China ACDC TIG, Mitech 160 malfunction

[junkweld]

I took a look at the transformers in my machine. Some have the winding numbers and some don't. The 220Vac is rectified using module rectifiers to create an output DC. Then it is chopped using Toshiba K2698 MOSFETS and a SG3525 type PWM controller to create a 100KHz AC output. That is fed to some Sanken FML 33S diodes to create an output DC. That is fed to an H-bridge to create the weld voltage and frequency using IRFP264 MOSFETS and a controller board. This machine only does TIG and MMA. No plasma cutting. I think the first stage of AC created is around 100KHz because the transformers look to be too small for 20KHz.

On the front of the machine there is a pot labeled "5 Hz - 2.2 KHz". I am guessing that it does adjustable output AC frequency. It doesn't have the peak and basic current knobs so I don't think it is for pulsing. It only has one current knob.

The HF start board has a transformer, small coil, high voltage caps, and copper contacts for the spark gap. Yes, the copper points will wear and need adjusting. I am surprised none of the manuals talk about adjusting them.

One of the structure plastic supports holding the large alum. heatsinks was broken. I am fixing that when I put it back together. The parts won't be here until Monday though to fix the PCB with the FML 33S diodes on it.

EDIT: Changed 20KHz to 100KHz

[junkweld]

Update:

Installed the new 1.5K (1500 watt) transient voltage supressor (TVS) and two small diodes after work yesterday. I ohmed out the PCB output terminals once they were installed and they read a diode drop of voltage as they should. I installed the PCB on the alum. heatsink and connected all the wires. I had labeled the wires well when I took it apart - *VERY IMPORTANT*. I double checked all the connections and ohmed out the output of the PCB. It was still good (not shorted).

Applied 220vac and it came up properly without the OC LED on. Everything seemed OK. I stuck a 6013 rod I had laying around and struck an arc. It welded properly. I was happy. I also have a set of spare diodes and TVS I will tape inside the machine. Overall, it was $2 worth of parts and $7 to ship them. Test equipment was just a simple continuity/ohm meter.

I will try to make a little diagram of where the parts were and label a pic. If you look at the Super Series Manul at: http://www.everlastgenerators.com/manuals.php
Go to page 10, You will see a transformer that is connected to two diodes and then two inductors. The diodes in the pic are the LARGE output welding diodes........not two small diodes I am about to mention. The diagram in that manual is the same as the Mitech internals. On my machine, after the two inductors, each inductor has one small diode (400v,1amp) connected to the inductor at the anode and the cathode is connected to the single 1.5K bi-dir TVS. The other side of the 1.5K TVS is connected to the center tap of the transformer.
The two small diodes and TVS are there to clamp any high voltage transients. The parts in my machine were shorted so the machine would not come up properly because it detected this short.

[junkweld]

Two pdf files:
One with a simple schematic showing the fault
One showing a low res pic (limit on forum is 100Kbytes) of the fault area.

[Jeffw5555]

nice write-ups on your diagnosis. I wish others would do this on their US made machines. I bought my Syncrowave 350 really cheap because it had a "blown" main board, and the previous owner's local dealer told him it was $1,000 for a new board. I used to design and repair various electronic gear for a living, now I just do it for fun.

The main control board had a shorted tantalum cap. When it shorted, unfortunately it took out a trace on the PCB along with several resistors. I simply jumpered over the blown trace, and replaced the cap and resistors for about $2 total. Took about 2 hours all said and done to fix it, including diagnosis. I also replaced the bearings in the cooling fan because they were noisy.

The welder works perfectly now.

Oh and BTW, Sundown is often an argumentative troll. Please ignore him when he rants about stuff that he finds uninteresting that others like to hear.

[Jake_SS]

Good work.
Usually Mosfets operate at a much higher frequency than IGBTs, but IGBTs are more reliable.

[junkweld]

nice write-ups on your diagnosis. I wish others would do this on their US made machines. I bought my Syncrowave 350 really cheap because it had a "blown" main board, and the previous owner's local dealer told him it was $1,000 for a new board. I used to design and repair various electronic gear for a living, now I just do it for fun.

The main control board had a shorted tantalum cap. When it shorted, unfortunately it took out a trace on the PCB along with several resistors. I simply jumpered over the blown trace, and replaced the cap and resistors for about $2 total. Took about 2 hours all said and done to fix it, including diagnosis. I also replaced the bearings in the cooling fan because they were noisy.

The welder works perfectly now.

I like fixing up stuff that most people throw away because they think it is too expensive to repair. You can get a lot of nice stuff cheap.

[junkweld]

Good work.
Usually Mosfets operate at a much higher frequency than IGBTs, but IGBTs are more reliable.

Yes, the MOSFETs usually fail quicker. Mouser.com carries the MOSFETs. They are part number 2SK2698 (500v, 15amp, .35 ohm) made by Toshiba. The chopper uses 16 of them. They are $2.57 each. So for about $50 the machine can be fixed if they ever blow.

The 12 DC rectifier diodes on the lower PCB are about $2-$3 each. The H-bridge FETs are IRFP264 (or IRFP 260, I forget) at about $4.50 each to do the AC weld freq. output.
Overall, if any of the power components go it is a pretty cheap fix for a 160 amp AC/DC TIG machine.

[Ira]

Thanks Junkweld, I am working on a Mitech right now....it is throwing HV arcs all over the pulser board behind the front panel and your information has been very interesting.
Mine is almost the same as yours although it is a 'Super160' it has two large relays where you pointed to the blown diodes.
HV gen looks like one from a CRT tv set and is very simple.

I need to decouple the H blocks and check out the LHS boards to find out if there are any issues that5 could cause the HV getting back into the pusing controller......maybe protection caps, movs damaged...Dunno yet.
Thanks again.

[epjohnso]

Good story. While my welders are working well, it's nice to know that there are some out there who will take the time to talk about fixing and repairing welders or any electronic gear.
I don't really understand some of the logic, but I'm sure if I had someone who'd sit down with me and show me some of the tricks and isolations your doing, it would be a great lesson.

[Ira]

Good story. While my welders are working well, it's nice to know that there are some out there who will take the time to talk about fixing and repairing welders or any electronic gear.
I don't really understand some of the logic, but I'm sure if I had someone who'd sit down with me and show me some of the tricks and isolations your doing, it would be a great lesson.

Hi, Yes its a little easier for me because my background is in electronics anyhow.
In the Mitech super 160 there are different sections. As Junkweld says, the left hand side and is 220-240v rectifiers and a regulation system to bring the volts down to 60 volts DC. Then if you want AC, there is a system that converts it back using a switching system and some waveshaping.

I have a problem getting that AC output and I think it is generated from the pulser board behind the front panel, then fed to the wave-shaping circuit so that the 60v can be used in AC mode. Because I had some of the start voltage which is very high, on the pulse board, I think some semiconductors have been damaged. I will change some of them and see......wish I had a circuit diagram. So-called 'Mitech' in China has not assisted me yet......not holding my breath.

I still don't know how I got this arcing effect on the pulse board but I hope to find out soon.

There are some who think we are nuts to play with these things but some of us just can't justify going and spending 2000/3000 dollars Australian for US made goods...even if they are better made.

I got this one really cheap and was using a home made one up until now....I made my own tig welder in 2008.....HV start and all, but only DC operation.

Cheers, Ira

[vicegrip]

Another old thread dug up somehow.
And I'm glad too, I missed it when it was currant.
==============
Great trouble-shooting mind-set !!
Of cource I have no clue what most of it means.
But the detective work is the same as mine on mechanical canundrums.
vg

[Ira]

Ummmm, Junkweld started this one some time ago...it was only when I did a search on the net to find out about the welder I got that I came across it.

All cred to Junkweld for the input up to now.

Hello Junkweld.....

Cheers, Ira

[Ira]

OK. So here is the thing. For anyone who wants to know or anyone who cares.
Never buy an inverter style welder!
Now that I have looked at this in some depth, I cannot understand the stupidity in trying to get away with it for any length of time........Unless your design is absolutely bullet proof and you use only the best components, over engineering the s**t out of it from the get go.

Reason: Take 240v, (Australia) rectify it and drop it to 60v (DC) with the ability to draw 160Amps. This means disposing of 160v at 160 Amps....25000 watts.

No wonder they need two fans.....what they do is use a 'pulse width modulator'
just a fancy way to say they switch it on and off a few times per second and then just add up the ON bits and store it. (integration) so that the average voltage is now much less.
But switching 240v DC on and off at 160 amps takes quite a bit so they use 12 mosfets on a heatsink with a whole lot of filtering and smoothing just to get 60v DC at 160A.

****! a transformer is so much easier and more reliable. My old stick welder is 60v at 160A and it gets a bit hot but is dead reliable......all those little plug packs that power mobile-phones, intercoms, calculators, all run on the switch mode design instead of wire transformers...they have decided that the transformer is old fashioned and too environmentally unacceptable.......
Bl@@dy Greenies!

But I am having a rage....I will just pull this to bits for the parts and go back to my home made DC only TIG that cost me peanuts to build and goes like S**t out of a Shotgun!

Stay away from Chineese inverter cr@p! it is too complicated and unreliable QED!

Cheers.

[B_C]

Man I wish I had some PIZZA.......

[wmgeorge]

K. So here is the thing. For anyone who wants to know or anyone who cares.
Never buy an inverter style welder!
Now that I have looked at this in some depth, I cannot understand the stupidity in trying to get away with it for any length of time........Unless your design is absolutely bullet proof and you use only the best components, over engineering the s**t out of it from the get go.

I'd say never buy a Chinese inverter to be correct. The US made ones as the Miller Maxstar I own, seem to be reliable. If you check over the Home Shop Machinist board a gentleman there repairs welders... he rarely has an inverter in the shop.

I'm sure when the Chinese buy a Miller and then try to reverse engineer and copy it they will have problems.