What's new
What's new

Initial pinch rolling tips needed

M. Moore

Titanium
Joined
Jun 8, 2007
Location
Vancouver Island, B.C. Canada
The new (old) Webb pinch rolls are set up and working well.
Now that we have done a few jobs I have some questions about the process now that I understand it a little and I mean a little bit.
I have done some research and watched videos to see how the process is done but there is not much out there and what is there certainly is not instructive in any way it just shows the process usually without any explanation. There are some really cheap and cheesy home hobby shop type instruction videos that are not worth watching.
So that said we generally are just rolling plate into curved sections, it would be rare for us to have to roll a full cylinder but it may happen so tips on that process are very welcome.

The main issue is trying to get to the holy grail of settings " xx on the gauge for 28" radius " in one shot kind of thing. We are sneaking up on the shape by rolling the part several times and in order to avoid an overbend we usually underbend, repeat, repeat and then finally hit the mark.
What I need to understand is why the part will roll a tighter curve on the first bend and then take way more bending roll adjustment to get it to bend even a little bit more?

I also need to wrap my head around the fact that a 4" wide piece of 1/4" flat bar will roll to xx radius at xx setting and a 12" wide x 1/4" piece will be a different (larger) radius when rolled on the same setting, both pieces being done one after the other.
I have heard it said that there is an "art" to rolling and to me that just means you need to stand at the roller for two years learning as you go.
I do get it that one piece of plate will roll differently than another but there must be some tips and tricks that help get to the right radius faster.

I have discovered that I can unroll an over rolled part if I can feed it in from the front and have the bend roll fully retracted, it works quite well to roll it almost back to flat. This only works for gentle radius segments and shorter lengths as anything longer with a tighter radius will hit the ground before it goes into the rolls from the front.

So lets have some input from all you very knowledgeable roll hands out there as I need to continue my PM apprenticeship training.

Thanks,
Michael

ps, we have to do some rolling on thursday and I will try to take some pics in progress if I remember....
 
If you have many pieces to roll and you want to set up for a one pass run try this. Roll one part and measure then set that part aside. Adjust the rolls and roll a second part, measure and set aside. Roll a third part and measure. If you are good then this is your setting for the remainder. If not adjust again and set the part aside.

When we are setting up multiple parts we usually have at least two trial parts set aside before production begins. During production if there are any parts not rolled to tolerance set them aside. If the parts all begin to trend out of tolerance make a small adjustment and continue.

Once the all parts are rolled you can go back and re-roll the under rolled parts using more or less the same procedure. Your end setting will be greater than the parts rolled with one pass. (see further down about stress)

To answer some of your questions there are many seen and unseen factor affecting the rolling process.

One unseen factor when rolling cut plate is which way is the grain of the material. You can get two parts cut off the same plate but roll very different because one was cut at 90 deg from the other. If you are cutting the parts then try to cut from the sheet or plate so the grain in aligned the same for all. Another factor is the age of the plate. Material that sits goes thru a stress relieving process by itself. In many cases there is nothing you can do except be aware when rolling parts cut from different lots of material (or from different rolling mills).

Now on the subject of stress relieving remember that the process of rolling or bending add stress to the part. This is why a part that goes thru the rolls resists rolling with each added pass. Much like when your bend a coat hanger wire back and forth until it breaks a plate will become more and more work hardened as it's rolled. A good roll operator gets to know how much these factors affect rolling and can reach the desired shape with fewer passes. If you try to sneak up on the size using small adjustments you will soon get frustrated.

As to the answer to why does width affect the rolling just pull out a length of 1/4" x 1" flat, grab both ends and try to bend it. Do the same with a piece of 1/4"x 4" flat. Not only will the parts resist bending differently but the spring back is also different. Your rolls are stronger but the resistance to bending and rate of spring back does not change.

Walter
 
^ the once through adjust and move to the next is the only way to not drive your self nuts chaseing your tail, its much the same on a press break. I do like sharpie marker pen marks on the adjustment knobs too, can be as crude as Fuck but can get you right back to the same spot so fast and easily.

Other thing is material, some black flats vary so much its horrible, good material can be one hell of a saving on complex bending - rolling jobs. Even thinner bright cold rolled sheets often vary enough thickness wise to really throw things off.

Hence why more or less anyone that has to do it considers rolling part art form, because there’s just so much out of your control material wise that has such a massive impact on the final part.

Easy way of checking part radius is cord height, take a known length stick of anything with square ends and measure the apex of the inside of the curve will give you a very good idea as to radius with out messing with templates. Its a bit of maths or a quick sketch in cad but it can be a really fast and easy way to check things.

Then of cause theres what the part measures on its side when gravity is no longer pulling the curve flatter!!!
 
Get yourself a recipe card box with some recipes cards in it. Each time you roll something note the initial settings to prebend ends as well as following settings and number of passes.
Organize cards in box according to thickness and material if you are doing stainless, aluminum etc.
Each time you do a job you will be able to dial in faster. You will never get that holy grail number because plate even changes in a stack you are given.
Forgot to mention that you will be making a lot of home made sweeps so have some thin gauge handy to cut then punch a hole in one end, mark with a big felt marker and hang it up for next time. A shop that has been doing a few years work will have hundreds of sweeps.
 
trial and error is the only thing that has worked for me.
obviously, sneaking up on the desired radius is best for one offs, and rolling two or three samples is, as mentioned, usually needed for production runs.
every machine is a bit different. every mill run, even from the same mill, can be different.
When I do big runs, usually at least 5% to 10% of the parts will need tweaking. Sometimes this is easy, sometimes its not.

interestingly enough- I once visited a factory that makes rolls- in this case, angle rolls, but the principles are the same. They make a full on cnc angle roll- and, to make accurate bends, you have to run at least two sample pieces thru first- the machine actually measures the resistance- and then, and only then, will it be able to accurately give you a desired radius, but only in that particular profile, size, and mill run.

even the computers have to use trial and error.
 
FABULOUS replies!
Excellent info contained in this thread and no bickering!
I really appreciate the time you guys are spending to answer these questions.
Hopefully I can pay back to others sometime.

I didn't include in the initial post the question of pre-bending each end for a full rolled part.
So lets say I am doing a 24" diameter cylinder from 1/4" x 30" wide material. I want to pre-bend one end and then turn the part around and bend and roll it to diameter. That is how I understand the process, the question is, how do I know how much to pre-bend the first end when I don't know the final setting for a finished roll of that diameter? Is it also just a guess?
The other time this might be necessary is when material is tight and you want the curve to look good all the way to the end of the part, this alone could save us quite a bit of material as right now I am cutting off 3-4" of each 3/8" plate.


Michael
 
You prebend the ends to the required radius. The part you are prebending is the part that spans the bottom rolls and fails to get bent so it never does get bent further when you are rolling the rest of the plate.
You do your best then pass it so the poor fabricators. Avoid the lunch room if you screw up :'))
Keep in mind that a rolled piece dropped on edge will open up or if it is shipped by truck the vibration will release the tension.
 
A lot of shops use plywood patterns that are pieces of 1/2" or 3/4" plywood, long strips that have the radius desired cut in one long side. You make em as you need em, and slowly accumulate a library. Then you can hold up a 24" diameter pattern to your prebend, and get it pretty close. This is especially handy for bigger radii- a 4' or 6' long piece of a 15' radius is very handy, as its much harder to estimate how far off you are on bigger circles.
 
Hobbyman,
What is a sweep?
I really like your idea about the recipe box and notes for all the different specs.
I understand that the prebend doesn't get rolled, I don't have any way to unbend it if I over prebend it so it usually comes out under-bent at the moment, thus my question. I will have to experiment on parts that I have extra material until I get the hang of it and more comfortable with the process.

Ries, I like it that the CNC machines have to use trial and error too!
I already have a wall full of radius gauges that I use for my cove legs, 1/2" and 3/4" MDF is a good material as it sands smoothly and easily, I have a well equipped woodworking shop to make my templates. I can make the templates much faster than the rolling at the moment.

Adama, I really hope that gravity has no effect on a 30" wide piece of 3/8" plate rolled to a 29" radius, if it does then the material is very suspect! I would assume you are talking about gauge plate in the 12 and thinner range? I think the formula for radius from cord length is H squared + (1/2Cord length) squared, divided by 2 x H (Height) does the trick and I use that formula all the time in my work however it is much faster and easier for me to use a wood pattern.

One of the quirks that I have to deal with on the Webb machine is that there are good gauges on the bend roll (at each end) and a nice easy to see vertical gauge that rides on the back roll gear link. The quirk is that the gauges are marked with the same scale and are all direct reading, the vertical gauge rides on the gear link which travels in an arc meaning that the scale does not match with the scales on the roll ends which travel straight up and down. I will have to make a new scale for the vertical gauge so it matches the end gauges as it does annoy me each time I use the machine. Hard to believe that such a quality piece of machinery has this glaring mistake as I noticed it the first time I used the machine.

I did add a remote pendant style switch which is really nice to use, it was an old crane control with east/west and up/down buttons which were perfect for the forward/back rolling and up/down on the bend roll.

Pics to come next week on the next rolling job.

Michael
 
I would expect gravity to have a affect on everything. Your really stupid not to, yeah may only be a 1/8" but a circa 5' pipe will oval a fair bit under its own weight. It will mesure diffrent depending on just were the joint is too. All final gauging for high accuracy really needs to be done on edge, not horizontally.

All depends on what tolerances your working too and even then if it has to clear something in reality, but trust me, it will have a tape measure level measurable difference.

The coments about truck ride and slight drops also hold true, 3/8" is floppy. Thickest i have personally bent - rolled is 60mm that was a small part hence rigid, but even larger parts of 1" + once rolled often have way more flex than you expect. I have seen the above catch several people out, can be really expensive mistakes too! Especially on something as simple as a cylinder.
 
Adama,
True, gravity has an effect on everything. I doubt it is causing my 3/8" thick plate to unbend by itself especiallly when it is pinched on one end and suspended out the front of the machine. I could put my whole body weight on that plate and not unbend it a "measurable" amount. We are talking about a 29" radius on a 36" long x 27 wide piece of 3/8, the arc height is only about 6".


We did some more rolling yesterday and the helpful tips did help with the job but more practice is needed.
The initial roll was quite close but the fina rollings still need some practice as we erred on the safe side so it took a few swipes to get it right.
I forgot to take pictures.

Michael
 
^ If your roller has the range its very easy to unroll a bit if you go to far, sorry thought you were rolling a cylinder. Should not be overly hard to unbend it a bit at that thickness though, just turning it upside down and jumping on it should take a suprising amount out, you can normally get a bit back do to the fact that with shallow bends your only taken so much of the thickness of that 3/8" plate past its yeild point, taking a bit out your not bending anything like a 3/8" thick plate, its one of those quirks of bending.

If you have one its also easy to bump the bend out a bit on a press brake on shallow curved parts like that if you don't have the travel on the rolls.
 
I will have to try stomping on the plate to unbend it a little, I never would have guessed it possible with 3/8" thick material. Funny that I didn't see any videos of that technique as I could see it being quite popular.......
We don't have a large press so that option is not available yet, I do have a 50 ton press that we use for smaller stuff, it has a 32" width.
 
^ that press will piss it. 50 tons if you can set up even a crude wide V which can be as simple as bits of flat, will prove more than enough to straighten even some heavy plate. i can happily straighten 8x1 bright flat with my 30 ton press at a fraction of that 30 tons capacity with just a crude v barely 7" wide.

Now don't get me wrong, you won't flatten it by jumping on it, but you can unbend a surprising amount, especially on larger shallow bends simply be just taking some of that yielded material back into yield using the tension in the stretched but unyielding areas in the cross section. Same effects really noticeable in tube bending and most other bending so long as your not coining the full material thickness into yield like you can border line do with mandrill pipe bending or closed die forming. The larger the radius the more you can unbend for a given force angle wise.
 
I'm late to this thread, but you might save a little time with your formula
> H squared + (1/2Cord length) squared, divided by 2 x H (Height)
because it is equal to
(1/2)H+(C squared/8)/H.
You can take any piece you use for the chord, say it is 20 or 21 inches long, and calculate (C squared/8) just
once ever and write it on the chord piece, 50 or 55.125 whatever. Then you don't have to square H any more,
just take half of it, and divide the number on the chord piece by H, and add.
 








 
Back
Top