MIG Robot Welding Carbon Steel Lamp Posts

LINCOLN YOU LIGHT UP MY LIFE:

One customer I reacently assisted manufactures carbon steel street lamps 11 to 7 gage. It’s a simple manual welding application. On the end of the lamps they weld a floor-mounted flange. The flanges are typically >13 mm. They also weld around an access box located on the post, near the flange.

This application became unnecessarily complex the day they decided the parts should be welded with a robot. The company ordered a Fanuc ArcMate 100 robot. The robot came with the Lincoln PowerWave, pulsed, 450-amp power source.

The robot system was sold by AGA who had the technical support of Lincoln / Fanuc. Almost two years after the robot was installed the robot had never come close to it’s daily weld production quota.

When the robot was installed it was placed on the production line, however numerous weld issues occurred and the robot was quickly moved to another part of the plant. The robot was moved to a location where personnel could “play around” with the settings. With the assistance of Lincoln, Fanuc and AGA they played for almost two years.

The robot welding issues generated were numerous and the plant personnel viewed the robot as a liability, not capable of meeting the weld production goals. In the 24-month playing around period, AGA and Lincoln personnel made numerous visits to try to assist the plant. In desperation the company actually paid Lincoln to send in one of their more experienced technical reps..

EXCESSIVE ROBOT GUN TOUCH SENSING.

To locate the two welded parts, the robot first used its touch sense feature. The robot was programmed to use 28 touch points for one square and one rectangle part. The touch sensing took 50 seconds or 20% of the 4 minutes 10 seconds robot cycle time.

I believe that in the attempt to resolve the robot “weld process issues” too much focus was placed on the robot bells and whistles rather than on optimizing the weld process. The amount of touch-sensing was beyond excessive.

Weld gaps were an issue partially addressed by the touch sensing. However the real gap issue was generated by the assemblers. When these guys tacked the parts they did not evenly distribute the gaps. Once I provided these instructions the gaps were within 060 (1.6mm), which is acceptable for this application.

YOU SELECT THE WRONG WIRE SIZE AND THE WELDING PROBLEMS BEGIN.

Two years previously, when the first robot welds were made, someone recommended the pulsed MIG process, using an 045 (1.2mm) wire. The resulting pulsed welds were too hot. Someone on the Lincoln /AGA team then for some strange reason recommended an 045, E71T-1, all position, gas shielded, flux cored wire.

When I arrived at the plant the following daily weld issues occurred,

[a] Inconsistent weld results, requiring constant robot program changes.
[b] excessive weld burn through,
[c] lack of weld fusion,
[d] excessive undercut,
[e] slag entrapment.

THE WELD $AVINGS BEGIN.

The first thing I checked out was the reliability of the Fanuc robot touch sensing equipment. It worked fine. We reprogrammed the robot to touch the parts only six times. We tested the repeatability it was OK. The touch time cycle per part was reduced from 50 seconds to 10 seconds.

WELD CONSUMABLE LOGIC

Any one who has read my MIG books would be aware of the fundamental fact that the most superior wire size for parts less than 3/16 (4.8 mm) is an 035 (1mm) wire. Especially when the parts have weld gaps.

I replaced the 045 flux cored wire with the 035 MIG consumable. I welded the parts first using a combination of short circuit and spray transfer. I then set the Power Wave to produce pulse welds on the parts. All the welds produced were optimum from a quality and deposition rate perspective. Proponents of pulsed note. The differences between the pulsed and traditional mode welds were not measurable.

MAKE SIMPLE ROBOT MOVES.

For the robot to weld all the way around the square end flange weld joint, which by the way was in a fixed position. The robot made two welds, one each side, each move wrapping halfway around the square joint. The robot moves were complex, stretching all 6 axis to their limits. In stretching the robot reach limits many of the welds did not have the weld gun positioned for optimum weld control.
From a programmer perspective, when robot program points become complex, they eventually will require correction. In these circumstances the robot operator may find it difficult to duplicate the initial program moves..

I had the flange and box parts reprogrammed and produced four simple straight welds.

AFTER 3 DAYS AT THE PLANT THE PAY OFF.

With the new robot program in place, the 035 wire and new weld procedures, we reduced the total cycle time by 50%. We welded 20 lamps and did not have an issue with a single weld. The bottom line, the customer now had a stable process and could now produce in two shifts what they were going to produce in three shifts.

Conclusion: When the 045 MIG wire was recommended it required pulsed weld parameters that were simply too high for the thin gage parts and gaps.

IN MY BOOKS PART OF THE RECOMMENDED ROBOT WELD PROCESS CONTROL PROGRAM, IS EVALUATE THE SIZE OF THE WIRE AND THE WELD CURRENT USED ON THE APPLICATION

Recommending a “deep penetrating” 045 all position flux cored wire for the pulsed power source made little sense. Some of the flux-cored welds were made in the “vertical down” position. For the flux cored wire to function on the thin gage post, low weld setting had to be used. The low flux cored welding parameters, and vertical down weld positions ensured lack of fusion for the thicker flange side of the weld. Also vertical down resulted in slag entrapment.

In contrast to the 045 wires, the 035 MIG wire was the key due to its current range compatability with the application. This part did not require the pulsed transfer mode, however as the company had purchased an unnecessary pulsed power source, I left the settings in the pulsed mode.