I have 20kw of solar, 8 of it is AC coupled with Fronius inverters. The other 12 is MPPT450/100, I have 2 x 48/5000 inverters in parrallel, 5 x 48v 14.8kw batteries, cerbo GX, 2 x EV chargers. Apart from the Fronius everything is victron. everything is connected to the Cerbo including the Batteries (JK BMS) The system has no grid connection. I do have Home Assistant monitoring and turning on loads (AC’s)when the battery is full in summer or winter.
I recently added the second MPPT450/100 to assist over the winter months in Western Australia, with charging 2 x EV’s The system was performing fine with the initial 12kw (8kw Fronius, and 4kw MPPT450/100). However now that I have added the extra 8kw and have another 4 still to be add, I have noticed an issue when charging now. The DVCC charge limit is set to 240amps, the house base load during the day is around 1200 to 1500w. When the battery gets over the 90% mark, I periodically get Hi charge current alarms form the BMS. The 5 batteries are all set to 0.2C or 57 amps each. So a max charge of 280 amps. It seems to be exacerbated by cloudy conditions, ie sun comes out from behind the cloud the solar’s all ramp up next thing the hi alarm and a system briefly trips the ACPower then recovers. There are a couple of UPS’s in the house that beep when this happens so we can audibly hear it.
Is there a setting or an option in the Victron eco system to reduce the DVCC charge limit when the battery hits say 90% soc? I have tried writing a new value from HA but it just goes from 240 to 1 or 0 not to the preset value of 110.
I have checked all the batteries when it happened yesterday, the SOC was 90% and the batteries were all under 100% ranging 85% to 95%SOC and taking charge no issues. We need to grab all the charge we can when the sun is shining during the winter time, as the EV’s suck it all up, but the month of August now is coming into spring, and there are some glorius days and some partly cloudy days, it on these days were power isnt an issue we can reduce the charge rate back to 175 where it used to operate fine. But I dont want to have to remember to manually do this every time
Remember DVCC limits tend to be ignored if you export power.
Typically, this is the job of the BMS, if it is correctly configured.
Ideally the battery should lower it’s CVL, at the very least it should drop its CCL near top of charge, this is pretty standard behaviour, assuming the system is managed by the BMS and setup right.
HI Nick, thanks “offgrid” so no exporting. The JK BMS / batteries are dropping off there required charge, however due to the large solar array, the Victron inverters / Cerbo are not backing off the solar and so it goes from full charge to nothing in 11min according to the trends. This happens as the batteries hit 55.5v pack so they stop taking current. But in these times and including cloud interruptions the solar is pumping out all it can. I have done a test over a couple days of fine weather, and if I reduce the Max charge current down from the 0.2C value of the batteries (280amps) to 180amps, the system seems to work a lot better as the Fronius units are already in the throttling curve at this lower charge rate. Which is perfectly fine for good clear sky days as we well and truly recover the consumption from the previous night. But in days like today where the its over cast cloudy conditions (and the next 4 days are the same rain / cloudy forecast,) and typically the panel are making 3 to 4kw of power when we do get a break in the clouds and the system ramps to 15kw, I dont want the “DVCC Max charge limit” limiting this charge current when the battery soc is 50% or 60%, I want to capture as much of these kw as possible.
The issue is on good clear sunny sky days were power isnt an issue and the batteries are typically charged by 11am, on these days (around 60%to 70% of the year) the lower charge current is perfect. But in winter time / overcast rain weather when we do get a couple of hours through out the day of sun shinning through I want maximum harvest, not to be throttled, unless the batteries are above 90% SOC.
Should I be turning on “Shared current sense”? In the manual section 11.1 has notes 3,4. ESS is not enabled in my system as we are off grid so note 3 is out, leaves note 4. which says it should be on. However the JKBMS is not in the supported battery list, even though it reports all data to Victron and controls the charge voltage etc.
I would drop your maximum charge voltage to say 54.4V then there’ll be a buffer at the top end so you don’t get high voltage warnings and bms dropouts.
Applicable only to systems not using ESS and not using a
managed battery, as charge control for MPPT solar chargers and
Orion XS is external in those cases.
Is the BMS connected via CAN or are you using any mods to access it?
Which chargers are not backing off fast enough - the Victron MPPT’s or the Fronius?
Is the Fronius on AC IN or OUT?
Has it been configured so that it is being controlled via modbus?
Are all the Victron chargers visible on the GX and under “external control”?
Unfortunately from numerous posts, the JK BMS seems to be problematic, or people struggle to configure them correctly.
There are a few topics open in the modifications section where DIY/unsupported BMS’s tend to be discussed, if you haven’t already that is probably worth a look.
Setting CCL to 0 isn’t a great way to control charging, and being offgrid puts extra demands on the battery which is why sizing requirements tend to be much larger.
I would consider lowering your voltages as suggested, it will have little effect on capacity.
Thanks for the reply, The BMS is the longer JK, and all 5 are connected via canbus to Cerbo. Cerbo can see the number of batteries connected and the cell highest lowest by total pack. MPPT’s are both showing “External” control and are connected via VE.Direct, the MPII are VE.Can connected. The 2 x Fronius are running using Frequency control not Modbus. I also have a victron Smart shunt connected via VE.direct, however the controlling device in the Cerbo is the JKBMS, the shunt is just as a backup.
I haven’t actually caught it when it happens to see if the Freq was up high enough to start slowing the fronius input down, but first time was about 5 minutes after I turned on the second MPPT, I didn’t realize the battery was already 100%. But even after this time it has happened several times. Yesterday I did a test using only the MPPT’s with the FLC set to 200A, and it all appeared to work ok. However there was no load on the system (300w) and the 2 x MPPT’s could only manage around 110 amps combined. I was going to try again today, but a storm front has arrived so the testing will have to wait.
Yes I can reduce the voltage I think that’s roughly where I used to have it 54.5 (Andy’s No’s) not sure how it managed to drift up. Quick maths says 3.46v per cell, think Andy uses 3.45. so maybe it has always been higher….cant recall.
So today this tripping on high current or voltage happened about 10 times in the space of 2 hours, we have patch clouds, with good sun behind them. The DC Amps as the sun comes out ere ranging up over 300amps. So the system is NOT shutting off solar production quick enough. These amps would remain above the 240Amps set in DVCC for a couple of minutes, which says even the MPPT’s are not limiting quick enough. This only happens when the sun has been behind a cloud and pops out again, and the system goes from 3 or 4 kw to 15kw in seconds.
So I tried setting Modbus control today, as per the documentation. section 4.3.14 confirming the changes in the web pages for the inverter, I also enabled the “Dynamic power reduction” and used the software power limit to 0.
This does not seem to be working, in the “Modbus” list on the Cerbo GX it only shows the 2 x Victron EV chargers. Does not show the Fronius units. It does however show an instance in the “Available Service’s” the Unid ID of 20 matches that of the Fronius 5kw, and the 21 is the 3kw
However the units don’t seem to be controlling via modbus, and the MPII is still ramping frequency all be it too slowly.
Any hints as to what I should see on either the cerbo or the fronius to show its working ?
In the short term I am running the ducted AC unit, draws 4 to 6kw that seems to take away the peaks nicely, and still allows the battery to charge and no alarms. On the upside the house in nice and cosy at 24 Deg C..
Yea, I have read the setup guide several times, but I cant get it working. Clearly I am missing a step some where. As soon as soon I set the software limit in the dynamic power to zero the Fronius inverter drops to Zero. The IP in a reserved IP for that address by the router. It always gets the same address. Screen shots below of the only two pages I can find to set …..that said the documentation always refers to ESS, I dont have an ESS system as I am offgrid, and as such ESS is not enabled.
I initially had the intention to turn ESS assistant on but it required me to set the Region in the MPII and I didn’t want to lock it to anything so just left it as Non ESS system. Is this something I am going to need to do to make this work ?
I do have the PV assistant on using the frequency shift which does work when the batteries are charging at a consistent pace, the patchy clouds is where I seem to get the most issues, with the solar on and off again rapidly.
It was working ok with the two AC coupled solar systems and one array on the MPPT, but recently I added another 8kw on the MPPT’s and have one more 4kw array to add, as we found in the winter months we were struggling to charge EV’s some weeks. So I think it was only working due to capacity limitation.
I see, the ESS will be of no benefit as when there is no AC-in present, the system essentially turns off ESS anyway, hence it not being necessary for an off-grid system.
Ok that is good that you have it set for fallback, but if we can get modbus working the the system should theoretically respond a bit quicker than with frequency shifting.
What do the CCL and DCL limits on the battery do when the battery is at 90%? Do they fluctuate drastically, make a big jump?
Another path you could follow if you just want to write values is using NodeRED to write the DVCC charge limit at certain SOCs to try and overide what the system wants, let me know if you want to explore that.
Thanks for picking up on the conversation. So I haven’t seen the CCL / DCL move (I have been trending it for a couple years as I had a JK that kept dropping in and out about 18m ago.), I trend it in VRM and it seems to be consistently as per DVCC settings. Well actually, I just checked the trend and over the past few days it has been moving. As per earlier in this post I adjusted the “Current sense” from off to on, and also reduce the charge voltage. The charge current has been self adjusting on two occasion in the past couple days. I turned it off at around 1300 today after all the trips, trying to go back to how it was and identify the cause. When this happened earlier today, the battery was around 60 to 70%, but as the sun was coming in and out from the clouds I was seeing swings over 300Amps on the Cerbo display.
As for Node red, I did use it some time ago in my camper on a PI for a better display than the Vic IOS, but haven’t touched it since. I saw another post this afternoon by a chap “Owen” who had a node red install doing exactly as you say altering the max charge limit. This is something I would like to look at, just not sure if the cerbo can take anymore load. Its currently connected to 2 x MpII, 2 x MPPT’s, 2 x AC coupled Solars, 2 x EV chargers, 6 JK Batteries, 6 Ruuvi tags, VRM, and HA all hanging off it. I have an IO extender here to control the battery room AC, just not installed yet…Will it be able to Run Node red and not falter on the coms?
Yeah this seems like normal behaviour for a BMS contolled battery. What we want to achieve is that when the battery starts becoming full, we should set a global charge limit low enough to accomodate spikes ie 100A I think, as you can see how the CCL jumps up and down, porbably due to cloud edge spikes from the MPPTs. If we set a hard limit that is lower than the spikes we wont destablise the system everytime a cloudcomes over and the battery is close to full, there will not be as much power that the MPPTs need to quickly dump with a lower limit.
As far as limits of the cerbo, I have a site with 6 quattros in a 3 phase parallel setup with energy meters, 3 Fronius AC coupled inverters and some other sensors and it has not slowed down one day running the Large OS for NodeRED, so I am sure you will be safe in that regard
Thanks for the reply and the confirmation on the Cerbo capacity. I am out in the morning tomorrow, but Ill get Node red installed after lunch and hopefully have a NR selector working. I am thinking the Values would be something like 200Amps under 80%SOC, 80-90% 170Amps, 90-100% 120Amps. At 90 % the Main house AC kicks in via HA script, as well as two smaller AC’s at 51hz and 51.5hz in an attempt to absorb extra energy and slow down the charge rate as is.
But I might need to add a reduction for production, ie if the Solar is over 10kw for 5min, and load is less than 2kw, reduce charge limit to 170A, reset on time ie 6pm, see how this goes.
I also need to not loose sight of the Frequency shift vs Modbus control. I need to solve this as well, to cope with cloud cover. I was thinking the frequency tables of the two devices dont match, and in the documentation it says dont worry about it as the MPII will cater for it by continuing to elevate the frequency. I think the victron starts at 50.2, where as the Fronius doesn’t start till 51 and ends at 52.7 if I remember correctly. If I mod the tables in the MPII assistant to match the Fronius, the initial moves will be picked up quicker as well.
What do you think of these settings / implimentations?
