Background
Over the last several years I have rebuilt several power tool battery
packs by replacing the internal cells with new ones. The method
factories and quality rebuilders use to connect the various cells
together electrically is spot welding a nickel strip to the metal ends
of the battery cell. These strips run from cell to cell to
connect them as required for the particular tool. The only way I
have had to perform this function is to solder wires from cell to
cell. This is really not a good technique, as the long duration,
high temperatures, and large heated area required cause a considerable
rise in temperature
within the cell. This can either degrade the cell or outright
destroy
it in extreme cases A spot welder operates at temperatures well
above those of soldering, but for a very short time and heats only a
very small area. The total heat rise internal to the cell is
actually
negligible.
I have recently watched a number of YouTube videos showing a wide
variety of home made spot welders built to weld these nickel
strips to battery cells. The designs shown vary widely in their
complexity and capabilities. The simplest I saw consisted of a
car battery with a starter solenoid attached. A pair of wires
went to a couple 12 gage wire electrodes which were placed on the
nickel
strip. The operator then quickly pressed and released a button
which activated the solenoid. The timing control was totally
dependent on the operator's ability to consistently press the button
the correct length of time. Other models had very complex timing
circuits driving banks of parallel connected MOSFET transistors to
switch the very high currents and low voltages involved.
Something about watching these videos triggered a desire to build a
battery spot welder with a simple, economical, but dependable
design. I really have no overriding need for one of these
machine, although it would be very useful when I do need to build a
battery pack. There really were three reasons that I decided to
go ahead anyway:
- The machine intrigued me and it would be great when I do
need it
- I wanted to design and build a fully featured, high quality
machine on a tight budget.
- I wanted to use this as a test case to learn to use Arduino
single board computers
Regarding number 3 above, I have found that if I want to learn a
new
technology such as a programming language, a computer program, or a
specific
computer, just starting to take lessons on-line makes it hard to
concentrate and provides little incentive to really "dig in". But
if I have a project that requires it, I start absorbing the knowledge
vigorously and enthusiastically. I just learn it faster and
better by immediately applying it to a project. I have done that
here to learn to use the Arduino hardware and its software.
The Arduino is a very small single board micro-controller. It is
in reality a small computer designed specifically to easily control a
number of analog and digital inputs and outputs. Included is the
capability to do logic and numerical mathematical functions and a
variety of elaborate control functions. Its programming is based
on the C and C++ computer languages and any C programmer (which I am
not) should be able to pick up the programming in a few minutes.
My Overall
Plan
I decided that using a transformer from a scrap microwave oven,
modifying it to remove the high voltage winding and replacing it
with a very low number of turns of a heavy welding cable would be the
most practical source of low voltage, very high current as needed for
spot welding. There are numerous examples of machine using this
power source on YouTube. They all appear to have more than enough
power and control the welding by suppling a timed pulse or pulses and
in some cases even controlling the current output. This would be fairly
easy to implement using an Arduino micro-controller driving a solid
state relay. This relay would turn the transformer primary on and
off as desired, eliminating the need for complex, high current switches
which directly control the several hundred amp welding current. I
have
recently watched a number of YouTube videos showing a wide
variety of spot welders designed specifically to weld these nickel
strips to battery cells.
This approach would require me to design and build a suitable housing
and connection scheme along with the hardware and software to allow the
operator to control all the timing setting of the welding pulse(s).
Work is
Started
I started looking for a junk microwave oven. The local scrap yard
stated that the get them all the time, but that all they currently had
were in bales. I wasn't sure if he had said "barrels" or
"bales". I shortly discovered that this meant that they were
crushed and bound in tight bales about 3 or 4 feet on a side.
There was absolutely no way to retrieve anything from there.
I then went to our local Habitat Restore where I found an old microwave
for $10. I bought it but really wished that I could have gotten a
scrap one. I then did some quick figuring that if the scrap yard
charged 50
cents a pound (as they do for steel) it probably would have cost just
as
much, as I am sure a microwave is at least 20 pounds. I was then
happier!
I disassembled the microwave, keeping the transformer and several other
pieces. I then tore into the transformer. Fortunately most
microwave oven transformers have separate coils for the primary (120
volts) and the secondary (about 2000 volts) which sit side by side on
the transformer core instead of being wound over each other in one full
length assembly. This allows removing one coil without affecting
the other. I removed everything but the primary coil.
![](swTrans%20clearing.jpg)
To remove the high voltage winding, I sawed one side of the loop of
wires flush with the core. I had hoped the rest of the
coil would just pull out from the other side - it did not! I then
started pulling groups of strands out until the rest was
loose enough to pull as a unit. There was also a low voltage
winding consisting of several turns of an orange wire
which I removed at the same time.
![](swTransformer.jpg)
In the space vacated by the high voltage winding, I made a 1 1/2 turn
secondary using number 2 gage welding cable.
This provides a secondary of about 2 volts and several hundred amps -
ideal for spot welding! The original primary coil
can be seen at the bottom. I was very careful not to damage this
when removing the other wires.
On a side note, every video on YouTube I saw that made this
modification stated that they used a 2 turn secondary. If
they truly had, the second lead would have exited on the back
side. They really had 1 1/2 turns just like this one.
\
In addition I just found a second microwave transformer at a scrap
yard. (It turns out that it cost just as much for just a
transformer as scrap as I paid for a total working
microwave at a thrift store.) I thought I would share here how I
removed
the windings form this transformer. It was much easier and
cleaner than my first one. If anyone reading this wants to do
something similar this might help them.
What I did was to cut the high voltage coil with a sharp chisel instead
of sawing it. This produced a much smoother cut and
allowed me to drive the coil out of the core with a small block of wood
cut to just under the cross section size of the coil.
![](swTransformer%232.jpg)
Here is my second transformer (for future use, not this project).
After cutting the coil on both sides with a chisel, I was able to
drive the coil out of the core with many, many blows using a small
wooden
drive tool. Even the 2 1/2 turn 3 volt coil came out fairly
unscathed. I was forced to use this method as the base flange of
the transformer did not give me a clear shot at sawing the coil.
It
turns out that it was a much better method!
Now that I have my power source I will concentrate on controlling
it.
I bought a "starter kit" for Arduino which contained an Arduino
micro-controller and many accessories. These included a
breadboard and connection wires to make connecting circuits easy, and
many sensors, controls, and a display. With this I was able to
breadboard test circuits and try out various programs to control
them. I started taking YouTube lessons from the same excellent
teacher, Paul McWhorter, whose 13 part series on using
Fusion 360 I took earlier. He also has a series of lessons on
using the Arduino
- 36 of them! Using his wonderful tutorials I learned from the
starting basics on to some very complex setups using input and output
devices of various types. I followed along on each lesson
duplicating his setup and trying to solve the problem before he reached
the point of telling us how. Several of his lessons used devices
I did not have, so I just followed along on those. Pretty soon I
had the knowledge to operate an incremental resolver for my user input,
an LCD display to follow in making my adjustments and the overall logic
to make it all work together.
My welding profile is to fire the system for a short preliminary pulse,
wait a short time, then fire a much longer pulse to actually complete
the weld. Information I found describing the techniques for this
type of welding recommended this approach as the preliminary pulse
breaks down any surface dirt or oxidation and provides a good contact
between the electrodes and the metal. After a short delay, the
main pulse is then more repeatable as the contacts are already solid
when it
starts. I really don't know how much difference this
makes, but with the Arduino it is simple to do, so I did!
At power on, the display lights up and shows the status of the
welder. By default the display shows the
settings for the pre-pulse, the delay, and the main pulse and is ready
to operate with those parameters. If it is desired to change any
of the values, pressing the resolver knob steps through adjusting each
of the settings and then back to operational mode. While in the
setting mode, turning the resolver slowly will change the settings 1
millisecond at a time. Turning the knob rapidly will change them
10 ms. at a time. After stepping through the 3 setting modes and
back to the operational mode, it is ready to weld with the new
parameters.
After breadboarding each step of this operation separately, reading the
resolver,
writing to the display, and writing the code which ties it all
together, I put together a breadboard which had all these steps
operating together, and finalized my operating code.
![](swBreadboard.jpg)
This is my breadboard of the control system. The resolver (with
the knob) and the push button send the user input to the Arduino on the
left
and it provides output to the display and the solid state relay which
will control power to the transformer.
Building
the Welder Case
OK, now with a power source and a working controller, I need to build
and wire a box to hold all the parts.
I quickly did a rough layout to determine how large a box I needed,
then went to my old friend Fusion 360 and started to design it.
I did something I had not done before, I built a multi part assembly
shown all together, then I 3D printed each part separately. It
worked very well with only one exception. After getting the base,
the transformer, and all 4 sides done without a hitch, I started on the
top cover. For some reason which I have yet to figure out,
I could
draw the basic part, but when I tried to extrude the ribs and cut
the holes it would not work. I re-did this several time
using different methods of defining my working plane and my extrude
settings, but with no results. I finally built the part flat on
the X-Y plane with everything working as is should. I then moved
the completed cover up to its correct position and all was well.
![](swFusionLayout.jpg)
This shot of my computer screen shows the Fusion 360 design. I
have turned off the left side panel to show inside. (Actually, I
never completed
the left panel as I realized it was identical to the right panel.
I later added a cutout for the USB port on only the right cover, but by
turning
off the final step, I can still print a left cover.) After
completing the design, I was able to select one panel at a time and
prepare it for 3D printing.
![](swPrintedPanels.jpg)
This is the result of printing all the case panels. The ribbing
on the plastic panels makes the very thin 1.5mm. plate (about 1/16th
inch) very stiff!
The bottom
(not shown here) is a panel of just under 3/4 inch thick MDF.
![](swHarness&panel.jpg)
This shows the logic wiring harness; The pin headers on the right
all plug into the Arduino, then looking clockwise, the 4 pin header
drives the panel display,
the
2 loose wires, the solid state relay, the 2 pin header connects to the
push button and connector, and the 5 pin header, the resolver.
On the right is the front panel with the connector and push button
(upper left), the power switch and resolver (upper right), and the
display on the bottom.
![](swCompletedWiring.jpg)
The left image shows the completed AC wiring. It also shows the
mounting of the Arduino with the USB port extending through the right
hand wall. This will allow any
reprogramming that may be required or desired in the future to be done
without any opening of the case.
The right picture shows the completed
unit looking from the top-back. Only the top cover and handle are
missing
here. I wanted a carrying
handle on the top, but I did not want to trust the thin plastic to
carry the weight of the heavy transformer. The steel black cross
solves
that. It is mounted to the base with
pipe standoffs and threaded rod (see left picture). When the top
cover is in place the handle screws pass through it and into this
assembly. The 12 volt power supply for
the Arduino is mounted to the bottom of this cross.
In designing the case I had
initially
thought about having my welding probes coming out of the front of the
case and somehow moving up and down to make the weld. Upon
further thought I felt I would be more comfortable welding with a hand
held unit attached to the cables. I therefore have the cables
exiting the case through the front. As I write this portion, I
still have to design and
build the electrode holders to attach to these cables.
Initial tests:
Up until now, I have only tested the power of the transformer trying to
weld
some fairly thick sheet metal. I was able to get weak welds with
materials about 10 times the final thickness I will be doing.
This showed me that I have plenty of power to make the types of welds I
want.
Finally, my roll of nickel connecting strip arrived and I can try it
with realistic materials As I have not yet built (or even
designed) my electrode holder, I kludged a setup using a couple of
bolts, 2 lengths of 12 gage electrical wire held together with a couple
of zip ties, and a piece of cardboard to keep them from shorting
out. This worked well enough to determine that the system works
very
well! My first test on my .15 mm thick nickel (about .006 inch)
was to a piece 1/32 inch stainless. This is far thicker than a
battery case, but my first attempt stuck the strips together
nicely. This was at only 60 ms. and I was readily able to pull
the strips
apart, however the nickel failed around the weld and left holes where
the weld had been. I increased the main pulse to 110 ms. and got
a much more
substantial weld. When I pulled it apart it left much larger
holes around the weld.
I then used a AA battery and welded to it still using 110 ms. and
received a beautiful looking weld. I could not pull it apart by
hand. I think welds like this are what I want to see each
time! I won't do any more testing at this time, as I need to have
a suitable electrode holder and electrodes
![](swTestResults.jpg)
.
On the left is my test setup showing my cobbled together electrodes and
the battery that my final test welded to. I triggered each weld
using a foot pedal
I
attached to the triggering connector.
The right shows a close up of the welding test strip and the finely
constructed electrode holder. The two small holes on the left of
the nickel strip are my first weld. Pulling
it apart left two tiny holes. The center of the strip shows my
second test with a longer pulse and larger residual holes after tearing
the weld apart. My attempt to weld to
a AA battery was totally successful! I could not pull the strip
off the battery with my bare hands and the weld looks very
characteristic of factory battery packs.
Operation:
Operating the spot welder is quite simple. The only controls are
a power switch, a "weld" push button, a connector for a remote
button or pedal, and a control knob. When the system
is powered up, it is set to provide default values for a short pulse, a
short delay, and a
main pulse much longer than the short one. Hopefully I will find
that a certain value of each of these will provide consistent, strong
welds. Just turning the control knob has no effect unless it is
first pressed. Pressing it puts the machine in the mode to adjust
the length of the short initial "pre-pulse". Turning the
adjustment knob will
raise or lower the value of this pulse by one millisecond if the knob
is turned slowly, or by ten milliseconds if the knob is turned
fast. Pressing the knob again sets the machine to adjust the
delay between pulses in the same manner. A third press allows
setting the main pulse time, again as above. A forth press of the
knob places the machine back in operational mode. There are 5
display screens to indicate the results of these settings.
![](swDisplayScreens.jpg)
These are the 5 screen displays provided: The first one is the
standard operational screen. It displays all the current
settings, and is the only screen which allows
a weld to be made. Turning the knob has no effect while in this
screen.
The next 3 screens are shown sequentially as knob presses call them
up. As the setting knob is turned, the number changes to indicate
the new value to be used. Pressing
the knob again after the 3rd setting screen returns it to the first
screen.
The final "welding" screen shows momentarily as the weld is actually
being made.
The Electrodes:
For the electrode holders, I decided to make a hand held straight line
unit - the cables come in the rear and the electrodes come out at the
front at nearly right angles. I will have 2 brass bars holding
the
electrodes at one end and connecting to the power cables at the
other. They will be spring loaded downward and pivoting at the
rear. This will allow each electrode to adjust its pressure if
the holders are not perfectly aligned.
![](swElectrodeHoldersAssm.jpg)
The electrode holders were machined from 1/2 inch hex brass I already
had. (Yes, it is nice to still be able to make parts the old
fashioned way - by cutting metal!) The flats at the rear are where
I bolt the cable lugs and the hole and slot at the front clamp the
electrodes. The holders are mounted using 3D printed blocks which
provide insulation, guidance and a pivot point. The orange parts
are
prototypes and the black ones are the final version.
The electrodes are angled 15 degrees toward the front and about the
same toward each other. The gap is adjusted by changing the length of
the protruding electrodes.
As for the electrodes themselves, most YouTube video presenters just
use household electrical wire. This seems to work fairly well,
but this wire is soft and bends easily. I wanted fairly hard
copper so the electrodes would keep their adjustment well. A
number of users recommended copper roof nails as an ideal
electrode. I got a small box and tried them with mixed
results. They are marginally too short, are not very round, and
the cross section is rough where the serrations are to make the nail
grip
well. In the end I bought a couple lengths of precision 1/8 inch
diameter hard copper sold as raw material for knife blade pivots.
These were my material choices for electrodes: Home wiring #12
gauge wire shown - available in #10 but harder to find, copper roofing
nails, and
precision hard copper rods. Shown also are electrodes made from
the nail and the rod.
![](swCaseIterations.jpg)
Not having any idea what I wanted for a case or how to make it, I
started simple and made many iterations until the case I wanted
emerged. The
front prototype needed only a couple minor changes to produce the final
black one shown below.
![](swCompletedUnit.jpg)
And here is the completed unit!