I have 3 units of VMII Revo kind of inverter, I have tried to repair it to some extent but still showing fault 04. Please I need assistance from the gurus on the platform. Thank you allg

2 hours ago, Abbas Olatayo said:

VMII Revo kind of inverter

These appear to be Axpert VM II work-alikes, but with their own firmware. Fault code 04 seems to be bus soft start fail, presumably similar to the Axpert fault code 09. I have no idea if they changed the circuit slightly or considerably.

It's likely that there are at least one pair of IGBTs failed, shorting the DC bus. In Axperts, this often this causes half or all MOSFETs to blow as well. You have to be careful to test and repair all the gate drivers before testing with a current limited power supply.

See the AEVA Axpert repair and modification thread; it may be of help, or it might be just too different inside.

20 hours ago, Coulomb said:

These appear to be Axpert VM II work-alikes, but with their own firmware. Fault code 04 seems to be bus soft start fail, presumably similar to the Axpert fault code 09. I have no idea if they changed the circuit slightly or considerably.

It's likely that there are at least one pair of IGBTs failed, shorting the DC bus. In Axperts, this often this causes half or all MOSFETs to blow as well. You have to be careful to test and repair all the gate drivers before testing with a current limited power supply.

See the AEVA Axpert repair and modification thread; it may be of help, or it might be just too different inside.

Thanks for the wonderful support, it is bus soft start fail which is 09 on axpert inverters and it also shows on the manual. The MOSFETs have already blown. 

Hi, I also have a Revo VM II 5.5kW and I have a problem with the Igbt transistors that keep going out. There are four dc/ac outputs. I checked almost everything. The first two have a different sequence of wake-up signals and I don't know if it's OK. Can you confirm the correctness of the difference between these signals? Or measure these signals on a functional converter? Thanks.

12 minutes ago, nogers said:

Can you confirm the correctness of the difference between these signals?

I can't confirm. I can't see the timebase on those tiny photos. I'll assume that it's 10 ms per division.

It is normal for one pair of IGBTs to switch at 50 Hz; that's the pair that drives the neutral output. The pair that drives the active is PWMd to generate the sine wave. So that likely explains most of the gross differences.

I don't know how exactly you are measuring these signals, but the slope on the neutral output pair is concerning. But that could be a DSO earthing issue.

I am measuring igbt transistors G-E and the collector is connected via a 230v40w bulb to protect the transistors.  Furthermore, the 20 micro F capacitor at the output is also removed to allow the load.  It is clear that the output will not be a pure sine wave, but it should not affect the input control signal.  The only thing I can think of is the quality of the new transistors.  They are from Aliexpress.  Could they be of poor quality and causing the problem of burning and shorting the Igbt?

4 hours ago, nogers said:

I am measuring igbt transistors G-E

It's tricky measuring these, as the potential of the emitters can be very high compared to earth.

4 hours ago, nogers said:

The only thing I can think of is the quality of the new transistors.

I assume you're talking about the driver transistors. It's hard to imagine that they aren't up to the job, unless their Vce rating is very low. Is the power supply to the gate driver circuitry perhaps sagging? That could cause the IGBTs (which are also transistors) to die, because the gate turn-off voltage is not enough to hold one IGBT off when the other is on.  But if your scope readings are correct, it's getting as much negative drive as positive. It should get about +15 V for turn-on, and -5 V for turn off. Something seems totally wrong with your gate drive.

Perhaps check the voltages statically (bench power supply for ±12V, no bus voltage, no control card, no load), and individually short the driver inputs to control card ground to check the IGBT gate to emitter voltages with a multimeter.

I know how the first problem with the converter arose.  It was active over the network at night, and when replacing the fuses with another, better type, I connected the input cable of the panels.  Panels de-energized and disconnected.  I checked all excitation transistors as well as voltage +12, -12v, 5v.  Everything seems fine.  I replaced only 4 Igbt transistors.  The only other thing I can think of is some connection between the low voltage and the high voltage for the Igbt.

4 hours ago, nogers said:

when replacing the fuses with another, better type, I connected the input cable of the panels.

If you accidentally shorted one of the PV inputs to earth, that shorts BUS- to earth, which connects one of the IGBTs across the DC bus.

So measured. Twice -5.5v/ +7.7v QB2,QA1 and twice 0/ 12.1v QD2,QC1. R21, ZD4, C55 is not originally installed on the board. Strange. According to this wiring it works correctly. If they were to function the same, these components would have to be supplemented. Igbt transistors are FGA60N65 in my case. No error found.

12 hours ago, Coulomb said:

If you accidentally shorted one of the PV inputs to earth, that shorts BUS- to earth, which connects one of the IGBTs across the DC bus.

Ouch, something very important to remember 👌

17 hours ago, nogers said:

R21, ZD4, C55 is not originally installed on the board. Strange.

That's possibly the reason for the slopes on your DSO signals.

Does it look like they were never installed from the factory? Or could they have been blown off, or taken off by someone else, or ... ?

These parts were always missing. It's so original. I am the first to disassemble the converter. It was working normally and properly before. It's mine and it worked for about a month until I connected the ground and input from the panels. Is there a pattern of the correct signal (picture)?

4 hours ago, nogers said:

Is there a pattern of the correct signal (picture)?

I'm not aware of one. It's actually difficult to capture; usually you need two inputs and the "subtract" facility on your DSO. I don't even know how you're getting yours; maybe you should explain, as the signal may be exhibiting artefacts from the connection.

The transistors are now tentatively just grabbed from the other side and connected via a light bulb for protection. I measure directly with a hand-held oscilloscope. See photo.

3 hours ago, nogers said:

I measure directly with a hand-held oscilloscope.

Ah, OK. If you operate that inverter with real bus voltage, keep your hands well away from the scope, as its reference (negative) input will have high voltage square waves with respect to earth, and it could bite.

I know the dangers very well, I have been dealing with electronics for a very long time and I am an electrician by profession. However, it doesn't solve my problem. do you think that if I complete the components according to the channel with 5-a+7v, the problem with excitation will be eliminated? I can basically also use an external source for the bus voltage with limited current for safe testing. I have a 30v 10A power supply.

8 hours ago, nogers said:

do you think that if I complete the components according to the channel with 5-a+7v, the problem with excitation will be eliminated?

I don't understand "5-a+7v". But certainly, completing the circuit would be a good start. I suspect that was the cause of the problem in the first place. It must have been quite marginal all along. I think you should have roughly -5 V and +15 V at all gate inputs, with respect to the emitter.

The lack of C55 possibly explains the slope of the gate drive.

Edited November 12, 20223 yr by Coulomb

So I adjusted the excitation so that it was the same for all four igbt transistors. C55 is and was present (nicely described). But it's only -5.5V/+8. An increase to the recommended approx. 15V is not realistic without total adjustments to the source. I have one more question. What is the approximate self-consumption of an inverter powered by a solar panel? 20, 50, 100W, More?

16 hours ago, nogers said:

What is the approximate self-consumption of an inverter powered by a solar panel? 20, 50, 100W, More?

It's roughly 1% of rated power. So about 50 W for a 5000 W inverter.

16 hours ago, nogers said:

But it's only -5.5V/+8. An increase to the recommended approx. 15V is not realistic without total adjustments to the source.

That might work OK. So is the +12 V rail on this inverter close to correct? Axpert inverters have feedback to set the +12 V rail, and other power supplies (such as these isolated gate drivers) get voltage proportional to the 12 V rail.

Is the gate drive flatter at the tops and bottoms now? It used to droop to roughly half voltage by the end of the PWM cycle.

Hi Coulomb. You're right. The resource is designed exactly as you say. 12V is fine. Only it's not exactly Axpert but some kind of Revo II clone and it's not 100% the same in everything. The curve seems a bit better on the N side and I'm not sure at all on the L1 side of the phase. It's strange to repair (modify) something that has no damaged parts and still doesn't work properly. On the other hand, it is a pleasure to consult such a matter with someone who has such extensive knowledge as you and is still on the other side of the globe.

5 hours ago, nogers said:

I'm not sure at all on the L1 side of the phase.

It looks messy because of the PWMing. What does it look like at about 20 μs per division, single shot mode? I'm guessing it will be a reasonable square wave, with pulse width and amplitude varying depending on what part of the 50 Hz cycle you catch it at.

That slope is still concerning. Does the 12 V supply perhaps have a trapezoidal waveform on it, averaging to 12 V? It's as if something is loading the power supply more and more as the 50 Hz cycle progresses. The power supply should be operating at many kilohertz, and this should be keeping the gate drive constant at the 50 Hz timescale.

Hi. At 20μs it is exactly as you say. And the 12v output source on the transformer is trapezoidal with oscillation. also measured transformer tx7 input and output, these oscillations gradually increase. I attach a photo.

On 2022/11/09 at 11:23 PM, nogers said:

Hi @nogers and @Coulomb , There are unpopulated components on the IGBT PSU sections that just does not make any sense to me. For example some key unpopulated  components are R158,ZD4 and R21 on the board-under-test as shown above .  The partial schematic depicts the relevance and importance of these components. Could it be that the board was working before for the wrong reasons?  EDIT : and there could be more unpopulated components as the third PSU is not captured onto the photo above,.

 

Edited November 19, 20223 yr by BritishRacingGreen