question

@Jona Aal

Ideally the person who installed should be troubleshooting.

Failing that local boots on the ground will be needed. Google victron where to buy and use your location.

An internal faliure is a pylontec problem. Not a Victron problem. Nothing can be helped there from this side. You will need to figure out which one has the issue and (possibly) remove it. But start with swapping which on is the master.

You can swap who is the master on the pylon stack.

Switch off battery life.

I assume by the mention of the generator that you are off-grid? One thing I have seen is keeping them charged can solve the internal error message. Might be awkward if you are off-grid.

The other option is to remove battery Comms from the system, try seeing if it wakes up then.

The other thing I have seen is the Comms cable, so a simple one. One way is to only connect the can h l and ground (so only 3 wires). Pins 1 to 3 should be empty, you get funny issues of they are not on the premade cables.

Thanks @Guy Stewart (Victron Community Manager) I'll go and check the DC bus and fuses and take it from there..My initial fault finding has been batteries and BMS snce those are the alarms i've had. I have only had a small amount of time on site, with no tools so will go back and pull it apart bit by bit.

Also @Alexandra thanks for the suggestions. I am the installer so am doing the troubleshooting (!) I have been pretty thorough in checking cables are OK. I hav'nt managed to get any of the internal faults to be repeatable though they are frequent.

I'm sure thet keeping everything charged fully would be nice and over the last year that seems to have kept the internal fault issue at bay (we had it a lot after comissioning), off grid though you kind of need the batteries to be cycled!

The system is all on OEM cables, no mods and it has suddenly started behaving like this with no obvious cause. Pylontech are not particularly helpful and there is no foum dedicated to their batteries, so this forum is the only place where there is any level of experience.

Will post back with results.

Thanks again for your help

JOna

This exact issue with almost a 3V difference (MPPT was 3V higher than all other devices) has been demonstrated on this site by a user in Peru. Your configuration is nearly identical to his.

https://community.victronenergy.com/questions/196650/why-is-mppt-reporting-nearly-3v-higher-than-pylont.html?childToView=210866#answer-210866

I spend extensive hours helping him troubleshoot, but he got to be too much for me to deal with. One can only give a couple dozen hours of free support before one tires of going around in circles with an argumentative/indecisive person.

SVS is forced off with Pylontech batteries, so each device will report its own measured voltage, and they will all be a little different, and everything but the BMS or a battery monitor will be influenced by charge (higher voltage)/discharge (lower voltage) currents.

Recommend you take a shortcut and conduct the testing described herein:

https://www.victronenergy.com/upload/documents/Pre-RMA_bench_test_instructions/114199-Pre-RMA_bench_test_instructions-pdf-en.pdf

Page 24, section 6, paragraph 6:

Important this is done at 0A current.

Confirm your meter measures same as BMS and VE.Bus reported voltages, then check the MPPT voltage at the battery terminals.

The way you can achieve higher SoC is to disconnect the MPPT from the Cerbo, disable BMS management of the MPPT in VC and set an absorption voltage 3V higher than you actually need.

Obviously, you would want to be conservative in your approach to confirm the 3V offset happens at all times, and you don't over-volt the battery. Start with 1V, confirm, then 2V, confirm, then 3V or whatever the offset actually is.

Pylontech are 15S batteries, and charging to 3.45V/cell can typically attain very high states of charge, i.e., 51.75V.

@Guy Stewart (Victron Community Manager) @Alexandra

Update, I have found the issue: An poorly crimped lug at the DC bus from MPPT +ve, which I was able to wriggle off with a pair of pliers. The lug is the only one of its type used in this install so I can only assume incompatable with the crimping tool which seems to have done a good job on all other connections I have just inspected. Also, it wasn't checked carefully enough prior to installation.

All other connections were very tight so I can only assume that the lug to stranded copper connection was heating and cooling over the last 16 months, maybe the last few weeks have been more significant as the batteries have been cycling and PV supplied full-bore to recharge daily, as opposed to the summer when lower loads and long-day generation were allowing batteries to remain near to 100%.

One thing I noticed is that the spring washer on that connection is discoloured (from the heat) and totally flat, unlike the others. I wonder if the heat has annealed the copper and allowed it to permanetly deform to flat so not providing any force to keep the nut tight. This is academic as the cause is clear but I wonder about the wisdom of spring washers as opposed to lock nuts?

Lessons learned: Always check the qualoity of the crimped lug after forming. I have several sites to check that have had this crimp tool used on, which I will book in for an inspection over the next few days as a priority.

@Guy Stewart (Victron Community Manager) absolutely spot on with the diagnosis. Another lesson learned: always check the simple things before you go looking for complex data connection issues and the like.

Q to Guy, did you notice the voltage discrepency just looking at the graphs, or did you follow it up after I mentioned it as a possible symptom?

Off to the wholesaler now for a length of cable and considering how I might break the news to the client that I accept responsibility...