Testing and Verification of Intercell Welds

Tony Schröer, Jeff R.Snell

Digatron / Firing Circuits

One of the more critical processes in battery manufacturing is the inter-cell weld. At the
same time inconsistencies in lead material and less than reliable testing methods
complicate process control and verification efforts. Time proven methods to verify welds
such as X-ray, Gamma ray and ultrasonic testing are either not suited for lead welds or
impractical when considering the production environment. Only electrical tests are suited
for 100% production lot testing of inter-cell welds.

This paper attempts to highlight the weaknesses inherent in traditional electrical testing
methods when using a tester integrated with the inter-cell welding machine.
Unimpeachable data regarding the quality of inter-cell welds cannot be achieved using
traditional electrical test methods.

More reliable results can be obtained with an improved measurement system, which is
separate and installed after the inter-cell welding machine. An optimized inter-cell test
process and measurement techniques are discussed.

1. Introduction This intercell weld is one of the more

critical and difficult processes used in
To achieve a battery voltage of 12V, six manufacturing batteries. Thousands of
cells must be connected in series. batteries are destined for scrap if any
Connections from cell to cell for one of the six intercell welds per battery
automotive batteries are made within the (Fig.2) does not fulfill the specifications.
polypropylene container through holes
punched in the partitions. The positive -
strap terminal on one side and the
negative strap terminal on the other side
of the hole are pressed together and
welded (Fig.1).

Fig.2: Automotive Battery

The following will focus on the

resistance welding process of lead,
Fig.1: Positive and Negative Strap Terminal specifically lead-alloys.

Testing and Verification of Intercell Welds -1-

There is very little literature available
which discusses the behavior of lead
materials when welded. What literature
exists, only considers welding by open
flame.

There are two reasons for the limited

literature about lead welding:

- Only the battery industry employs the

process of welding lead to Fig.3: Two-Phase-Diagramm
manufacture it products

- Lead welding is one of the most Lead is a very soft metal with a relatively
difficult processes to control. It low melting point at 327 deg.C
requires control of several variables
to achieve reliable results. With respect to the resistance welding
process the melting point can be
controlled by an appropriate amount of
energy. But the control range is small
Over the past decades the welding and if the energy applied, slightly
process have been more focused on exceeds the maximum limit lead
improving efficiency. It may not be becomes liquid and will squirt out.
practical to eliminate all subtle material
and process variations.
Another important parameter is the
In some cases this leck of process hardness.
control may result in marginal intercell
connections which could lead to battery One cannot state the general hardness
failure. of a lead-antimony alloy. The rate of
hardening depends on the cooling
Why is it difficult to weld lead? process and the percentage content of
antimony.
Instead of pure lead (Pb), lead-antimony
(PbSb) alloy is used to increase the Fig.4 shows hardness versus age.
hardness of lead. But properties and
particulary mechanical properties of low
antimony alloys do vary and depend
mainly on material history

Fig.3 to fig.6 will reflect the

characteristics of lead-antimony alloy
with 3% antimony.
Fig.4: Hardness vs. Age

Testing and Verification of Intercell Welds -2-

It is obvious that newly casted lead As a result internal resistance
behaves somewhat differently than lead measurement systems can only be used
which was stored. Newly casted lead is to compare resistance data within a
very soft and can be easily formed production batch and do not accurately
together. Stored lead – roughly speaking measure the real resistances of the
one day or more – is brittle and breaks intercell welds.
with even minor deformation.

The Oxidation curve (Fig.6) indicates

Usually welding machines receive newly that oxidation begins immediately after
casted lead without the necessary contact with air. Once a thin oxidation
hardness. The welding clamps will pinch layer has been developed further
the two cast on strap flanges and just oxidation is inhibited. The film has an
the force of the clamps will join the parts isolating property.
and create an electroconductive
junction. In accordance with equation

d=m/OxV where
Even though welding in this state is
possible the process energy must be m = oxidation [mg],
increased to compensate for the low O = 5400 mm2 and
resistance junction. It is possible to v = 0.252 mm3/g
conduct the welding procedure without
welding the material. And the internal in our test a thin film of about 0.4µm to
resistance measurement will display 0.6µm was developed after a few
values within the limits and will let the minutes.
battery pass.

The electrical resistance curve (Fig.5)

typically shows the highest gradient at
the melting point.

Although most welding machines ensure

rapid cooling after the high current
welding process the electrical resistance
may vary significantly due to the
temperature gradient. Fig.6: Oxidation vs. Age

Any measurement system whether

integrated or separate from the welding
machine should have probes which
break through the oxidation layer to
ensure accurate measurement.

All these variables make it very difficult

to control the welding process for lead
Fig.5: Electrical Resistance vs. Temperature alloys and to consistantly produce
quality welds.

Testing and Verification of Intercell Welds -3-

Despite these issues the battery industry
has proven that lead welding is possible.

Todays welding machines consist of two Contact detection

hydraulic clamps for interconnection 1,
via voltage
3, 5 and interconnection 2 and 4.
These clamps are positioned by measurement
automated x/y-positioning systems [1]. Water cooling

Power electronics are used to provide

welding currents in kA range using
Weld electrodes
programmable controllers and a PLC is with projection
used to control the conveyor station and forming tools
to position the battery underneath the (thrust adapter)
clamps.
Fig.7: Welding Clamp Details
Fig.7 shows details of the welding
clamps.

The voltage sense connection provides Fig.8 shows the strap connections
the feedback for current control and also before and after projection. If the initial
detects the initial contact made with the contact area is too small the resistance
parent material. will be too high and the welding process
It is important to control the time will generate too much heat, causing
between inital contact with the parent lead, to squirt out at the beginning of the
material and the start of the welding welding process.
process. This time period determines
the size of the contact area and the If the initial contact area is too large a
initial contact resistance prior to welding. very low resistance electroconductive
Because this pause time is so critical junction will result. This will always
special contact detection modules are create a condition for cold welds.
used to precisely measure initial contact.

Strap position before

forming of the projection

… and after projection forming

Fig.8: Projection Forming

Testing and Verification of Intercell Welds -4-

The entire welding sequence (Fig.9) will
be explained in the following.

Fig.9: Welding Sequence Fig.10: 1000 Hz MF-DC Inverter Power Unit

When the battery is positioned, the Fig.10 shows the principle schematic of
clamp contacts the strap terminal and a 1000Hz DC inverter power module
closes the projection forming electrodes using pulse-width-modulation.
(phase 1).
The 3-phase AC signal is rectified to
Phase 2 starts with a signal from the provide a DC link voltage for the IGBT
contact detection module. If the inverter. The IGBT inverter is controlled
programmed pause time (phase 2) is by 1kHz frequency and provides 0.5 ms
complete, the weld current starts (phase control cycle. The welding transformer
3). Current amplitude and current time generates the high current output and
are controlled by the welding controller. after rectification by the two diodes a
The lead will melt and due to pressure true DC signal is applied across the
applied to the strap terminals the welding electrodes.
material will be pressed through and fill
the partition hole. Even considering the equipment and
process improvements developed over
In phase 4 the hold time allows the lead the recent years there remains
to cool (phase 4). significant variation in both, process and
materials which can result in
After the hold time the clamps will open substandard welds.
and the contact detection module
indicates the end of the welding process As difficult it is to weld lead, it is more
(phase 5) difficult to verify lead welds.

Dynamic regulation is required to control The next pages will investigate and
weld energy and particulary weld analyse existing methods of quality
current. As a result switched mode inspection for weld processes for use in
power supplies are typically used in battery assembly lines.
modern welding machines.

Testing and Verification of Intercell Welds -5-

2. Experimentals

Methods used in the industry to verify

welds are:

X-Ray, Gamma-Ray, Ultrasonic,

Electric Inspection such as
Resistance Measurement or
Resistance vs. Welding Time and
Destructive Inspection.
Fig.12: Twist off tool kit with torque display

Strap terminals are twisted off using a

2.1 Destructive Inspection
torque wrench to measure the maximum
torque value prior to failure (Fig.12). It is
assumed that the welding quality
correlates to the torque value.

Fig.13 shows torque data, deviation and

distribution of 28 samples

Fig.11: Adapter to prepare sample weldings

Destructive testing of welds is performed

on a periodic or sample from a lot
bases. The results can be used to adjust
the intercell welding machine. Fig.13: Production lot test results

A special adapter is used to produce

weld samples to avoid wasting entire
cell packages (Fig.11).
Another criteria to verify the welding
A fixture keep the strap terminals in the process is optical inspection.
right position for welding. There are
cavities in the upper platform to position
the terminals. Different battery sizes can
be accommodated by the same fixture
using the telescope.

After welding the operator lowers the

upper platform to get access to the strap
terminals.
Testing and Verification of Intercell Welds -6-
Destructive Inspection, optical evaluation

Fig.14: Ideal Welding Fig.15: Corona welding Fig.16: Blow holes

Defective areas can be seen clearly as displayed in figures 15 and 16

2.2 Non-Destructive Inspections Unlike X-Rays or Gamma-Rays,

Ultrasonic waves accurately detect the
smallest of air gaps, which create an
There are several non-destructive unbridgable barriers completely
inspection methods typically used in the reflecting signals.
industry.

As a result the high degree of

absorption, X-Ray and Gamma-Ray
cannot be used for inspection of lead
welds. The reason is the attenuation
effect whereby the atomic number is
cubing and multplied by the density of
the material in question. As a result the
degree of absorption for lead is much
higher when compared to other metals.
It is technically inpossible to use X-Ray
or Gamma-Ray to find irregularities in
the intercell welds.
Fig. 17: Ultrasonic inspection,
pause time too short

Ultrasonic inspection would be an Fig.17 shows an ultrasonic image of an

ideal method to determine irregularities intercell weld where the pause time was
in intercell welds. The main advantage is too short.
the straight-lined propagation of
ultrasonic waves through the medium One can see in the image that there is
and very good diffusion [2]. In this no coherent area of welding. It is
regard ultrasonic waves are comparable assumed that blow holes have been
to X-Rays or Gamma-Rays. generated.

Testing and Verification of Intercell Welds -7-

In general blow holes will be generated Fig.19 shows the result if the welding
if there is too little material available. current is too high. The lead squirts out
This will be the case when pause times and leaves blow holes.
are too short and when the welding
process starts immediately after the two
terminals are in contact with each other.
Welding current below the minimum
required will result in an incomplete and
poor quality weld.

Fig.18: Ultrasonic inspection,

pause time too long

If the pause time is too long the material Fig.20: Ultrasonic inspection,
will develop an electroconductive welding current too low
junction and the resistance will be too
low to produce the necessary welding
temperature. Fig.18 shows a corona In Fig.20 it can be seen that the areas at
welding at 1000 ms pause time. 1, 4 and 7 o‘ clock are not welded
properly.
The next two figures will show the
results when varying the welding
current.

Fig.19: Ultrasonic inspection, Fig.21: Ultrasonic inspection,

welding current too high welding current with ramp
Testing and Verification of Intercell Welds -8-
Fig.21 shows a substantial result.

Whenever current ramps were used

better quality welds resulted.

The major advantage of a current ramp

is the preliminary warming of lead which
will create uniform material and process
conditions. Thus the influence of
material hardness can be minimized.

The current time was set to 240 ms

whereas 180 ms were programmed as a
ramp.
Fig.23: Ultrasonic inspection,
current time too short

The next figures show the variation of

current time.
If the current time is too short the weld is
incomplete. Fig.23 shows welded
material in the center only. Outside the
center the weld is very shallow.

The next figures will show the

importance of correct pressure when
welding the strap terminals

Fig.22: Ultrasonic inspection,

current time too long

If the current time is too long the energy

applied to the weld will be too high and
material will squirt out. After coagulation
blow holes and hair line fractures will be
found. As it can be seen in Fig.22 only a
small area between 11 and 12 o‘clock is
welded.

Testing and Verification of Intercell Welds -9-

Fig.24: Ultrasonic Inspection, Fig.25: Ultrasonic Inspection, Fig.26: Ultrasonic Inspection,
pressure 2.5 kN pressure 3.8 kN pressure 4.7 kN

Fig.24 to 26 show results generated at establish a good ultrasonic junction

2.5 kN, 3.8 kN and 4.7 kN. between the testhead and the intercell
weld.
Too little pressure will not create the
desired electroconductive junction, In typical applications water, grease or
where resistance is low enough so as oil are used as a transfer medium to
not to produce excess heat. achieve the necessary contact. It is
The lead will squirt out and will leave difficult to believe that this could be
intercell welds with blow holes (Fig.24). implemented in an assembly line for
batteries.
Beyond a certain force applied, the
results will be consistant with Fig.25 and
Fig.26. This supports the theory that
increasing pressure should be applied Beyond destructive inspection
up to that point just before structual techniques, electrical testing is the only
damage to the partition or hole might suitable method to verify intercell welds,
occur. in production environments.

Several battery manufacturers have

The ultrasonic inspection of intercell implemented two tests:
welds yields significant and
reproducable results without damaging The first using a tester integrated into
the test object. the cell welding machine and also the
second test after the battery has left the
But problems occur with regard to the intercell welding machine.
testheads used to transmit and receive
ultrasonic signals. Testheads are only
available for larger objects. The space In the following the advantages and
required to properly position a testhead disadvantages of both systems will be
is not available. Another problem is to discussed.

Testing and Verification of Intercell Welds - 10 -

 The internal microprocessor calculates
Voltage sense
the resistance which - at this stage –
V does not represent the resistance of the
intercell welds. First, the parasitic
resistances must be eliminated from the
Intercell weld
calculation.

Parasitic resistances are:

Welding current

Fig.27:4-Wire Current-Voltage Measurement - electrode resistance

- contact resistance
- resistance of the lead material
The general principle is shown in Fig.27.

The principle is based on the 4-wire

current-voltage measurement. A power
source will provide constant current to
the test object and voltage will be
measured across the test object. A high
impedance voltage input is required to R1: electrode resistance
measure voltages in the microvolt range. R3: contact resistance
The microprocessor will then calculate
the resistance. R6: material resistance

R5: contact resistance

R7: material resistance

R4: contact resistance

R2: electrode resistance

Fig.29: Parasitic Resistances

Voltage sense

Fig.28: General Assembly V

of a Welding Clamp

R1 R3 R6 R5 R7 R4 R2

The general assembly of a welding Welding current

clamp is shown in Fig.28. Fig.30: Real Resistance Model

The voltage sense leads are connected
to the clamps. After the welding process
and a cool down period an additional
measurement current is programmed to The real resistance model of the
pick up the voltage drop across the integrated resistance measurement is
clamps. shown in Fig. 30.

Testing and Verification of Intercell Welds - 11 -

Some machines measure the offset The parasitic resistances mentioned on
resistance and store this result for use the previous pages will form the major
as a compensating value when part of the total resistance. Irregularities
calculating the intercell resistance. in the intercell weld will not represent a
significant portion of the total resistance
However in production conditions the and will typically be within the tolerance
offset resistance may increase over time band.
due to contamination that accumulates
on the contact surfaces of the thrust
adapter (Fig.31). Another fact to be considered is the
resistance measurement takes place
when the the test object is subjected to
pressure from the welding clamp. Since
lead is a soft metal the pressure applied
is capable of creating a low resistance
elctroconductive junction even without
welding. Experiments confirmed that
resistances of both welded and
unwelded strap terminals can be within
cleaning cleaning the tolerance band when using a
clamping pressure of above 3.5kN.

Subsequently the integrated resistance

Fig.31: Contamination of Thrust Adapter measurement can only be used as
comparative data. It does not reflect the
real resistance of the intercell weld.

Contamination of the thrust adapter will Another option would be to measure and
result in increased resistance between record resistance during the welding
the electrode and the lead material process. This is commonly done when
because of a layer of contamination welding steel or sheet metal [3].
accumulates between the electrode and
lead material acting as an insulator.
Additional heat will be generated at the
electrode/thrust adapter, not at the lead
material. Therefore this heat will not
contribute to the welding process.

Heat and high pressure join to create a

thin lead film on the upper thrust
adapter. Any accumulation of lead will
tend to accelerate this plating process.
Poor intercell welds will result.

Fig.32: Master resistance curve for steel and

sheet metal welding
It is necessary to periodically clean the
surface of the thrust adapter to ensure a
consistant welding process.
Testing and Verification of Intercell Welds - 12 -
Welding steel and sheet metal master The results of all the records are shown
curves are used to control the welding in Fig.34.
process.

The first part of the curve should be

ignored as the data will vary significantly
depending on material conditions under
pressure.

The important part of the curve is the

shaded area. This area indicates the
acceptable limits of time and resistance
that will result in an appropriate level of
energy applied to the weld.

The question is:

Can this control method be used for lead
welding as well?
Fig.34: Intercell Weld Resistance Curves
To answer the question 3000 curves (yellow curve)
have been recorded.

The resistance curves for lead are

almost constant throughout the process.
There is no significant event during the
process that could surve as a direct or
indirect indicator of a good or bad weld.

Table 1 shows the deviations in voltage,

current and resistance as measured
during welding of 3.000 samples.

Fig.33: Expected Resistance Curve

Process 0.01s 0,02s 0.09s 0.17s
time

The typical resistance curve is a Voltage +/- 12% +/- 10% +/- 5% +/- 5%
combination of contact resistance and deviation
material resistance vs. welding time. At
the very beginning in the welding Current +/- 5.9% +/- 5% +/- 5% +/- 5%
deviation
process contact resistance will be very
high generating heat up to the melting Resistance +/- 9.3% +/- 8.2% +/- 5% +/- 5%
point of the lead material. Contact deviation
resistance will decrease, as material
resistance increases due to rising Table 1: Deviation of voltage, current and
temperatures(Fig.33). resistance

Testing and Verification of Intercell Welds - 13 -

The flat resistance curve can be The integrated resistance measurement
explained by two variables which tend to system cannot reliably determine the
compensate for each other: Although quality of strap terminal welds. In
the resistance of lead increases with addition this method will not differentiate
higher temperatures, resistance also between welded and non welded strap
decreases as the distance between the terminals. It should only be used to
electrodes is reduced. capture possible trend data such as the
resistance increase due to
contamination

3.2 Separate Resistance Measurement

3. Results
To measure the strap terminal
3.1 Integrated Resistance Measurement resistance more accurately a separate
system is required. This system ensures
real 4-wire measurement (Fig.35).
To summarize the investigation of the
integrated resistance measurement one
can say the advantages are:

- High pressure contact allows for high

test currents which generate higher
voltages across the clamps improving
signal to noise ratio measurement.

- No additional equipment investment is

required

The disadvantages are:

- Resistance measurement takes place

when the test object is subjected to
pressure from the welding clamp. Since Fig.35: Separate Resistance Measurement
lead is a soft metal the pressure applied
is capable of creating a low resistance
elctroconductive junction even without Contact is made by two current probes
welding and two voltage sense probes. The
current probes are spring-loaded to
- No real 4-wire measurement. Parasitic ensure safe contact. However, unlike
resistances will form the major part of the integrated resistance measurement
the total resistance. Irregularities in the system, this method does not clamp the
intercell weld will not represent a terminals together. In contrast the
significant portion of the total resistance probes apply force in the opposit
and will typically be within the tolerance direction attempting to separate a poor
band weld.

Testing and Verification of Intercell Welds - 14 -

The voltage probes are also spring-
loaded and have tapered contacts which
penetrate the oxyd film layer to
accurately measure voltages.

Five of these contact couples are

connected in series assembled in one
common testhead.

The machine interface should allow fast

change over of testheads to adapt to
any battery type. Fig.37: Effective Welded Area 104 mm2

The maximum current is limited by the

design of the spring-loaded current
probes.

Fig.38: Effective Welded Area 145 mm2

Fig.36: Testhead of Separate Resistance
Measurement

Tests were conducted to determine the

optimum current amplitude required to
generate significant variation in
measured resistance which then
correlates to percentage completeness
of welds.

Welded terminals were modified by

Fig.39: Effective Welded Area 245 mm2
drilling holes of increasing diameter
through the weld nuggets.

It is clear that test currents well in

excess of 100A are required for
meaningful results (figures 37 to 39).

Testing and Verification of Intercell Welds - 15 -

Unimpeachable data regarding the In general the addition of a separate
quality of intercell welds cannot be Welding Resistance Tester would be an
achieved using the integrated electrical enhancement to any battery assembling
resistance measurement line.

Digatron / Firing Circuits has literature

More accurate results can be obtained available describing the Welding
with an improved measurement system, Resistance Tester Product. Please
which is separate and installed after the contact [email protected] or
intercell welding machine [email protected]

References:

[1] Oliver Roehl, Bielomatik Leuze GmbH + Co; Intercell Welding Machines
August 2004

[2] Burghard Danch, Study of the Intercell Welding process,

October 1993

[3] Matuschek Messtechnik GmbH, Adaptive Control for Resistance Welding

in the Automotive Industry,
August 2001

Testing and Verification of Intercell Welds - 16 -