I am not an expert welder but did convert an AC buzz box to DC for a bit of
fun and since DC seemed to be preferred or at least more popular than AC. I
found that using DC made it easier to get the arc started. I can pull the
rod further away from the material and still maintain the arc but that is
not really useful. You can also use electrodes that are DC only although for
general purpose work on mild steel I doubt if that is a big benefit since
the common ones are available in either. If you want to have the electrode
positive or negative for penetration reasons, with DC it is always positive,
not switching back and forth 60 times a second.
If you are welding typically in a corner and the steel is magnetized then
you will get arc blow using DC (the arc is deflected by the magnetic field)
so AC is required.
Since I built my rectifier I haven't used AC but I only weld steel.
Probably not the easiest way but I used 12 35Amp 600V bridge rectifiers in
parallel. To use diodes in parallel you have to be careful to ensure the
load is balanced or else one takes most of the load, fails, then the next
one takes most of the load, fails, and so on until they all fail.
I brought the AC leads from the welder into an old desktop computer box. I
was fortunate to have some old electric furnace lugs which made connecting
to them easy. I wired the 2 input terminals of the bridge rectifiers to the
lugs using 18" of #18 wire all cut to the same length. These leads are more
than just connecting wires they are actually 0.01 ohm resistors to balance
the load. I wired the outputs of the bridge rectifiers to the output
terminals using the same length and size of wire thus adding another 0.02
ohms for balancing. So each bridge rectifier has in effect 0.04 ohms in
series with it to balance the load.
The open-circuit voltage of the welder is about 85V AC but when the arc is
broken collapsing magnetic fields can cause much larger spikes. That is the
reason I went with 600V diodes and the extra cost over 200V is little. Each
diode in each bridge is parallelled with a 0.01uf 1kv capacitor for
transient suppression and each diode bridge has a 200V surge protector
across the dc terminals. Personally, I think it is overkill and if I were to
do it again I would leave out the 0.01uf capacitors since they were
relatively expensive because I needed 48 and a pain to solder in place.
I mounted the bridge rectifiers on a large aluminum finned heatsink I
happened to have. I also used 2 small fans (muffin fans) to blow air across
the heatsink. I don't think it really gets very hot but better safe than
sorry. What you do want is to keep the diodes at close to the same
temperature so their characteristics stay similar. Also I bought all the
diode bridges at once so they have the same manufacturer and date code. That
way their characteristics are more likely to be similar than ones from
different batches. Makes the load sharing less likely to be any problem.
Note that for my normal welding only 3-4 of the bridges are required to ge
about 100-140A. The others are there for higher power and to handle current
if the rod sticks down when the machine and also a safety factor.
Now we have pulsating dc but we don't want to current to go to zero when the
voltage goes to zero. I did this by building 2 inductors which I placed in
series in the positive output lead. Could be in the negative lead if you
want. I got the cores out of the biggest junked microwaves I found at the
local dump. Microwave transformers are cheaply made thank goodness. All of
the E shaped laminations are stacked up and all of the I shaped laminations
are stacked up, the coils placed on the Es and then the Is are welded on. A
better way is alternating the Es and Is but they are hell to get apart. So
just hack saw the weld and you can pry the E and I stacks apart. Pull or
cutoff the existing coils. I used #4 welding cable and put something like
about 9 turns on each core. You can't put the I part back on tightly since
the high current will cause the steel to saturate and you will lose the
inductance. I put something like about 1/8" masonite between the E and the
I pieces to create a gap.
I couldn't easily find any connectors for the rectifier box that were
compatible with the Miller connectors and weren't worth a fortune. I bought
4 sets of generic welding connectors and made a set of adapters with the
miller plugs on one end. I put the other ends (males) on the leads which are
the input to the rectifier box. I mounted the female parts of the last 2
sets in the box and put the males on my welding cables. If I want to switch
to AC welding I can disconnect the welding cables from the rectifier box and
the box from the welder and connect the welding cables to the welder.
Other ways of doing this, like I said this wasn't the easiest way. Ideally,
you can find 4 big 400A diodes and you only have to deal with 4 pieces not
12. You also don't have any load balancing issues. I don't have ready access
to surplus places so I didn't pursue this route but it is the best way to go
if you can. You still require the inductors. Note that long welding cable
leads will provide some inductance as well and so just how critical it is to
have the inductors - I don't know. I often think of measuring what the
current is when the voltage is zero but have never done it. I did buy some
6010 which is a DC only rod and it worked just fine so I am happy.
Another cheap source of diodes is from car alternators. These diodes are
probably rated at around 100A so you would need about 3 sets. Problem is
that they are pressed into plates and are I believe a 3 phase full wave
rectifier configuration with 6 diodes per alternator. You have to be able to
mount them to create a bridge.
The car alternators I have seen use 9 diodes in the pack. 6 provide 3 phase
full wave rectification and the remaining 3 provide field current for
excitation. They are high current devices but may not have a high enough
Thanks for the info. My knowledge about using alternator diodes for welders
came from some posts by others using them. I would guess that the voltage is
high enough because of a lack of information to the contrary. Years ago it
was not uncommon to see rectifier diodes with voltage ratings of 50V but now
you rarely see less than 200V or perhaps 100V sometimes.
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.