Hi there,
I was having a similar problem a while back with a SAW beam welder.
I finally figured out that the wire stickout from the contact tip was crashing into the tack holding the weldment together and causing an awful mess of spitting and sputtering for about 6 to 8 inches past the tack. Training the fitters to place low profile, elongated tacks rather than a big rounded tack solved the problem.
Hope this helps...
Tim
I don't know that I am familiar with what you material you are welding. For instance, if it is a casting, there may be structural discontinuities such as gas pocket or inclusions which the arc encounters that temporarily destabilize the arc process. When the sputtering event occurs, I tend to think that the entire arc / weld pool interface has gone into uncontrolled oscillation, in the same way mechanical compliant systems can. The changes you may need to make may be slight adjustments of inputs (stickout, amperage, voltage, travel speed) in order to improve overall system stability and reduce system sensitivity to transient conditions tending to provoke oscillation.
A philosophical (maybe non) answer to your question, based on the perspective of my experience with SAW, is that process stability with SAW can be looked at as kind of unstable equilibrium, or quasi static equilibrium. Consider the harmonic resonance of a simple spring, or waves in water, then consider that the electical current, power source inductance, the gas shielding envelope, the transfer of metal across the arc, the fluid weld pool, the slag blanket, the travel carriage, etc, are each compliant systems with their own compliance properties, the interaction of which would probably require some all too too supercomputer to model. In this context, arc process stability is based upon the appropriate combination of system inputs which tend toward the greatest process stability in order to produce quality welds consistently and to have the greatest resistance to normal transient events that would act to destabilize the process.
A lot of formal and anecdotal research has focused on obtaining the greatest deposition rate possible wtih SAW, but this is normally constrained by workmanship and quality issues. In the end, it seems, merely some kind of balance between repair rate and productivity through the welding station is achieved. Figuring out why the arc process is unstable, or identifying which balance of inputs tend toward improved process stability, seems not to be a priority in the commercial context. As soon as a relative stability is achieved, and who knows why it happens to be stable, the motivation becomes to drive the process faster, and in doing so provoke the conditions for process instability.
GRoberts, We have run in to the same problem on occasion and have removed it most of the time by moving, cleaning or adding a ground to the other end. We have even had to stop and de-mag a few times.
Other times it has been a little to much melted flux gets in the path. We change the lead angle of the electrode or move the arc slightly down on the radius (opposite of the direction the work travels). You use a little more flux (sliding down the side) but it keeps the weld cleaner.
Thanks for all you replies.
As far as the tacks go, we actually deposit about 1/2" of SMAW weld in the root before the two parts can be self-supporting enough to be put in the horizontal position for sub-arc welding. On this particular unit, we didn't happen to have a single problem for the first 1 1/2" of weld, although it has happened in the root before. We had the most problems of all on the first unit (in addition to the problem described). The original problems included undercut, which hasn't happened since we put a second ground on (one for each end of the part). We also checked the second unit for magnetic fields, and they were within the customers requirements. (We do a de-mag after every mag-particle inspection).
As far as the arc location, we start out at about a 16" diameter, (using 3/32" metal cored wire at 450-500 amps, 36-38 volts) and use a 2" offset from the top center of the part. The bead shape looks good, does it sound about right?
We are welding castings together (1 1/4 Cr, 1 Mo, 1/4V composition- 450F preheat), but the weld is just as likely if not more-so, to start a flash/sputter episode when we are welding on top of previously deposited weld as it is when tying into the casting wall. (the castings are also UT inspected before welding) As for parameters, we do need to keep the heat input relatively low (hence the 3/32" wire), but we have tried adjusting the parameters within reason. We have fiddled with things such as 1 1/8-1 3/8" stickout (longer stickout tends to get ropey beads that move around too much), 450-550 amps (our welding head won't push the wire out much faster than 550, and the beads get too small under 450) 34-40 volts (too low gets high/ropey beads and too high doesn't look bad, but I wonder about fusion for such a wide bead). Travel speed is kept between 20-25 IPM to keep heat input relatively low.
I can definately relate to the productivity vs. quality thing. No one ever sells process stability when they are talking about how fast their new welder/process can put in the filler metal. Our customer keeps saying that automating as much as possible can get rid of a lot of problems. Well, how do you develop the tolerances/quality/equipment/controls needed for automation when you are only doing 6 parts, and the customer certainly isn't paying any development costs. They want their first part to be a production part.
Also note that I'm not saying it can't be an equipment problem. I just can't afford to go out and over-react and buy a $60,000 system and lo-and-behold, still have the same problem. It would be nice to know what the problem is before spending money on by-guess and by-golly.
The off set sounds right but on that size OD you could get more. Just to where the flux wants to roll down the side. We use a finger to catch the flux and keep it piled up to cover the arc.
If the pile gets to skimpy of flux to much of melts and begins trys to travel with the work. On the other hand if you get to far down the puddle tries to roll back over it.
Hope This is of some help. Each time we encounter this and make changes I can't help but wonder if we really fixed or did the Gremlins just got else where to plague some one.
We have to use a wire brush to keep the flux in the groove since we are using a flux tube that deposits the flux in front of the arc instead of concentric. I know what you mean about the gremlins. Part #1 had the most problems, but the end quality was ok. Part #2 went just about perfect, and the quality was great. Part #3 & #4, the problems start coming back. What gives? Must be the gremlins you speak of.
Hi, GRoberts, RonG,
I haven't been around a sub arc machine for about 25 years, and then we used it to weld plate girders, no circumferential welds. I have an old chart that shows the suggested off center distances for the electrode, for circumferential welds, and it indicates that for diameters of 3" to 18", the electrode dispersement ahead of vertical center is 3/4" to 1". This seems to be a rather broad diameter range, especially since you both have indicated using 2" or possibly more on a 16" diameter. The chart also indicates that the electrode should be positioned nearly perpendicular to the surface of the work, particularly on smaller circumferences, and goes on to say that positioning the electrode at too great an angle of incidence may produce a distorted bead and may cause difficulty in striking the arc. Also, excessive or inadequate displacement produces a distorted bead. It is not my intention to disagree with anything that has been posted, I'm just wondering as to whether any of this might have anything to do with the ugly welds mentioned in GRoberts initial post.
Yes we have the same references. My point is the the larger the OD the nearer to flat the work is.
If I recall correctly Mr Robert said he was using 3/32" metal cored wire. The wall thickness of that wire is very thin and does not always burn off nice and straight and uinform thus you have the same potential as with SMAW when you get to much angle on your rod.
I forgot to mention I always try to keep my wire pointed at the center of my work.
In my experience with SAW (I have a few years at useing a #500 roll of wire with out ever stoping once) rotating grounds have caused more problems that any thing else. We use some high priced ground connections that work well but like every thing else "Out of sight out of mind" and if its working good its easy to forget it needs a little maintenace now and then. Its one of the simple things that slip by when your thinking about all the other varibles.
We have another shop on the West coast that uses a webbed copper strap with weights looped around the work for a ground and its constantly sliding on the work. I don't care for that approach but it works for them.
I don't know where the reference came up with the 3/4-1", but it seems like it would vary quite a bit depending on your variables. Just tonight, near the completion of the weld, where the OD ends up being about 26" (we have about 2" of weld left for an overall OD of 30"), the operator changed rolls of wire and did not check his electrode placement off center and it ended up with a slighlty drippy looking weld, and we found that the electrode was only about 1 1/2" ahead of center. We moved it back to 2", and the weld flattened right back out. If we were running a slower travel speed or smaller weld puddle, then less offset might work, but we haven't been experimenting with that.
The company president called me today and asked if I could get ahold of Lincoln to see if they have sub-arc experts they can send out to take a look at our process and see if they can come up with any improvements on how we are operating. Anybody here know a good place to start on who to contact. My local Lincoln rep is rather inexperienced, and new to the company.
If you don’t mind me asking why did you choose to use a metal core wire in lieu of a solid wire for this application? Have you tried to weld any of the parts using a solid wire? If not, would it be possible to weld 6 or more parts using a solid wire to see if you are still getting the same problem.
While metal core wires have a place, I’m not sure this is one of them.
I share your concerns about the metal cored wire for SAW. If I had been able to find a solid wire four the application, I would have used it, but as of yet, I have not been able to find one of the 1 1/4 Cr-1Mo-1/4V composition. We only use about 700lb at a time, so ordering a whole heat of solid wire didn't seem practical at the time of PQR development. I personally believe that the metal-cored wire, although not the cause of the problems, may contribute to it due being a less robust process. I just don't have an alternative at the moment that I know of.
And thanks to Chet and Bonniweldor? for the helpful contacts by e-mail
I’m not sure if you looked at any of these wires but, as a suggestion you might try using the following: Metrode weld wire SA 1CrMo (1.2 Cr – 0.5Mo) with the LA121 Flux (also from Metrode) or Bohler Thyssen S2CrMo (EB2R) wire with UV420TTR Flux.
While both of these wires do not contain any significant amounts of V they might be applicable to you application unless you are bound by your customers requirements.
In order to have the same creep strength at 1175F (service temperature), the customer is requiring a matching weld metal with a normalize and temper PWHT. I don't belive that most non V low alloy creep resistant electrodes would meet our customer's requirements. The metal cored wire is basically a chemistry match to Oerlikon OE-N125, which also meets our customer's specification. Metrode and BTW also make a SMAW electrode of the same type.
Do you have any capability to independently monitor/record the wire feed speed and the travel speed? Say, affix an encoder wheel in the right places and run the output to a computer? Do not use OEM systems for this as they are likely filtered to give a stable reading. This is valuable data when attempting to divine a cause.
Try radiographing the metal core to see if there can be seen any evident variation in the core packing density. Lay wire sections side-by-side to get the greatest length in the picture. Do a few images so a reasonble length of wire can be sampled.
Also, if the operator notes any characteristic instabilility, stop immediately and take a sample of the wire for investigative purposes. See if the process exhibit instability upon re-start. [ if this is at all feasible within the context work and specification ]
Sometimes when we get arc instability, it will actually arc back to the tip and stop welding. We have been taking samples of the wire from that spot to verify that there is metal fill, and so far there has always been. However, I had't thought of the X-ray idea. That would be fairly easy to do. How hard is it to spot variation using this method?
So far, we have a couple things we are going to try for smoothing out the wire feed before we start again. We have to use several extensions on the nozzle in order to get the machine up high enough to avoid some extensions that stick off the part. The generic extensions we have been using have a 1/4" ID, but we are going to try some with a 1/8" ID so the wire doesn't have as much room to move around, and should also add to the column strenth of the wire. We are also going to change from using the 1/16"-3/32" drive rolls, to using the 3/32"-7/32" drive rolls. We think this might get a better grip on the wire without increasing the force on the drive roll itself. We are also going to try a different wire straightener as suggested, as the one we have from Lincoln is for solid wire and seems to be relatively innefective on the metal cored wire. We will probably be starting up again in the last week of July. Hopefully, one of these ideas will help with our project.
Thanks again for everyones helpful ideas.
Just guessing here- a wire with that much current generates a substantial magnetic field. If it moves the field moves with it. Moving magnetic fields generate electrical currents in anything conductive that they surround, which generate magnetic fields of their own. Those fields could exert forces on the wire causing it to move again. Without some damping force this could go on forever.
I think your idea to exert better control of the wire in the extensions is a good one. It might help to make the extensions of a non conductive material (plastic or ceramic, I don't know the temperatures the parts need to survive). On another theoretical track (that the disturbing magnetic field has an external origin) making the extensions out of magnet iron could help.
Remember, Just guesses.
Bill
Forum
