Rolled Cylinder

[kevb]

My company is involved in the rebuild of a magnetic
drum separator.Why are we having trouble with the
actual "can"?

This drum is slightly more than 8 feet long and is
30.25 ID.Material is 3/16 thick stainless plate

It is very difficult to get an accurately rolled can.
We do have SOME latitude as to actual diameter,but
as it rotates around the magnetic element, concentricity
is crucial.

I feel the rolled product,whether with one seam or two
should be welded with seams aligned,obviously.

Then open the rolls and slip the tube in and re-roll.
Is my understanding of rolling procedures faulty?

Are we not adequately specifying when we order ? Can
we expect this to be round within say,1/8 inch.Also
this re-rolling should remove the flat at the seam/s

As this is a fairly costly item,any comments or
tips would be appreciated.

Thanks in advance.

 

[MikeHalloran]

Standard practice includes pre-curling the ends of the blank to approximately the right curvature, usally with a big urethane hammer and a pipe sawhorse, sometimes using a press brake with a gently curved punch and a compliant urethane die.

Re-rolling after welding is tricky because the HAZ has different physical properties from the remainder of the article, so it will behave differently.





Mike Halloran
Pembroke Pines, FL, USA

 

[dvd]

I would recommend that you turn the o.d. concentric to the centers after welding, otherwise you are hoping on a random process to end up with what you want.

 

[arunmrao]

agree with dvd's suggestion. It is the method I adopt whenever I need to form a rolled cylinder or even join 2 or more lengths of cast cylinders .

Save water Drink Beer

 

[MikeHalloran]

It's like shaving a few thou off the od of a beercan... except, it's a very big beercan.



Mike Halloran
Pembroke Pines, FL, USA

 

[kevb]

To those who suggest turning the OD,I gotta say give your
head a shake.Impossible

This "can " is 100 inches long and 30+ inches OD. 3/16 wall
thickness.The runout can easily exceed the thickness of mat'l.

Since it is rolled from plate, a nice OD means a nice ID.
Since "actual" diameter is somewhat flexible whats really
at issue is a cylindrical object with straight sides.

We received a shell that is not too bad,still a fairly
large flat area at the seam though.

Thanks to the responders

 

[MikeHalloran]

I can think of two (high-cost) ways to correct the parts, offered in the hope that maybe someone can think of a low-cost version.

- Slip the tube over a nice piece of thick-walled pipe, of say (okay, guess) 28" or 26" diameter, and get a big guy who doesn't tire easily to wail on the seam with a big urethane hammer. He'll need a set of custom radius gages to gage his progress. He can tighten the radius by beating on the outside, but if he goes too far, it will be difficult to undo.

- Same thick-walled pipe, presuming you guesstimate the radius right from some springback tests. Secure the rigid pipe and the workpiece with seam up in a pit, and drive something heavy and slightly compliant (like the tire of a Hummer) along the seam a few times.

If you were making a bunch of parts, I'd suggest starting with a shorter, thicker, and slightly larger cylinder, and ironing the workpiece down onto a collapsible mandrel, sort of like making a soda can but on a much larger scale.





Mike Halloran
Pembroke Pines, FL, USA

 

[dvd]

It may be "impossible" but it is routinely performed.... I have seen it done on longer cylinders and larger diameters.

 

[MikeHalloran]

Agreed, it's possible, but you need a monster lathe, and a monster mandrel, neither of which are likely to be just lying around. I think it's fair to say that for a lot size of one, it's _practically_ impossible to turn the part.



Mike Halloran
Pembroke Pines, FL, USA

 

[bvasa]

If you don't mind going a completely different process direction, you might consider flowforming. You'd start with a thicker walled, short tube at approximately the finished ID, then extrude it using an internal mandrel and external rollers to reduce the wall thickness and extend the length. I'd offer to assist, but my flowformers won't handle this part length.
Google flowforming or tube spinning. You'll come up with 2 or 3 companies in the US and a couple in Germany that could handle this.
It may be tough to justify the cost for a one-off part, but it still may be cheaper than finding someone who can bore and OD turn the cylinder you described.
Good luck!

 

[budt]

Could you take two pieces that are Half the Circumference of the finished part PLUS 2', Roll them to the desired Radiuus, Cut off the Flatted Ends then weld them together?


Bud Trinkel, Fluid Power Consultant
HYDRA-PNEU CONSULTING

 

[Jarery2000]

Impossible? Give my head a shake?

Maybe in your facility its not possible. In ours we do this every day. We have cylinders rolled up to 150" long, anywhere from 20" to 100" diameter from plate ranging from 1/2" to 2-1/2" thick. The products have a T.I.R. specification of less than .030".

There are two ways of making the seam accurate. Some places pre bend the ends then roll the can. Others leave a flat on each end, roll it, then cut the flats off. You will typically get more accuracy with the method employing flats as the material stretch does not have to be calculated since they cut the total length after rolling not before.

Once rolled it is then rerolled. Grinding of the seam prior to rerolling is usually not done. Diameter accuracy is usually +/- .125". Circumference accuracy, or diameter if measured with a pie tape is +/- .060".

We then machine the od of each shell to meet T.I.R. and diameter requirements. Our maximum T.I.R. is typicaly .030" with .010" as commonly achieved during final measurement. All of our shells are custom in lot sizes of 1 so we dont have permanent set ups. Just large machines with highly trained employees

Never assume something is impossible just because it is outside the scope of your capabilities

 

[Metalfixer]

Jarery2000,
What is your company name and location?
Thanks, Larry

 

[Jarery2000]

We use Cessco in Edmonton for our rolling of cyclinders with a flat profile. They also are able to hold the same tolerances. All of our rolls we have are crowned for creating a cylinder with a peaked center which wouldn't work for your case.

http://www.cessco.ca

is their website. You would also need ot start with thicker material, likely 1/2" which will raise your cost a lot with SS.