PDA

View Full Version : Twisting square tubing



Ken_Shea
08-18-2009, 10:36 AM
Any one have any success (deliberatly :D) twisting square tubing with uniform results, such as you see with some square wrought iron railings?

Ken

BobWarfield
08-18-2009, 11:03 AM
Ken, twisted quite a lot of rod in my blacksmithing exploits, but I had never tried square tubing. I'm assuming you're talking pretty large cross section, else it is easier just to use the rod. I've done 1/2" or so, but for tubing, I'd assume you're looking at 1" or larger.

Gonna take some elbow grease, leverage, and heat!

Cheers,

BW

Black_Moons
08-18-2009, 11:04 AM
Iv seen a twister before... http://www.metalcraftmachinery.com/Metal-Twister/index.htm these guys sell one. Also got some nice pics of underclothed women playing the game twister for my effort in finding that pic in google search, so all is good.

Ken_Shea
08-18-2009, 11:14 AM
Bob,
Largest would be 1 1/4" in 12 Gauge.

Rats, don't like the heat idea, too difficult to get uniform results I would think.

Just a nice smooth uniform twist, would not have to be even a complete turn, just something to give some aesthetic form.

I have access to an old and large lathe that I was thinking on using with a plug and 4 jaw chuck with some form of tail end support for the square tube. Years ago I saw an old gentleman use one for twisting solid square bar stock. In back gear you can fall asleep waiting for a full revolution.

Would be nice if it was an easy setup for a test run.


Ken

Ken_Shea
08-18-2009, 11:16 AM
Iv seen a twister before... http://www.metalcraftmachinery.com/Metal-Twister/index.htm these guys sell one. Also got some nice pics of underclothed women playing the game twister for my effort in finding that pic in google search, so all is good.

BM,
That appears to be for solid bar .

Thanks
Ken

oldtiffie
08-18-2009, 11:17 AM
Evan has a terrific very strong, long and rigid power-driven twisting machine that he used for square section steel bar.

I don't know if he used it for square tube.

My guess is that he will read this thread/post and reply.

Ken_Shea
08-18-2009, 11:31 AM
Evan has a terrific very strong, long and rigid power-driven twisting machine that he used for square section steel bar.

I don't know if he used it for square tube.

My guess is that he will read this thread/post and reply.

Well, it's been an hour :D

Ries
08-18-2009, 12:16 PM
I have done it- with my twisting machine, which is mondo.
Its a Hebo, made in Germany, 3 or 4 hp motor geared down to about 10rpm, incredible torque, and it has an electric brake so you can stop on a degree. Any degree.
http://www.usahebo.com/
The trick is that there is a tiny moment, when the square tube twists, but before it collapses. This is why you need the CNC control my machine has, and the control down to one degree increments.

Otherwise, it just collapses in on itself if you go too far, and its pretty much impossible to Un-collapse.

I have done even bigger stuff hot- I have an arbor in my yard whose posts are 3" square, 1/8" wall- we heated up about a foot with a rosebud, then used a big homemade wrench, with one end in a vise that is bolted to the floor.

The short answer is- if it was easy, everybody would be doing it.
The machine, twisting a small square solid bar.

http://img.photobucket.com/albums/v132/rniemi/twister.jpg

Twisted solid 1" square-
http://img.photobucket.com/albums/v132/rniemi/moretwists.jpg

And a piece of 2" round tubing, twisted, showing the collapse- square does this pretty easily.
http://img.photobucket.com/albums/v132/rniemi/guns4.jpg

Boucher
08-18-2009, 12:25 PM
An old guy that made ornamental iron type burglar bars and decorations here had a Rigid pipe threader mounted on a rail and he twisted 1/2" bar. Doubt that would work on tubing. I broke the gears in my Rigid pipe threader the cost of repair was aprox $400. Also saw a hand powered version where the "hand wheel was the wheel off of an old farm implement. It was not in operation but appeared to use the same technique the old man did in that one end was fixed and the other end slid through a square die. The old man had a counter to count the number of revolutions turned.

Ken_Shea
08-18-2009, 12:36 PM
Ries,

Actually looking at your picture, some controlled collapse can look pretty good.

winchman
08-18-2009, 12:45 PM
Is there a tendency for the tubing to collapse at each end right next to where the twister grips it?

Roger

Evan
08-18-2009, 01:00 PM
It's nearly impossible to twist hollow tube sucessfully. The reason is how the twisting is regulated. When a piece of bar is twisted it strain hardens with essentially zero strain at the neutral axis and maximum strain at the outer fiber. The distribution of strain is kept even by the material itself in a solid bar. As any particular portion experiences slightly more strain than the rest it becomes slightly harder to twist so the strain is taken up by the weaker areas. This automatically distributes the strain equally throughout the length of the section.

With a hollow section the section winds up and some parts may become closer to the neutral axis than the rest. As a result those parts are easier to wind as the torque available depends on the twisting radius. This leads to a runaway condition that causes that section to absorb all the strain and collapse. Better results can be had by twisting angle iron which will wind up evenly until it forms a hollow square tube. This was something entirely unexpected that I found when playing with my twister. While my twister isn't CNC it is powered and I have very good control of the power via a belt type slip clutch.

This is angle iron twisted until it closes ( the piece standing up).

http://metalshopborealis.ca/pics2/pickettwister7.jpg

Ken_Shea
08-18-2009, 01:27 PM
Better results can be had by twisting angle iron which will wind up evenly until it forms a hollow square tube. This was something entirely unexpected that I found when playing with my twister.

This is angle iron twisted until it closes ( the piece standing up).

That is very interesting and unexpected !

EDIT:
Evan, is there a way to calculate the shrink rate?

Evan
08-18-2009, 01:41 PM
If you mean shrinkage in length, it isn't allowed to. The starting length and the finished length are the same.

darryl
08-18-2009, 02:00 PM
I would tend to think you'd have to keep a lot of tension on it while twisting. I don't know how much that would help prevent random collapsing of the tube, but I'd think the tension would have to be high.

A quick thought that came to mind is to surround the square tubing with a close fitting round tube while doing the twisting. Still, I don't think that would prevent random collapsing of the square tube, though it would probably help.

Cheeseking
08-18-2009, 02:14 PM
Would filling the tubing with sand or other media help the collapsing issue??

Ries
08-18-2009, 02:56 PM
It might work better to put a piece of round bar inside the square tube- that would limit collapse. Of course, it might be difficult to get it out.
Grease it up, and then use a come along or a hydraulic puller, maybe.

Really, though, unless you want to use the collapse as an aesthetic feature, its a lot easier to just twist solid square. When twisting square, it appears bigger- a square fills a bigger area when twisted, visually, as it fills out a circle the same diameter as its major axis. So if you twisted 1", it would look as big as 1 1/4" or so untwisted.

Evan
08-18-2009, 05:47 PM
The only thing that might work well is to heat the tubing close to red, maybe 900 F or so. The problem is with all the material acting as the outer fibers there is no balancing material to spread the strain around. Filling the tube may prevent collapse but it won't even out the twist along the length. Even twisting depends on the material resisting further twisting the more it is twisted. That only works with certain geometries. The case of the angle iron was predictable if you think about it. As soon as the edges close that part becomes much harder to twist any further.

boslab
08-18-2009, 06:55 PM
how about filling with 'cerromatrix' or some other tube bending alloy?

John Stevenson
08-18-2009, 07:16 PM
The twisted brass tubing you see on old showmans traction engines is made by pulling a length of chain thru the tube.
Don't know how this would work on steel, would think it would have to be a very light gauge.

Ken_Shea
08-18-2009, 07:28 PM
how about filling with 'cerromatrix' or some other tube bending alloy?

At about $100 a pound it could be cost prohibitive to fill a 36" section :D

kyfho
08-19-2009, 12:25 AM
I have documentation for an elegant twister project that you can build that may work for you. Alas, it is at work. I will retrieve and post it for you tomorrow.

Ken_Shea
08-19-2009, 07:59 AM
I have documentation for an elegant twister project that you can build that may work for you. Alas, it is at work. I will retrieve and post it for you tomorrow.

That will be interesting, thank you.

wierdscience
08-19-2009, 08:35 AM
I've twisted both solid and square by the ton.Tube will start and go into a spiral,then as a full wrap is reached it will starighten out,past that it will go spiral again and then straighten out as it reaches 2 full turns.

That being said,it is easier to twist solid bar.Bar straightens from jump and the twist is very consistent.

1/2" and 5/8" bar is doable in a 14" lathe set in back gear.I've done 1" and 1-1/4" in a 20 lathe,old conehead.

1-1/4" solid bar requires about 1100ftlbs of torque,tube takes more.

Ken_Shea
08-19-2009, 08:44 AM
Well I guess I was warned, Ries did say

"The short answer is- if it was easy, everybody would be doing it."

Still the twisted angle Evan showed looks promising.

Ken_Shea
08-19-2009, 08:55 AM
1-1/4" solid bar requires about 1100ftlbs of torque,tube takes more.

Is the math behind those figures complicated, for a + - * and / guy :D

kyfho
08-19-2009, 01:10 PM
Ok, here's what I found, but after reading further on the physics of twisting tubing, I am inclined to think that it may not work.

http://www.metalprojects.no-ip.com/twister.html

http://www.metalprojects.no-ip.com/twister/twister1.JPG

Evan
08-19-2009, 03:50 PM
It's not only nearly impossible to twist tubing cleanly but like Weird said it is harder to twist. That's because it is already work hardened from drawing. That also doesn't help the physics because it needs the progressive work hardening to control the twisting.

If you anticipate doing much twisting then plan on motorizing it from the start unless you really like sore arms. It's a lot of work to twist steel.

I used a garage door opener with some pulleys and sprockets to take it down to about 6 rpm. The motor is mounted on a hinge and is pulled back via a handle (not shown) to tension the drive belt. It gives excellent control. The final drive is chain.

http://metalshopborealis.ca/pics2/ptpower.jpg

http://metalshopborealis.ca/pics2/ptpower2.jpg

I made a counter using a dollar store pedometer and a supermagnet. The pedometer operates on an internal reed switch. No modification required, just put the magnet so it passes near the pedometer.

http://metalshopborealis.ca/pics2/counter2.jpg

wierdscience
08-19-2009, 08:19 PM
Is the math behind those figures complicated, for a + - * and / guy :D

Not complicated at all,I stuck a piece in a vise and twisted it with a 1" drive torque wrench,ratchet,adjust,ratchet,adjust,click=1100ft lbs:D

wierdscience
08-19-2009, 08:21 PM
Oh,forgot to mention,one of the side effects from twisting is in Evan's pics.The mill scale comes off:)

Ken_Shea
08-19-2009, 08:31 PM
Not complicated at all,I stuck a piece in a vise and twisted it with a 1" drive torque wrench,ratchet,adjust,ratchet,adjust,click=1100ft lbs:D

I shoulda known :D

Dr. Rob
08-19-2009, 11:55 PM
I've done it.

1 1/4" square tube with perfect results.

I put of lot of work into it though. It wasn't easy.

Smaller tube like 16 mm is easy, you can do that by hand.

Gotta run, i'm late. More later.

.

Dr. Rob
08-22-2009, 01:27 AM
Sorry for delay.

Had grand intentions of writing a big pictorial how-to, but the camera just died so too bad. Had very nice pics of an 800 mm long picket, 30 x 30 x 3 mm tube with a 400 mm 360 twist in the middle. Really indistinguishable from solid, save for the tube's small corner radius.

Anyway, i set about doing this several years ago, using criteria that it should be easy enough to show someone else how to do it (no black magic), no tube deformation except the twist, no collapsing or uneven helix, straight, and done cold. Using heat or black magic sort of defeats the purpose or goal of inexpensive, lightweight repeatability.

So I made very good clamping and fixturing dies to fit the big lathe (i really don't like using the lathe for this btw) that fit the tube as well as possible, and strong as hell.

There is of course a trick to it. Ries was very close when he suggested inserting a solid rod into the tube... Indeed, as the helix closes, the inside of the tube when viewed axially appears circular. So, insert a tube very close in outside diameter to the ID of the square tube. This helps keep the tube walls from collapsing, and helps keep the twisted part straight and aligned.

Further, one aspect to consider is the ratio of tube size / wall thickness. Thats why tube like 15 x 15 x 2 mm is easy as pie without any interior supporting tube, but a 40 x 40 x 2 tube will likely turn to mush immediately.

Having said all that, i unfortunately didn't have the chance to tun it into a documented science... always got derailed by something else. Does it matter how hard you twist, or if the tube is formed hot vs cold, or welded vs seamless, or long vs short ??? Idunno. You find out.

But, "The short answer is- if it was easy, everybody would be doing it". Yup. Agreed. And Evan's angle iron trick is a pretty neat trick; that's a new one.

Gotta go.

Doc.