View Full Version : grinding a weld inside a pipe 8' from the opening....how?

Rob Frink
11-12-2002, 01:42 PM
I need to butt weld 2 sections of steel tubing that are 8 feet long, 4" ID. The function of these tubes requires that the ID is ground smooth and flush at the circumferential weld. The welding process will generate and internal weld bead about 1/16" inside the tubing bore which I need to grind/finish.

So here is the million dollar question: How would you grind the weld seam inside a 4" tube that is 8 feet from the opening.?

The tubing is seamless DOM.

Whatta think?


Dr. Rob
11-12-2002, 02:31 PM
Piece o' cake. Just cut the tube off close to the weld, so you can get at it with conventional grinding tools, and weld it back together again when you've finished grinding. Ta-dam, finished!

You do realize this might not be seriously intended..?

Treven Baker
11-12-2002, 02:57 PM
Is there any posabililty you could make a connector sleve and then do two lap welds after sliding the ends of the pipe into the connector sleve?

11-12-2002, 03:10 PM
1) 4" opening is large enough that it's conceivable to make a customized pneumatically powered grinder jig with you can send down the 8 feet length. I have no doubt that such a machine have been built before for specialized applications.

2) Machine a sleeve which will go over the two pipes. Using MIG, you may be able to weld the two outter seams without distorting the ID of the tubes.

3) Machine a round copper bar who's OD is the same as the ID of the tube. Slip the copper in the tube where the butt joint will be and carefully MIG weld around the tube. The copper will minimize the bead information inside the tube. Knock the copper bar out once the welding is done.

4) Electron beam welding may work but the 8 foot length may pose a problem http://bbs.homeshopmachinist.net//biggrin.gif


11-12-2002, 03:22 PM
The scenario described seems very similar to one Guy Lautard presents in Machinist's Bedside Reader (either # 1 or #2), wherein he depicts a homebrewed scheme for boring out large cylinders. It had a kind of racheting fed cutter traversing back and forth thru the bore. I don't have that handy now, and can't recall exact details. So don't know how it might be adaptable to a grinding setup. I mention this mainly to possibly help stimulate other ideas from any familiar with Lautard's article. Of course 16 ft is a pretty long traverse!

Tho Dr Rob's idea has merit I guess. After all, one could cut one of the pipes into 16 six inch pcs, weld on one piece, grind that bead, then weld on the next, grind, ...,weld, grind, ...etc. (joking of course)

11-12-2002, 03:53 PM
Here's another possibility: There's a company called "Extrude Hone" and specialized in getting a smooth surface in areas not normally accessable to grinding tools. Their process involves a putty like substance with abrasive materials mixed in. They then squirt that through whatever needs to have a smooth surface under high pressure. (GM uses them to polish the air intakes for the Corvettes.) You could either get them to do the clean up for you, or figure out a way to do it yourself.

11-12-2002, 04:37 PM
If you're not committed to the welding process that leaves the internal 1/16" bead you need to remove, you might consider laser welding. Lasers give an extraordinary control over heat and penetration. It's possible you could do the weld to your strength requirements without the internal bead. It's probably worth asking a shop in your area. Good Luck!

Rob Frink
11-12-2002, 05:09 PM
Great feedback ...thank you!

The design of the workpiece is fixed. It is a high volume production piece. The scenario of 2 pieces welded together is economically the most feasible as well as meeting all of the design goals. I expect the cost savings of the butt joint/internal grind design to far surpass the cost to develope a technigue and tooling to make the parts.

So for the fun of it, let's say that there is an unlimited budget to develop a process and tooling to grind the ID.


11-12-2002, 07:01 PM
Check with Ladish Tri-Clover, or other manufacturers of sanitary or dairy tubing products. They do this regularly, heli-arc with purge, grind and polish.
I am thinking something like a cylinder hone on long shaft, but they may be able to supply the equipment or provide the assembly.

11-12-2002, 07:51 PM
How about an air powered die grinder. Mount excentric in a sleve with acme thd on OD - sleve is in holder less than 4"OD - holder has 3 cam locks on OD that lock the holder in position and centers it. Moving the holder to the center of the pipe puts the grinding wheel in contact with the pipe just in front of the weld. The grinding wheel drag on the pipe ID and Weld causes the sleve to rotate moving forward on the acme thd. The grinder has a rotating coupling so as to not wind up the hose. May have to put a drag break on the sleve with the die grinder turning 8-10K

After I went through this I realilized every thing could be on the out side on the end of an 8' arm. That way the feed could be controled.

11-12-2002, 08:24 PM
Good ideas all. What about using a cylinder hone on a long extention. Using a coarse stone and spray a coolant inside with a long tube

11-12-2002, 11:03 PM
Rustybolt's reply is like several others, it misses the point. That is that "all" the cutting must be done on a single point, the weld bead. A hone cutts over a cylindical surface defined by the lenght of the stones on the hone. This is a good way to finnish up after the weld bead is cut out.

This bouncing back and forth is what I miss. I got down sized, right sized, out sized several months ago. Most of what I posted was ideas from all the previous posts in a different order. Keep it going and one of us will get it right.

11-12-2002, 11:12 PM
I think Albert had the most logical and lowest cost solution - a machined copper plug. However, you may have to water cool it to be able to extract it in a reasonable time as the welding heat will expand it and you want a snug fit for a smooth bore.

This will allow full weld penetration, BTW. If the finish is critical then it could be bored out with a gunn drill.

The problem with that is the abrasive would affect the entire tube instead of just the weld. I was invited to their boot at IMTS to see their new micro-tech stuff (had a dance with a doctor instead). It is a neat process - Ultra-slow waterjet ! http://bbs.homeshopmachinist.net//biggrin.gif

[This message has been edited by Thrud (edited 11-12-2002).]

11-12-2002, 11:29 PM
Maybe you could use a cylinder ridge reamer with a long extension, and bring it down to where a hone could finish it off?

11-12-2002, 11:48 PM
OK. What about something like a 16 ft wide pyramid roller with the cutter/grinding wheel sandwitched in the middle of the top roll.The top roll would be the high speed spindle and the lower rolls would be the low speed regulating rolls.

Rob. what are these some kind of long air cylider?

11-13-2002, 12:06 AM
I would suggest a line bore type of set up cutting about .ooo5 under size on the radius. Clean up and control surface finish with a hone.

Neil Peters

11-13-2002, 01:50 AM
Am I missing something here?? Why not just use 1 x 16' tube instead of 2 x 8'?? http://bbs.homeshopmachinist.net//confused.gif

Dr. Rob
11-13-2002, 04:00 AM
Thanks, Tel. That's the best thing I've heard all day!

11-13-2002, 07:04 AM
Dunno Doc, if it HAS to be welded I liked your scheme http://bbs.homeshopmachinist.net//biggrin.gif

11-13-2002, 09:52 AM
Rob, I had a customer who did this all the time using a Plasma arc. The ends of the tubes must be clean and come together very well. The tubes are then blocked at both outer ends and filled with argon gas. The plasma arc power and speed of the weld are chosen to allow flowing together of the parent metals, no filler is used. The tubes are rotated past the arc. This leaves a smooth finish with no internal bulge. However the size may be very slightly reduced due to shrinkage if this is a problem back to square one.

George L

11-13-2002, 10:27 AM
How does this sound...

Make two or more cylinders machined to the same OD as the ID of the pipe (4”). Cut a 2” dia. hole in the center of each cylinder. ” (If you’re going to use it a lot, line the central hole with bronze bearing material.) Take an 8 ½ foot section of stiff 2” OD, thick-walled pipe. Weld or otherwise attach a boring head to one end of the pipe. (Make sure the boring head is perfectly aligned with the pipe prior to welding.) Chuck the new boring tool in a lathe and use the cylinders to keep the shaft centered in the pipe you just welded. Now, cut away! If you are really ambitious, you could lengthen the pipe, have it extend past the head of your lathe, drill holes near the boring head and attach a fitting for cutting fluid. Pump the fluid through the pipe to flush as you cut.

Obviously you would have to secure the welded pipe past the bed of the lathe (unless you have a very large lathe) and made arrangements for feed. But, it might actually work . . .

[This message has been edited by tonydacrow (edited 11-13-2002).]

11-13-2002, 07:48 PM
Would a reamer .001-.002 undersize work?

11-13-2002, 10:47 PM
We all have ideas but what about evaluating someone elses. ShavingMaker said to use a cylinder ridge reamer. Think about that for a minute. Some one did all the work way back when, (in the good old days engines got a valve job every 5000 miles and a ring job every 10 and the tools were simple to use). The weld bead is no more that a cylinder ridge just in the center of the cylinder. Don't re-invent the wheel just give it a big KISS.

Gordon Campbell
11-13-2002, 11:06 PM
I would use a chill ring for a four inch boiler tube. That would keep you from getting a wedding band on the inside of the weld seam. The root pass could be run with a tig torch, if you are good enough to control the amount of filler metal deposited, in stead of a chill ring. Gordon

11-14-2002, 04:39 AM
Gordon's right - just get a B pressure welder from Alberta and you won't have to worry about any inside obstructions.

The Pipeline gods have spoke - "get thee a real welder" http://bbs.homeshopmachinist.net//biggrin.gif

kap pullen
11-14-2002, 09:42 AM
The naval gun factory used a "wood reamer" for semifinishing gun bores.

It is simply a slab of iron, say 3.75, by 1, by 10.inches long. Put a driving shank on one end,centerdrill the other and slot for a hss tool bit on the long narrow side.

On each flat side, bolt a piece of hard wood, and turn the wood by your shank for a snug fit in the bore.

Put a tool bit in the slot in the slab. sharpen like a boring tool and set even with the wood od surface.

If you grind it with an 80* lead angle and a little flat like an end mill od you may be able to ream it with a t handle driver by hand.

Or set it up on your boring mill table and feed with the machine.

Than hone the last couple thou out.

I have made this tool and trued some cannon barrels after drilling this way.


11-14-2002, 03:24 PM
It's quite a few years ago now, but I went through the Heil Co. plant in Milwaukee and saw how the sized the ID of the tubing that they used for the hydraulic cylinder bodies.
They had a steel "ball" that they inserted into the bore, with LOTS of oil, and used a hydraulic press to push the "ball" through the tube. The tube ID had a nice burnished finish after the treatment. The "ball" had several flat spots on it, as I recall.

I'm not sure this'll help in this application, but it might be worth looking at.
If it was my design, I'd sure look at gett the tubing 16 ft long and skip all of these problems!

David Hafnorske
11-14-2002, 07:42 PM
I don't know why it would have to be so precise. But if the only concern was to get clearance for another part I would try the following. First, how thick is the wall of this pipe? Lets say it is a 1/4" thick. This would leave enough pipe to bevel the inside of the pipe, to allow for the weld. Just like you would bevel two pieces of plate for a but weld. Except it would be on the opposite side of your weld. Then use a cooled copper plug as mentioned before, tapered like a torpedo on one end, so it doesn't binde when you drive it out. Then follow up with a cylinder hone. If the ID of the seam was SUPER critical than why wouldn't one just buy full length pipe and eliminate the whole problem? It seems silly to beat yourself up getting the ID of the seam to +or- .001 if the rest of pipe is only manufactured to +or- .010.

[This message has been edited by David Hafnorske (edited 11-14-2002).]

David Hafnorske
11-14-2002, 09:07 PM
Is this an application like a steam tube? Where the inside of the pipe has to be smooth so that the weld does not cause turbulence and erode the pipe like a waterfall? If so, it will take a smarter man than I to tell you what to do. Know anybody that is a pipefitter?

11-14-2002, 11:54 PM
We did a whole bunch of pipe welds with a AstroArc power supply and a Diametrics Weld head. This is an automatic welder. The weld head has tracks and crawls down the inside of the pipe. The person doing the weld looks at the position of the weld head via a fiberoptic eyepiece. When it is in position he activates the clamping rods which extend out and grips the pipe on the inside. The pipe is beveled on the inside. The pipe is fitted up with an consumable ring for weld material clamped between the two lengths. The area is purged with Argon. The person light off the torch and now comes the really confusing part. There are two fiberoptic lines one watching the front of the arc and the other looking at the back of the arc. The person controlling the weld enters weld/current/voltage parameters into the AstroArc controller so that the arc varies depending on whether it is welding overhead on down on the bottom. The left eye sees the arc approaching and the right eye sees the arc leaving. Each eyepiece was fitted with a weld lens. With a little practice the weld is almost as smooth as the inside of the pipe. Our welding was on the coolant outflow piping on a nuclear reactor.
The Diametrics Head and AstroArc Power Supply was $95,000. These are old technology today.

You said that we could spend all the money we needed for this job...!

Rob Frink
11-15-2002, 08:03 AM
Wow! this is great!....tons of excellent ideas!

I've learned a bit more from the customer about the application and process.

The final assembly is a high performance structual component that will see a good deal of cyclic stress...Such as an aircraft or automotive chassis. The reason for removing/finishing the weld is too reduce and eliminate stress risers from changes in section. Rather than a "cutting" application, they want a "contact wheel" belt grind finish such that there are no tool marks. The tube is hydro-formed in sections and welded together. One section is bent and contoured such that there is no way to get inside ....the other section is straight ...but 8 feet long. Hydro-forming tube is like blow molding a plastic bottle, The tube is placed into a die, then the ID of the tube is filled with hi-pressure water untill the tube expands to the shape of the die. It is felt that a 16' length is impractical for the HF process and machinery.

I truly favor the gun bore tool described with the wooden block pads.....it is truly very simple.

Thanks for the input and keep 'em coming,

[This message has been edited by Rob Frink (edited 11-15-2002).]

Ragarsed Raglan
11-15-2002, 09:06 AM

I would have thought from a cyclic stress point of view an undressed weld would have been far more acceptable than a machined weld (and I mean machined in the context of any form of metal cutting, be it abraded or whatever), which would be darned difficult to inspect without recourse to expensive borescopes.

Change of section (at the low percentage factors of weld bead) is less detrimental in cyclic stress than 'notch embrittlement' (even micro notching can be disasterous) as a result of some machining process.

I'd get to know more about the end use, and talk closer to the engineers on this one.


[This message has been edited by Ragarsed Raglan (edited 11-15-2002).]

11-16-2002, 01:15 AM
Sounds like an aerospace pressure duct to me which means both pressure fluctuations and also mechanical stresses and vibrations. In this type of situation the weld bead can be relatively thick when compared to the wall thickness and thus be a stress problem. What is important here is both the finished weld height and its blend into the wall surface. I would think a surface finish of 16rms would do the trick and a finished height relative to the wall of .000 to .005 would work. The height would keep the weld in the boundary layer and thus not materially affect the flow. Do not let the weld surface go below the wall surface or the stress person will come unglued. This should be a three step process, rough bore to size, coarse flex hone to form a blend contour, and a fine flex hone to control the surface finish. The finish should have a cross hatch lay. Been there, done that, and got the t-shirt as the saying goes but in reality it was usually a coffee mug with the project logo.

Neil Peters

11-17-2002, 01:50 AM
Perhaps I've missed something but why can't the engineers use mechanical tubing which is available in lengths from 7 to 24 ft . The Ryerson Steel catalog lists such a variety of I.Ds, O.Ds that one would think that a selection could be found.

Or is this just another case of "Engineer's Tunnel Vision?

11-17-2002, 06:54 PM
Sounds like an "oops" that was "fixed" with little forethought. But then, we have not seen what is really going on either...

Peter S
11-24-2002, 08:11 PM
Warning: Not much help to the above problem, contains history and may cause boredom...(but does show how they did it in WW2 in 40'long, 3"ID pipes)

I have just been re-reading some articles by W. Brian Taylor about the "PLUTO" ("Pipe Lines Under The Ocean"), where the UK developed and laid fuel supply pipelines across the English channel during WW2.

When the allies landed in Europe in 1944 - how did they supply fuel to the invading army? As early as 1941 the British had realised that sucessful re-invasion of Europe depended on a supply of fuel. They therefore began in 1942 to develop a method of laying pipelines across the channel, pipelines that were eventually to follow the armies right into Germany. They realised that advance would have to be done without the prior capture of any oil ports or oil storage facilities.
Initial fuel supply after D-day was managed by the "Mulberry" harbours, however within 2 months, 4 pipelines were laid and operational, eventually there were 17 pipelines crossing the channel. Pipes were laid at night, without loss.
Once on land, the pipes were continued by conventional means, ie short lengths of pipe joined with flanges, these pipelines reached out in several directions. By VE day, one pipeline was pumping petrol from the Isle of Wight in England to the outskirts of Frankfurt.

How do you lay a pipeline quickly over 30 miles of water? First thoughts were to use a similar method to the allready sucessful cable-laying techniques - ie a 3" ID lead tube, wrapped with its many layers of armouring and protection - a hollow cable. This makes interesting reading, they had to manufacture, store and lay this pipeline with the lead tubing pressurised at all times to prevent collapse.

But the other method was altogether more novel - welding together 3" bore steel pipes into a continuous pipeline of say, 70 miles length and winding it onto huge floating drums (35' diameter x 60' wide), much as thread is wound onto a cotton reel. Then anchor one end of the pipeline in England and tow the drum (known as "Conun-drums') towards the French coast, the free-to-rotate drum paying-off pipe in it's wake....
Would it work? Could pipe be rolled up in this way? Tests quickly established that this was possible, but then how to un-coil the pipeline - would the pipe remain bent, or maybe lash-out like a released clock-work spring? Well, there was no problem here either, the pipe could be spooled and un-spooled without kinking or any spring-like problems.

The steel pipe was initially made in 40' lengths. Each 40' length was butt-welded to another length, and made into one long, straight pipe, 3/4 mile long. These long lengths were stored in a specially made trough, one end at the butt-welder inland, the other end near the port where the huge drum was waiting to coil up its pipe. The 3/4 mile long lengths were then fed onto the drum (which was under powered rotation for coiling), stopped to allow for butt-welding on another 3/4 mile length , then repeated until about 70-80 miles of pipe were spooled onto the drum.

Butt-welding caused both external and internal finning. The external fin was easy to remove and the internal fin at each 3/4 mile join was considered insignificant, however the internal fin at each 40' join would impede fuel flow, and so had to be removed, without leaving any swarf in the pipe.

Here is how Brian Taylor describes the process....

"The internal de-finning stage led to the nose-end of a 45-ft long hollow reamer being equipped with a feature akin to a mop-head plus a stiff-wire bottlebrush..
The prototype welding-line comprised three stages, all arranged on a common axis. Each 40-ft length of pipe entered the welding shop via a guide-tube set in the wall at the RH end of the building. After having both ends machine-trimmed and cleaned, the pipe entered a 'V'section roller-conveyor located along the rig's axis.
The first length of pipe to enter the vacant welding line had it's leading end (the LH end) temporarily plugged by a hammered-in wooden bung. The team then pushed the pipe from right to left along the line, until its trailing end (its RH end) could be held in the correct welding position by the butt-welders LH jaws. At that stage the next trimmed and cleaned legth of pipe advanced along the 'V' until stopped and held captive by a catch, a yard or so to the right of the welder. Then the long reamer entered the captive pipes RH end. It penetrated beyond the catch, across the yard-wide gap and into the pipe-end held in the butt-welders LH jaws.
It stopped with its tool-head well down the pipe and the cutter blades safely beyond the ensuing weld-point. Then, released by the catch, the captive pipe advanced into the welding machine where its RH jaws clamped the pipe in the correct welding position.
As the welding current flashed between the pipe-ends, its intense heat fused the metal.
Once plasticised, the heating cycle ceased and hydraulic rams forced the pipes together to ensure complete fusion. Due to the pipes LH end being sealed by the wooden bung, a blast of compressed air fed down the hollow reamer left the pipe at its only exit - the RH end. Thus cold air blew across the hot joint from left to right. Although not its main purpose, it hastened cooling.
With the pipe still held by the pipe welders jaws, the reamering unit then rotated, trimming the fin as the whole shebang slowly withdrew from the pipes RH end. Simultaneously, a sustained blast of compressed air forced the metal swarf - plus any odds bits of rogue-scale lossened by the bottlebrush - to be blown out of the pipe at its RH end. The bore gaining a final wipe from the reamers tail-end mophead."

The story then goes on to describe the external fin/flash-removal machine, and how the pipe then progressed in 40-ft per cycles to reach its 3/4 mile length.

The whole project was difficult and a large undertaking, and involved interesting solutions to other problems, but I won't go into any more detail.
When the drums were fully spooled they contained about 80 miles of pipe weighing 1600 tons. Each drum was towed by two powerful tugs, with a third tug behind to act as a rudder.
There were also 2 vessels carrying 100 miles each of 3" lead pipe (pressurised), 1 vessel with 30 miles.

11-24-2002, 08:46 PM
Peter s: not a bit boring. How did they ever get the endangered species out, make the job safe? get funded? all those things needed to get a real job done? THe war only lasted from what- 1937 to 1945? Must have had some real computers to not lose the paper work involved. HAts off to them. http://bbs.homeshopmachinist.net//smile.gif

You thnk we could do the job as quickly today?

11-24-2002, 11:03 PM
As my dad used to say "You can dig a 6' deep foxhole in seconds when some bastard is shooting at you." He also told me that guys that had to be dragged out of the fart-sack in the morning could jump 12' fences when being shot at.

You would be amazed what you can do when you are about to be toast and it is all or nothing. The Brits are like Badgers. Canadians are like Wolverines - we would chew our own legs off to get to beer. Oops! Told our secret... http://bbs.homeshopmachinist.net//biggrin.gif

11-25-2002, 08:12 AM
Speaking of fast jobs: according to M Taylor, (author of "worlds strangest Aircraft" and writer for Janes), Hitler Goering decided late summer of 1944, that they needed a Fighter airplane, simple to build and fly and capable of winning air battles. The presentation to manufacuterers was first week in Sept, contact awarded, 9 Sept, prototype shown 20 sept, accepted about 1 Oct, first model flew 6 dec 1944. A person cna do a lot in a short time when they are being shot at from one side, and no impediments beteween him and his goals.