Author Topic: Homemade IGBT MIG welder with ATX PSU parts. [blown yet again][blown count: 5]  (Read 28562 times)

0 Members and 1 Guest are viewing this topic.

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts.
« Reply #75 on: July 04, 2019, 10:55:26 am »
Getting a little tight in the ATX case but everything should fit just fine.
I might be able to fire this up either today or tomorrow.

Ps: the plywood is bowing but that's only because it's just a mock-up to see how everything fits.
« Last Edit: July 04, 2019, 10:57:13 am by Refrigerator »
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: Homemade IGBT MIG welder with ATX PSU parts.
« Reply #76 on: July 05, 2019, 10:10:00 am »
<snip>
 All the cheap MIG welders have problems (flexible plastic rubbish held together with self tapping screws).

Really? what utter junk. Chinese I assume ?

Of course.  They are very cheap but you get what you pay for.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts.
« Reply #77 on: July 05, 2019, 11:05:28 am »
Just tested the beast out, but not on 230V yet.
I connected the 12V supply to all the control circuitry as well as the bridge rectifier for the inverter and ran a quick test, shorted the output with with my DMM set to amps and got 1.3A.  ;D
Now i can hook up the scope and set my deadtime, after that everything will be packed into the ATX case and closed shut.
Also i'm not sure whether or not to put a choke on the output of the rectifier, i found a nice FBT core that has brass bolts holding it together and i can fit 3 turns of my 35mm2 welding wire in it.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Homemade IGBT MIG welder with ATX PSU parts.
« Reply #78 on: July 05, 2019, 11:41:37 am »
<snip>
 All the cheap MIG welders have problems (flexible plastic rubbish held together with self tapping screws).

Really? what utter junk. Chinese I assume ?

Of course.  They are very cheap but you get what you pay for.

Of course you do, but if it only lasts a short while because it's junk, is it even worth buying ?

My Father always said to me, buy the best you can afford, that way you will only buy it once, and save money over buying junk again and again.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts.
« Reply #79 on: July 05, 2019, 12:19:05 pm »
<snip>
 All the cheap MIG welders have problems (flexible plastic rubbish held together with self tapping screws).

Really? what utter junk. Chinese I assume ?

Of course.  They are very cheap but you get what you pay for.

Of course you do, but if it only lasts a short while because it's junk, is it even worth buying ?

My Father always said to me, buy the best you can afford, that way you will only buy it once, and save money over buying junk again and again.

Does it even have a spring or does it rely on the flex in the plastic to keep the pressure constant?
And that motor looks wimpy as. For a spoolgun this sort of thing would be ok-ish but not for anything else.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #80 on: July 05, 2019, 01:59:08 pm »
Getting quite messy but if everything goes to plan i'll have it running today.
With the DTC cranked all the way up i was able to get 4.4A on the output (that's running on 12V) but i toned it down to limit max duty cycle to about 85% just to be safe.
Also came up with a neat way to connect the welding cable to my diode bridge.
My old LPG injectors have a coil resistance of 4Ω, which is going to draw 3 amps from my 12V supply. My 12V supply can handle up to 23.5 amps so that's not a problem but i'm worried the solenoids on those injectors won't last long with 36W going into them.
Good thing the coils are removable so i can perhaps retrofit a coil from a different solenoid such as a solenoid from a washing machine for example.
« Last Edit: July 05, 2019, 02:04:49 pm by Refrigerator »
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #81 on: July 05, 2019, 09:48:48 pm »
The maiden voyage is going to be tomorrow but all the tests are now complete.
The way i connected my voltage pot meant that if i disconnected it the voltage would be regulated to 2.5V (my reference voltage).
And that's almost exactly what i got at the output when testing on 12V + 100Ω load (because it would max out at 4V with the pot).
This means that my voltage is properly regulated.
All that is left is to finish is the mains wiring (circuit breaker, cable, plug), then i also need to add a resistor to load the inverter output a little for stability reasons.
After that i'll make a driver thingy for the gas solenoid (probably a small MOSFET on a wire) and the welder will be ready to make some sparks (the good kind, i hope).

I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 29810
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #82 on: July 05, 2019, 09:56:08 pm »
..................
After that i'll make a driver thingy for the gas solenoid (probably a small MOSFET on a wire) and the welder will be ready to make some sparks (the good kind, i hope).
Maybe not, this connection needs improvement IMO:



Decent lugs are not that expensive and I have on occasion used small diameter copper pipe crimped onto the welder lead with the edge of a nut as the anvil and squashed tight in a vise.
Likewise the lug flat is formed in a vise and then drilled to requirements.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #83 on: July 05, 2019, 10:57:22 pm »
^ to me it looks fine, all the strands were locked in place, the copper plates were cleaned and the M8 bolt was cranked as far as the aluminum would let me.
Perhaps the washer is a little too small but i have mo worries about this connection.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #84 on: July 06, 2019, 11:09:14 am »
It blew up.  :-BROKE
Well, if failure is the best way of learning then i'm in for a whole lot of knowledge.  :palm:
Basically what happened is everything was fine until i blipped the trigger and it instantly blew three IGBT's, two low side and one high side.
Took the opportunity and cracked one apart to see how big the die was and it came in at about 3.5x3.5mm with two bond wires per leg.
Both low side IGBT's completely blew apart, but the one high side IGBT only had a shorted gate.
I'm thinking perhaps the high side IGBT failed first and stayed on then while the two low side IGBT's were turning on the amount of power they had to dissipate was astronomical and they blew up. (that's what my inner Sherlock says anyways)
Now, why would the high side IGBT fail in the first place? Damn voltage spikes i bet.
I'll be remaking the IGBT board to incorporate more protection against voltage spikes and i'll also have to test the remaining IGBT to see if it's still good.
If it's good then i'll still have two remaining IGBT's and will have only one per side, but that should still be enough.
If not i still have some Toshiba MOSFETs that would get the job done as well.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #85 on: July 06, 2019, 11:14:53 am »
It blew up.  :-BROKE
Well, if failure is the best way of learning then I'm in for a whole lot of knowledge.  :palm:

I think failure is the ONLY way to learn, at least it always has been for me :)
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #86 on: July 06, 2019, 11:42:10 am »
Is there no way of gradually ramping the power up and watching the waveforms on a scope?
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #87 on: July 06, 2019, 12:35:15 pm »
Slight change of plan.
I cracked open a failed (but genuine) RJK5010 MOSFET (20A, 500V) and it has a die size of 6x6mm = 36mm2.
Assuming that 20A/36mm2 = 0.55(5)A/mm2 then the IGBTs i had would be 12.25mm2 * 0.55(5)A/mm2 = 6.8A (yikes).
6.8A is a far cry from the advertised 20A so i will not even try to plug these back in but rather i'll go the MOSFET route.
BTW i knew the IGBTs were not genuine but i didn't think they were THIS bad (i expected something like 10A).
My Toshiba MOSFETS, on the other hand, are genuine, i know because i pulled them from a LG plasma TV.
They're 2SK2698 (500V, 15A, 0.35Ω), i had four but used one to replace the blown RJK5010 mentioned above (and fixed a plasma TV  :) ).
So i'll have only one per side but that's not a problem as long as my gate drive is correct.

I also checked the switching transistors from the ATX power supplies i took apart and they're like 10 or 12A rated so in theory using them would have been better than the fake IGBT's!

I actually might make another welder just to test these transistors out, mini spot welder for battery tabs perhaps?
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #88 on: July 06, 2019, 12:37:45 pm »
Is there no way of gradually ramping the power up and watching the waveforms on a scope?
My scope only goes to 50V max (10x probe) but i do have a small variac that i will use for my second test.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline strawberry

  • Super Contributor
  • ***
  • !
  • Posts: 1199
  • Country: lv
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #89 on: July 06, 2019, 02:22:46 pm »
RC snubbers for IGBT and rectifier
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #90 on: July 06, 2019, 04:44:00 pm »
RC snubbers for IGBT and rectifier
I already had those but i assume they were insufficient, now i have a new board with a RC snubber and a 470V MOV right on the MOSFETs.
Also i soldered on much bigger screw terminals because the smaller ones had a problem with stripping the threads or being a pain to fit the wire in.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline profdc9

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: us
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #91 on: July 06, 2019, 06:48:41 pm »
I have made boards for building half and full bridges.  They are intended to be used with dual resonant solid state Tesla coils, but they can work for other IGBT based bridges as well.  For example:

https://github.com/profdc9/DRSSTC-PCB-Pack/tree/master/full-bridge-transistor

The whole project is at

https://github.com/profdc9/DRSSTC-PCB-Pack/

There are spots for the IGBTs, DC link capacitors, snubbers, TVS diodes, gate drive transformer inputs, etc. to help minimize the inductance of the layout.

Dan
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #92 on: July 06, 2019, 09:56:58 pm »
I tried running the inverter on 12V with my new MOSFETs and was getting completely nothing.
Turns out at least one secondary diode is shorted, could this be why the IGBTs blew?
Well as they say when it rains it pours and i happen to be absolutely drowning right now (in crap diodes and IGBTs that is).
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #93 on: July 06, 2019, 10:33:57 pm »
Current mode control won't save you from counterfeit transistors*, but it will save you from failed diodes and shorted loads.  I don't recall a control scheme being posted in this thread so I'm reiterating this.

*If you have to ask, they're counterfeit.  If they're too cheap to believe, they're counterfeit.  If they're from eBay or Ali or Amazon, they're probably counterfeit.  It's very little labor to grind down the face and laser a new label.  Anything from 2N3904s to 2N3055s to power modules and ICs.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #94 on: July 07, 2019, 08:09:52 am »
Current mode control won't save you from counterfeit transistors*, but it will save you from failed diodes and shorted loads.  I don't recall a control scheme being posted in this thread so I'm reiterating this.

*If you have to ask, they're counterfeit.  If they're too cheap to believe, they're counterfeit.  If they're from eBay or Ali or Amazon, they're probably counterfeit.  It's very little labor to grind down the face and laser a new label.  Anything from 2N3904s to 2N3055s to power modules and ICs.

Tim

Actually the one diode that was the closest looking to the original did give different measurements, but i suspect the TO-247 ones are probably some 40V 20A shottky diodes or something along those lines.

I'll pull them apart to see how the die looks.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #95 on: July 07, 2019, 09:05:49 am »
The plot thickens...
Pulled the diodes apart and they're very different inside.
The diode closest resembling the original had a die of 5x7mm and was connected by four bond wires.
But the TO-247 diode was entirely different, it had a die of 6x6mm and was connected with a big copper tab bonded directly to the die.
So in terms of current handling the TO-247 diode should actually be the better choice because it not only has a bigger die (36mm2 vs 35mm2) but also the better connection to the die (copper tab vs four bond wires)
Also for a quick comparison, assuming he bond wires in the RJK5010 MOSFET and the aforementioned diode are the same then then it having four would mean that it can only handle about 80A.

But then again the TO-247 didoes aren't that great if they short out and die before any real current gets to go through them.

Meanwhile i had more of those TO-247 diodes arrive so i might test the reverse breakdown voltage on them.
« Last Edit: July 07, 2019, 09:09:37 am by Refrigerator »
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
Re: Homemade IGBT MIG welder with ATX PSU parts. [almost finished]
« Reply #96 on: July 07, 2019, 10:25:08 am »
Did a quick reverse breakdown test and the voltage is 35V at 135µA and 65V at 400µA.
That makes it a 30V diode, no wonder it blew, in a half bridge configuration this thing would see at least double the output voltage.
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline strawberry

  • Super Contributor
  • ***
  • !
  • Posts: 1199
  • Country: lv
ESAB Arc251i uses current mode cotroller but they do not use this function by some reason. but rather relay on custom 30A semikron module (blows time to time).

sherman and similar chinese use sg3525 or tl494,  2..or more paralel IGBT in bridge configuration. even seen paralel output transformers.

without output inductor. if welding wires tangle/twist inductance will drop below minimum and overload IGBT
 

Offline RefrigeratorTopic starter

  • Super Contributor
  • ***
  • Posts: 1593
  • Country: lt
ESAB Arc251i uses current mode cotroller but they do not use this function by some reason. but rather relay on custom 30A semikron module (blows time to time).

sherman and similar chinese use sg3525 or tl494,  2..or more paralel IGBT in bridge configuration. even seen paralel output transformers.

without output inductor. if welding wires tangle/twist inductance will drop below minimum and overload IGBT

Good thing that i put the inductor in then.

What about the rectifier diodes?
I have a blog at http://brimmingideas.blogspot.com/ . Now less empty than ever before !
An expert of making MOSFETs explode.
 

Offline strawberry

  • Super Contributor
  • ***
  • !
  • Posts: 1199
  • Country: lv
https://elektrotanya.com/esab_caddy_arc-251i_810-xxx.pdf/download.html#dl
rectifier 3pcs dual diode module (not popular brand)
Esab arc 251i only drawback is wimpy plastic case

KEMPPI mastertig 2200
controller protected and filtered I/O
discrete diodes dual TO-247
IGBT module
transformer ~4x U core
big heatsinks

Miller maxstar 200 SD
custom controller (bunch of OP and logic)
single custom integrated module semikron SKIIP (IGBT , 3F rectifier, tyristor, PFC)
DC link voltage 810VDC

Migatronic navigator 3000
controller ML4828IS
rectifier two pcs DSEI 2x101 06
IGBT bridge can be seen in picture under PCB (4pcs)

Sherman MIG 200
controller sg3525 and lot of LM324 and wires
Sherman MIG 261M/4R
no triger circuit protection (control borad VCC overvoltage can happen) , weak funktion switches prone to corosion
bunch of discrete IGBT
 
The following users thanked this post: Refrigerator


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf