Automotive Tier 1 Robotic Welding Assembly - Bottleneck Easement Weld Shop Examples

Situation: A stamping and welding plant consisting of 150000 square feet, 350 employees had a number of internal throughput bottlenecks which resulted in a delivery record of 97%. In automotive tier 1; the expected minimum metric is 99.99%. The plant contained 10 side frame presses ranging in size from 400 to 2000 tons. The majority of the presses were coil end fed but could be configured differently based on the product. The welding shop contained 50 robots with 25 performing mig welding and 25 units spot welding.

The plant suffered from the following problems:

Press shop up time of 18-25%. A well running press shop will run at 85 to 90% up time.
Excessively long change over times of up to two days. Keeping in mind that some shops change their presses in minutes, a 4 hour changeover for this shop would be considered acceptable.
An epidemic of smashed stamping dies which resulted in a tool room running chronically 100% over budget.
Weld shop rework in excess of 60% in an industry where 2% on a bad day is considered acceptable.

Weld Shop Case Examples

Sport utility gas tank protective cover robotic spot weld cell
Bumper mounting bracket weld cell
Shielding gas changeover
Tailgate spare tire mounting bracket weld cell
Car Bumper weld assembly cell
Seat belt anchor spot welder
Spot welder deformed welding tips
Mig welding process standardization

1. Sport utility gas tank ground shield robotic weld cell (special cause quality problem):

Situation: A very large $2 million robotic weld cell having a total of 6 robots, spot welded a variety of brackets onto a gas tank shield for a sport utility vehicle. Two robots loaded and unloaded raw stampings and finished assemblies. Four more robots inside the cell performed a total of 35 spot welds on the two assembies. The cell had the following operational issues:

A chronic problem with an average of 10 cold welds per assembly.
A high level of downtime. The cell required the constant attendance of a robotics technician to clear low pressure faults and restart the cell.

Analysis: Observations of the cell concluded that the cold welds were probably due to interruptions in the weld cycle due to pressure faults or insufficient air pressure. Pressure gauges observed from a distance at each pressure multiplier were fluctuating wildly indicating a lack of sufficient compressed air to maintain pressure during the spot welding process. As further evidence of an attempt to fix the problem, there was a disconnected 30 HP piston compressor sitting on the floor nearby. The compressed air from a central compressor was distributed through the following items:

2" main appliance shutoff ball valve,
Air receiver of size 3 feet diameter by 6 foot length,
2" main applicance compressed air pressure regulator,
1 1/4" branch pressure regulator installed on each zone,
1" individual pressure regulator at each spot welder pressure multiplier.

There were two problems with this arrangement:

The air receiver was installed up stream of the first regulator and should have been installed downstream of the first regulator. The first regulator is intended to buffer the robot cell from fluctuations in the plant air supply pressure. The receiver was moved to the downstream side of the 2" regulator.
There were two many regulators in the air delivery piping and as many orifices which acted as pressure restrictions in the air supply piping inside the cell. The zone regulators were removed.

Outcome: 1. A 90% reduction in pressure fluctuation on all gauges was observed when spot welders were engaged. 2. Pressure faults were eliminated. It was no longer necessary for a robotics technician to be present to keep the cell running. 3. Most importantly there were practically speaking, no more cold welds; Product quality in terms of nugget weld penetration was now excellent.

2. Bumper mounting bracket weld cell (common and special cause quality problems):

Situation: This cell was reported by the materials manager to be the number one material delivery problem in the shop. The cell was producing 300 pairs per shift and needed to produce 600 pairs per shift. Overtime on weekends was not sufficient to make up the shortage. The part was an "L" shaped angle bracket of size 4.5" by 1 1/4" on the legs of the L by 4.5" long. The finished part had a 4.5" stiffener rib welded to the middle with two 1" by 1/8" fillet welds on each side of the plate. Two nuts were projection welded onto one of the L legs. Two threaded studs were pressed into punched holes on the second leg.

A turntable with four stations was installed inside the fenced controlled entry cell:

Manual load.
Fillet welding of the stiffener bracket by two robots.
Projection welding of the nuts.
Unload by robot to a bolt press insertion station followed by mechanical ejection after threaded stud installation.

The raw material was cold rolled low carbon steel of thickness approximately 10 gauge. The welding process use flux core wire of size 0.030" and was unshielded.

Analysis: Weld quality was consistently poor and out of position, with visually poor quality due to flux spatter, and poor adhesion. The cell suffered a high level of downtime due to the constant need to reteach the robots to address out of position welds. Further projections nuts were also out of position with winking holes. The parts required 100% manual rework with a hand held un-shielded mig torch off line. Expensive flux core wire was being used. Due to no mechanical ventilation from this cell and the toxic fumes from the flux core wire, there was also a respiratory breathing hazard present in the shop which had gone unnoticed and unaddressed. The cell was running with 100% rework.

Actions:

An experienced welding industry consultant recommended that the welding process be converted to metal inert gas shielded arc welding using hard wire of the same gauge thickness. The flux core wire was deemed unnecessary and the cell should be converted to use a mig process with hard metal wire. Power supply voltages should be increased to achieve droplet transfer from globular transfer. Droplet transfer was considered to produce a far superior weld quality in the industry. A 13% carbon dioxide argon mix shielding gas was recommended. The plant was using a purported 9% mix. Note that carbon dioxide is used to create an exothermic weld puddle and a weld with good penetration. The specialist indicated that 9% mix produced lower penetration quality.
It was observed that none of the fixtures had jam nuts to secure and lock the fixturing positioning adjustments in both the x and y directions. The adjustment screws were quite loose. If the fixtures were moving, this would cause positioning problems with both the fillet welding and projection nut welding processes.

The cell was shut down for five days:

All four fixture stations were adjusted to present a consistent position of parts at the fillet weld station. Jam nuts were installed. The table was rotated to verify the part presentation consistency at the fillet welding station.
Bottled 13% shielding gas was installed at the cell to facilitate the test. Coils of hard wire were installed in the feeders.

Outcome:

Within days of making these changes, the cell was running at 750 pairs per shift.
Weld quality was very consistent, far superior and had excellent penetration. Note, during the changeover, the robotics technicians were not able to achieve acceptable weld penetration with the centrally piped 9% CO2 shielding gas and achieved good penetration when the cell was rigged with gas bottles containing 13% CO2/Argon blend shielding gas.
The cell was now running with less than 10% manual rework.
A 40% savings in weld wire consumable was realized.
Delivery performance on this component achieved a consistent 100% on time delivery to the downstream bumper final assembly weld cell.

3. Shielding gas changeover (common cause quality problem):

Situation: The plant purportedly used 9% carbon dioxide argon mix when experts recommended that a 13% mix produced optimal metal inert gas weld quality. The plants 25 gmaw/mig welding robots installed on 10 or more weld cells were not producing consistent weld penetration in the lab where welds were sectioned and acid etched each shift.

The plant purchased both gases in bulk liquid form. Evaporators deliver both gases in gas form to a central mixing station containing pressure regulators and fixed orifices that allowed for the delivery of a consistent mix through an overhead pipe distribution system to each cell.

Given the vastly better penetration achieved by switching the bumper mounting bracket weld cell to 13% CO2 gas, it was felt that the central system should be adjusted to deliver the same mix to the whole shop. Accordingly the liquid gas vendor was contacted to bring a meter and assist with the mix adjustment.

On the day of the changeover the gas vendor measured the existing carbon dioxide concentration to be less than 3%, not 9% as previously thought. The concentration at the mixing station was then adjusted to 13%.

Outcome: All mig weld process quality penetration problems were eliminated everywhere in the shop. The temporary employees (students) responsible for sectioning welds indicated they were getting penetration where that had "never seen it before". This is an example of a common cause quality problem in an automated welding process.

There were other root cause implications with this problem. It was unknown for how long the shop had been running with a very low level of CO2 in the shielding gas. Given the importance of the shielding gas to achieving adhesion and penetration in this process it is reasonable to conclude, that this explains why the shop started using sub-optimal mig welding process variables, such as oversize wire, flux core wire, and metal core wire to achieve acceptable mig weld quality.

4. Tailgate spare tire mounting bracket weld cell (special cause quality problem):

Situation: A small weld cell having one robot welding mounting brackets to the back side of a tailgate spare tire mounting bracket was experiencing a high level of burn through. The process was experiencing a high percentage of scrap in the order of 25% due to burn through or blow out. The mig welding process was using wire thickness 0.045".

Analysis: The material being welded was 0.010" thickness. All other low carbon steel metal stampings of similar thickness in the shop were being welding with 0.030" hard wire. The wire was deemed to be to thick for the material. The filler wire was switched to 0.030" mil thickness. Appropriate changes were made to other welding parameters such as stickout distance, power supply voltage and metal feed rate.

Outcome: A 99% reduction in scrap was achieved with this trivial change.

5. Car bumper final assembly robotic weld cell (cost savings opportunity):

Situation: A weld cell having two identical and parallel flow paths to achieve needed throughput, welded the brackets in case two above to the inside back of a automotive bumper. On the car, this steel bumper would form the inside backing of a painted polymer facia on the front or rear of an automobile. Four robots two on each flow path used metal core wire to mig weld the mounting brackets to the inside cavity of the main bumper stampings.

Analysis: The material being welded was 10 gauge low carbon cold rolled steel. The metal core wire costed approximately $4.50 per pound at the time and was deemed unnecessarily expensive compared to hard metal wire which cost $0.45 per pound. All four mig welding robots were switched over to the hard wire.

Outcome: A 90% savings in weld wire consumable was realized.

6. Seat belt anchor spot welding machine (special cause quality problem):

Situation: A banana shaped stamping made of 10 gauge cold rolled steel approximately 4.5" wide by 24" long and having a "T" bracket spot welded on it was being weld assembled on a small PLC controlled, two on, spot welder. The "T" bracket was welded to the stamping with two spot welds about 2" apart. Two parts were fixtured on a table. The operator stood infront of the welder and activated the process with capacitive style swipes. Two electrodes would lower and weld the first part. The electrodes would retract. The table would shift to locate the second part under the welding guns. The guns would descend and weld the second part.

Analysis:

The process produced parts with a lot of cold welds. Often one cycle of a total four welds would have 3 cold welds. A cold weld is one which has no fusion and pulls apart with ease.
The process was running with an unusual level of weld expulsion. Weld expulsion is literally an explosion of liquid metal from the spot welding process. Employees' clothing would be full of holes after a shift. Employees were forced to wear special aprons and face shields to protect themselves from the constant expulsion. Nobody wanted to run this welder.
The welding machine required frequent robotic technician attention to adjust and re-secure proximity part presence sensors.
Part shipments to the end customer were chronically late.
Given the high level of cold welds, there was a high probability that this problem was being leaked to the end customer.

The root cause stumped all of the site staff. At one point in frustration, staff were in the process of organizing a completely manual off line spot welder.
After observing the welder run for a period of two hours, it appeared that the guns might be coming off the part early before the spot welds were completed. A spot welding process employs very high amperage currents of 25000 to 30000 amps for a very brief period of 6-10 hertz (1/10th of a second). If the low voltage high amperage current was still on as the welding tips pulled away from the part, an electric arc having a temperature of 20000 degrees celsius would be created. This would be more than enough to vapourize any metal grounding the arc. A great deal of expulsion would result.

A PLC programming specialist was engaged to examine the program. The following problems were found with the program:

There were no time delays between the completion of the weld cycle and the removal of the guns with the result that an arc could be pulled for a brief period.
The outputs used to control the air valves that lowered the weld guns were not latched once the limit circuit which included a variety of sensors was made. This allowed the guns to pull off the part if for example a part presence or table in position sensor fluttered. Given the frequency of loose sensors and the hourly need to tighten them to keep the welder running, this would also be the source of an even greater amount of expulsion. Further, the air over hydraulic spot weld guns clamped the stamping with tons of force to ensure intimate contact between the two parts to be welded. Together with the thermal stresses from the spot welding process this would result in warping of the sheet metal stampings and loss of contact with the part presence sensors. Any of these conditions would result in a gun pulling off the part during the spot welding process resulting in liquid metal expulsion from the arc created.

Actions:

The PLC programmer was asked to create a new program starting with a blank sheet of paper. This program would include process timers separating the start and stop of the spot welding power from part presence and machine motions (guns down and up etc.). Critical outputs would be latched to eliminate any chance of guns being withdrawn while the welding current was on.
Jam nuts were installed on all of the inductive proximity sensors on the fixturing.

Outcome:

The welder started up with a complete elimination of cold welds. In inspection, the parts pulled apart, with a consistent adhesion nugget being pulled out of the parent material indicating excellent and consistent adhesion.
There was a complete elimination of expulsion.
The part proceeded to run with 100% on time delivery.

7. Spot welder cold welds and deformed (mushroomed) copper tips (common cause quality problem):

Situation: The shop contained about 25 spot welding robots or guns in total. There was a daily common cause problem associated with cold welds occurring at seeming random.

Investigation: With many of the above crisis type problems addressed, focused attention could now be devoted to this problem. When such problems occurred it was common practice to change out the copper tip. Copper tips cost a few dollars each, are hollow for water cooling and are threaded for easy installation onto the tip of a spot welding water cooled copper electrode.

Inspection of 30 of the mushroomed tips removed from the process found that there was a thick layer of calcium on their insides. It was theorized that the calcium was increasing the resistance to heat transfer resulting in a hot tip which would mushroom after a relatively short term in service (a few shifts).

The spot welders received cooling water from a centralized cooling water distribution pipe system. A review of the water softener installed on the makeup line to the system indicated that the softener had probably not been in service for at least a decade.

Given that the plant was located in a city near the Niagara Escarpment and that the town got half of its water from artesian wells, a water treatment company was engaged to test the water in the city supply. A hardness of well in excess of 300 ppm (soft water <17 ppm) was found (off the meter's scale) was found.

A new water treatment system was installed. At the same time, the water treatment supplier was engaged to acid flush the system to the spot welders with sufficient acid to remove the calcium scale. This process took a full two days on a weekend involving over 20 hours of flushing before the re-circulation water pH started to drop. It was estimated that several hundred pounds of calcium sludge was collected in the welder point of use filters. Filter elements had to be changed several times before they stopped fouling during the first week of operation after acid flushing.

Outcome: There was an estimated 95% reduction in the incidence of mushroom tips with a corresponding reduction in cold weld frequency. Case examples one and six above are special cause problems. This case is an example of a common cause problem in a spot welding process.

Mig welding process standardization: In this robotic weld assembly plant, all of the stations were essentially welding the same type of material, cold rolled low carbon steel of thickness 0.100". A specialist from a leading welding industry vendor, after reviewing the shop, concluded, that there were as many different mig welder process setups as there were processes. He recommended that they conduct an automotive mig welding process best practices class at their office.

At this three day class, welding engineers, robotics technicians and key shop leadership were taught the current state of the art in running an automotive stamping and weld asembly welding shop. Key to this were the following recommendations:

Change the shielding gas to 13% carbon dioxide.
Standardize the weld material process to droplet mode from globular transfer. Set all power supply voltages to a standard setting to achieve this.
Set all other weld parameters such as stick-out distance, wire feed speed, shielding gas flow, and torch travel speed the same settings, needed to achieve acceptable single pass bead size, penetration and weld quality.
Outfit all cells with hard filler wire of size 0.030".

After completing this class, the robotics technicians proceeded to launder the shop setting up all of the cells to a similar set of process parameters using only one type of wire. As this process was completed together with the completion of the above projects the following improvements accrued:

A marked improvement in visual mig weld quality was achieved having good penetration everywhere with very low levels of spatter.
The level of manual rework with hand held mig torches dropped from over 60% to under 10%.
A $40000 per month savings in weld shop consumables was realized.
The on time delivery record for the site moved to 100% consistently.
Internal scrap dropped to nil.
Premium freight of $70000 per month was saved.
The shop absorbed a 20% increase in sales without the need to hire anyone.
Bottom line profit almost doubled from 12 million to 23 million per year.

The general manager in trying to ensure a best in class welding shop had recruited a number of engineers with a welding major. In dealing with welding problems in this shop these individuals, with great pride, had tried to fix basic welding problems by changing the welding processes by for example, using bigger wire, more amperage, use of metal core wire and flux core wire and finally a less than optimal shielding gas. Further, the shielding gas, which was determined to be the most important variable in achieving good weld penetration, was not being routinely monitored. As a result the concentration had dropped to an extremely low level, resulting in poor or no penetration consistently. The result was a shop with a chronic rework level of over 60% and one which was failing to deliver on time.

All of the solutions implemented here were process related and did not require any prior welding knowledge.