this problem as well so I am hoping you, or someone else comes up with an elegant solution.

https://www.duramag.com/neodymium-magnets-ndfeb/?_ctcoderef=goog-ppc&gclid=Cj0KCQjwnubLBRC_ARIsAASsNNlujYrzxo8NTIEj7maA JAUilT6ODMF1TyJ_7P_tsXQTVswFvX1xJFIaAncXEALw_wcB

T. Cotten
07-27-2017, 05:58 PM
Well Folks,..

First, I trial-and-error cut some stock until a rubbing showed great contact.
Then I took BigLakeBob's advice, and set up a direct indicator, and adjusted the compound to follow the bevel with the toolpost.

Zero math is a good thing.

Both seem quite accurate, but they do not agree. Not even close.

PS: So I take a bunch of time and measure out three Large Port nuts: The OEM, and two popular reproductions, one park'd and one cad.
The disparity among them wouldn't bother me so much if the measurements didn't reproduce. Looks like a spread of eight degrees.

Hope to average more OEMs before I have to set up for other things..

T. Cotten
07-28-2017, 01:16 PM
Ignorance is bliss, Folks.

Now I know things I wish I didn't.

Using BigLakeBob's method, apparently accurate to the half o' degree on my Logan, I busted Liberty's R&D time budget logging the bevels found in H-D models in my stash and inventory.
And regretfully, it appears I shall be measuring all that come in. A couple of degrees variation of the OEM examples is perplexing enough, since I have to cut seals with no room for failure, but the eight degree spread for the reproductions is disturbing.

Indian models will be another forensic time-sucker,
Thanks to all...


07-28-2017, 03:29 PM
Just a thought, hot glue from your wife's hot glue gun then unscrew it and measure?
I use it for all kinds of quick molds to measure and fixture fragile parts for machining.

T. Cotten
07-28-2017, 04:43 PM

Most nuts have a relief inside of the threaded portion, so anything cast-in would be locked from un-screwing. That led to my comment about breaking out sealing wax.

BigLakeBob's suggestion solved the problem most eloquently, although opening up many other questions.

Like: How have seventeen years of my own seal productions succeeded inspite of reproduction nuts!

PS: I've got all of the glue guns, stored away. She doesn't use them either.
PPS: My seals were cheaper in 2000. Now they must go up again.

07-28-2017, 04:52 PM
You don't have to fill entire area just a section hold nut at 45* give a couple squirts let harden then peel out.
I'm sure 1000 ways to skin it as long as it works.
Glad you got it!

T. Cotten
07-28-2017, 05:05 PM
Its a matter of tactile control measuring what's left, Bob!

Eyeballing a sliver of anything with a protractor isn't accurate enough.
(Needle points are a treat.)

With the nut chucked up and an indicator on the toolpost, the lathe does the rest. Don't even have to turn it on.
The operator is pretty slow though.


07-28-2017, 09:00 PM
If it's that critical send it to me and I'll put it on the CMM and send you a digital drawing of it. B

T. Cotten
07-29-2017, 09:00 AM
If it's that critical send it to me and I'll put it on the CMM and send you a digital drawing of it. BI appreciate the offer, Bob!

But I get a wide range of manifolds constantly (four different ones this week), and you never know what nuts they will show up with.
Its hardly practical to mail nuts off and wait for them to come back, and I couldn't expect you to do it for free from here on out!

From the sampling I already have, its obvious that neither OEM nor reproduction nuts (for Harleys anyway) can be assumed to be perfectly consistent.
BigLakeBob's simple suggestion is quite accurate, and I only have to get up to speed where I don't lose much more time that I can't charge for: I'm too expensive already.

(But its obvious that my prices must rise eventually, if only to cover the rising cost of material.)

PS: I googled "CMM", and found I could neither afford one, nor find space for one. Once again, low-tech rules!
PPS: Its now six different manifolds this week, and the UPS man isn't here yet. I'm going to have to lock the doors 'til after Davenport...

T. Cotten
07-29-2017, 09:41 AM
If Bob and Mark and others would like to contribute, Folks...

We could all start sampling various nuts, and compare our findings here.
Then everybody could benefit.

PS: As a follow up, it now appears that the failure I mentioned earlier in this thread was due to a very high angle (whatever the lathe says it was...) nut used with a seal that was most likely patterned off of originals at a much lower angle. Such geometry biases the force of the nut against the inlet nipple at the very outside edge of the seal. The critical sealing is needed inside against the spigot.

08-06-2017, 09:21 AM
Sorry to jump into this thread late into game. I have been away visiting the spirit world for a week by having a ketamine infusion pain control treatment. Boblakes quote sure has a lot of validity after going through that treatment.
Now to get back on track, I will call the angle side the front side of the nut. One thing I have noticed is that the readings, using the flat on the back side of the nut, to take the measurements on the front side of the nut will give you a trigonometric sine error because the angle machining process is machined from the front side of the nut. Now this would be fine if the front side and back side were completely parallel, which I am sure they are not because the back side is not a critical measurement and does not have to be held critical, to the front side of the nut, while being cut off. My whole point being is that the front side needs to be indicated in flat to get a proper angle measurement, this can either be done on each individual part before the angle measurement is taken or check each part first to see if the front and back side are completely parallel with each other. The backside must also be burr free by either running a new flat Norton stone across the surface or by running a fine file lightly across the surface. If the front and backside are parallel, then the backside can be pushed against a stop to measure the angle.

If this makes sense to people on this thread, I will continue on how to double check boblakes compound measurement mathematically. It is very easy to do, but you have to use a scientific calculator.


T. Cotten
08-06-2017, 10:44 AM

When the nuts are held in a three-jaw chuck against their 'back', the run-out of the 'front' face is readily apparent.
Most all have been quite parallel, front face to back, and if there is any run-out on the front face, it only takes a gentle nudge to true it within the jaws.

Once again, the mass-produced seal's failure was due primarily to reproduction nuts with gross error on the angle.
Considering how many of these nuts I fitted with my own seals without any reported issues, I must attribute success to my different parameters for PEEK.

Now I can measure all the various nuts for a much more accurate fit.

It was surprising that fresh brass seals from the same producer as the nuts were similar in error, so as long as you didn't mix sources of nuts and seals, they had a chance.
(At least as much chance as any brass seal, anyway...)

The age-old question comes to mind: If somebody is going to the trouble to make a part, why not make it right?

PS: I hope to measure OEM and reproduction nuts for Sport Scouts, Chiefs, WLs, etc. soon.
If nobody else wants to contribute their findings, I guess I shall keep my own as "proprietary".

08-06-2017, 12:52 PM
I sure would report findings, but I do not have any parts to check.

08-08-2017, 07:41 AM
20352Sometimes I'll use magnets to hold a part at the front of the jaws, quick method that saves a lot of indicating, assuming the front of your jaws are in decent shape. Just wanted to add that if I want the part to not be flush with the jaws, I put a gage block (or equal thickness of any material) under each magnet. This way if I want to machine .05 off a part I might stick .100 blocks under the magnets. I find this method to work well on thin parts.

08-08-2017, 08:13 AM
DANG ! That is a good idea, Bob. The only downside I can see, is that it would magnetize your chuck jaws and collect chips. However, the benefits would outweigh the deficits.

08-08-2017, 08:30 AM
DANG ! That is a good idea, Bob. The only downside I can see, is that it would magnetize your chuck jaws and collect chips. However, the benefits would outweigh the deficits.

I have not had that problem, the small magnets I use don't seem to have enough to induce magnetism into the larger mass of the jaw.

08-08-2017, 08:49 AM
That is a new, great idea, to me as well Bob. Thanks for the tip!

T. Cotten
08-08-2017, 11:24 AM
DANG ! That is a good idea, Bob. The only downside I can see, is that it would magnetize your chuck jaws and collect chips. However, the benefits would outweigh the deficits.Just need a big de-Gausser, that's all Eric!

PS: I have two of these, and have never tried them. But I de-Gauss floatvalve needles with a jeweler's coil.