PDA

View Full Version : Impossible Task?



BigBoy1
09-02-2011, 09:09 AM
I need to produce a tapered steel rod, 12" long with one end being a 0.5" diameter circle with the other end being ellipitical shaped with the ellipse having a major axis of 0.375" and the minor axis of 0.125". The taper has to be smooth from one end to the other.

The tools I have are manual, a lathe with taper attachment and mill, each with a DRO. I don't know of any way to do it with what I have. Are there any suggestions on how to make it with the tools at hand or am I embarking on an impossible task? Thanks.

Chris S.
09-02-2011, 09:27 AM
For the ellipse: Can you do a series of rough plunge cuts to remove the bulk of material, finished off with masterful filing? ;)

Chris

Toolguy
09-02-2011, 09:49 AM
Simply center drill one end to put on the live center (.5 diam. round) and chuck up the other end in one of those cheap Chinese 3 jaws everyone is always grousing about and take a cut. You'll probably come out spot on.:p

How about this - Same setup with a 4 jaw, use indicator to offset the right amount in 2 directions, cut to size. Rotate part 180 degrees, cut other 2 sides. Might need to draw file to blend transitions.

Scottike
09-02-2011, 09:57 AM
You didn't specify the distance between the axis on your ellipse, so I'll just assume it's .5" (.375r + .125r ?).
I think you can forget the lathe and taper attachment, It's all on the mill.
(your stock is .5" dia.?)
Basically, your profile is similar to a small sailboat mast, so if you take a look at a copy of Chapelles Boatbuilding it would tell you how it's done in wood.
The difference is you'll be using a mill and end mills rather than a plane and spokeshave.
Depending on the accuracy and tolerances you need, you could layout the profile of the ellipse on the end of your stock and "eyeball" your rough cuts to shape or sit down and do the math to get your tangent angles, depth of cuts, taper, etc.
Once you have the profile roughed out, then it's a matter of smoothing the high points with a file/abrasive paper.

edit: aren't you afraid a martini stir stick made of steel will break your glass pitcher?

macona
09-02-2011, 11:33 AM
An ellipse? You are going to need a cam set up for something like that. Or do you mean eccentric?

john11668
09-02-2011, 11:40 AM
Relieving lathe ???

No way of doing a half decent job on a centre late

macona
09-02-2011, 11:43 AM
Probably be best done on a cnc mill with a 4th axis.

Weston Bye
09-02-2011, 11:54 AM
Simple enough job for a Rose Engine, (Google it) or a lathe set up like one.
An eliptical cam on the spindle coupled to the cross slide would move a tool post grinder in and out to generate the ellipse. Working between centers, offset the tailstock to get the taper.

Or maybe couple an eliptical cam on the spindle to the taper turrning attachment.

Mcgyver
09-02-2011, 12:01 PM
eons ago there was a build of an elliptical turning device in Model Engineer, its primary intended use iirc was file handles, that sort of thing. If you mounted one end in that at the chuck and the other on centre in the tailstock, you'd get what you want using the taper attachment....in theory! Given how long and spindly it is, with seemingly no way to use a steady, it probably is impossible. A part like this almost assuredly isn't a mating part, correct? So figure out, even if by clamping to the table and milling steps, how to rough out on the mill then file and sand to finish

For the elliptical chuck if you want to look at that, there are good indices for ME out there so finding it isn't as impossible as it might sound....wait...i have the index here! here you go

vol 112 issue 2799 A Novel Elliptical Chuck, C.E.Saunder: A Device for Oval-to-Round Turning Between Centres.

Gravy
09-02-2011, 06:17 PM
McGyvers post triggered a thought:eek:

Instead of turning a taper and then somehow creating an ellipse on one end, would it be better to start with straight elliptical stock and turn a round taper on it?

Creating the straight elliptical stock is left as an exercise for the student.:p

Forrest Addy
09-02-2011, 06:30 PM
Some jobs are not suited for machining.

You're describing a linear transition from 1/2 dia to a 1/8 x 3/8 ellipse over a foot.

You can screw around forever trying to sculpt semi-organic shapes with machine tools or you can forge the shape you describe. I've seen good blacksmiths in action. A good smith with an eye for form could knock out that shape (or a dozen) in less than a hour (twenty minutes each for a dozen.) The finish if the smith takes care will be black and glossy ready for fettling if needed.

Actual time to forge depends on the tools available. I see extensive use of rounders and flatters here but I'm not much of a blacksmith. Lazlo? Didn't you just complete a basic blacksmithing course?

If I had to make one to three, I'd forget machining. I'd start with a 1/2 dia rod and rim punch the ellipse in one end. Then off to the 6" x 48" belt sander I'd go taking with me a 12" scale to track straightness. I think it would take me about an hour each and a couple of belts of different grit. Take a little time and care the final shape will be acccurage and the final appearance will be glossy smooth and free of sanding and file marks ready for polishing if desired.

Gravy
09-02-2011, 07:15 PM
I'll take a wild guess here: this sounds like a shift lever for a car. If I'm right, belt sanders are your friend. It just has to LOOK good. If I'm wrong and it needs to be a mathematically smooth transition, make friends with a CNC shop.

darryl
09-02-2011, 10:09 PM
One fairly obvious problem is that the piece is so long, but slender. It's not going to resist normal cutting pressures without support. A travelling steady rest would work if you could make the toolpost follow a guide bar that is driven at one end by a cam. You would basically have to face the end of the rod and advance towards the headstock bit by bit. I can see a way to do it but it's a lot of fooling around with temporary fixtures.

A straight guide bar would be situated behind, but at the same height as the workpiece. It would pivot at the headstock end and be driven fore and aft at the tailstock end. A follower moves the cutting tool fore and aft. The tailstock end could be driven by an eccentric machined on the end of the workpiece, which is supported by a live center. After all the machining is done, you would remove this end.

The cutter itself is going to be a trick- as it moves fore and aft, it will also have to rotate through several tens of degrees back and forth so it presents a workable rake angle to the elliptical shape at all times as it's cutting it.

Don't forget to video this process when you get it all up and running- there's going to be a lot of compound action going on :)

darryl
09-02-2011, 10:47 PM
It has also occurred to me that if you took a line down the rod at any point in the rotation, it would be straight. This means that you could set up on a mill, rotate the workpiece bit by bit, make a cut along the length each time, and raise or lower one end of the rod by the right amount each time so that end gets the final shape you want. The cutter would probably have to look like a saw blade, but which is tapered towards the rim leaving only a nose radius on the teeth. It would cut a groove down the length of the workpiece at each pass. Once you have gone through 360 degrees of rotation, the deviation from the desired shape would consist of the tiny scallops produced by the nose radius. You would then turn the piece in a lathe and file until all those roots just barely disappear.

If we assume that you would turn the rod 4 degrees each time, you would need to make 90 passes under the cutting wheel to get the job done. For most of those passes, you could get away with fixed support blocking under the rod. Once you reach a point where the machined portion needs blocking, you will have to reposition that support each time you rotate the rod. Tedius, but can be done.

You would need support under the bottom and at both sides of the rod if you are to keep it in place under the pressures of machining. Only the first cut would require bottom support only- after that the pressures are downwards and sideways.

This is going to be considerable work for the cutter since it will have to go in 3/16 at one end of the cut at its deepest. I might be a bit optimistic suggesting 'only' 90 passes-

Cheeseking
09-02-2011, 10:54 PM
Okay, I have no clue how to make what you described on your manual machines. But please do me and everyone else here a favor and divulge what the friggin heck this proboscis is/used for...unless of course it involves national security.

The Artful Bodger
09-02-2011, 11:53 PM
No, it is not impossible, but it might be quite a bit harder than a lot of things we might attempt.:)

Then on the other hand it might actually be quite easy!:)

Make a device that will mount on your tail stock, it will have a centre but around the centre there will be an eccentric and on the eccentric there will be a link bar to a fixed point on the lathe bed.

The work piece is mounted between centres with the addition of a lathe dog at the tail stock end that will drive the eccentric. As the eccentric rotates the link bar causes the tail stock to be more or less offset (of course you loosened the tail stock offset to allow this.

The work piece will be circular at the head stock centre and elliptical at the tail stock end.

I know someone will tell me that that is only an approximation of an ellipse to which I make the pre-emptive response that you can make any damn shape you like by replacing the easy to make eccentric with a suitable cam and cam follower.:D

DATo
09-03-2011, 03:42 AM
How about turning the primary taper on the lathe and then milling the ellipse with a home made form tool ground to the dimensions of the ellipse while holding the part in a dividing head (supported at the other end with a tailstock), and angling the dividing head in the Y axis to the required taper?

Weston Bye
09-03-2011, 06:45 AM
Something so long and slender will not be turned easily. That's why I suggested grinding in my previous post. Many light passes of the grinder will deflect the part less than gouging away metal with a tool bit.

The work will have to be done between centers with the tailstock offset, unless the taper attachment is used.

A pseudo ellipse can be obtained by making an offset center to be held in the 3-jaw or 4-jaw chuck, presuming that the small end of the part would be produced at the headstock end of the lathe. This would require two operations - grind one side of the ellipse, then turn the part 180 and grind the other. Or offset the center in the opposite direction but leave the part in the same relationship to the spindle. Some kind of special lathe dog would be required to work with the chuck rather than the faceplate.

That would still only approximate an ellipse.

As I suggested in my previous post and The Artful Bodger also proposed, there would need to be a coordinated movement between the spindle and the cross slide, or the part in relationship to the cross slide.

A past issue of HSM featured a true ellipse milling attachment for making small oval boxes. Probably wouldn't do the OP much good though.

BigBoy1
09-03-2011, 07:04 AM
I didn't think it would be an easy job. What I'm making is a spike bayonet for one of my rifles. The original bayonets are like hen's teeth but from pictures and the socket on the rifle, I figured I could make a good approximation so from a distance it will look like the original but not pass a close inspection (maker's marks, proof marks, etc.)

Cutting the taper, 0.5" dia. on one end to 0.375" dia. on the other is doable on my lathe. The belt sander and milling ideas seem to be the way to go. I'll have to look into them further. Building a camming mechanism for just one item seems to me to be overkill.

Thanks for the suggestions and help.

Weston Bye
09-03-2011, 09:02 AM
Now that we know something about the original part, the question has to be asked:

What would George Wilson do?

RussZHC
09-03-2011, 11:23 AM
Gotta go with a mix of Forrest and gravy's suggestions.

To me all the time spent fussing with whatever device you come up with to do this by machine and you could be done (assuming its a single piece you are after...)

Find a computer app that will generate the cross sections you want so as to make gauges for checking [say every half inch or .25"], mark the raw stock the matching intervals and go at it with really coarse belts, progressing until you are close and then finish to degree you want/need.

Scottike
09-03-2011, 11:55 AM
I still think the mill is the way to go, (tools on hand and all that) but just another thought here - using a piece of 1/2" flat stock would allow the piece to be more securely supported than 1/2" round stock and would allow three sides to be milled before being cut from the stock for final finishing.
Forrest's blacksmith thoughts sound the best, but does the OP have access to one, or the tools?

DR
09-03-2011, 12:55 PM
If a universal mill (swivel table) or a CNC mill is available it's a piece of cake.

Lay it out in 3D CAD, divide the circle into segments, divide the ellipse into the same number of segments. Connect the segment ends to give straight line cutter paths. Use fine enough segments so faceting is minor. Hold the blank in an indexing head with a center at the other end.

If neither of the machines above is available, mount the indexing head and center on a swiveling sub-table to adjust for the angles of the straight line cuts.

Obviously because of flex over the 12 inches there's a problem. Come up with a way to eliminate flex and you've got it.

If the flex is too hard to overcome, do the part in two lengthwise halves. Use small screws to atttach the halves to a rigid backer. Mill each half and silver solder together. Sand as needed to remove facets.

precisionmetal
09-03-2011, 01:47 PM
If it needs to be perfect, just drop it out of a 12" long piece of material in one shot on a wire edm machine that has 12"+ of Z.

About 18 sq. inches of material removal... a bit difficult to flush due to height, but completely possible. Estimate 1.5 hours of wire time on a modern machine.

darryl
09-03-2011, 04:30 PM
Hmm- Seems that dimensional tolerances wouldn't be critical at all, just the final look would need to appear realistic. Seems like you could turn a smooth taper first, leaving one end at 1/4 inch diameter (use a travelling steady and the compound to machine an inch or two at a time) then heat and hammer the tapered end to the width and thinness that you want. The rest is filing, belt sanding, etc.

BigBoy1
09-04-2011, 06:41 AM
Hmm- Seems that dimensional tolerances wouldn't be critical at all, just the final look would need to appear realistic. Seems like you could turn a smooth taper first, leaving one end at 1/4 inch diameter (use a travelling steady and the compound to machine an inch or two at a time) then heat and hammer the tapered end to the width and thinness that you want. The rest is filing, belt sanding, etc.


I agree that the basic form factor is not hard to make (a tapered rod) but the finishing details is were the work will have to be done. I was trying to do the job myself on my equipment but will definately keep an eye out for a CNC machine to use. Appreciate all of the inputs.