Yes, that. It's egg-shaped before sawing.
But, regardless of when it got egg-shaped, it's egg-shaped, and it's sitting on my shop floor. I suppose I could make a fuss with the saw shop and/or the supplier. I'll do that with the next batch. I have neither the time nor the inclination to reverse course on this batch.
It turns out that the only diameter dimension that matters is an inner machined 2.878" bore, which has to be +/-.005" size and roundnness. The actual tube OD and ID are an order of magnitude less critical so the OD ranging even as much as 3.00" to 3.04" is within spec, for this customer.
This is how I'm doing it:
The size is actually uninteresting, since it's a new-ish Brother Speedio and easily interpolates within tenths.
Roundness is where all the trouble happens. Squeeze the tube, and it deforms of course. The material doesn't shrink though, so if you squeeze it .005" in one axis, it gets bigger in the other axis by ~.005".
4-point conformal contact vise jaws (actually a complete fixture, but it acts like a vise). The contact patches are ~1/2" wide, and ~90* apart. The shape of the contact patches is 3.02" and round. The rest of the jaw surfaces are relieved. In CAM, I made entities that emulate a deformed 3.02" circle, close to an ellipse 3.015" x 3.025". I believe (and the results tend to prove) that this is the deformed shape of the tube when fixtured. The fixture closing pressure is very low, just enough to keep things from shifting or chattering.
After a little practice, and a few iterations of the ellipse in CAM, I'm able to hold the bore round within .002".
Regards.
Mike
Frustrating that you do not also have control over the next step - the assembly of the 2.878" "item" - an anchoring fitting or such, if that's wot it is - INTO the tube.
If that could be inserted before de-clamping, the assembly stays a fit and task is done.
More control over the process, yet, it might be made a swage fit, even less machining.
With what I have to-hand, all-manual shop, "Old Skewl" the minimal mess-with or spend would be the Hardinge 2-Jaw scroll, tooled top plates - prolly Baltic Birch, HDF, or linen-bakelite - then faced with a stiff elastomer to preserve "as had" ellipse.
First op offline to sharpie-mark hi/lo and set them into the fixture same axis of the ellipse each go. Lathe work, IOW.
Same chuck, mill's table, ELSE the James Morton Phase 2, tiny Chandler Duplex "J" boring head is too much like work... but.
Piece count is high enough, one mought fab a bespoke-size piloted cutter for one-pass on-size bore & ream.
A bit more care as to including a clamping mandrel, this could be made into a portable tool to work "down in the ditch" or "up on the tower" for field-use, no machine shop even on the same South Pacific, Atlantic, or Caribbean island. You'd have to know remote satellite and submarine cable stations.
I guess we are on the same page, given the need to
not move the ellipse lest the bore go out-of-round on de-clamp.
But it is all SUCH a
nuisance vs simply starting with ROUND(er) stock.
2CW
