Powermax 1000 dead IGBT Q7 advice + help identifying parts needed

[e01f]

This is my first post, so hey everyone!

I got a defective powermax 1000 to repair and need some help identifying one or two odd components that are too far gone to be identified.

My background: hobbyist welder and EE (still in education, 6th year) from Germany

Tools&Resources available: Multimeter, LCR-Meter, Soldering Iron, Oscilloscope, Lab Power Supply, ...

The only description I initially got is "it broke, won't turn on" - so I took her apart and it looks like I got myself a burnt IBGT Q7. - Please see below photos.

Everything apart from the damage seems to be in good condition, seems like the unit didn't get that much use. Was connected for 3x400V (32A breaker) in Germany.












It seems to be very similar to the damage in this thread, the damage is just a little more aggressive on this unit. And this seems to be a "Gen 2" PCB.

I managed to clean up all the charring and made a list of all evident damage:

Designator	State	Type	Value/Label	Characteristics
Q7	1/2 defective	IGBT	2MBI200S-120	Dual 1200V 200A, 24nF, #2 configured G-E always-on
Q6	possible damage, needs more testing	IGBT	2MBI100PC-140	Dual 1400V 100A, 10nF
L2	functional, light damage	Inductor	n/A	n/A
R98	burnt, arc damage	Resistor	Brown-Black-Gold-Gold	1 Ohm 5%
R99	burnt, arc damage	Resistor	Red-Black-Black-Yellow-Brown	2M Ohm 1%
R100	burnt, arc damage	Resistor	Red-Black-Black-Yellow-Brown	2M Ohm 1%
R101	possible damage	Resistor	Red-Black-Black-Yellow-Brown	2M Ohm 1%
R66	burnt	Resistor	Brown-Black-Black-Gold	10 Ohm 5%
R67	burnt	Resistor	Brown-Black-Black-Gold	10 Ohm 5%
R28	burnt	Resistor	Red-Violet-Black-Gold	27 Ohm 5%
R20	burnt, arc damage	Resistor	Violet-Green-Silver-Gold	0.75 Ohm 5%
R??	burnt, arc damage	Resistor	Yellow?/Gold?-?-Black-Brown OR Brown-Black-?-Yellow?/Gold?
Q2	case exploded, 0.3 Ohm short on pin 2-3, open circuit on 1-2, 1-3	Transistor/BJT?	…12 // -140

My plan is to first tackle all the visible damage and then function-check all the components in the PFC section for invisible damage and replace if necessary.

This machine has quite some power to it, so I am going to take precautions and double-check things. If something goes electrically wrong at this kind of power, it makes for a very bad day.

For a start, I have ordered a new PFC IGBT (was around 45€) and may order a new Q6 IGBT next as some have suggested in the old thread. All the other components in the PFC section look to be standard off-the-shelf parts that can easily be sourced if only I could identify them.

I have also checked the input and output diodes and all other IGBTs and have found no further damage there.

It would be very nice if someone knew what Q2 (last image) or the anonymous resistor (see red circle in photo) was or could provide a photo of it in non-defective state or give any further advice.

I will happily post my solution to this damage (if I get her up and running again, that is) in exchange.

[ronsii]

I'm guessing you already found this thread... https://weldingweb.com/vbb/showthread.ph...sehold-breaker

[e01f]

Yes I read that one, but thanks anyways. Unfortunately, neither Q2 nor the badly burnt resistor (red circle) can be identified from the photos in that thread. Was hoping someone with experience in these units or the same model of plasma cutter is still around and could have a quick look.

[ronsii]

Yeah, I know I looked though the pics.... you might try PM'ing the op on that one... haven't seen him around but you never know... Quite a few guys have the 1000 so maybe someone will pop in with help.

[e01f]

A quick update: The OP replied but unfortunately has no more info on the unit since he got rid of it back then.

I also found this: https://www.cnczone.com/forums/hyper...nyone-can.html and sent a PM to the OP there.

In the meantime, I had a quick look at the circuit and it seems actually two sections have failed:

Section 1 - Flyback System Power Supply

The blown Q2 seems to have been the main switcher for the nearby transformer T1 which then generates all the "harmless" logic supply voltage rails for the rest of the board (it's basically more or less a standard SMPS). It operates off the intermediate voltage rail, V_BUS.

Section 2 - PFC/Intermediate Voltage Rail (V_BUS)

The IGBT Q7 is used as a switcher and diode combination (thats why E1 and G1 of the dual package have been shorted during manufacturing, doing this gets you a powerful, relatively low loss diode) and together with the massive choke in the back right section of the unit forms a boost converter that is generating another (very high) voltage rail labeled V_BUS. V_BUS is 750VDC +/-15% according to the service manual and is backed by the two beefy capacitors. It is used as the intermediate voltage for the output inverter and other parts of the system, the Flyback System Power Supply, Section 1.

Initial thoughts

As to what happened to my unit, I summarize:

That a powerful arc that has formed between the E1 terminal (which hardwired to the rectified mains negative) and the rectified mains positive present on J9 during operation some time during the failure.

That a (smaller) arc has also formed between the rectified mains positive (J9) and V_BUS positive terminal, terminal C1, of the IGBT Q7.

That the Q7 IGBT failed. The E1-C1 IGBT (the one configured as a diode) is definitely damaged, tests with the second IGBT in the package, C2E2 have been inconclusive to this point. The arc may have been the reason or a symptom of Q7 failing, I have not come to a conclusion about this yet.

That the failure also propagated through R98 into the drive circuitry.

That the failure also propagated through the gate/emitter connector J8 into the drive circuitry.

This means that the driver ICs in the section may be damaged as well.


I have yet to find a plausible explanation for all this to happen. V_BUS would not have far exceeded the allowed limits without the protection tripping and preventing the unit from damage.
Maybe there was EMI present, maybe a massive voltage excursion / HV transient on the mains made its way through the input filters?
There could have also been conductive dust present.

[e01f]

Good News!

Today I started off by desoldering all the dead components from the board and putting them in individual labeled boxes for further reference. Then came the hardest part: identifying the unknown components.

IGBT Q2

After putting some more thought into the components, I took some dimensions and finally made a guess as to what Q2 must have been. Measuring its physical size with a caliper, we have a TO3PF case. This case is typical for MOSFETs and also IGBTs. So with the bits of part number that were left ant knowing that it runs off the intermediate voltage of 750V, my guess was it could have been a 1400V 12A MOSFET or IGBT.

I did a bit of research online and after a couple of tries, a magic part number popped up: 1MB12-140, a 1400V 12A IGBT made by Fuji Electric Co., the same MFG as with the Q7 IGBT!
All available photos matched, and also the little "B" letter at the bottom left corner and the mold marks align.

To try and confirm this some more, I then googled "1MB12-140 Hypertherm" and found this thread: https://tractorpoint.com/forum/weldi...erm-600--.html at tractorpoint. Seems that same device is also used in the Powermax 600. Makes perfect sense to me for Hypertherm to reuse their logic power supply in all devices they make... why reinvent the wheel every time. So yeah, dead part #1 identified!

A quick search at digikey returned no results for the 1MB12-140 and all available listings on eBay or other places show ridiculous prices of 30-40€ for this part (I would say it is worth maybe 5-10€), but I am sure that I will be able to replace it with something much better (in terms of specs and price) if required with not much effort.

Anonymous burnt resistor = R30

So after sucking off this one, it revealed the label underneath: R30. I put what was left of it under the microscope and identified: Yellow - UNKOWN (maybe Green or Gray) - UNKNOWN (maybe Gray) - Black - Brown which makes this a 4xx Ohm, 1% resistor. The xx is yet to be figured out.

After further investigation, I found that R30 is used in the over current protection circuitry of the DC-DC driver IC U5 (marking FA13844). It forms a low pass filter for the current feedback signal derived from the shunt resistor R20 (0.75 Ohm) together with C36 (2.2nF).

This is very good news! This means that the exact value of R30 does not matter all that much as long as it is in the right ball park (of 4xx Ohm). I had a look at the datasheet for the FA13844 and the reference design mentioned there uses 1k Ohm and a 100pF capacitor which gives about 1.6MHz cut-off frequency. Hypertherm used 2.2nF for C36 and with say 470 Ohm, this would give 154kHz cut-off frequency. According to the datasheet, U5 can operate from 10..500kHz so it all comes down to what frequency U5 is set to operate the boost converter at. I will have a look at the circuit around U5 to determine the design frequency and select R30 accordingly, but in summary 470 Ohm should be just fine.

Next step is obviously to check for invisible damage in the whole PFC and supply section of which I am sure there will be some and then place some orders for parts. My replacement IGBT for Q7 has already arrived. So far, I would estimate parts will be about 50-60€ if I find no further major damage.

Some over at the other Powermax 1000 thread at weldingweb have suggested to also replace the inverter IGBT, so I guess this would be something to double-check (or maybe just replace it anyways as a precaution).

Will keep you posted!

[Insaneride]

The part number for Q2 maybee facing the heat sink. Unscrew it and check.

The resistors are precision so they use fourth band multiplier.

[ccawgc]

In my experience, When L2 , and R98 blow. There has been a failure in the PFC (boost ) circuit.
The failure dominoes all across the pcb.
The repair has always been both IGBT blocks and the power pcb. Replace them.
In the beginning People tried to just replace the PFC IGBT and the circuit pcb, only to have the main inverter IGBT blow.
Even if it tested good with an IGBT tester. When I saw this failure, all most all of them had single phase 230 volt plugs on them.
Your power pcb should have tubular resisters all around the large buss caps.
Good luck on your repair attempt. Be prepared for it to blow again.

[e01f]

Good news and perhaps my final post on the Powermax!

The summary is I checked 45 or so and replaced 20 or so components on the PCB. I just made a couple of test cuts in 8mm flat stock at 50A setting and she seems to work just fine. Will have to do more extensive testing to confirm everything holds but still.



Due to the extent of my testing I will not comment on it in detail but only summarize.

Below you find a table of each and every component I checked and/or replaced and some small comments:

BOLD RED means the component was replaced because it was defective

BOLD BLACK means the component was replaced but for other reasons than a defect

NORMAL FONT means the component was checked or otherwise found to be intact

Designator	State/Comment	Type	Value/Label	Characteristics	Replacement	Cost
Q7	1/2 defective	IGBT	2MBI200S-120	Dual 1200V 200A, 24nF, #2 configured always-on	2MBI200S-120-50	45,00 €
Q6	likely damaged, driver U8 is blown	IGBT	2MBI100PC-140	Dual 1400V 100A, 10nF	2MBI100PC-140	161,64 €
L2	functional, light damage	Inductor	n/A	n/A
R98	burnt, arc damage	Resistor	Brown-Black-Gold-Gold	1 Ohm 5% 1W	ROX1SJ1R0	0,31 €
R99	burnt, arc damage	Resistor	Red-Black-Black-Yellow-Brown	2M Ohm 1% 2W	2x ROX2SJ1M0	0,48 €
R100	burnt, arc damage	Resistor	Red-Black-Black-Yellow-Brown	2M Ohm 1% 2W	2x ROX2SJ1M0	0,48 €
R101	possible damage	Resistor	Red-Black-Black-Yellow-Brown	2M Ohm 1% 2W	2x ROX2SJ1M0	0,48 €
R66	burnt	Resistor	Brown-Black-Black-Gold	10 Ohm 5% 1W	ROX1SJ10R	0,31 €
R67	burnt	Resistor	Brown-Black-Black-Gold	10 Ohm 5% 1W	ROX1SJ10R	0,31 €
R28	burnt	Resistor	Red-Violet-Black-Gold	27 Ohm 5% 0.5W	ROX05SJ27R‎	0,17 €
R20	burnt, arc damage	Resistor	Violet-Green-Silver-Gold	0.75 Ohm 5% 1W	RR01JR75TB	0,28 €
R30	burnt, arc damage	Resistor	Yellow-Green??/Gray??-Gray?-Black-Brown	4xx Ohm 1% 1/2W (Meas 220k?)	ROX05SJ470R	0,17 €
Q2	case exploded, 0.3 Ohm short on pin 2-3, oc on 1-2, 1-3, replacing with similar type IXYH10N170C	IGBT	1MB12-140	1400V 12A, 2nF, configuration GCE, TO3PF	IXYH10N170C‎	7,29 €
U5	case cracked, original unavailable, using similar UC3844N as replacement, requires adjusting R27 and C16	IC	FA13844 D3701	CMOS current mode control IC for DC-DC conv.	UC3844N	1,40 €
R27	out-of-circuit test: ok, needs to be changed due to U5 replacement	Resistor	Red-Black-Black-Yellow-Brown	2M Ohm 1% 2W	680k	0,48 €
C16	out-of-circuit test: ok (11.9 uF, ESR: 1.9Ohms), needs to be changed due to U5 replacement	Capacitor	12µF 63V 105°C LXY	12µF, 63V, 105°C, Nippon LXY series	100µF 63V 105°C Nippon LXZ	0,50 €
C36	looks ok, double check!	Capacitor	2n2J100	2.2nF 5%, 100V? Plastic
Q3	looks ok, tests ok, replacing as a precaution	BJT	D44H8	NPN 60V, 10A	D44H8	0,90 €
Q4	looks ok, tests ok, replacing as a precaution	BJT	D45H11	PNP -80V, 10A	D45H11	1,00 €
U10	in-circuit test: suspicious, out-of-circuit test: ok	IC	A3120V	2.5A peak IGBT-BJT Output Optocoupler 630V	HCNW3120	4,06 €
U8	in-circuit test: suspicious, out-of-circuit test: output short ground	IC	HCNW3120	2.5A peak IGBT-BJT Output Optocoupler 1414V	HCNW3120	4,06 €
R84	in-circuit test: suspicious, out-of-circuit test: failed open	Resistor	Brown-Black-Black-Gold	10 Ohm 5% 1W	ROX1SJ10R	0,31 €
R83	in-circuit test: ok	Resistor	Green-Blue-Gold-Gold	5.6 Ohm 5%
R??	in-circuit test: ok	Resistor	Brown-Black-Red-Gold	1k Ohm 5%
R96	in-circuit test: ok	Resistor	Brown-Black-Gold-Gold	1 Ohm 5%
C53	out-of-circuit test: ok	Capacitor	105 C5K	1µF
U11	out-of-circuit test: ok, replacing as a precaution	IC	TL780-05C	1.5A +5V POSITIVE VOLTAGE REGULATOR	L7805CV
D15	in-circuit test: ok	TVS Diode	1N6280A	24V 45A suppressor, 1.5kW capable
D17	in-circuit test: ok	TVS Diode	1N62xxx
R9	in-circuit test: ok	Resistor	Brown-Black-Red-Gold	1k Ohm 5%
D2	in-circuit test: ok	Diode	MUR460	4A 600V rectifier diode
D3	in-circuit test: ok	Diode	MUR460	4A 600V rectifier diode
D11	in-circuit test: ok	Zener Diode	1N5226	3.3V 500mW
R68	in-circuit test: ok	Resistor	Brown-Black-Black-Gold	10 Ohm 5% 1W
R87	in-circuit test: ok	Resistor	Brown-Black-Red-Gold	1k Ohm 5%
R56	in-circuit test: ok	Resistor	3.9Ohm J	3.9 Ohm 5%
C10	out-of-circuit test: ok	Capacitor	.068K1600	68nF 1600V MKP
D6	in-circuit test: ok	Zener Diode	BYV26G	Avalanche Diode, 1400V 1.05A
R19	in-circuit test: ok	Resistor	Brown-Green-Orange-Gold	15k Ohm 5%
C23	test skipped	Capacitor	.1 400	100nF 400V MKP
D30	in-circuit test: ok	Zener Diode	BYV26G	Avalanche Diode, 1400V 1.05A
C36	test skipped	Capacitor	2n2J100	2.2nF 5% 100V
U9	in-circuit test: ok	IC	HCNW3120	2.5A peak IGBT-BJT Output Optocoupler 1414V
U6	in-circuit test: VAO/VSENSE feedback voltage wrong, out-of-circuit test: VSENSE pin abnormal impedance	IC	UC3854BN	Advanced High-Power Factor Preregulator	UC3854BN	4,41 €
D12	out-of-circuit test: ok	Diode	UF…

While checking, I drew some mockup schematics of the sections of the PCB that had damage in them to better understand. I would be willing to post them here but am unsure if this is ok with Hypertherm.

Due to the unavailability of high cost of some components I substituted some suitable replacements. This in turn meant that some parts of the circuit had to be redesigned, especially with U5 where the startup-resistor value was lowered and the buffer capacitor needed to be of higher value to prevent erratic startups. This in turn means that the unit startup time now is about five seconds.

R30 was replaced by a 470 Ohm resistor as described before.

All parts except the two main IGBTs are available from Digikey. For the inverter and PFC IGBT I sourced the original part since I could not find a replacement with the same mounting post dimensions and suitable characteristics on digikey. This should however not make that much of a price difference and saves me the hassle of possibly having to redesign the drive circuitry to accomodate for the new parts.

After all this was done I did a quick power up and then carefully re-coated the entire board with spray-can type 85kV/mm dielectric strength conformal coating to prevent any metal dust from shorting out things in the future.

There was also a deep cut in the work lead so I cut off a small section of it and reterminated it using a 10-16mm2 cable shoe.

I then performed all tests from the official Hypertem Maintenance Manual and to my surprise it seems all have passed.

One final word of warning: please do not attempt this repair if you are not sure what you are doing. There is 750Vdc buffered by capacitors of substantial power as well as rectified mains present in the unit, both can easily kill or severely injure you!

Thats it so far. Thank you for your support and I will report back in a while if it held up!