MultiRS 3 Phase System - AC PV Enphase Micro Inverters POWER CUT

Q&A and troubleshooting DESS
, , , , , ,
1

Hi Everyone,

A few days ago we had a power cut, and so the house with the solar micro inverters on the output had still power because of the UPS function.

3x Multi RS off-grid: DESS holds battery charge rate at 0 W, system frequency-shifts the AC-coupled PV off for 2 hours.

SYSTEM

  • 3x Multi RS Smart 48V/6000VA/100A, one per phase (L1/L2/L3), firmware v1.31
  • Cerbo GX MK2, Venus OS v3.79
  • Enphase microinverters on AC-out, read in as a SunSpec PV inverter, 6.48 kW total devided over 3 phases.
  • Grid meter: Energy Meter VM-3P75CT
  • Battery: 840 Ah bank, JK-BMS over CAN. Active battery service = SmartShunt 500A
  • Dynamic ESS (DESS) active, ESS minimum SoC 15%, AC input current limit 25 A
  • Multiphase regulation: “Total of phases”

All figures below are VRM logged data at 1-minute resolution.

WHAT HAPPENED
The street lost power at 11:01. The Multi RS units kept the house alive without
interruption - that part worked perfectly. But the Enphase microinverters produced
almost nothing for the next 2 hours 12 minutes, because the system kept shifting the
output frequency to 52.7 Hz.

11:01 Grid lost. Active input goes to “not connected”. Enphase was making 1375 W
11:02 Output voltage 240 → 229 V. Output frequency jumps to 52.7 Hz. Enphase drops to 17 W
11:04 “Grid lost” alarm
11:02 - 13:12 Frequency cycles between 50.0/50.1 Hz and 52.7 Hz (peaks to 53.3 Hz)
roughly every 2-3 minutes. 56 of those 131 minutes are at or above 52.0 Hz
13:15 DESS releases its charge restriction. Frequency returns to 50.0 Hz and stays
there. Enphase ramps to 3-5 kW - while still off-grid
14:00 Grid returns

PV energy produced during the 2 h 12 min of frequency cycling: 0.10 kWh.
PV energy produced in the 33 minutes right after 13:15, still off-grid: 1.71 kWh.

WHY I THINK DESS IS THE CAUSE
Throughout the whole outage, DESS commanded a battery charge rate of 0 W (logged in
VRM as “DESS Charge Rate”). It had been 0 W since before the outage started.

So the system was in this state:

  • No grid
  • House load only 100-400 W
  • Battery at 18.9% SoC, BMS charge current limit 360 A, charge voltage limit 55.2 V,
    battery voltage 51.7 V - plenty of room
  • DESS saying: charge rate 0 W
  • Enphase wanting to deliver 1.4 kW

With nowhere to put the power, frequency shifting is the only lever the Multi RS units
have left. The battery was actually discharging at 1-12 A for the entire off-grid
period, and SoC fell from 18.9% to 17.9%.

The moment DESS changed its mind at 13:15, everything snapped back to normal - with no
change on the grid side:

  • DESS Charge Rate: 0 → 5137 W, ramping to 16600 W by 13:27
  • DESS Restrictions: 3 → 0
  • DESS Strategy: 3 → 2
  • Battery current: -4 A → +7 A in the same minute, +55 A three minutes later
  • Output frequency: flat 50.0 Hz from that minute on
  • Enphase: 2 W → 526 W → 3-5 kW

The grid did not come back until 14:00, so for 45 minutes the system ran off-grid with
the Enphase at full output and a rock-steady 50.0 Hz. That is the control experiment:
same off-grid conditions, only the DESS charge restriction removed.

WHAT I RULED OUT

  • No alarms or faults on any of the three RS units for the whole period: error code 0,
    mode “On”, no overload, no DC ripple, no high/low DC or AC voltage alarm, no short
    circuit, no “sustain” active, no “energy meter missing”. State stayed 252 “Ext.
    control” throughout.
  • No PV inverter power limit was set. The Power Limit attribute was 0 / not in use all
    day; my own limiting setup only went live two days later.
  • Battery was not full and could accept charge - proven by the +55 to +88 A it took as
    soon as DESS allowed it.
  • Not a “battery can’t keep up” situation - load was 100-400 W against 1.4 kW of PV.

WHY THE ENPHASE NEVER RECOVERED ON ITS OWN
52.7 Hz is above the over-frequency trip point of the Enphase grid profile (~52.0 Hz
for EN 50549-1 OF2). So instead of derating on a frequency-watt curve, the
microinverters trip off completely and start their reconnect timer. The ~2-3 minute
cycle in the data matches that reconnect delay: they come back, ramp up, the RS units
immediately shift to 52.7 Hz again, and they trip out. Repeat for two hours.

MY QUESTIONS

  1. Is this expected behaviour? Should DESS’s charge rate command still be enforced when
    the grid is lost and the system is islanded? I would expect DESS to stand down and
    let the system run as a plain ESS/off-grid system.
  2. If it is expected, is there a setting to make DESS release its restrictions on grid
    loss - or does this have to be handled externally (I run Node-RED and can force it)?
  3. Is the frequency shift range configurable on the Multi RS? Jumping straight to
    52.7 Hz hard-trips any PV inverter whose profile trips at 52.0 Hz. A range ending
    around 51.0-51.5 Hz would derate the Enphase instead of knocking it out.
  4. Has anyone else seen 3x Multi RS + AC-coupled PV on AC-out + DESS behave like this
    during an outage?

Raw 1-minute CSV of the event is attached if anyone wants to check the numbers.

02_detail
02_detail1820×1008 128 KB

01_overview
01_overview1820×1890 333 KB

BTW the energy prices were positive all day long.

2

Time Out freq PV Enphase Battery SoC DESS charge rate

CEST Hz W A % W

GRID PRESENT — normal operation

10:30 50 873 11 18.5 0
10:35 50 1075 9 18.6 0
10:40 50 1041 2 18.7 0
10:45 50 857 5 18.7 0
10:50 50 1013 17 18.8 0
10:55 50 1048 -1 18.9 0
10:56 50 1086 0 18.9 0
10:57 50 1064 10 18.9 0
10:58 50 1103 -2 18.9 0
10:59 50 1277 -2 18.9 0
11:00 50 1367 -1 18.9 0

GRID LOST — frequency cycling, Enphase knocked out

11:01 50 1375 5 18.9 0
11:02 52.7 17 -5 18.9 0
11:03 50.1 2 -5 18.9 0
11:04 52.7 105 -2 18.9 0
11:05 50.1 2 -5 18.9 0
11:06 50 3 -5 18.9 0
11:07 52.7 93 -3 18.9 0
11:08 50.1 2 -5 18.9 0
11:09 50 27 -4 18.9 0
11:10 52.7 105 -3 18.9 0
11:11 50.1 2 -5 18.8 0
11:12 52.2 298 0 18.8 0
11:13 52.7 2 -5 18.8 0
11:14 50.1 2 -5 18.8 0
11:15 52.7 209 -1 18.8 0
11:16 50.7 2 -5 18.8 0
11:17 50 3 -5 18.8 0
11:18 52.7 40 -2 18.8 0
11:19 50.1 2 -7 18.8 0
11:20 50 3 -9 18.7 0
11:23 53.3 134 -2 18.7 0
11:25 50.1 2 -4 18.7 0
11:29 52.7 213 2 18.7 0
11:30 50.1 2 -3 18.7 0
11:35 50.1 2 -3 18.6 0
11:39 53.3 108 -1 18.6 0
11:40 52.5 2 -3 18.6 0
11:45 52.7 264 -9 18.6 0
11:50 52.7 41 -2 18.5 0
11:55 53.2 49 -3 18.5 0
12:00 50 3 -5 18.4 0
12:03 53.2 124 -3 18.4 0
12:05 50.1 0 -5 18.4 0
12:10 50.1 2 -5 18.3 0
12:15 50.1 2 -5 18.3 0
12:17 52.7 229 -2 18.3 0
12:20 52.7 112 -3 18.2 0
12:22 53.3 116 -3 18.2 0
12:25 52.7 235 -2 18.2 0
12:30 51.1 160 -3 18.1 0
12:35 50 80 -2 18.1 0
12:38 53.2 79 -2 18.1 0
12:40 50 2 -3 18.1 0
12:43 53.3 163 0 18 0
12:45 50.1 2 -3 18 0
12:50 50.1 2 -3 18 0
12:55 50.1 2 -3 18 0
13:00 50.1 2 -3 17.9 0
13:05 52.1 2 -3 17.9 0
13:07 52.7 150 -1 17.9 0
13:08 50.1 2 -4 17.9 0
13:09 51.6 151 -1 17.9 0
13:10 52.7 2 -4 17.9 0
13:11 50.1 3 -4 17.9 0
13:12 52.7 121 -2 17.9 0
13:13 50.1 2 -4 17.9 0
13:14 50 3 -4 17.9 0

Grid Back — DESS allows charging, Enphase runs again

13:15 50 526 7 17.9 5137
13:16 50 1974 34 17.9 5137
13:17 50 2190 37 18 5820
13:18 50 3158 55 18.1 6243
13:19 50 3393 59 18.2 6505
13:20 50 3344 55 18.3 6870
13:21 50 3075 54 18.4 7310
13:22 50 3856 68 18.6 8337
13:23 50 3040 53 18.7 8822
13:24 50 3034 55 18.8 10458
13:25 50 2591 37 18.9 12108
13:26 50 2930 54 19 15383
13:27 50 2279 43 19 16600
13:28 50 2040 35 19.2 16600
13:29 50 1619 26 19.2 16600
13:30 50 2283 39 19.3 16600
13:35 50 2939 50 19.7 16600
13:40 50 3208 49 20.2 16600
13:45 50 3645 67 20.9 16600
13:50 50 914 10 21.4 16600
13:55 50 1152 15 21.5 16600
13:59 50 2296 21 21.9 16600

14:00 50.1 846 9 21.9 16600
14:01 50 759 90 21.9 16600
14:02 50 752 292 22.3 16600
14:05 50 1380 293 24 16600
14:10 50 895 271 26.9 16600

3

I’m trying to get the same situation. but without grid connection the solar panels still charge the batteries.

image
image1411×680 60 KB

Does someone have any idea what happend?

4

Hi George, unfortunately I can’t help you, but I came about your post just now while searching for a replacement of my 3x MultiPlus-II 48/5000. Thinking about getting the MultiRS 3 Phase System as well as a replacement but I also have an AC PV Enphase Micro Inverters configuration. Hope someone can clarify the situation that you saw, although reading your last post it seems that you can’t reproduce?

Have you ever had the MultiPlus II prior to the MultiRS? I’m looking for this replacement as my current setup makes to much noise (low humming noise) which turns worse when it loads/unloads fully. My wife wants me to get quiter ones … any thoughts?

5

It’s also possible it’s because the battery is above 95% and usually DC PV finishes battery charge when offgrid as it is easier to control.
This would happen with or without DESS when offgrid and frequency shifting

6

Hello Anthonny,
I’ve already made this change myself because of noise and efficiency, but for a single-phase system.
It works perfectly with the Multi RS.
Best regards, Dennis

7

Hello Georg,

In the MultiPlus, this could be configured via VEConfigure, but in the Multi RS it is fixed.
I have a single-phase Multi RS with a SolarEdge on the AC output (and DC PV connected to the Multi RS).
The off-grid test worked perfectly, by the way, and the frequency remained stable at 50 Hz until the battery was nearly full, at which point the frequency increase began.
However, the Multi RS manual states in section 3.9 that the Multi RS does not fully charge the batteries via AC-PV.

Best regards,
Dennis

8

No, the SOC of the battery was around 17%

image
image458×457 16.2 KB

This was also my thought, but this seems really strange.

9

Never, the only thing is that the MultiRS is quite expensive. but overal, better efficienty and not that much noice. it will pay itself back when trading energy!

10

any other ideas?

11

I have an Enphase system connected to the output of my Quattros, and I also use DESS. I don’t have the Enphase modbus connection in my geo.

I don’t think DESS charge rate plays a factor in how the units charge from AC coupled PV while islanded. I believe that’s only used to for when a charge from the grid is needed. That parameter is zero most of the time for me, and the Enphase units continue to produce as expected while islanded.

Do you have DVCC enabled with a smart BMS? If so, what was your dynamic charge current limit / charge voltage limit at the time? This is available in the VRM data.

12

image
image1455×856 75.2 KB

just the normal 55.2VDC DVCC enabled, and yes there is a smart BMS used: Jk inverter bms.

13

Yes, however the Victron units will dynamically adjust their charge rate depending on what the BMS requests: 11. DVCC - Distributed Voltage and Current Control

So if your BMS is sending a DCL of zero amps, then that will cause the Victron units to increase the frequency to 52.7Hz (to throttle off the Enphase units).

For example, here’s a scenario where my BMS was telling the Cerbo that it can’t accept any more than ~426A. This can happen for a whole host of reasons (e.g., high state of charge, low temperature, bad configuration, a bug, etc.). It’s possible that your BMS was requesting a lower charge voltage / current at the time of the issue.

image
image2542×1610 182 KB

14

image
image1407×886 90.6 KB

here I see no real issues with the CCL.

15

I found it a strange thing that there is no alarm or whatever that tells why the output of the inverter is raising the frequency

16

Just for safe measure can you also plot Charge Voltage Limit?