Exactly how does an "inverter welder" work

[theoldwizard1]

Arc welding (stick, MIG or TIG) is done by creating an arc from one electrode to the work piece. This requires very high current but not necessarily very high voltage. For years, a very large "step down" transformer converted "high voltage" (120V, 240V or higher) to lower voltage (and higher current) via a transformer. Many stick welders work this way today. Lincoln used to (still does ?) sell a rectifier kit (some diodes and maybe some other components for filtering) that would convert that high current AC to DC.

The latest in stick welders are inverter welders. These are DC only output and are very small, meaning they don't have the huge transformers like they used to,

I do have some background in electronics (which may be part of the problem !) but to me "inverter" means converting DC back into AC (a lot trickier than AC to DC).

So what kind of "black magic" electronics is going on inside those tiny inverter welders ?

Forget about all the fancy controls, I'm talking basics.

 

[RPH]

Look into induction heating technology. This will give you the basics needed. I work with induction and its not as hard as you think. In fact, induction heating was the first commercial use of electricity.

 

[nehog]

The mass of the transformer is very dependent on the frequency. The core, (iron) is heavy and to magnetize that big core takes a lot of wire. As frequency increases the core gets smaller (there's a complex formula for this...) until you reach a point where no core at all is needed (that's radio frequencies, however.)

Basically they work like this:

the input voltage (120 or 240) rectified into high voltage DC. Various bits of magic are applied to compensate for 120 vs 240 volts, but in the end that high voltage DC is typically about 340 volts. This high voltage (still a relatively low current) is then converted to high frequency AC (line frequency is 60 Hz, while that converter converts it to a frequency of perhaps 30,000 Hz). At that high a frequency the transformer can be relatively small, and more efficient too. The primary of that transformer is 340 volts and the secondary is perhaps 30 to 40 volts. That low voltage, high current AC is then rectified to DC. Again because the frequency is very high, only small filter capacitors are needed to supply relatively clean DC voltage.

 

[theoldwizard1]

The mass of the transformer is very dependent on the frequency. The core, (iron) is heavy and to magnetize that big core takes a lot of wire. As frequency increases the core gets smaller (there's a complex formula for this...)

That is the detail I was missing ! Got a link to more details ?

The input voltage (120 or 240) rectified into high voltage DC. Various bits of magic are applied to compensate for 120 vs 240 volts, but in the end that high voltage DC is typically about 340 volts.

That sound high, especially for 120V AC input. IIRC, peak-to-peak voltage is only Vin * √2.0

This high voltage (still a relatively low current) is then converted to high frequency AC (line frequency is 60 Hz, while that converter converts it to a frequency of perhaps 30,000 Hz).

Modern switch power supplies are faster than that !

At that high a frequency the transformer can be relatively small, and more efficient too. The primary of that transformer is 340 volts and the secondary is perhaps 30 to 40 volts. That low voltage, high current AC is then rectified to DC. Again because the frequency is very high, only small filter capacitors are needed to supply relatively clean DC voltage.

Make sense now.

 

[theoldwizard1]

WOW ! I think I strained my brain !! I had no idea how complicated transformer design is !!!

Check the attached graph. Optimal switching frequency (for pot style core) is between 250K and 400K Hertz.

More specifically, the peak magnetic flux density Bmax (flux density has to do with overall efficiency of a transformer; higher frequency = higher efficiency) , should not reach the core material's saturation flux value Bsat

Bmax = Vrms×108/4.44N×Ac×F​

F is frequency and is a multiplier in the above equation. It is easy to see how increasing frequency increases magnetic saturation and efficiency.

 

[RPH]

The multiplier on the incoming ac is1.414. The peak of a standard 120 vac is about 187 vac. This is what the rectifier works on. In my typical system 480 vac works out to be a690 vdc bus that my transistors will now convert to high frequency upto 450 kHz. Depending upon application. Transformer saturation is always a concern.

 

[theoldwizard1]

The multiplier on the incoming ac is1.414. The peak of a standard 120 vac is about 187 vac. This is what the rectifier works on. In my typical system 480 vac works out to be a690 vdc bus that my transistors will now convert to high frequency upto 450 kHz.

Ahh ... just what are you designing ?

 

[RPH]

High frequency induction power supplies to the 3 gigawatt range. Bonding applications, brazing, hardening and tempering, shrink fit, and tube welding are some of applications.

 

[justsam]

Am I the only one that played with all that WWII surplus aircraft electrical "stuff"?

It was all 400Hz in order to reduce mass of transformers, motors, relays, etc

I am not sure what modern "weight reduced" systems use since it is much easier with today's semiconductor devices to do the conversions.

 

[nehog]

...

That sound high, especially for 120V AC input. IIRC, peak-to-peak voltage is only Vin * √2.0
...

For 120 volts, it works as a voltage doubler, in 240 volt feeds it works full wave. Switching between the two modes is automatic.

 

[bczygan]

MY HEAD HURTS TOO!

 

[bczygan]

Is one type better or worse than the other besides size and weight?

Bill

 

[matt_i]

Is one type better or worse than the other besides size and weight?

The transformer machines are heavy and use a lot of current but are as bulletproof as the insulation of the windings.

The inverters are lightweight and just as powerful as far as the output but if you lose the electronics it can be a major % of the initial cost to repair it.

 

[sberry]

I have both. I believe Hobart discontinued the ****** Stickmate. Being able to do real work from 120 is outstanding. Wish they had that when i was a kid.
I ran a Maxstar back to back against the Lincoln Dc buzzer today. The buzzer is really good. The Max has a lo ocv safety frature, it's good but I really think I got to give the buzzer an edge on starts and restarts. It's so good a guy couldn't tell it apart from machines cost 5x as much if it was on the other side of the wall and couldn't see it.

 

[Lelandwelds]

Is one type better or worse than the other besides size and weight?

Bill

On the classic machines, you picked a bigger hunk of iron to wrap wire around or ramped current around a nearby hunk of iron. Fine tuning was hard.

Inverters have high part counts but it is relatively easy to add double extra inductance for SS welding or zero for aluminum. The low end can be a perfect stable five amps but leave the high end perfect also. You can have lots of dig amps when the arc length gets shorter. Or, not. Both a stick and mig in the same wrapper. Any voltage or phase of input or output is possible (if not reasonable).

All by flipping a switch or made mostly invisible to the user if that is the design goal.

 

[Lelandwelds]

I ran a Maxstar back to back against the Lincoln Dc buzzer today. . . . I really think I got to give the buzzer an edge on starts and restarts.

So, are you comparing Lift Arc starts against the classic high frequency starts? That's really high praise if you think that highly of both and think they are close.

It gets interesting when you compare different amps or metals . My skills were not as sharp as the guys who did it for a living. My answers were more," Yup, works good."

 

[sberry]

I am a yup it works good guy too. No, these were not tig but sticks. Some 6011 and 7018. Super common for equipment repair. I have done this a lot, i am still impressed as to how good this welder really is. I am not in current practice. If i was welding regular for a couple weeks could really refine it and be critical.
With small electrodes on smaller material, critical small note pipe etc the current can be fussy within a couple amps and even influence real deposits quality, even some grain structure if the electrode is operating at exact the right current.
When I was a daily driver could tell some minute changed in some of it, not much between machines.
This doesn't include dowhill pipe specialists that like a bit different setting but vertical and overhead structural and pipe.

 

[joe_padavano]

This is a very timely discussion. I'm agonizing over pulling the trigger on the AlphaTIG vs the new Eastwood TIG200 Digital, and the whole issue of inverter technology vs. old-school transformers is of interest. Of course, that also brings up the question of IGBT vs MOSFET technology...

 

[theoldwizard1]

This is a very timely discussion. I'm agonizing over pulling the trigger on the AlphaTIG vs the new Eastwood TIG200 Digital, and the whole issue of inverter technology vs. old-school transformers is of interest.

I am pretty certain that both of those use inverter technology. All the "fancy extras" require inverter technology AND a microprocessor to implement.

Of course, that also brings up the question of IGBT vs MOSFET technology...

IGBT = Insulate Gate Bipolar junction Transistor
MOSFET = Metal Oxide Silicon Field Effect Transistor.

IGBT is really a two stage transistor with the first stage being a MOSFET and the second stage being a bipolar junction transistor. Good for high power.

MOSFETs are good for high power where low loss is important (power supplies), but the more power you need, the more they cost, and in many cases, multiple MOSFETs are used in parallel (sadly, when one goes they all go).

IGBT are also good for high power, but a single IGBT and handle a lot more power than a single MOSFET. Probably more cost effective for really high power where you can give up a bit more losses.

 

[Lelandwelds]

I am pretty certain that both of those use inverter technology. All the "fancy extras" require inverter technology AND a microprocessor to implement.


IGBT = Insulate Gate Bipolar junction Transistor
MOSFET = Metal Oxide Silicon Field Effect Transistor.

IGBT is really a two stage transistor with the first stage being a MOSFET and the second stage being a bipolar junction transistor. Good for high power.

MOSFETs are good for high power where low loss is important (power supplies), but the more power you need, the more they cost, and in many cases, multiple MOSFETs are used in parallel (sadly, when one goes they all go).

IGBT are also good for high power, but a single IGBT and handle a lot more power than a single MOSFET. Probably more cost effective for really high power where you can give up a bit more losses.

Thirty years ago when Powcon was the new thing, IGBT and the rest would pop and leave you stuck. Service could not diagnose and fix. Parts less labor costs more than the entire machine.

Today, they are routine and becoming standard and ordinary. Service can fix without drama. When something dies it doesn't kill everything across the entire unit. One little corner of the guts needs to be swapped. I don't care what is under the hood as long as it has the right feel and is reliable. Most "features" are unimportant as long as it has the right arc and longterm reliability.

Some are dogs. Some are great. Kinda like vehicles, restaurants, and movies.