Thanks very much for your help Coulomb much appreciated.

I'm trying to find a replacement for the faulty components the number on the device is ost75n65hm but I can't seem to find one anywhere, anyone know where I can get these or an alternative, thanks 

On 2022/07/09 at 10:02 PM, AndrewM314 said:

trying to find a replacement for the faulty components the number on the device is ost75n65hm

I assume that these are the MOSFETs [ edit: incorrect assumption; see next post ] on the battery side? If so, Weber suggested some improved spec parts in this post:

http://forums.aeva.asn.au/viewtopic.php?p=64358#p64358

and the 2015 post that it links to. That's assuming it's a 5 kVA model. I don't believe that they used higher spec MOSFETs on the 5 kW models (back then, it was 5 kVA but 4 kW), but I'm not certain. The part number above suggests 75 V and 65 A; I suggest at least 80 and preferably 100 V for replacements. 75 V is really very little margin. Is this perhaps a clone?

Edited July 10, 20223 yr by Coulomb

16 minutes ago, Coulomb said:

I assume that these are the MOSFETs on the battery side?

They are actually the IGBT's QA1 & QC1, so do you think I could use any IGBT as long as they are better than 75A/650V and would I have to change all 4 if I'm using different devices or wouldn't that matter

6 minutes ago, AndrewM314 said:

They are actually the IGBT's QA1 & QC1,

Oops, sorry. Yes, nearly any IGBTs should do, except there are a few about with extremely high gate charge/capacitance. Just check that the gate capacitance isn't over 10,000 pF (10 nF). That's the limit for the gate drivers.

https://forums.aeva.asn.au/viewtopic.php?p=91985#p91985

Edit: I would always replace upper and lower pairs of the same full-bridge leg. But I recently checked that QA1 and QC1 are both for the neutral output, so just replacing those two should be fine.

Edited July 10, 20223 yr by Coulomb

Quick question can the main board be tested without the mppt board and par board connected? 

1 hour ago, AndrewM314 said:

Quick question can the main board be tested without the mppt board and par board connected? 

I believe so, yes. I don't know which board you mean by par board, but either (the paralleling board or the SCR board) can I believe be left out and the inverter should function.

20 minutes ago, Coulomb said:

I believe so, yes. I don't know which board you mean by par board, but either (the paralleling board or the SCR board) can I believe be left out and the inverter should function.

In the service manual it shows the par board to the left of the mppt board. Is that board usually known by another name 

30 minutes ago, AndrewM314 said:

In the service manual it shows the par board to the left of the mppt board.

OK, that's what I think of as the SCR board; PAR sounds too much like the "paralleling" board.

Not much is known about this PAR board; I believe it somehow brings down the changeover time from 20 ms to 10 ms by paralleling the relay contacts with SCRs, or something like that.

On 2022/07/09 at 2:34 AM, Coulomb said:

Yes, the gate driver components. Check at least all nearby diodes (including zener diodes, tested as ordinary diodes) and resistors, including the internal diode in the high side gate driver chip.

I've checked all the components around the high side gate driver  all seems ok except there is a short circuit between QC1 emitter and QD2 emitter with all the IGBT'S removed, that doesn't seem right?

4 hours ago, AndrewM314 said:

all seems ok except there is a short circuit between QC1 emitter and QD2 emitter with all the IGBT'S removed, that doesn't seem right?

Looking at the IGBT circuit again the emitter short is ok, it connects to the negative side of the DC which is correct . The only other anomaly is D20 is short circuit, I removed the diode to test it and the diode is ok, anyone know what D20 does?

 

 

15 hours ago, AndrewM314 said:

The only other anomaly is D20 is short circuit, I removed the diode to test it and the diode is ok, anyone know what D20 does?

On my old model, D20 is one of four through-hole 1N4007 1 A diodes. But that's for the 230 VAC power supply, which recent models don't have.

Can you describe it or post a photo?

34 minutes ago, Coulomb said:

Can you describe it or post a photo?

I've attached a picture of D20's location 

Ah. I'd say that's D40 on this schematic:

HFPW+ (High Frequency PoWer +) is a square(ish) wave power connection from the main power supply. Rectifying and filtering it produces +12 V. The main power supply is a flyback design; these are intended to provide power in only one "stroke". The "intake stroke" ("induction stroke"?) fluxes the multi-winding inductor; during the "power stroke", the inductor de-fluxes and pushes current into the loads. D20/D40 is intended to ensure that power only flows during that "power stroke".

TX7 provides isolated power for the IGBT drivers: the two high-side full-bridge IGBTs, and the third drives the three BUS- referenced transistors: two low-side full-bridge transistors, and the buck transistor. Each winding provides ≈+15 V and -5.6 V for the gate driver output side.

Thanks Coulomb for the explanation and diagram, looking at this diagram and the hfpw diagram it looks like one side of D20 goes to ground through the windings of TX7 and the other side of the diode goes to ground through the windings of TX9 so it makes D20 look like it's short circuit but in actual fact I was measuring the very low resistance of the windings.

So hopefully everything should be ok, just waiting for replacement IGBT'S to give it a try.

Out of interest with my diode being numbered D20 instead of D40 could that my inverter could be a clone?

23 hours ago, AndrewM314 said:

Out of interest with my diode being numbered D20 instead of D40 could that my inverter could be a clone?

That by itself, no. See Do I Own a Clone? for clone checking.

Edit: most newer models don't have the AC-in power supply, so they would not have D20 and its neighbours. So re-using designator D20 seems pretty reasonable. Sadly, no-one seems to have traced and published a schematic for any of the later models.

Edited July 14, 20223 yr by Coulomb

On 2022/07/14 at 3:43 PM, Coulomb said:

That by itself, no. See Do I Own a Clone? for clone checking.

After reading the above post everything appears ok except that it doesn't have the little date label.

Also I've Just noticed it doesn't actually have Axpert in the model name.

Does this sound like it could be a fake inverter.

 

5 hours ago, AndrewM314 said:

everything appears ok except that it doesn't have the little date label

That's actually pretty telling, though of course they may have changed their manufacturing habits.

The label also seems too far from the front panel.

5 hours ago, AndrewM314 said:

Also I've Just noticed it doesn't actually have Axpert in the model name.

I don't believe that's necessary or significant.

5 hours ago, AndrewM314 said:

Does this sound like it could be a fake inverter.

There is something about the label that says to me that it's not quite right. In particular, the font and the height of the barcode, and spacing of the bars.

Also, I have this theory, NOT confirmed, that there are no genuine Axpert MKS IIs or VM IIs made for the last several years; all the genuine models* have at least a removable display, if not the even newer round colour display. The above label suggests a manufacturing date of 2022/March, which according to this theory is way too recent. If anyone has evidence to the contrary, I'd love to hear about it. So that unconfirmed theory suggests that it's fake.

* Edit: now that I think about it, I do believe that there are still some Axpert MKS models made, being those with the 145 V max PV voltage. These would still have the oldest non-removable monochrome displays, so I guess it's not a big  stretch to imagine that they also still make the Axpert MKS II, which seems to be considered part of the Axpert MKS I family, if you like. It appears with the other Axpert MKS I models on the Voltronic Power web page (see the Specifications tab).

Executive summmary: It's hard to say.

Edited July 22, 20223 yr by Coulomb

I managed to repair this unit by changing the two faulty IGBT's, it's now working ok and I have just connected via WiFi, the unit ID is 5535553..., Which I think I've read before is another pointer to a fake unit, is that correct.

12 hours ago, AndrewM314 said:

the unit ID is 5535553..., Which I think I've read before is another pointer to a fake unit, is that correct.

Yes. It's been known that genuine Axperts have had this number; it can happen if the EEPROM loses its values, for example. But generally, it's another strong indication that this is a clone.

On 2022/07/13 at 3:11 PM, Coulomb said:

Ah. I'd say that's D40 on this schematic:

HFPW+ (High Frequency PoWer +) is a square(ish) wave power connection from the main power supply. Rectifying and filtering it produces +12 V. The main power supply is a flyback design; these are intended to provide power in only one "stroke". The "intake stroke" ("induction stroke"?) fluxes the multi-winding inductor; during the "power stroke", the inductor de-fluxes and pushes current into the loads. D20/D40 is intended to ensure that power only flows during that "power stroke".

TX7 provides isolated power for the IGBT drivers: the two high-side full-bridge IGBTs, and the third drives the three BUS- referenced transistors: two low-side full-bridge transistors, and the buck transistor. Each winding provides ≈+15 V and -5.6 V for the gate driver output side.

Thanks for this part of the diagram. Do you have the complete schematic. My inverter had the ZD4 open, it seems that the one of the ceramic capacitors in parallel with the electrolitic capacitor was shorted. Any way it is working now, but I would like to have the complete schematic. In the service manual does not appear.

It's the first schematic in the repair and modifications topic:

https://forums.aeva.asn.au/viewtopic.php?t=6007#schematics

11 hours ago, harrimansat said:

In the service manual does not appear.

The schematic is an unofficial thing, with most of the hard work done by forum reader Maxo. They don't want to give the clone makers too much help, even though the clone makers seem to have access to the schematic, or they just trace it themselves.