Robotic Welding
I heard some things from a good friend that I would like to clear up and see if they are true or not. I'm guessin most of it is since he's been a welder for 10 years now. He says that robotic welding is accurate for short welds but a longer weld it is not as accurate as a good/excellent human welder because the robotic weld cannot feel the welding surface (ss, aliminum, etc) getting hotter and it will continue at a steady pace where as a human will start slower and gradually get faster. So pretty much the weld that the robot is making doesn't take into account the metal caracteristics as it gets hotter and hotter is this true?
true. look through this post. i think it says something about his company not even using the robotic welder they have because he can outweld it
https://honda-tech.com/zerothread?id=1030027
https://honda-tech.com/zerothread?id=1030027
<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by 98AccordEX »</TD></TR><TR><TD CLASS="quote">true. look through this post. i think it says something about his company not even using the robotic welder they have because he can outweld it</TD></TR></TABLE>
Not true. Technically.
The plant I'm working in uses a fair amount of robotic welding. I'd put good money that the reason Engloid can outweld the machine is because he works in an area that does a lot of non-standards.
We have one machine, directly down from my office, that runs all day long. It's purpose is to weld muffler bodies (Semi-truck mufflers, think 4 feet long and 3 thousand dollars) down the seam. It runs a 4 foot seam bead all day, 23 hours a day, and never produces scrap.
Several more machines down the way, there is a robotic welding department with machines that do tighter, more intricate stuff. They only produce scrap when Human error is involved. It is the same thing every time though, that's why humans can SOMETIMES outweld machines. We also have an entire weld department with people welding. They do the non-standards or short orders, and the inside baffling that the machines can't reach.
Not true. Technically.
The plant I'm working in uses a fair amount of robotic welding. I'd put good money that the reason Engloid can outweld the machine is because he works in an area that does a lot of non-standards.
We have one machine, directly down from my office, that runs all day long. It's purpose is to weld muffler bodies (Semi-truck mufflers, think 4 feet long and 3 thousand dollars) down the seam. It runs a 4 foot seam bead all day, 23 hours a day, and never produces scrap.
Several more machines down the way, there is a robotic welding department with machines that do tighter, more intricate stuff. They only produce scrap when Human error is involved. It is the same thing every time though, that's why humans can SOMETIMES outweld machines. We also have an entire weld department with people welding. They do the non-standards or short orders, and the inside baffling that the machines can't reach.
I know nothing about robitic welders, but they could be programmed to increase speed as they moved along, it wouldn't be difficult to do. I'm not sure if their are any that do that though.
Originally Posted by Goullish
I'd put good money that the reason Engloid can outweld the machine is because he works in an area that does a lot of non-standards.
So far, I agree with nearly all the answers above, in this post.
Yes, you can program most of them to increase speed. Actually, all of them I've used, you could. However, you sometimes can't time how long it will take, with given amperage, to form a puddle the size you want, or how well it will dissipate as you travel along. Granted when welding the same parts over and over, you can get pretty close though. When welding manually, you can hammer the heat and back it down when the puddle is formed.
You can also have issues with the program (or programmer). If the programmer isn't very good at optimizine the program, you can be overlooking a lot of time savings that could otherwise be gained. In other words, the programmer should program to get a good weld, then increase speed (and other parameters, accordingly) until quality diminishes, then bump back town. That ensures getting it as fast as possible. Most places I've worked, they had programs running a lot slower than they had to be.
For one, there's differences in machines. Some will run a set program every time. You pick the program, hit go and it welds. Others, it will sense the arc gap and adjust to any changes in the surface to keep the gap constant. All machines have certain ranges that they will work in. For example, it may only be ble to move the tungsten within a 1/4" range. If you start the machine at the lower end of this range and then it needs to drop the tungsten closer, it may exceed the limits it can move in...and just stop in middle of the weld. A human will not do this.
Some machines, you can watch them and adjust them on the fly...bumping it left, right, up, down...more amperage, less, etc. These are nice for fitups that are not always precise.
The big problem with any machine like these is that the fits are of great importance. However, it's not really dependent on the length of a weld. I recently worked at Heil, where they make the tanker trucks. They have a seam welder that runs the entire length of the tank with one pass. It will put out hundreds of welds with no mistakes...only if the operator sets it up right, the fitup is correct, and the operator makes the proper bumps to keep it inline as it welds. Likewise, say you take a weld like the one below:
When these are done, there is no bevel. It's just 2 pieces of pipe butted together. If there's a gap of about 1/16" anywhere, the appearance of the weld will be affected. It will either be low in that point or even blow through. If I was welding this by hand, I could add a little wire to it at the gapped area.
So, if you have a lot of parts to do, can get to them with a machine, and the fitup is consistent, it will usually work fine. The versatility of a human is why humans will never be out of the welding industry. As for the comments I made about my outwelding the machine...in that case it is very true. The fitups were not consistent enough to use a machine to weld them. In order to get precise fitups that these welds would need, they'd have to machine the bevels. The cost of the prep work would be more than any labor savings on the welding end. Also, you have to consider that the machine has to be set up on each fit. WHen a weld is messed up due to small issues in fitup, they can be time consuming to repair.
You also have to consider that in most applications where these machines are used, a person's time is consumed by either operating it or waiting on it, then setting it up for the next part. Therefore, since you already have one person's labor time consumed with it, you have to save significantly somewhere to make it worthwhile. This could be saved in consistency, quality, speed, or another way.
Lets say that you're welding the end cap on a stainless steel muffler. Just welding a corner joint. If you can put it in a machine and make that weld in a minute seconds, without repairs...you still have a man occupied by loading the next part and starting the machine. Quality can be good and consistency can be good. Now compare to manually welding: You can put it in a turntable and weld it by hand with the same minute, with equal quality and a lesser failure rate, because the human can add wire to fill any small gaps. Sure, the appearance may be slightly lesser, but well within the quality standards. You have likely saved a lot of time on the fitup since it doesn't have to be perfect...and saved the money you would have spent on the machine.
Another example: Quality and speed is good with human or machine...but if you want the good quality by hand, you have to pay a $18/hr welder. You can achieve that same quality with a $10/hr man that operates the machine. In a 40hr week, you've saved $320, for a yearly savings of $16,640. The machine will pay for itself if production continues long enough...and it's a lot easier to replace a $10/hr operator than an $18/hr welder.
There's many factors that play a part in the decision of whether to use a "robot" or to weld by hand. There's really no set rules of what's best or what criteria to determine it by.
I've rambled long enough...
oh, and for those intrigued by these machines, they will range in price from about $10k-70k and even higher...
<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Engloid »</TD></TR><TR><TD CLASS="quote">Another example: Quality and speed is good with human or machine...but if you want the good quality by hand, you have to pay a $18/hr welder. You can achieve that same quality with a $10/hr man that operates the machine. In a 40hr week, you've saved $320, for a yearly savings of $16,640. The machine will pay for itself if production continues long enough...and it's a lot easier to replace a $10/hr operator than an $18/hr welder.
oh, and for those intrigued by these machines, they will range in price from about $10k-70k and even higher...</TD></TR></TABLE>
This is where the particular plant I work for makes it work.
For the seam welder (there is actually 2) it takes one man to operate it. Typical load time of a body is 10-12 seconds, and then a minute, a minute five to run the bead. During that 50-55 seconds of free time, that man is actually doing other things to the bodys that just came out, such as loading core assembly's, or baffle assembly's, to hand over to the human welders right next to him. Then he unloads in 6 seconds, throws another one in, and keeps going. The guy I share my office with did a time study on that particular machine, with the guy fully chuggin and using a single program, it will put out 63 muffler bodys in an hour. They require 35 to maintain the line. Show me a human that will put out 40 4 foot welds in an hour, for even 4 hours straight, and I'll be impressed. (Engloid, you don’t count, after looking at your work I think you're just a robot wrapped in skin)
They save money with the machine on speed and consistancy. It does the same thing every time, and doesn't get tired. They also save money on personel because that one man is able to do 2 jobs.
On the end side, they save money with the other welders (the ones that weld the end caps and sometimes side tubes) because they use 2 men per 3 machines. It loads from the front, and kicks the part out the back. Plus, it does 2 - 4 welds at a time.
However, the big thing about these though is that it's all very simple. The welds are not complicated by anyone's imagination, part uniformity is huge, and they change very little, they need a new program maybe once every month, month and a half.
oh, and for those intrigued by these machines, they will range in price from about $10k-70k and even higher...</TD></TR></TABLE>
This is where the particular plant I work for makes it work.
For the seam welder (there is actually 2) it takes one man to operate it. Typical load time of a body is 10-12 seconds, and then a minute, a minute five to run the bead. During that 50-55 seconds of free time, that man is actually doing other things to the bodys that just came out, such as loading core assembly's, or baffle assembly's, to hand over to the human welders right next to him. Then he unloads in 6 seconds, throws another one in, and keeps going. The guy I share my office with did a time study on that particular machine, with the guy fully chuggin and using a single program, it will put out 63 muffler bodys in an hour. They require 35 to maintain the line. Show me a human that will put out 40 4 foot welds in an hour, for even 4 hours straight, and I'll be impressed. (Engloid, you don’t count, after looking at your work I think you're just a robot wrapped in skin)
They save money with the machine on speed and consistancy. It does the same thing every time, and doesn't get tired. They also save money on personel because that one man is able to do 2 jobs.
On the end side, they save money with the other welders (the ones that weld the end caps and sometimes side tubes) because they use 2 men per 3 machines. It loads from the front, and kicks the part out the back. Plus, it does 2 - 4 welds at a time.
However, the big thing about these though is that it's all very simple. The welds are not complicated by anyone's imagination, part uniformity is huge, and they change very little, they need a new program maybe once every month, month and a half.
Trending Topics
Who is Mr Robot?
iTrader: (2)
Joined: Jul 2004
Posts: 21,474
Likes: 10
From: ATL - Where the Pimps and Players dwell
<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Goullish »</TD></TR><TR><TD CLASS="quote">(Engloid, you don’t count, after looking at your work I think you're just a robot wrapped in skin)
</TD></TR></TABLE>
After looking at his pics numerous times, i have often thought this myself, lol
</TD></TR></TABLE>
After looking at his pics numerous times, i have often thought this myself, lol
Lets see if I can explain this without hurting anyones feel
ings
The only robotic Arc welding I have been around was used for dangerous **** like Lithium etc. and it worked fine but at a very high cost per weld. That is done with heat censoring unit in the torch which move the tungtin in .002 step's.
ok now automatic stright line (airline welder) or circumference weldor's weld at least 10X faster then the fastest man that's right 10X and make perfect welds every time.
When the money people come up with a price for a welded part they figure 1in per
minute by hand also for perfect welds that are going to pass any of the weld tests including x-ray.
Most if not all good welding jobs are under uncle sam's rules and that means you have to be hand weld certifed before you can make one tack weld and he don't care what your using automatic,hand feet whatever
ingsThe only robotic Arc welding I have been around was used for dangerous **** like Lithium etc. and it worked fine but at a very high cost per weld. That is done with heat censoring unit in the torch which move the tungtin in .002 step's.
ok now automatic stright line (airline welder) or circumference weldor's weld at least 10X faster then the fastest man that's right 10X and make perfect welds every time.
When the money people come up with a price for a welded part they figure 1in per
minute by hand also for perfect welds that are going to pass any of the weld tests including x-ray.
Most if not all good welding jobs are under uncle sam's rules and that means you have to be hand weld certifed before you can make one tack weld and he don't care what your using automatic,hand feet whatever
<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Goullish »</TD></TR><TR><TD CLASS="quote">Show me a human that will put out 40 4 foot welds in an hour, for even 4 hours straight, and I'll be impressed. (Engloid, you don’t count, after looking at your work I think you're just a robot wrapped in skin)
</TD></TR></TABLE>
I'm trying to think of how I could do 160 feet of weld in an hour... Hell, I don't think I have ever used a mig gun that could run high enough wire speed to do that. Even 600inches per minute of .045" wire isn't even going to do that.
MXnut:
The welding you're thinking of is just a difference in the definition of "robotic." To some here, the definition of robotic means that the welder (person) is pushing buttons, rather than holding a torch or gun. Personally, I figure that if you have to do anything other than tell it to start (bumping, adjusting, etc), it's an automatic machine, not robotic.
I recently worked at Heil Trailer company. They spent $500k on a robotic welder to weld the rear axle frames for trailers. It was fully programmable. You hit the button and could walk away... granted, it would ALWAYS put the welds in the same place, even if the weld joint wasn't there.
It was a high dollar piece of junk. It didn't actually sense anything...it just welded where it was told to each time.
One of the "automatic" welders I used would adjust the arc gap by sensing voltage (resistence) at the arc gap.
There are relatively few places that robotic welding machines are going to be useful and cost-effective.
</TD></TR></TABLE>
I'm trying to think of how I could do 160 feet of weld in an hour... Hell, I don't think I have ever used a mig gun that could run high enough wire speed to do that. Even 600inches per minute of .045" wire isn't even going to do that.
MXnut:
The welding you're thinking of is just a difference in the definition of "robotic." To some here, the definition of robotic means that the welder (person) is pushing buttons, rather than holding a torch or gun. Personally, I figure that if you have to do anything other than tell it to start (bumping, adjusting, etc), it's an automatic machine, not robotic.
I recently worked at Heil Trailer company. They spent $500k on a robotic welder to weld the rear axle frames for trailers. It was fully programmable. You hit the button and could walk away... granted, it would ALWAYS put the welds in the same place, even if the weld joint wasn't there.
It was a high dollar piece of junk. It didn't actually sense anything...it just welded where it was told to each time.One of the "automatic" welders I used would adjust the arc gap by sensing voltage (resistence) at the arc gap.
There are relatively few places that robotic welding machines are going to be useful and cost-effective.
Thread
Thread Starter
Forum
Replies
Last Post
Catch 22
Road Racing / Autocross & Time Attack
9
Dec 29, 2004 06:25 PM




