1 hour ago, wael_fathe said:

... by replacing all caps in the smps with 1600mf 25vdc....never put1000mf

Those caps can cause so many issues!

The exact capacitance likely doesn't matter much; it will probably be the Effective Series Resistance that will make the difference. Higher value capacitors (thos with more microfarads) usually have lower internal resistance.

18 hours ago, Coulomb said:

Those caps can cause so many issues!

The exact capacitance likely doesn't matter much; it will probably be the Effective Series Resistance that will make the difference. Higher value capacitors (thos with more microfarads) usually have lower internal resistance.

Truly...from now on...i will automatically replace all caps and the 7912 automatically....they solo reaponsible for many faults...they effect the longevity for the repaired device..

Is it possible to increase the power capacity?

On 2025/01/08 at 6:41 AM, Coulomb said:

Interesting. Are these the capacitors in series with the transformer windings? With a diode across them?

i still have inerter with 08

i have the cpu board ...new one

can you indicate me the capacitors in question so that i can change both

cpu board and caps

and get rid of this stubborn error?

see the picture ?

which cap?

@Coulomb

8 hours ago, wael_fathe said:

can you indicate me the capacitors in question so that i can change both

We've been talking about at least two sets of capacitors. For fault code 08 and 51, it seems that the ceramic capacitors in series with the MOSFET gate transformers go low in value. These are 10μF 50 V SMD capacitors; it may be hard to find them in SMD form. They are not through-hole components, so I cant show them to you on your photo. On the PIP-4048MS traced schematic, these are C74 and C31, see this post for other parts that should be checked. The transformers are small, usually near the MOSFET heatsinks. Look for large SMD capacitors near the transistor end of these transformers. Possibly bottom left of your photo. The driver transistors look like tall TO-92 through-hole parts.

This post confirms the value of C31 and C74 as 10μF. It's unusually high for an SMD part.

All these links are findable from the repair and hardware modifications index.

The other capacitors that cause big problems are the power supply capacitors; see this post and scroll down.

Edited March 17, 20251 yr by Coulomb

On 2025/03/11 at 4:52 AM, falahaboamar said:

Is it possible to increase the power capacity?

i installl 2200 instead off 1000 micro faraad

today i solved mysteries charging problem

inverter cant reach to floating of 27vdc

or way exceed 27vdc ...with strange huge battery charging current ...battery that is around float cant pull huge current

when turn grid ac off i noticed that there still charging current suggested by screen

go figure ...charging current with ac input off

i replaced current coil sensor.....no use

i tested lm7912 voltage and the dude give me 8vdc-

big ehhhhaaaaa

i replace all caps and lm7912

and the imaginary charging current went to 0 (with input ac off) and once ac input on the inverter locked at float 27vdc with logical 3 amps of battery current

more and more 7912 and caps failure with multiple symptoms ranging from all kind of measurement problem to igbt random failures

pictures later stay tuned all

@Coulomb

Edited March 25, 20251 yr by wael_fathe

Hi Powerforum and @BritishRacingGreen,

thanks for this repair thread and all the useful information.

I also had to repair an MKSIV inverter and replaced

eight MOSFETs and some of the 22 ohm resistors.

Now the inverter is working again, but not fully.

The second MOSFET string, which was most damaged,

have a strange rectangle switching signal, see attached picture.

Does anyone know where this behavior is coming from?

Which issue or faulty component cause this?

Thanks, regards and vy73

Hans

On 2025/06/04 at 10:42 PM, dc1rjj said:

The second MOSFET string, which was most damaged,

have a strange rectangle switching signal, see attached picture.

I assume that this is at the high voltage end of the DC-DC converter. [ Edit: Sigh. It seems I didn't consider the waveform photo carefully. ]

My wild guess would be some of the push-pull transistors between the SG3525 and the isolating transformers. Or possibly those small MOSFETs that are supposed to turn the DC-DC converter off when not needed.

Edited June 29, 2025Jun 29 by Coulomb

Hello everyone, I have a strange behavior on one inverter I have. There is a high level noise on the power supply line of the upper UCC23513 when I turn on the inverter. If it is turned off and charging through solar panel then the signal is very clean DC signal (~20V).

When I turn on the inverter and starts outputting 220V , a lot of ripple pulses starts showing only this part. Other UCC23513 have a very small ripple which conclude that the problem is from above part.

Sometimes a big ripple occurs and reset the main DSP because it affected by somehow the 5V line (Detected by the DSO).

I have changed every single component of the upper side, even I changed the IGBTs. I don't know where those ripples are coming from.

Any clue ?

I forgot to mention that the output voltage is also very clean sinewave signal

The noise is clearly visible, also you can see the zoomed in form

On 2025/06/08 at 3:32 PM, Coulomb said:

I assume that this is at the high voltage end of the DC-DC converter.

My wild guess would be some of the push-pull transistors between the SG3525 and the isolating transformers. Or possibly those small MOSFETs that are supposed to turn the DC-DC converter off when not needed.

Hi,

tnx for your answer. No, this is not on the high voltage side, but on the 48 V battery side.

I also measured this strange rectangular signal coming from the NSi6602 dual channel MOSFET gate driver chip. The input signal to this chip is ok (but with much noise...)

Was such behavior seen before? What can be done to get the rectangle right?

regards

On 2025/06/20 at 6:07 PM, Ahmad_k said:

I have changed every single component of the upper side, even I changed the IGBTs. I don't know where those ripples are coming from.

I assume that "every single component" did not include the pulse transformer TX7. Having eliminated everything else, I think it has to be that transformer.

Perhaps there is an insulation breakdown, without being a total shorted turn. Or perhaps a dry joint or similar sometimes-open circuit in that winding.

Consider re-soldering the pins of the transformer. I assume that all the other parts have been re-soldered.

A strange one, to be sure.

On 2025/06/24 at 9:26 PM, dc1rjj said:

No, this is not on the high voltage side, but on the 48 V battery side.

Duh, sorry. I don't seem to have considered your waveform photo carefully.

So this is the output of one of the half bridges containing the second string of MOSFETs? With respect to battery minus? And CH2 on battery minus? The amplitude seems too high, at around 75 V P-P.

Is the other set of MOSFETs replaced in the same half-bridge? If not, the other halves should be replaced (i.e. all 16 MOSFETs), as they likely sustained some damage when the first set saw massive over-current.

On 2025/06/24 at 9:26 PM, dc1rjj said:

I also measured this strange rectangular signal coming from the NSi6602 dual channel MOSFET gate driver chip.

What part of the circuit is this in? Sorry, I'm not as familiar with the MKS IV.

On 2025/06/29 at 7:13 AM, Coulomb said:

I assume that "every single component" did not include the pulse transformer TX7. Having eliminated everything else, I think it has to be that transformer.

Perhaps there is an insulation breakdown, without being a total shorted turn. Or perhaps a dry joint or similar sometimes-open circuit in that winding.

Consider re-soldering the pins of the transformer. I assume that all the other parts have been re-soldered.

A strange one, to be sure.

Yes I did, I replaced it from a working inverter and still have same issue. That's a weird, this is the first time I faced this problem.

On 2025/07/02 at 4:03 PM, Ahmad_k said:

Yes I did, I replaced it from a working inverter and still have same issue. That's a weird, this is the first time I faced this problem.

I am also having same problem, one set of MOSFETs on DC side are heating a lot even at no load, don't know why this is happening. If you find the solution please tell us here, also I have not yet checked the signals with oscilloscope.

Also one more thing I want to add is that this solar inverter is having 6 MOSFETs on DC side and I have installed only two MOSFETs on each channel for checking purpose.

Edited July 15, 2025Jul 15 by tanveerhabib

Good day. Newbie on the forum and in inverter repair. I was working with phones and other electronics. And now, in three days, two of my inverters have burned out. Photo of one of them.

This is a sako inverter. a capacitor c138 of unknown value caught fire, presumably shunting high-frequency interference at 350v. from the burnt rectifier diode, two 75g60, and two k100e10e1 on the low side.
I am trying to check the signals by short-circuiting the output of the optocouplers, 350v rises, but I cannot get signals to the sine transistors.
please help, maybe there are ideas on where to go next.

On 2025/07/26 at 8:17 PM, Tihones said:

This is a sako inverter. a capacitor c138 of unknown value caught fire, presumably shunting high-frequency interference at 350v. from the burnt rectifier diode, two 75g60, and two k100e10e1 on the low side.
I am trying to check the signals by short-circuiting the output of the optocouplers, 350v rises, but I cannot get signals to the sine transistors.
please help, maybe there are ideas on where to go next.

Clean all ..remove igbts ..start by primary side replace 10 ohm resistors and bad igbts...this design have 12 fet 8 for inverter operation 4 for reverse conect protect...repalce all bad fets

Go to secondary side ...remove all igbt clean all ...now crtical this inverter igbt turn off at 0 voltage so no 5.6 zeners all zeners in red orange color are a 16vdc zener...with all igbt off and fets are fited u should get error 05 and 370 vdc on big caps ..put ur meter on vdc and test all voltage on zeners u should get 16vdc if this low then ur coplers are shorted u must geet 16vdc on zeners by the way there are only 3 zeners 2 underneath board one is so far near smps...strange it feels 3 zener for 4 igbts but that is. True

Now if u get 16vdc on zeners ..put ur meter on voltage scale and turn on the inverter is all ok u should get 1.5 to 2.5vdc that appear momentarily between gate and emitter for igbt one by one in all igbts ...if that happen put all igbts in solder well pay attention to 10 ohm or 47 oh. R if any open...and u should now get 220ac

I get 2 inverters that tell me deeps truths about repair of thsrse things

One have shorted only 4 fets so i decided to replace ..only bad ones ..turn on the device it work fine however 12 reg is only giving 8vdc and there is crazy measurment problem.like showing once i put it in charge. Mode charging powrr 2k...

Replacing ..12vdc reg still bad..so i decided to replace all the fets and reg. Stablize 11.5vdc..

But soon when turn on 2 igbt dc dc burnned i replaced all dc dc igbtt and thieer zener and set of 2sa1020..and other 2655 and all back to normal....

In second inverter i replaced 2 bad ac. Igbt and the inverter work for while drawing 2.6a from 24vdc ps and fet get shorted...

I replaced all fets and replaced all zeners5.6 18vdc z. And all coplers ...all l44 diodes ..

The inverter work and draws around 1.6 a from sp loaded it 100 watt for hour no issues..but reg ic inly 11.2 vdc

In untshell it seems that structures that about to fail will stress 12vdc and draws high currents from ps

But do u think 11.2 is stil normal @coulomb

Hi all, recently got a faulty Mastervolt Combimaster, which i have found quite similar in architecture to the inverters discussed here, And thanks to the posts I have managed to get it repaired by replacing burnt out High Voltage Capacitors, IGBTs, Gate drivers and their associated zeners. Inverter appears to work fine, however some concern with it getting quite warm when charging in 'float' mode, i assume that is when the battery is full and the unit is trickle charging.

I have a few questions to ask, first of all is whether this is normal behavior, The unit draws around 50W from the mains while in that float charging state with zero load, and gets quite warm, which is suspicious since when its inverting on a larger load of 200-300w it stays cool. Is this a typical consumption for float charging LiFePo batteries (I have 2 12v 300AH batteries). I would have assumed the float charging rate would be pretty low, with occasional jumps to keep the battery level, not a constant 50W

I am suspecting that there might be some inefficiency thats happening? maybe an inefficient charging stage, or something amiss with some circuitry that is only in oparation during AC in mode (Power supply for the logic maybe?)

Im not expecting a straight forward answer to this, other than some of the obvious (float consumption , warming), so i would appreaciate very much if someone could point me an explanation on how these inverters work , I think i have a grasp of the inverting mode of operations, However the charging one is where i would love some insight, i read somewhere that the AC side IGBTs act as the input rectifier using the body diode of the IGBTs? is this true? i would have assumed they would be actively switched? also does this result in lower efficiency? Also how does the buck converter section work during charging mode?

On 2025/07/29 at 9:38 PM, wael_fathe said:

But do u think 11.2 is still normal @coulomb

No, that seems too low to me, assuming that the voltmeter can be trusted to within one percent or so. The voltage of the 12 V rail is determined by the width of the pulses in the main power supply. I would suspect parts around the 3-terminal voltage reference and the associated opto-coupler.

But of course it could also be that something is still drawing too much current from the 12 V supply, and that's overloading it. Those sorts of faults can be a nightmare to fix.

15 hours ago, soulfreeze said:

an explanation on how these inverters work , I think i have a grasp of the inverting mode of operations, However the charging one is where i would love some insight,

The AC full bridge (four IGBTs generating L and N outputs) are bidirectional. They can push or pull power depending on the phase of the sine wave with respect to the utility's, and push or pull VARs (reactive power) depending on the amplitude of the sine wave compared to the utility's. So for utility battery charging, or indeed any charging with these models, the firmware manipulates the phase to cause the power flow direction and amplitude (power amplitude, number of watts) required.

But that power ends up in the bus capacitors, since that's at the other end of the full bridge (a DC <-> AC converter). If nothing loads the bus capacitors, they will quickly overvoltage, and other parts of the inverter have to step in and trip everything. But before that happens, the DC-DC converter, which is also bidirectional. pushes power towards the battery, keeping the bus voltage withing reasonable bounds.

But the DC-DC converter has a fixed voltage ratio, set by the transformer turns ratio. Typically this is 8:1 or 7:1, though more recent models use something like 7.2:1. And the bus voltage is being buffeted not only my the full bridge running in reverse, but also the PV boost converter(s). The bus voltage can't be more than a volt or so lower than the present PV voltage. So the bus voltage is fluctuating. How to regulate the battery charge current in this situation? That's what the buck converter is for.

When power flow is from the bus to the load, this converter is "straight through", i.e. the DC-DC high-side voltage is very near the bus voltage. But when the power flow is the other way, i.e. towards the battery, i.e. charging the battery, the buck converter is controlled to regulate the charge current. But the buck converter can only buck, never boost, so the transformed battery voltage always has to be less than the bus voltage. In other words, the bus voltage mustn't be allowed to fall too low. This can be done by fine tuning the exact power demanded from the full bridge. You can see that the inverter firmware is juggling a lot of balls at once.