Practicality of making an inverter welder. / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Practicality of making an inverter welder.

previous 1 2 3 next

Move Thread

LAN_403

Steve Conner

Sat Nov 13 2010, 03:30PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Arcstarter wrote ...

That makes me wonder... Would it be best to use a hall sensor CT, or a linear shunt plus analog meter? I have both, but i did not even think about a current meter! Is there any real advantage of using a hall effect CT over a typical passive CT for over current limiting?

The only advantage of a Hall effect CT is that it can measure DC current, which regular CTs can't. The output of the welder is DC, so you can't use a regular CT.

Hall CTs have nothing but disadvantages in all other areas, but you put up with those because you want to make an isolated DC measurement.

You can't accurately infer the welding current from the current sensed on the primary side, because you can't see the current flowing during the freewheeling phase, so you don't know if it is continuous or discontinuous. You also have the magnetizing current of the transformer. Nevertheless, it might work OK, if you used the peak value to allow for the freewheeling phase.

Electra

Sat Nov 13 2010, 07:16PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

You can't accurately infer the welding current from the current sensed on the primary side, because you can't see the current flowing during the freewheeling phase, so you don't know if it is continuous or discontinuous.

Yes of course you are completely correct Steve, I did wonder how well that would work, though I just assumed it did, but I didnâ€™t see what the catch was at the time.
Or I suppose you would end up with constant power instead.

I never had the requirement use a hall C.T. for anything yet, so canâ€™t comment on using them.

Arcstarter

Sun Nov 14 2010, 04:30AM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

Well, i found some stud diodes that could possibly work, because i think they came from a welder. Problem is, i have no idea what the part number is.

They are made from IR, and they where made in Italy. It has 93 30 on it, but i think that is some sort of date code or something, it sure isn't a part number. There is also 155778 written on it.

There where 4 mounted on a fairly large heatsink, around 6"x6"x4 inches depth, each with a .01uf 1kv ceramic capacitor in parallel, and the heatsink had one of those thermal switches on it with a neat .1uf 600v capacitor co-packaged with a 47 ohm resistor in parallel. That was not connected to the diodes in any way, it must have been for the primary side of the device.

Given the 8 or so gauge wire spot welded to the diode, it was passing quite some current, not to mention the heatsink size. The cable connected to the heatsink was a 1 inch wide flat braided copper wire.

Any idea how to find a datasheet? I would like to use 2x two in parallel for fullwave rectification, using the center tap on this transformers secondary as the ground.

Electra

Sun Nov 14 2010, 11:50PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

I could take a guess that the code stood for 93 Amps 300v, a lot of SCRâ€™s and diodes use a similar coding scheme with amps followed by voltage in 10â€™s but often with more letters at the beginning / in-between.

They most likely are intended for line frequency 50/60Hz though, from what you describe with capacitors connected in parallel. Couldnâ€™t find a data sheet either, so cant say for sure.

Look for some fast ISOTOP package types (never cheep) or a number of TO-247 oneâ€™s with the best amp rating you can find, that say they can be paralleled in the datasheet.

The split secondary is the most economical for use of diodes, and for lowest voltage drop, remember each diode has to stand twice the open circuit voltage, plus a generous safety margin now though.

Arcstarter

Tue Nov 16 2010, 07:01PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

Yay! I found some rectifiers that will work fine. I have two DSEC240-04A diodes. They are SOT-227, 200 amps@25C and 400 volts. Here is the datasheet for the 600v version, the only datasheet that i saw when i typed DSEC240-04A.

One question i do have about these, is how to mount them. It is common cathode, which is connected to the mounting tab. What is the best way to mount it? I always use thermal paste, but i guess in this instance that would not work. Using the heat sink as the output 'wire' is what i intend on doing, i even have the perfect heatsink with the holes already drilled and tapped for the mounting of the IGBTs, plus a very thick wire for the output connection.

I would like to use just some basic logic for the PWM, such as op amps for over current and over temp etc, a simple oscillator with a 74hc14 with variable duty etc... Is that a bad idea? The way i have it in my very odd brain, it would be easier and cheaper to make it that way instead of using a TL494 or other PWM IC, mostly because that is what i have.

Electra

Wed Nov 17 2010, 01:48AM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

Those diodes sound just the thing, the datasheet shows them as common cathode to mounting tab, are you saying yours are the other way round? If using a centre tap transformer then yes what you said is fine, use the heatsink as one output connection, this is often done.

If your diodes are the wrong polarity why not consider using them as a negative voltage rectifier, so the centre tap would now become the + terminal, whilst this is not the normal convention I doubt it would really make a great deal of difference in this application, and better than trying to insulate them from the heatsink.

As the heatsink is going form one of the output connections, I would be inclined to use a separate heatsink for the IGBTâ€™s to maintain isolation should they fail in a disastrous way.

Nothing wrong with making you own PWM circuit, though I wouldnâ€™t say because it would be easier or cheaper, but is good for understanding what is going on as you have access to all the points in the circuit. All at the risk of it throwing up unexpected problems, if it doesnâ€™t work exactly the way you imagined mind you.
But then Iâ€™ve learned more by what Iâ€™ve done wrong than by what Iâ€™ve done right.

Arcstarter

Sun Nov 21 2010, 01:38AM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

I have a few more questions which i can not find anywhere on the web.

What is the point of having variable frequency? Those schematics all seem to have a pot for freq adjust, and i am not sure how necessary that is, or what it would help.

How about the output inductor. Any idea on what their value that should be? I would like to just parallel these two E core inductors i have, the E core is around the size of a SMPS core from a large TV and each have 4 turns of 10 or 8 gauge litz. I can have a couple strings of 2, i have 4 inductors in all. I think this is what they where used for in the first place.

I will be using this fullbridge of 75 amp 600v bricks. The PDF Electra posted shows a welder with a halfbridge of 100 amp devices. What i found to be intriguing was the fact that the gate driver was a fullbridge of 200ma transistors! So, i think i will use a couple UC3710T gate drivers, they can output 1 amp steady-state.

As for cooling, i am not too sure... I am using 240v input, the fans i have are all 120v.

Also, i found a TL494, so i will try using that.

Electra

Tue Nov 23 2010, 10:22PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

I measured the inductor in my welder and its about 50uH, looks about 10 turns of thick copper foil on an EE 50mm ferrite core, no doubt with a big air gap.

You can probably find the calculations in a good power supply design book if you know, all the other factors like transformer secondary voltage, switching frequency output current, ect.

I canâ€™t say I know the exact reason for including a freq trim pot, other than they must have designed everything for an optimum switching frequency in mind.

Arcstarter

Tue Nov 30 2010, 02:42AM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

I have been messing with assorted circuits, and i am about to be ready to start putting things together to see if i can make a working welder.

Alright, so, one question i came up with is regarding the arc starting... This transformer has a winding of about 5 turns, and 14 gauge litz which would be about 100v. What might that be for? Is that somehow used to start arcs, or is the typical HF resonant arc starter the only/best way? Using that winding would require a big relay or big triac to bypass once the arc has been struck. And 100v does not sound like much, so i cannot figure that one out on my own.

Also, i will be using an over voltage protection circuit, which will consist of a voltage divider into an op amp. That being the case, how can i ensure the op amp is protected from HV and HF from an arc starter? I am getting alot of my ideas from that very helpful schematic Electra posted, and it seems that a 100nf bypass cap on the output is the only HV HF protection it has.

One last question. Is the capacitor on the welder's welding output's only purpose to bypass hv and hf, or does it need to be a certain kind of capacitor. If not, i would just use a ceramic disk cap.

I know i have alot of questions, but this is a rather large project, and ultimately, could help me get a job in the future. Thanks to TheBoozer, i have this idea stuck in my head of getting a job at a local welding supply company fixing welders, even if i just start with janitorial. A homemade welder would be very handy for proving that i am worthy :P.

Thanks everyone!

EDIT: This is the circuit i came up with with for over/under voltage shut down.

(will open a Java app). The clock is the oscillator input. On the undervoltage shut down, i will add an RC time constant of a couple seconds, because when the arc is started the voltage will drop to near zero volts and i don't want that to shut it down. Im not sure about the undervoltage shutdown being under ~12v, but it sounds good. The overvoltage shutdown is about 49.5 volts.

For overcurrent, i will just use another op amp and a CT on the primary. The secondary's current i do not care so much about, as long as the primary is limited. This is how the welder schematic i have been viewing and loosely basing my ideas on works, anyhow.

I have not made up my mind on temp cutoffs for transformer and bridge, i can either use a thermistor and op amps with the advantage of variable temp cutoff, or a bi-metal thermal switch. I think i will use a bi-metal switch for the IGBTs, and the op amp and thermistor for the transformer. The reason behind that is that i have a bi-metal switch that was on the heatsink my IGBTs came from. As for the transformer, i would like to play with the temp cutoff, because i am not sure what i should be comfortable with.

Any suggestions, corrections etc are much appreciated!

Electra

Mon Dec 06 2010, 12:05AM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

Far as I can tell there are 3 methods of arc starting are commonly used a.) scratch start, basically no extra starting circuit. b.) lift start, the electrode is touched and when lifted itâ€™s sensed and the welder gives a brief burst on high power setting, and then settles to the preset current setting. (at least I think it works like that). c.) HF/ Hv start where a capacitor is repeatedly discharged into a transformer, the secondary of which is wound with wire capable of handling the welderâ€™s current, and superimposes the hv pulses on the output voltage. If you search you should find some examples of this type of circuit. Yes some capacitors are needed to protect the diodes, ect from Hv. some polypropylene ones would probably do as they come in a few hundred nf values, and you can put some ceramics in parallel too.

You should be able try out the Hv start bit on itâ€™s own to experiment with the turns ratio to see if it can generate enough voltage to jump a small gap.

I couldnâ€™t open that schematic of yours; just found a list of examples?

But you can use diodes to clamp the output of your potential divider to the supply rails to protect the op-amps if that helps.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob