Dear Karthick M.,

You must avoid all pin holes in resistance spot welding, as it is prohibited in all weld acceptance specifications that I have ever seen.

The few cases that I have seen are caused by low tip force, either during the whole weld, or failing near the end of the weld.  
Since I know nothing of your application, or even what units you use in your location, I will give an example of U.S. automotive welding. 
A typical weld of electroplated galvanized mild steel of 1mm to 1mm, written as "G1.0G/G1.0G" would require: 
(Actually this is for metal of 0.80mm to 1.09mm)

GMT
Weld Force
K Amps
Pulses
Pulse Time
Cool Time
Tot Wld Time
Hold Time

mm
Lbs    kN
kA
#
Cycles   ms
Cycles   ms
Cycles   ms
Cycles  ms

0.80-1.09
585     2.6
8.0
1
9       150
0       0
9       150
2       40

Notice that the tip force should be 585 pounds, and this should be bumped up to 810 pounds force if this is a robot or a manual gun, 
because the approach angle is not always the same as it would be with a "hard tool" such as a press welder.  
This would use a tip with a contact diameter of 5mm to 6mm.

If you have set the tip force with an accurate force gauge, there are still some things that would cause a loss of tip force during welding:
1. Slow cylinder, not reaching full tip force in the allotted squeeze time.  This can be checked with a specialized "Squeeze Analyzer" 
   or with a peak reading force gauge such as the Sensor Developments unit that is popular in the US auto industry. (Ask for details - it is complicated.)
    
2. Poor follow up, the cylinder cannot move quick enough to maintain full force as the tips indent into the part.  This can be due to:
  a. Long air lines between the regulator and the cylinder.
  b. Restrictions caused by a small weld valve.
  c. Hydraulic cylinders.
  d. Oversize cylinder running at less than 40 psi to give the required force.
  e. Undersized cylinder running at full plant air pressure, therefore subjected to any "dips" in plant air pressure. 
  f. Binding cylinder, could be a bent shaft or lack of lubrication.
  
3. End of travel reached during welding, caused by:
  a. Cylinder rod too short on a "C" gun, or "link" is short on a "Scissor" gun.  See if the gun fully closes with one tip off.
  b. Air-oil cylinder running out of travel on "Intensifier" stroke, caused by:
     i.	Selection of Air-Oil cylinder with less than .25" intensifier stroke.
    ii.	Intensifier stroke started too early because of lack of squeeze time between "extend" and "intensify" operation.
   iii.	Some (or all) of the intensifier stroke used up pulling a gap together.  The part must pull together with 50 pounds or less of tip force.
    iv.	Low oil supply in the reservoir.
    
4. Lastly, I have seen logic problems cause this.  Check for anything in the logic that could cause the guns to drop pressure during welding, 
   such as a part present switch that is included in the logic "rung" for "gun stroke forward", which I have seen many times.  
   Once it is established by the logic that it is time to weld, nothing should drop weld force, as this could create a great hazard with 
   excessive flash and hot metal explosions.

Please tell me if this helped, as I provide these answers to help you, and your success is my reward.

Steppers are used to compensate for the current density reducing as the tips mushroom, reducing the heating action.

Chrysler publishes a chart of current density, showing 200,000 amps per square inch for a normal contact area of ¼' and 10,000 amps.

When the tips grow to 3/16", the current density falls to 80,000 amps per square inch, and it quits welding.

You will get a decreasing button size with tip wear, so you must add current.
You cannot tolerate the shrinking button, as you are on the "slippery slope" of a chart that equates button size to weld current.
Such as this one:

The independent variable is along the bottom, which
illustrates our "Slippery Slope" best.

As the tips mushroom, expressed as left-to-right,
along the bottom, the results are a smaller button, 
which is expressed along the vertical axis on the left.

This illustrates that the button is quickly lost as the
tips mushroom, and this is due to the current
density falling off to the point where it quits welding.

So it is important to add current as the tips wear,
to avoid the "Slippery Slope".  

Don't accept a smaller button as the tips wear, 
add current to stop this.

You may find that you have to add current at the rate of one Amp per weld.  This would give a current profile like this:

Of course these "Boosts" are added to the weld
current so the actual current might go from
10,000 Amps to 14,000 Amps over the 4000 welds.

Before Pertron introduced "Linear Steppers",
we had "staircases" for steps, so we divided up
however many steps the control allowed into
even amounts.  Square D had ten steps, so you
would use that with 400 amps boost every 400 welds.

With the Linear Stepper, you could set the whole
thing in one "step" that simply boosted 4000 Amps
over 4000 welds.  This is the least confusing way
to do it, but you miss some features if you do.

One of the features is the "Near End of Stepper"
warning that is announced by the controller while you are in the last step.  For a Medar, this is step 5, Pertron used 4 steps.
So with a Medar, we like to have this announcement for about ½ to 1 hour.  If you make 500 welds in an hour, use 500 counts.

Another feature is the ability to handle tip sticking while the tips "wear in" to give full contact area to the part.  This generally
takes about 50 welds on brand new tips.  You might have to start at a reduced current of 500 amps to stop problems due 
to high current density if just the edge of the tip touches the part in the beginning.
To handle this, you must reduce your "Base Heat" (What you have in the schedule) from 10kA to 9.5kA, then put 500 Amps
boost in step 1 with a count of 50 welds.

So now we have step 1 and step 5 set to give us the best results, what do we do with steps 2, 3 and 4?  Whatever you want,
they are usually divided up into whatever is left, and given the same boost rate as the rest, in this case 1 Amp per weld.
The chart below shows that:

Now the stepper boosts and counts add to each
other, so what you would enter into the control
would look like this:

You can see that it could be simpler,
but the control companies haven't done this.
In fact, Medar made it even more complicated, because you can enter the "% boost" and the "Amps boost" for every step.
That is what the seemingly unused zeroes are in the stepper programming.  Don't worry about this now, save it for later.

This gets even more complex, as Medar made the stepper work in conjunction with a tip dresser to tell the robot when to
move over to the tip dresser, and when to stop for a new tip.  We save this for way later…

I looked up "alimentation", which translates into "nourishment", so I think you mean the "feed" cable or power service.

So I will try to answer with that in mind.
The kVA is the size of the transformer, as far as power handling capability.  
Your welder may not use the full capability of the transformer, so the service cable is usually sized for the job.

So you must have the following information:
1. The current it carries i.e. it may carry 10,000 Amps for a typical spot weld)
2. The duty cycle of that current
      a) How many cycles it is carrying current in one minute of production (i.e. 20 welds of 12 cycles each = 240 cycles)
      b) Divide that by how many cycles there are in a minute (i.e. in the USA it is 3600)
      c) Use decimal form of duty cycle (not %) (i.e. the above is 240/3600 = 0.067  (also expressed at 6.7%)
3. The square root of that duty cycle (i.e. the above would = 0.258)
4. Multiply the current by the value in (3) (i.e. 10,000 x 0.258 = 2580 this is your equivalent Amps, continuous)
5. Refer to a sizing guide that shows length and continuous current 
(i.e. http://www.flexcable.com/images/uploads/downloads/AirCooledCableSizingGuide.pdf)
And look up the proper cable (i.e. if you need 16", use 1200MCM.
6. Remember that Air cooled cable REQUIRE that each end be bolted to a water cooled connection block.
7. Also there is a different table for water cooled cables, kickless cables, and leaf shunts.

Please tell me if this helped,

One place where many get confused is on the surface classification, and there are basically three surfaces:

The class A surface, where the surface is in plain view, like the side of a door, or the hood,

The class B surface, where an opening exists, like around the door opening or the trunk lip.

The structural surface, which is everything else.

So the welds are Class I, II, and III, and the surfaces are A, B, and "other". 


 
First there are many types of welds, both resistance welds and plasma welds.  I specialize in resistance welds, 
so I can pursue that, hoping that is the area you are asking about.
 
Then there are many types of resistance welds, in order of popularity in the automotive industry:
Spot welds, projection welds (nuts, studs, or embossments), seam welds (Overlapped and mashed), flash butt welds, and tube welds.  
There are many other types also, such as wire welding, hot upset staking, and more, but they are very rare in the automotive field. 
 
Each company has its own specification for welding, so I may need to know what company you are asking about.  
I know GM the best, but I am familiar with Chrysler, Ford, and some of the "Imports" (to USA) too.  
So to answer in general, I will refer to GM's specifications.  
 
Hoping you meant "Spot" welds, I shall continue with that:
 
At GM, the spot welds are classified in specification GM4488M as:
"Structural" welds, and "Processing" welds, and reads this way:
3.2.1 Structural Spot Welds. Structural spot welds are
installed for performance of the product. All spot welds are
structural, unless specifically noted as processing welds on the
weld drawing. All structural spot welds shall be assigned to
patterns.
3.2.2 Processing Welds. Processing welds are installed to
facilitate in-process assembly, but are not required for structural
performance of the product. Processing spot welds must
be approved by Product Engineering and shown on the weld
drawing.
 
GM4488M is the basic specification for a structural weld, and then there are constraints placed on the finish for three classes of welds.  
This takes another specification to describe, GM----- (Hmm, can't find it right now…)
 
GM4488M is a long document that spells out what is acceptable for: (with paragraph numbers)
 
4.1. size, 4.2. shape, 4.3. cracks, 4.4. holes, 4.5. missing welds, 4.6. edge welds, 4.7. location tolerance, 4.8 distortion, 
4.9 indentation, 4.10. extra welds, 5. spot weld patterns, 6. spot welding classifications, 7. pattern conformance, 
8. weld surface finishes, and repair requirements and procedures.
 
Since you asked about "Classes" of welds, I have cut and pasted the section on "Spot Welding Classification":
 
8 WELD SURFACE FINISHES.
8.1 The following classifications are used to communicate
the desired appearance of spot welds in a finished assembly.
They do not modify any of the requirements in 4. The effect of
metal finishing on spot weld performance should be considered
when specifying Class I or Class II Weld Finishes.
 
8.2 CLASS I WELD FINISH. This classification covers
sheet metal surfaces that must be free of markings, indentations,
or imperfections so that they are not visible after
painting. It is normally used for sheet metal surfaces that are
visible from the exterior of the completed vehicle. This classification
is not normally applied to GM-PCP weld patterns.
Metal finishing will normally be required to achieve this finish.
 
8.3 CLASS II WELD FINISH. This classification covers a
minimally marked welded surface. It is normally used for
sheet metal surfaces that are occasionally visible in the completed
vehicles, or where some weld marking is considered
acceptable. Metal finishing is occasionally required to achieve
this finish.
 
8.3.1 For Class II Weld Finish metal expulsion or excessive
metal displacement (above normal surface) is not permitted.
The degree of marking allowed will be defined by Product
Engineering and normally controlled through approved visual
samples.
 
8.4 CLASS III WELD FINISH. This classification covers
normally marked surfaces where there is no need to define
special requirements for surface finish.
 
8.4.1 For a Class III Weld Finish where metal expulsion or
excessive metal displacement (above normal surface) are unacceptable
for safety, appearance, fit up or other reasons, the
drawing shall specify each weld with a flash free note. This
will require removal of the expulsion or metal displacement
when it occurs on the designated spot welds.

Do you want a list of all things that damage a kickless cable?

That could be a big list, and there are many obvious things, so I will list the non-obvious ones:
1. Overtemp.
2. Twisting a non-superflex type;
   Two types, Regular type cools easier because of better water contact, but shorts if twisted.
   Superflex type can handle axial twisting without shorting out, but doesn't cool as well.
3. Water flow in wrong direction, must flow in at lowest end, out at highest end, to purge bubbles.
4. Poor "dressing" on a robot causing more flex than needed.
5. Allowing ends to oxidize before installation.  Always clean just before installation, or silver plate the surfaces.
6. Loose ends, always use "Supertanium" bolts, they have the same coefficient of expansion with temperature, as copper.
   Do not use just any old stainless, many are weak, such as "316".  Being non-magnetic is not important.
7. Excessive broken strands, shorting across the cable, "leaking" current.

Does this help?

First of all the robot should be checked for:

1. Alignment to the surface of the weld to be done, must be within 3 degrees of perpendicular. 
2. Equalization, the gun must be free to move up and down at least 1/2" (13mm) from its weld position.
3. Repeatability, generally the robot is trusted to repeat position to 1mm, so check for how close the tips come to anything they can short to.  
   Also don't even attempt to place a part on a pin for a nut welder without special processing to allow for the tight tolerances involved, 
   or it will only fit some of the time.
4. Number of welds, it is a lot of strain to do any more than 20 welds in a minute.
5. Stepper how does it handle "end of stepper" from the weld control?  Best is to finish the part, move to a window in the cell guard fence, 
   and require a tip change.
6. Weld gun water cooling, does it flow more than 1/2 GPM with an in-line flow meter in each gun arm? (Do not use a bucket and a watch...)
7. Transformer size, will it overheat at maximum production using full stepper boost current?
8. Dressing of the hoses and cables, no tight bends, no dragging on the part. Is the primary side power wire easy to change?  
   (Sometimes the plugs on the ends are harder to deal with than just terminating to clamp connectors)  
   The secondary might be a kickless cable, is it easy to change?
9. Are all secondary copper connections silver plated? (Bare copper oxidizes to unusable in a few days...)
10. Are all secondary bolts grade 10 stainless? (Do not use "316" alloy, less strength than a cheap bolt)  
    People do not realize we use stainless because its thermal expansion matches copper.  Magnetics has nothing to do with it.
11. Nobody does this yet, but the water shutoff should be right at the gun, to prevent loss of water and an air-lock happening after tip change.
12. The air valve should also be right at the gun cylinder, as there is too much delay in squeeze time if a long length of line is 
    between the valve and the cylinder.  Measure squeeze time required to get to 90% of weld force.  This may require explanation of 
    how to do this, or you may know how.
13. Transformer capacity, will you have enough extra weld power to make up for line drops when at full stepper boost?  
    You must know a lot about your power before you can answer this.  Also are you able to weld above 50% for the lowest current weld 
    and with no stepper boost.  (Too much capacity is bad too...)
14. Force setting, does 80% air pressure give the optimum force?  Too much capacity is bad, Too little is a disaster.
15. Tip size and alloy.  I have my favorites, dome zirconium, 13 mm or larger.  Your may vary. 
16. Weld stroke over-travel, must be able to close gun with one tip off.

Auto companies in the US do a Production Part Acceptance Procedure, which consists of a couple hours or more of actual production, 
where parts are checked for dimensional integrity, and weld integrity (completely torn down)  
Usually the teardown is done on parts 1,3,5,7,9,19,29,39,49,74,99 and every 25 thereafter, 
and the dimension check on 2,4,6,8,10,20,30,40,75,100, and every 25 parts thereafter 
(Sometimes every 50 or 100, depending on the customer)
And failures must be corrected and (at the customer's choice) restart the test, or continue.

Tip sticking is caused by excessive heat at the surface between the tip and the metal. It is most common with zinc coated steel, as the zinc alloys with the copper, forming brass, and brazes the tip to the part. It is rare with bare steel, so I will continue with my answer as if you are welding galvanized.

At the start of a weld, the resistance of the interface surfaces is high, and the resistance of the metal is low, because it is cold. When we pass current through the stackup, the heat is formed where the resistance is, unfortunately it is at the surfaces in the beginning. This is because the rough surfaces have only contact with each other in microscopic points and bumps. This heat will warm the metal, but also flattens the surfaces so they lose resistance, and therefore don't heat up much more as we continue welding.

When the metal becomes warm it raises in resistance, and as we continue passing current the heat is produced from this resistance. This change occurs about 30-50 milliseconds into the weld process. From this point on, most of the heat is generated within the metal, right where we want it. The copper tips will cool the surface so that the metal stays firm on the outside and helps to contain the growing "nugget" as we continue to heat it to the point where the two metals join. This point can just be the plastic state in bare metal, but must be the liquid state in galvanized metal, in order to mix the thin layer of zinc into the nugget where it won't harm the bond between the sheets. If we just heat to the plastic state, which is fine for bare metal, the layer of zinc will still be between the sheets, giving a "stuck weld" where the zinc was not moved aside and just solders the sheets together.

So the heat starts at the surface, and then changes to the inner part of the material, but if we start out with too much heat in the beginning, we may braze the tips to the part. So we must find what is causing excessive heat at the very start of the weld, some of those things are:
1. Low tip force, and this can exist on one side of a robot welding a clamped up part, so center the equalizer*, and make sure it allows free movement of the gun after it closes.
2. High current at the beginning of the weld, starting with an upslope helps a lot here.
3. Insufficient cooling, you said it was OK, but how did you determine this?
4. Small tips cause this too, because the mass of copper is actually what cools the surface, the water cooling then carries the heat away. "B" nose tips put the largest amount of copper right at the weld, without increasing the contact size, therefore the weld size.
5. Small contact area, which almost always happens with brand new tips, because:
   a) They don't exactly meet right, but they do "break-in" over the next 25-50 welds.
   b) They were installed with a hammer blow distorting the contact surface.
If using a stepper, you may lower the current by 500 Amps, then step that back up again in the next 50 welds to compensate for this.

Another thing that contributes to the tip sticking is the alloy of the copper. True zirconium tips work the best, as the 1/2% zirconium strengthens the copper without adding much resistance. Any other material in the alloy adds to the resistance, making the surface heat a bigger problem. Many people will sell you "Class 2" tips with zirconium when you ask for zirconium, but there are other ingredients, which are not always the same as "class 2" is a strength rating, not a chemical description. Chromium is a popular one, but it contributes to tip sticking.

One exception to the "added ingredients" problem is the titanium coated tip from Huys in Canada. Although this coating contributes considerably to the surface heat, it also absolutely sheds the molten zinc. This tip cannot be dressed, so many people avoid it because they want to have tip dressers, but it will solve the tip sticking problem.

If you must have a dresser, and there are reasons that make us put up with this troublesome addition to an already complex industrial process, then you must very carefully optimize the 5 items above.

Please tell me if this helps, because I provide this service for the joy of knowing I helped, and if you don't respond, then I was deprived of my reward. If you were closer, I would just come to your plant, then my reward would be instant, plus I get an additional reward when the invoice is paid!

*EQUALIZER-
I find the equalizer to be setup wrong on every gun I ever worked on.
It is supposed to let the gun move up and down when it is closed on the part, because all parts aren't not clamped up exactly the same, the tips wear, and the robot programmer doesn't always allow the same distance from the lower tip to the part on every weld.

Most people will go to the gun and try to move it up and down when they hear the equalizer may not be set right, then adjust the air regulator to the balancing cylinder until it does. This is wrong, they should have closed the gun first, which will have some means of holding it in place, until the gun closes, then it allows this movement.

For all electrical devices you must measure:
1. The line current (in Amps)
2. The line voltage (in Volts)
3. And multiply them for wattage (in Watts).
4. Then you must measure the time it is drawing current (in hours)
5. And multiply hours with wattage to get Watt-hours.

There are two difficult parts in getting the measurements:
1. Some devices are on for a very short time
2. Some welds use different currents.

With resistance welding, we read:
1. The secondary current with a Miyachi (or similar) "Weld checker", and divide it by the transformer turns ratio to get line Amps.
2. The voltage will be line voltage, measured with a volt meter. Remember if there are significant voltage drops during welding, then you must find a way to measure the line voltage during welding. Many weld controls have methods to read "lowest line volts during last weld", use that number.
3. Multiply these to get wattage during the weld.
4. The Miyachi also gives time (in cycles or milliseconds). Convert cycles to time by dividing by the line frequency. Then convert to hours.
5. Then multiply hours by Watts to get Watt-hours for that weld.

If there are multiple heat settings you must do this for each setting, then add them all up for one hour, and you will have the Watt-hours used in an hour of production.

I imagine you would do the same with laser welding, but I can't tell you the best way to get the readings.

The strength of the metal will be the strength of the tensile test, because the weld is stronger than the metal. So you must find out the tensile strength of the metal. To do that, find the PSI from the material property sheets, which come from the steel producer.

Each steel producer will have different tensile strengths, as they process the steel differently. This is also known as UTS, Ultimate Tensile Strength.
This will be expressed as PSI or pounds per square inch. A 1/4" diameter weld will have more strength than a 1" wide sample of the sheet metal you are testing.
So divide the psi rating by the thickness, for example if the psi is 30,000 (Typical for mild steel) and your sheet metal is 0.100", then the UTS of a 1" wide sample is 3,000 pounds.

A 3/16" spot weld will be about the same strength, and will yield at about the same 3,000 pounds, because it will tear out of the steel just about the same.

Finding the thickness of a 12 gauge sample will be a problem as there are about five different gauge standards, so I would measure the thickness and not even bother describing the thickness in "gauge" as that is too unreliable of a measurement.

Hope this helps,

Dear Sir,

There are many steps you must do to be able to specify the gun, but we need a lot more information about the application (we need the underlined values).

1. You must find the tip pressure required, and specify that this must be achieved at 80 psi. If I knew more about the metal, I could look that up, but I need to know if it is mild steel, HSLA, or AHSS. I need to know if it is coated, as galvanized is often used in automotive work, that makes a big difference in the weld schedule.
Probably the tip pressure would be between 470 pounds to 670 pounds, but if it is a manual gun or robot gun it would be between 670 and 950 pounds.

2. We must find the maximum current the gun must produce, to get this we need to decide on the tip contact area, because that makes a big difference on what the current should be. To weld 0.8 mm, usually the weld diameter required is 4 mm. Usually a 6 mm tip is used to insure we always have at least 4 mm. Depending on the details needed on the metal, as just discussed, the current required could be between 8 kA to 10 kA, and if a heat stepper is used with a max weld count of 4000 welds, an additional 3000 to 5300 more Amps would be needed, depending on the steel coating, if any.

3. Once we know more about the metal, we can look up the schedule and get the weld time, so that we can calculate the duty cycle. We need the ratio of time ON/Time OFF during production. This can vary from 3% to 18% depending on the weld schedule and the number of spots per minute. This will help us figure the transformer KVA rating, but we also need to know the transformer turns ratio.

4. To get the turns ratio, we must find the secondary voltage required to produce the maximum current required for our weld. This depends on the impedance of the gun, which is related to the throat area of the gun, and the cabling between the gun and the transformer. This can be as high as 24 Volts for a large gun and a kickless cable, and as low as 3 Volts for a small gun with an integral transformer. Once we know the secondary voltage, we can calculate the turns ratio as primary voltage divided by secondary voltage.
When we have the turns ratio, we use it to figure the primary current by taking the secondary current divided by the turns ratio.

5. Then we can figure transformer KVA, as it would be the equivalent constant thermal primary current (ECTC) times the primary voltage. the ECTC is the actual primary current multiplied by the square root of the duty cycle. This will give us the actual KVA, however transformers KVA ratings are at 50% duty cycle, so we must take our actual KVA and divide it by 0.7 to find the nameplate transformer KVA.

So, if you can supply the information that is underlined we can figure out everything for you.

Hope to hear back from you soon.

First of all the robot should be checked for:

1. Alignment to the surface of the weld to be done, must be within 3 degrees of perpendicular.

2. Equalization, the gun must be free to move up and down at least 1/2" (13mm) from its weld position. Of course, this is only when the gun is closed. The equalizer must be locked in position when the gun is open.

3. Repeatability, generally the robot is trusted to repeat position to 1mm, so check for how close the tips come to anything they can short to. Also don't even attempt to place a part on a pin for a nut welder without special processing to allow for the tight tolerances involved, or it will only fit some of the time.

4. Number of welds, it is a lot of strain to do any more than 20 welds in a minute.

5. Stepper, how does it handle "end of stepper" from the weld control? Best is to finish the part, move to a window in the cell guard fence, and require a tip change.

6. Weld gun water cooling, does it flow more than 1/2 GPM with an in-line flow meter in each gun arm? (Do not use a bucket and a watch...)

7. Transformer size, will it overheat at maximum production using full stepper boost current?

8. Dressing of the hoses and cables, no tight bends, no dragging on the part. Is the primary side power wire easy to change? (Sometimes the plugs on the ends are harder to deal with than just terminating to clamp connectors) The secondary might be a kickless cable, it is easy to change?

9. Are all secondary copper connections silver plated? (Bare copper oxidizes to unusable in a few days...)

10. Are all secondary bolts grade 10 stainless? (Do not use "316" alloy, less strength than a cheap bolt) People do not realize we use stainless because its thermal expansion matches copper. Magnetics has nothing to do with it.

11. Nobody does this yet, but the water shutoff should be right at the gun, to prevent loss of water and an air-lock happening after tip change.

12. The air valve should also be right at the gun cylinder, as there is too much delay in squeeze time if a long length of line is between the valve and the cylinder. Measure squeeze time required to get to 90% of weld force. This may require explanation of how to do this, or you may know how.

13. Transformer capacity, will you have enough extra weld power to make up for line drops when at full stepper boost? You must know a lot about your power before you can answer this. Also are you able to weld above 50% for the lowest current weld and with no stepper boost. (Too much capacity is bad too...)

14. Force setting, does 80% air pressure give the optimum force? Too much capacity is bad, Too little is a disaster.

15. Tip size and alloy. I have my favorites, dome zirconium, 13 mm or larger. Your may vary.

16. Weld stroke over-travel, must be able to close gun with one tip off.

Auto companies in the US do a Production Part Acceptance Procedure, which consists of a couple hours or more of actual production, where parts are checked for dimensional integrity, and weld integrity (completely torn down) Usually the teardown is done on parts 1,3,5,7,9,19,29,39,49,74,99 and every 25 thereafter, and the dimension check on 2,4,6,8,10,20,30,40,75,100, and every 25 parts thereafter (Sometimes every 50 or 100, depending on the customer)

And failures must be corrected and (at the customer's choice) restart the test, or continue.

Yes you can cure this, but it will cost you!

This relates to question 20 on my test, on my website. We will use the changing resistance of the weld stackup to regulate the heat for us.
First you have to have a reasonable weld schedule that would weld this without the fit-up problem, and if you tell me the two materials, I will get that for you.

Next, you use that schedule three times, with long cool times between. For instance let's say you were welding G1.2mmG/G1.0mmG. This means the 1.0mm Galvanized sheet would be the "GMT" or "Governing Metal Thickness", The new guys at GM call this "DT" now, "Determining Thickness".

We look this up and find the force, time and heat to be: 600#, 10,000 A, 13 Cycles. But I would start with 9000 or 9500 Amps because new tips concentrate the current more than tips that have 50 welds on them. I would imagine you are using a stepper, so boost 500 Amps in the first 50 welds, and you will get through the troublesome "tip break-in" period.

Anyway, on to the cure, since the first time you apply this heat the part won't get all of it, as the 2mm gap prevents current from flowing through the weld area, but it might heat it up and bring it together. At this point the resistance of the hot metal will be about 150 micro Ohms, because it is not welded, and because it is still hot. Let it cool for 13 cycles, and at that point the resistance will drop to 10 micro Ohms if it welded, or stay at 150 micro Ohms if it didn't.

Then you apply the same heat again, and this time will probably weld it, but because "probably" isn't good enough, we let it cool another 13 cycles, and apply the heat again. In each of these cases, after the cool time, the resistance will be high if it didn't weld, low if it did. If it did weld, during one of these heat times, the low resistance of the weld, after it cools, will prevent excessive heat to be formed on any following weld attempts. After all, this is called "Resistance Welding" for a reason, we need resistance to develop the heat when we pass current through the part.

So usually we have three "heats" each one of them a full weld schedule, and two "cools" each one as long as the heat time was. Then you put normal hold time after that, which is usually half the heat time. Actually it is a little more complicated to figure hold time, but just use half the heat time for now.

All of this was counting on you using a proper squeeze time, because if the gun hasn't stabilized its force, it doesn't weld the same each time. If in doubt put in 90 cycles, and work downward from that until you get sparking at the beginning of the weld.

Now here is the cost:
1. Your weld time will be 5 times as long, upsetting the production manager, who says you fix this by "just turning the heat up". Hi is usually big and burly and doesn't want anything to slow his line down, and offers to "make mashed ta-taytas outta yer head".
2. The tips will last less than 1/3 as long as they should. If you were getting 3000 welds, you won't even get 1000 now, and here comes that big production manager again to tell you how to weld.
3. You must tell me how this worked for you, that is part of the cost.

This will be a very long answer, because there is so much to specify to get from the metal thickness to the KVA size.
This will take us through all the known calculations that have ever been developed by man, except E=MC2

1. A 1mm/1mm stackup needs a 4mm diameter weld.
2. A 4mm weld is best done with a 6mm weld tip, to allow for production tolerance.
3. We must figure the current based on the material:
    a) If this is in uncoated mild steel this will take 8,000 Amps for 9 cycles at (60 Hz)
    b) If galvanized mild steel it will take 10,000 Amps for 13 cycles
4. Next we must figure the voltage required, based on the secondary design:
    a) If this is a transformer/gun with a small throat area, it will be 3 to 4 Volts required.
    b) If this is a "hard tool" with 16" cables it will be 12 Volts required.
    c) If this is a "Portable gun" with a 10' kickless cable it will require 24 Volts
5. Then we can figure the INSTANTAINEOUS KVA Volts x Amps:
    a) The amount could be 3V x 8kA (24KVA) to 24V x 10kA (240KVA)
    b) Depends on material (3 above) and secondary voltage (4 above)
6. Be sure to allow for stepper boost required when figuring current:
    a) A prototype shop will dress the tips every 100 welds, so won't use a stepper.
    b) A production line may want to get 4000 weld per set of tips, so may boost 3000 Amps!
7. Now we must consider duty cycle, or "how many welds in a minute?":
    a) Add up the total cycles that current will pass in one minute.
    b) Divide by the total cycles in one minute. (60/sec x 60 seconds=3600)
    c) For example, 10 welds of 13 cycles each = 130 cycles. 130/3600=3.6 duty cycle
8. Now figure Equivalent Constant KVA:
    a) Multiply KVA (found in 5 above, and allow for the stepper!)
    by the square root of the duty cycle (7 above)
    b) For example, 240KVA x SqRt(3.6)=240 x .19=45KVA
9. Transformers are rated at 50% duty cycle, so divide your KVA by .7 to = 50%:
    a) 45KVA divided by .7= 65KVA.

And that is how you arrive at the KVA for the transformer.

BUT, I think you said; "Customer has 35KVA factory power" This could mean a lot of things.
So there are other factors:
10. Most likely this "35KVA" means the voltage multiplied by the current available.
11. It probably is the 3 phase power available, and the welder is a single phase load,
    and the single phase power will only be .7 x 35KVA available, or 24KVA.
12. It will support a lot higher KVA than this because the welder is not a continuous load.
13. Also many factories do not have the right voltage, and often have a transformer installed
    and the transformer must be selected to be a LOW IMPEDANCE transformer.

I really like the ARO guns that are marketed through ARO/Savair here in the US, but they said they use class 3 copper in the US, and class 2 elsewhere for the gun arms.  Very important to avoid class 3 because of the beryllium dust caused by grinding and sanding.  If class 2 works with its lower strength, I would recommend it.
Also MFDC is much better for materials that are affected by the high peak heat of every cycle, such as the new Advanced High Strength Steel, "AHSS".

If the material in the coating is harmful when included in the weld nugget, such as a Zinc (Galvanized) coating, then you must do something to minimize the amount that is melted into the center, “weld nugget”, area.
Many things can be done:
1. Use a thin even coating, such as Electro plated, Galvanneal (which is re-heated electro plated zinc) not G60, or even worse G90, two hot dipped standards that vary in thickness.
2. Use upslope to start your heat.  Carefully applied, it can cause the zinc to melt and squeeze out before the steel reaches forging temperature.
3. Use impulse heat, allowing the nugget to reach molten temperature without melting through the surface.  Note that uncoated steel is welded with lower heat, because you don’t have to reach melting temperature to fuse the metal together, just reaching the plastic state is acceptable.  This is not acceptable for welding zinc coated materials, you must melt it and get it to mix together, or there will be a thin layer of zinc caught between, making a weak weld.  Impulse heat uses 1 to 3 cycles of cool time, allowing the surface to cool without the center cooling much.  Usually a weld of 16 cycles is broken into three or four impulses of 4 to 6 cycles with 2 cycles cool time between. 
Makes a big difference in weld consistency.

This is very ambitious.  Usually the welds are qualified in “patterns” where about 10% to 20% of the welds can be discrepant, but the Pattern is still acceptable, because of these extra welds.  If you don’t do this, you will be constantly stopping production while you repair things, very inefficient. 
As far as patterns, they must be 100% to insure the integrity of the assembly.  If you were to judge by the individual spots, most assembly plants are between 95% and 98% acceptable welds.  Trying to do better than that may be very difficult, and may not have a good reason to do this.  Also you must never have a design that insists on a single weld, because the only way to insure 100% is to tear apart every weld. 
This is a concept that suppliers in Mexico do not understand, they want 100% good welds because the US customer tells them that. 
We must have 100% good parts, but a good part, if properly designed, can still be good with a missing weld. Many Automotive designers do not understand this important concept either.  If you are going to run a plant, you must not try to do the impossible, or you will never get any parts shipped.

There are very few formal classes anywhere for resistance welding.  Major universities in the US teach only arc welding and metallurgy in their degree programs.  So you can have a degree in Welding Engineering, from Ohio State, LeTourneau University in Texas, or Ferris State in Michigan, and have no proficiency with resistance welding.
In 1976, I was made the Welding Engineer at Chevrolet Gear & Axle, in Detroit, when the real Welding Engineer passed away.  My training was in Electronics, and I was working there as a Union Electrician, specialized in electronics, and I have been working exclusively in Automotive Resistance Welding ever since, 35 years.

I am surprised that this would change anything, as it does in arc welding, because there is no molten or vaporized material to flow from, or to, a tip in resistance welding.  Usually the tip life on an automatic or a robot is almost 4000 welds, and it is limited by contamination of the zinc more than anything else.  With steel welding, there isn’t much more life because the tips mushroom beyond usable diameter beyond 4000 welds or so. 
Tip dressers are tried here, but they are not used right at all, this is another whole weld subject to go into.

I will help, and you will get knowledge that you can’t get in a university.  I laugh because in the US they want a degreed Welding Engineer, then when he can’t handle resistance welding, they replace him with another one.  They never figure out that the degree means he studied arc welding, and non-degree people like me might have used their time learning resistance welding.  Unfortunately all these guys are self-taught, and that means they may not have a well rounded experience.  I have been in Automotive Welding, and I am good with that, but there are many other applications, such as wire welding, soldering, tube welding, small motor armatures, etc. that I do not know about.  Luckily I was involved in the growth of Pertron Controls Corp, and was exposed to some Aerospace Welding, as they were involved with many aircraft companies.  I got them into Automotive Welding, and with their precision control, they gained many customers, and grew to be the largest supplier in the US.  Square D Company bought them, in turn bought by Group Schneider from France.

No, I would be too busy.  It is better that just a few people find out about me, such as yourself.

This is on the hem welders, right? If true, I must tell you much more about hem welders, so you can justify why you must get rid of them.

The places I have seen copper plates is in hem welding, around the folded lip of a door.
The outer skin is folded around the inner skin, making a 3-layer stack up. The outer surface is placed face-down on an insulated, water-cooled copper block, and a 6 mm diameter welding tip comes down from the top, touching the inside of the folded lip. Another gun is brought from the top to contact the inner skin, as a “ground” gun.
Current flows from the welding tip, through the folded lip, into the inner panel, then across the inner panel to the “ground” gun.  No current flows through the outer side where the copper block exists.  I see that Question #Q24 was about this, but needed this drawing:

Hem welding with simply a copper plate between one tip and the part is really hard to do, even worse than the hem welding I drew a picture of above, which the car companies have now given up on. The popular way to handle this hem is with heat cured epoxy, cured in a few places around the hem with induction heating coils. This stabilizes the part so it can be sent to the paint operation, where the hot electrostatic "ELPO" dip will finish curing the epoxy.

Now the problems with hem welding:

1. The heat is usually controlled by forcing a controlled current through two sheets of steel. The current is forced into a round column by the two tips of equal diameter where they contact the steel surface. Since the water-cooled tips also cool the surface, the area that reaches fusion temperature is halfway between the two tips. Also the "nugget" size is controlled by the diameter of this column of current, which in turn is controlled by the contact area of the tips.
The hem weld arrangement does not control this current path very well, and it is hard to predict where the hottest spot will be, but probably NOT at the junction of the two metals. This position would change, depending on the distribution of resistance within the current path.

2. The resistance of the weld depends partially on the force between the rough surfaces as they are forced together by the tips. If this resistance is not consistent, the heat, therefore the weld, will not be consistent. This pressure is greatly influenced by how tight the hem is folded over on itself. If more "open", the tips will not push the inner surfaces together as much as if it were more "closed". And the hemmer machine is constantly touched up by the machinists and set-up people, with total disregard for any affect it may have on the weld heat. They are not "bad guys", it is just that nobody realizes that it affects the weld at all.

3. Because the hem weld is so unreliable, a heat-cured epoxy has been added in the hem, applied with "sealer robots" just before the hemming operation. This further affects the resistance of the work, making the weld even more variable. Further, the inconsistencies of the bead size and control of where it goes is seldom considered a problem by the non-welding people, so often they don't even use more sophisticated sealer dispensing controls, making even further problems, including getting the sealer on the surface of the part.

4. Sealer on the surface of the part greatly accelerates the erosion and contamination of the tips, causing further problems with controlling the weld heat.

5. The hem area is held very tightly, and is very rigid, so when the tips close down on it, any imbalance in forces will be felt as more or less force on each side of the sheet. A proper gun equalizer will compensate for this, but very seldom the gun equalizer is properly designed, or even properly maintained and adjusted. Since this is true of most of the equalizers in the whole plant, we can also be sure it is true of the hem welders.

6. Testing of the hem welds is very difficult, as there is never an inspection station between the hem welder and the point where the robot places the part on the conveyor to the paint shop. Further, gently pry-checking this area will produce ugly deformations on the finish surface of the part, unlike all the other structural welds that can be gently pry-checked. So the welds can get out of tolerance without getting "caught" as quickly as the rest of the welds, making the problem a big one when the poor welds are finally discovered. We have a term: "Weld Spill" that is often used here, where bad welds were not contained and repaired before the parts left the body shop. All other welds are well inspected, and repaired or contained in some manner, and it is a rare occurrance for one to escape all this inspection.
So with all these things going against hem welding, is it any wonder why we have replaced it with epoxy?