Possible causes of "#38 PV Input Shutdown"?

Q&A and troubleshooting
, , , , ,
1

My MPPT (250/70) shut down today with an “Error code: #38 - PV Input shutdown (no retry)”. This was 13:14 in the UK, on a variably-sunny day with a full battery, having reached 100% SOC at 12:44. This was just about peak irradiance for the day (666W/m2), but peak PV generation had been an hour earlier at 3.6kW.

It has never done this before so I had to look up the error. Everything seemed to be operating correctly and within spec (PV was 178V, battery 56.4V, house load ~200W.
Disconnecting PV and battery, then reconnecting battery and PV made everything go back to normal, but as this can only be reset manually locally, and I’m off on hols for 6 weeks soon, this is concerning. So I’d like to know why it might happen.

The system has been running happily since January, and nothing has changed in the last 2 weeks (before that I upgraded the original 35mm2 battery cabling to 70mm2 to be in-spec for the multiplus, and the wiring does now run cooler at peak discharge).

There is nothing obviously untoward in the graphs at the time of failure. PV was generating about 2.2kW, varying as clouds came in and out. Exporting 2-3kW. SOC 100%, battery voltage 56.37V.
Why would it suddenly turn itself off, shorting the PV contacts?
What criteria does it use?

Do I just have to wait and see if it does it again? Anything to check?
Is overheating a possibility - the unit gets quite hot, but then it’s passively cooled so that’s not too surprising (it is mounted with vertical airflow and a vent higher up).

2

@wookey

Possible causes of error #38 are described in the manual.
It is usually related to preventing battery overcharge protection.

If you click the link in the alarms logs it also takes you there.

So check that or settings. Check if the dC voltage overshot the battery CV

3

Yes. I read that. Where is this voltage limit specified? In the Multiplus ESS assistant numbers?
The battery BMS limits the voltage to 57.6V (3.6V per cell), and will turn off the charge FET if anything tries to drive it higher.
The peak voltages over the last 7 days have been:

   At MPPT     At BMS
Thu   57.3      56.25
Fri   57.1      56.31
Sat   54.9      54.93
Sun   56.9      56.27
Mon   57.3      56.34
Tue   57.2      56.34
Wed   56.8      56.23

Why would it fail today, but not Monday or Tuesday (or Friday or Thursday), all days with higher peak voltage, if high batt voltage was the reason?

So there is nothing in the logs to suggest an unusually high battery voltage.

4

What you have programmed. And also based on if it is can connected.

Ultimately you don’t want that happening on a regular basis.

It doesn’t have to be unusually high. Just higher than it should be based on programming or battery command.
Ripple can also cause issues where there doesnt seem to be readings to support the happening.

5

[quote=“LX, post:4, topic:34694, full:true, username:lxonline”]

What you have programmed. And also based on if it is can connected.

Well, yes, but there are a lot of settings in a lot of places - I was wondering exactly where to look to check what voltage the MPPT was set to use for its ‘voltage too high, issue error #38’.
I’m not seeing a voltage limit under the MPPT device configuration, but I guess one is implicit in the ‘use the 48V lithium’ switch selection, except that maybe that’s overridden when it’s operating under ‘external control’, which it is:
The MPPT is VE-Direct serial connected.
I am not using DVCC (due to bug reported in another thread) so the Cerbo/multiplus are in charge, running the MPPT under ‘external control’.

Interestingly the history the MPPT records says the highest voltage today was 54.23, whilst the VRM logged MPPT data said 56.8 . Which is inconsistent.

(The above reply has been pending for about 5 weeks, pending further investigations. Further investigations below).

We got another error #38 yesterday 14:06pm local time ohttps://www.zugfinder.net/en/train-ICE_9573n the 3rd very hot and sunny day in a row. I’m on holiday so ‘yay’ for remote alarm monitoring. Got a neighbour to go and reset the MPPT.
Battery reached 100% SOC an hour before (13:06).

During that 1h period (from 100% SOC to MPPT fault) the graphs show:
Battery voltage (measured by battery) was: 56.44V +/- 0.1V
System voltage was: 56.45V +/- 0.1V
VEBus voltage was: 57V +/- 0.35V
MPPT charger voltage was: 57V +/- 1.3V (so 58.3V peak)
Ripple was high during that time: 1V +/-0.6V. Before 100% SOC, ripple was typical 0.2V +/-0.05V
Power flow 3.7kW +/- 1.1kW

What exactly is the different between ‘system voltage’ and ‘VEBus voltage’?
I presume the MPPT voltage gets a bit higher than the battery voltage because it has to raise the voltage above the battery voltage in order to ‘push’ power out through the inverter. Is that how this works? And the ripple seems to appear on the MPPT voltage, but not on the battery or system bus voltages.

I’m not really sure how this works in detail, but any clues on whether this looks normal or not, or other info I should look up in order to debug the #38 error, would be welcome.

6

The system voltage with dvcc on will be the bms voltage.

The ve bus voltage is what is at the inverter battery terminals. This is not always an indicator of overcharging or anything as the terminal voltae can be higher feeding back or supporting loads.

More on ripple voltage here.

The mppts can be under external control if you have ess programmed.

If there is no dvcc control then it is ultra important to have the correct voltages programmed on the system, and they must all be the same on all the victron devices.

If you don’t have a compatible bms then yeah do that as a troubleshooting step but then really get a shunt and use that for measuring near the battery and have dvcc on and using the shunt for battery measurements. The ve bus voltage set point will be followed otherwise.