Making a welder out of an old transformer.
Seeing the post a while back about the burnt out 3 phase transformer has got
me thinking again.
After getting my old Forney welder, I have been playing around with seeing
how hard it really is to design your own welder to operate the way you want
it too. That has led me to experiments with a couple old 480V to 120/240
volt transformers I have.
I have decided to post a little article on possible methods to convert an
old transformer into a welder. Opinions on my ideas is welcome.
This post will be very long.
The type of transformer you are looking for is one with a large open core. A
single phase with two coils on a single donut core (two leg). Or a three
phase one with three windings and three legs that are bridged at the top and
bottom (three leg).
Those types have a large core that will have a lot of flux leakage when
driven by only one leg.
Transformers that have the E-I core with the return path going around both
sides of a single winding have very little flux leakage and you will have a
problem designing a reasonable current limited transformer out of it.
You get the current limiting action by arranging the windings on the core so
that there is a large leakage path between them. The farther apart you have
the primary and secondary, the more flux leakage. The closer you have them
located, the leakage and you will get less of a current limiting action.
The primary produces flux when it energized. That flux tries to take the
easiest path to form a loop. Be it steel, air, or water. The easiest path
for it to take is steel, when the coil is wound around a steel donut, then
the magnetic flux generally prefers to go through that steel donut to form a
loop. Path of least resistance. But if you put a shorted coil on the other
side of the loop, then it stops the flux from passing through the complete
donut. The flux will jump across the middle of the donut between the primary
and secondary to complete the flux loop. Longer, and narrower the gap that
the flux has to jump across, the more counter EMF the secondary has to
generate to force the primary EMF to jump that gap.
A tall but narrow donut with the primary on the right and secondary on the
left leg. With the primary and secondary side by side. It will take more amp
turns in the secondary to get the primary flux to bypass the secondary.
If you have a wide but short donut with the primary and secondary on the
right and left leg, the gap between the primary and secondary is large. Then
the gap between the top and bottom part of the donut where the flux has to
jump is short and wide. It takes a lot less current in the secondary to
force the primary flux to jump the gap between the primary and secondary.
And example of a simple transformer for making a welder.
A 240/480 to 120/240 single or three phase transformer.
Single phase has....
One 240V primary and a 120V secondary on each leg. The primaries are stung
in series for 480V or in parallel for 240V. The secondary windings are in
series for 120/240 center tapped. Or can be put in parallel for 120V. Often
the secondary windings are split between the legs to even out loading, and
reduce leakage. That means that there is two 60V secondary windings on each
side that is in series with the opposed one on the other side.
Three phase with three legs that are bridged at the top and bottom. Each leg
has one primary and one secondary. wye or delta on primary and secondary
If you are lucky, there is enough room between the existing coils to wind
your custom windings.
If it is tightly packed, then you may have to remove all, or part of, the
windings on one side to make room for your custom windings..
Pick a coil/side/leg for your primary. If you have a transformer that has a
mixture of damaged windings and good ones, then pick a good one. It should
have a 208,240 277, or 480V primary and one or more secondary windings on
that leg of the transformer.
Remove the damaged windings. Then start experimenting.
Example. (from the experiments I did with one of my transformers. 15KVA
240/480 to 120/240 single phase) Your transformer has a 240V primary on each
leg. Hook 120V to one primary. Run a single turn or a couple turns around
the core and measure the voltage across it. If you have three turns and you
get 3.6V AC then your transformer is running at 1.2V per turn. Since
designed operating voltage for the transformer is twice what you have the
primary hooked too, then designed operating level will be 2.4V per turn.
When testing, you want to leave the transformer hooked to the lowest
exciting voltage you have, that will make the testing safer, and the
currents developed for that core the lowest possible.
Lets say that you have the above stated situation. 120V hooked to a 240V
winding which yields you a 1.2V per turn. On a single phase transformer.
That tells you that the primary on each leg has 100 turns.
Some will have a secondary by it's self on each leg that has 50 turns.
Some single phase transformers will have each secondary split between both
legs, so you have two 25 turn secondary windings on each leg.
When you have the one primary excited with 120V you will have 120V on the
other primary and 60V or two 30V on the secondary windings on each leg.
Now to find out the current characteristics of your transformer.
There is two factors that limit current in the setup. The conductor
resistance. And the flux leakage. Conductor resistance makes the windings
hot, so we want the windings big enough that that isn't a factor.
Lets say that your transformer has an available current of 2040Amp turns
into a shorted leg when the other leg is excited by 120V.
If you had a single turn large enough to handle that current on that leg
then you would have close to 2040A on that turn when it's shorted, or 1.2V
when it's open.
(If you made that single turn with a 4 foot piece of 10g wire, the current
would be limited by the wires resistance. 1.2V across 4 feet of wire is
close to 300 amps. The wire will get hot very quickly. The wire will be
giving in, not the flux in the transformer. The voltage across the leg doesn't
drop, the wire is just forced to drop the voltage across it's length.)
Lets continue with that train of though.
If you had two turns of wire adequate for the current, you would get 2.4V
when open, and 1020A when closed. That would be in about the right range for
spot welding, but not for arc welding. So lets continue on.
10 turns would yield 12V at 204A
20 turns would yield 24V at 102A close but not quite.
30 turns would yield 36V at 68A getting real close to the right voltage.
40 turns yield 48V which does quite well at 51A with the 1/16 6013 rod.
50 turns yields 60V that also does quite well at around 41A
Considering that the two low voltage windings I have are 25 turns each. I
can have one winding plus 15 more turns of wire with it to get 40 turns, or
use both in series to get 50 turns.
So, my transformer without major modification can serve as a 40 or 50a
welder with the addition one 15 turns of wire on one side. You could use the
other unused 100 turn primary on that side for a 120V OC 20A welder. Now if
I could only find 1/32 welding rod! :-)
For the secondary windings left over on the side that you are driving as the
primary. since they are in the same coil as the driven primary, there will
be no real current limiting. So you could use them as a CV output to drive a
mig gun, If you hook them in parallel for 30V. Considering that my
transformer is rated at 15KW you should have 120A continuous duty on hand to
drive that mig gun.
That is what you could do with a very simple winding arrangement.
Now, lets get to the complex winding arrangements for an arc welder.
Lets go for a base 50V OC on all outputs.
Lets hook up the two secondary windings on the output (current limited) side
in parallel so that we have 30V OC with a short circuit output of 81A that
will allow us to make use of what is already there.
Take that as are base winding. The winding can handle up to 120A
continuously (two 60A windings in parallel)
For are 50A tap, run are base winding in series with 16 additional turn new
winding. That yields 49V at close to 50A
For the 40A tap, we can't just use a 25 turn tap in series with the base
winding because that will yield 60V OC. Remember we want a nominal 50V OC.
Here is where we get into the complex windings.
To do that, we have to have the additional 25 turn winding in series with a
buck winding on the primary driven side. A buck winding on the primary side
is close coupled with the driven winding so it will reduce the output
voltage but it won't reduce the current available at the output since there
won't be any current limiting action.
So, you will have a winding with 25 turns on the current limited side plus 8
reverse turns on the driven side. That will yield you 40A at 50V OC. (you
get the output current from 50 turns on the current limited side, and you
use 8 buck turns to get the voltage down to spec) Now, it won't be exactly
that because with the buck winding, you are not directly shorting the
current limited side. So you may have to take a winding may have to take a
turn off of each side until current comes up to spec for that tap.
Now lets go for the 60A tap.
Lets take the base winding and add a 9 turn new winding to the current
limited side.(9 turns of the new 15 turn winding for the 50A tap) And run
that to a 7 turn boost winding on the primary driven side. That will get you
60A at 50V. (34 turns on the current limited side plus 7 boost turns)
The highest we could wind this transformer for is 70A and still maintain
100% duty cycle, which would be 29 current limited turns plus 12 boost
turns. (base winding plus 4 with 12 boost turns)
Now if we went with interment ant duty we could easily go 80A which is the
base winding plus 12 boost windings. In actuality, that setup would be
closer to 100A or higher because the current limited winding will be driven
beyond shorted. It will actually be driven reverse polarity. (0V - boost
So, just experiment and move the winding around till you find the current
You can tailor it to get about any OCV you want. If you want a digging arc
or a rubbery arc, you can tailor it to fit your desires.
To get 100A out, at 50 OCV you will want to run it on a 50A 120V breaker.
If you drive the primary at full rated 240V you will have to arrange the
secondary and buck boost windings to compensate for a higher amp turns that
is available on the current limited side, plus the additional volt turns.
"Now wait a second" you say "how does all that help me find out how much
current my transformer will put out?????????"
Well. Put some test windings on your transformer and Either run you selected
transformer primary off of 120V or even 60Vor 30V from another transformer.
Have a current meter on the test windings and short them out for a second to
see what you get. Rearrange the test windings and try again.
When you get the current down to a relatively low level at 30 or 60 V then
you can step up to the next drive voltage level and check what affect that
has. Rearrange the windings, and step up another voltage level.
I would not suggest that you power the selected primary to full rated
voltage and then short a test winding on it with no prior knowledge of the
transformer in question. You may be in for a little bit more current than
you bargained on. Plus a lot of melted wires to boot.
One way of bringing up the power level more gradually is use a large variac
to slowly bring up the voltage level on a shorted test winding and see where
the current levels off. You will be able to see if the current is climbing
way faster than you anticipated with the voltage, and if you will need to
change your winding layout before bringing it to full voltage.
If you have a transformer that you have taken one of the coils off one of
the legs, then you will have a lot more freedom to arrange the windings.
If you have a three phase transformer with two bare legs then that opens up
a world of possibilities. When you short a winding out on one leg, then the
flux will shift to the other leg. You could use one winding on one leg as an
output winding and have a winding on the other that is connected to SCR's to
vary the current. When the SCR's are full on, then it forces all the current
to the output winding, and when they are full off, then none of the current
is forced to the output. (it bypasses the output winding through the open
Don't forget the secondary windings on the drive leg. They can be run in
series with the primary to reduce the volts turn drive level at a specified
Now it is getting late, and I am getting tired, so I am going to wrap this
up and post it to see what you people think.