diy solar

diy solar

Help troubleshoot PV strangeness when battery reaches 100% SOC

YAMLcase

New Member
Joined
Oct 13, 2021
Messages
36
new instsaller (me) with a new install been up around 1 week. My system:
- Sol-Ark 12k
- 2 strings of 6x 385W Mission Solar panels (I can go up to 8, so well below the max voltage)
- each panel has Tigo TS4-A-O optimizer
- 2x EG4-LL 48v 100Ah battery banks

first 3 days went fine, then 4th and 6th days I got arc fault alarms that tripped the PV off. Yesterday (the 7th day) I had a strange event happen: Around 4:00 PM Central Texas time my solar production dropped off a cliff and flat-lined at about 50% the peak production. From the graphs, I can summarize:

- Battery voltage shot up from 54.3V to 56.8V
- Battery state-of-charge reached 100%
- MPPT voltage went from ~190 to ~225
- Current dropped from naturally declining with sun exposure (down to ~9A) to flat-lined at ~4A

Update: I just went back to the graphs and noticed that with each arc_fault trip the SOC reached 100% on the batteries. Yesterday the voltage jumping to 56.8V was the first time, each other time the voltage dropped slightly or remained the same.

1654526448107.png

1654526341307.png
 
Last edited:
I had a strange event happen: Around 4:00 PM Central Texas time my solar production dropped off a cliff and flat-lined at about 50% the peak production. From the graphs, I can summarize:

- Battery voltage shot up from 54.3V to 56.8V
- Battery state-of-charge reached 100%
- MPPT voltage went from ~190 to ~225
- Current dropped from naturally declining with sun exposure (down to ~9A) to flat-lined at ~4A
That sounds like the charge controller decided the battery was full and backed off production. (It sounds normal to me).
The only thing that is not 100% in line with the battery being full is that the production did not go to zero. However, that can easily be explained if there was a load on the system. The charge controller continued to provide current for the load but no more current was going into the battery.

The load during the event looks pretty constant at 1KW so I am going to speculate that the 'always on' background load of the house is around 1KW.

Just as a quick background: When an MPPT controller decides the battery is in the absorption stage it will start reducing the current allowed in from the panels in order to reduce the current going to the battery and thereby hold the absorption voltage constant. As the batteries complete the absorption stage the current will continue to reduce till it hits zero. (This happens rather quickly with LiFePo4). However, the controller does not know where the current it is providing is going. If the current is going to a load instead of the battery, the controller just thinks the battery is not yet full and continues to provide current..
 
@YAMLcase did you ever solve your problem?

As @FilterGuy suggests it does look like your system is behaving as expected as far as your MPPT and batteries are concerned.

The arc-faults concern me slightly, as a Brit I have zero knowledge of these beasts although arc-fault breakers are now a "recommendation" in the UK regs. At present 99% of UK sparkies think they are an expensive white-elephant.

Does the arc-fault reset itself? It's possible it's just a side effect of the batteries going full, our Sofar AIO gives a "grid under-frequency" when the big irrigation pump starts (the grid is stable).
 
Arc fault detectors have a difficult task. They are looking for specific but subtle patterns in the current that signals an arc fault. This is complicated by having to exclude arcs that will occur any time a disconnect is opened or closed.

On the AC arc fault detectors, the companies have designed detectors that seem to be pretty accurate so you don't have a lot of false trips. However, have no idea how often the AC arc fault detectors don't trip when they should. (It turns out that it is somewhat difficult to design a test fixture that will reliably trigger one of the AC arc fault devices)

As witnessed by this thread, it seems like the technology for the solar PV arc fault detectors is not as far along (at least for some manufacturers). Part of this comes from the variability of PV power. A PV arc fault device must de be designed for a very broad range of voltage and current whereas the AC detectors at least have a relatively narrow range of voltage they have to operate in. This difficulty may also be part of why there are so few (none?) stand-alone PV arc fault products available.
 
What problem does the Tigo TS4-A-O optimizer solve in your setup?
This product description sounds like snake oil.
 
Optimizers have been around a long time and Tigo is one of the bigger names in the optimizer market. I have never worked with optimizers, but from what I have read they offer a couple of different benefits.

* The modern ones will implement a Rapid Disconnect protocol to meet the NEC requirements. (Sadly, the rapid disconnect requirements pretty much force some type of intelligent device at the panel)

* The original idea around optimizers was to put an MPPT controller at each panel (or small set of panels). Assuming they work well, this would have several advantages.
- Each panel will operate at its peak point. (Even without shade, each panel will have slightly different peak points)
- Shading on one panel does not impact the rest of the array.
- Having dissimilar panels in the array is far less of a problem.


Without the requirement of the rapid disconnect, I suspect you would be better off buying more panels than buying the optimizers. However, once you have to have something for rapid disconnect, the cost of also having an optimizer function may start to be worth it.

Hmmm.... as I sit and type this I am wondering what the presence of an optimizer does to the Arc Fault requirements. The optimizer would be considered a DC-DC converter and I don't know if there are arc-fault protection requirements on the output of DC-DC converters. On the input of the optimizer, the voltage may be low enough that arc-fault protection is not required. If the optimizer gets rid of the Arc-Fault protection requirement that would be another big benefit.
Note: If the optimizers eliminate the Arc Fault Protection requirement, I suspect it will be a short-lived loophole. The reason the requirement came about because of the large number of high voltage connections between all the panels and the likelihood some of those connections will fail. If the optimizers do not get rid of the connections so my guess is NEC will still view it as an issue.​
 
Back
Top