What's new
What's new

Highly Parallel Nuts

ClappedOutBport

Cast Iron
Joined
Mar 30, 2016
[Solved]

Hi all,

I'm designing a mandrel for machining some parts and need some nuts with a decent degree of parallelism, probably no more than 0.0005" runout on the face when running on the threads. Size is currently ~1", but I'll need other sizes in the future. I know about bearing locknuts, but they do take a special tool to really get them snug. Is there anything else out there, or would I have to have something custom made?

Thanks,
COB
 
Last edited:
Make a split mandrel from brass, just over the major diameter of your thread. Use a flathead cap screw and countersink the mandrel for wedging it further in operation (I like to turn a 60* angle on the regular 82* head).

Turn threads onto the mandrel (with tapered cap screw lightly tightened) for whatever nuts you want to use. Make them just able to thread on (not standard PD), but still properly running.

Run a nut onto the thread, give the tapered cap screw a little more tightening, then face the nut on both sides.

It's work, but it's likely the fastest/easiest way to get what you're looking for. You'll still have to test the nuts for accuracy, a sloppy job on the split mandrel or having the taper of the cap screw not on axis with the thread will make your life more difficult. Ditto buy the best quality nuts you can find for this program.
 
The threads won't support that much parallelism so ......

So... how do threaded chucks work then?

If you're tightening against a mandrel with a square shoulder, use a Super Nut. About the easiest to torque up nut you'll ever see, and they work great. I've made fixtures using these with a bunch of parts stacked on a mandrel, worked extremely well.

Superbolt Torquenuts - Robt. L. Rowan & Assoc., Inc

That's a lot more torque than I need. Basically I am holding a part between two taped washers. One end has a shoulder, the other a thread. The part is not 100% uniform, so I'd like it to be influenced by the nut as much as possible. I assume that's the right way to go about it.

Make a split mandrel from brass, just over the major diameter of your thread. Use a flathead cap screw and countersink the mandrel for wedging it further in operation (I like to turn a 60* angle on the regular 82* head).

Turn threads onto the mandrel (with tapered cap screw lightly tightened) for whatever nuts you want to use. Make them just able to thread on (not standard PD), but still properly running.

Run a nut onto the thread, give the tapered cap screw a little more tightening, then face the nut on both sides.

It's work, but it's likely the fastest/easiest way to get what you're looking for. You'll still have to test the nuts for accuracy, a sloppy job on the split mandrel or having the taper of the cap screw not on axis with the thread will make your life more difficult. Ditto buy the best quality nuts you can find for this program.

Thanks, I understand what you are saying. Is that going to give me much better results than just single pointing in one setup? It certainly seems like a lot more work. I only need one side to be accurate, and probably only one to two nuts.
 
Hi all,

I'm designing a mandrel for machining some parts and need some nuts with a decent degree of parallelism, probably no more than 0.0005" runout on the face when running on the threads. Size is currently ~1", but I'll need other sizes in the future. I know about bearing locknuts, but they do take a special tool to really get them snug. Is there anything else out there, or would I have to have something custom made?

Thanks,
COB

Don't ride the washer on the threads.

Stop the threads way short of "stack of parts" or whatever your doing.

Make a sliding fit "washer" that has plenty of length (or thickness)
to resist any out-of-squareness from the nut, and only guide on the smooth portion of the shaft.

Nut only provides end force, not squareness.
 
So... how do threaded chucks work then?



That's a lot more torque than I need. Basically I am holding a part between two taped washers. One end has a shoulder, the other a thread. The part is not 100% uniform, so I'd like it to be influenced by the nut as much as possible. I assume that's the right way to go about it.



Thanks, I understand what you are saying. Is that going to give me much better results than just single pointing in one setup? It certainly seems like a lot more work. I only need one side to be accurate, and probably only one to two nuts.

Threaded chucks pull the chuck up against a reference plane to achieve what repeatability they have (which isn't usually as good as what you are looking for).

Putting tape on the washers won't help with making anything parallel.

That is assuming an awful lot, considering you are trying to build some level of accuracy in to your turning fixture.

Any fixture that is turned and faced in situ (aka: without removing it from the chuck) is going to be as accurate as your lathe is capable of. Use the tailstock to press on a clamping element if you do not use a threaded section and a nut.
 
Don't ride the washer on the threads.

Stop the threads way short of "stack of parts" or whatever your doing.

Make a sliding fit "washer" that has plenty of length (or thickness)
to resist any out-of-squareness from the nut, and only guide on the smooth portion of the shaft.

Nut only provides end force, not squareness.


Sounds like a plan. Thanks Doug.

Threaded chucks pull the chuck up against a reference plane to achieve what repeatability they have (which isn't usually as good as what you are looking for).

Putting tape on the washers won't help with making anything parallel.

That is assuming an awful lot, considering you are trying to build some level of accuracy in to your turning fixture.

Any fixture that is turned and faced in situ (aka: without removing it from the chuck) is going to be as accurate as your lathe is capable of. Use the tailstock to press on a clamping element if you do not use a threaded section and a nut.

It's for EDM turning not lathe turning.
 
Thanks, I understand what you are saying. Is that going to give me much better results than just single pointing in one setup? It certainly seems like a lot more work. I only need one side to be accurate, and probably only one to two nuts.

I was thinking it might be simpler (especially if you want more than one nut of the same size) to take good commercial nuts and true the faces to the axis of the thread.

If you only need one piece of each size you're doing, and are willing to single point the internal thread and complete the rest of the nut's geometry, then sure, that's likely better.

Or did I misunderstand what you were asking?
 
Expanding a thread

i'd forget about perfect threads, change the design to use a precision boss and/or socket arrangement, leave the threads relatively sloppy . that's
how threaded chuck arbors work.

I have single pointed a thread to a expanding collet and then held on the pitch of the thread. See Rovi expanding collets on you tube to see how they work.
Jim Sehr
 
We have made literally hundreds of different nut arbors for gear cutting. If the face of the nut is not square with the thread, its surprisingly easy to actually bow the arbor itself, making any parts on the arbor not run true. It takes very little runout of the nut face to create arbor runout. Easiest fix we have found is to single point a mating thread thread, leave it a little big so that the nut turns tight on the thread. Mount the nut, and turn or grind the face. Or, make the nut, single point the threads and face the end in the same setup so that its dead square.
 
Can you make the mandrel with a shoulder and for a hand wring fit of your part? Then a sleeve between the part and the nut? You must be assuming that the bore of the part is the feature to work from?
 
The register diameter provides radial location.

No. I know that much. The "register" diameter is a red herring. It has been discussed ad infinitum here and elsewhere. The combination of the taper (30 degrees on the thread) and the flat back plane make for the alignment. Which is where I got my idea from. But others were correct to point out that the back plane is not necessarily flat on my setup.

Measure some commercial backplates. They are often 10-20 thou over the register diameter. That's what I make mine to and they repeat perfectly.

i'd forget about perfect threads, change the design to use a precision boss and/or socket arrangement, leave the threads relatively sloppy . that's
how threaded chuck arbors work.

I changed my design to Digger Doug's, so the backup washer has a greater approximately a 1:1 aspect ratio of D:L. That is a more tidy setup anyway, as I could nest the extra washer inside the part. I opted to go with a bearing locknut to alleviate any fears I had about junk nuts, and to help avoid it from being over-tightened.

So I think my question has been answered and my design more or less finalized. Thanks all!
 
Last edited:
I would guess that the accuracy would increase as the clamp nut length does so in an arbour arrangement the nut may need to be 2 or 3 times longer to iron out pitch error, the effective diameter would been to be fairly high or a close fit to the thread, I’ve seen an example but the threads were ground, the thing was hardened, and surprise it got stuck, you’d need scrupulous hygiene for something like that, like surgical, I’m guessing
Mark
 








 
Back
Top