Victron Multiplus running ESS.
DIY 17 cell 280ah lithium bank with JK BMS (no battery comms with Victron).
Battery SOC taken care of by a Victron smart shunt.
Victron MPPT.
MPPT settings are:
Absorption: 3.45v x 17 = 58.65v
Float: 3.375v x 17* = 57.38v
Fixed Absoption for 2 hours - tail current disabled.
The problem is when I hit 58.65 on the solar charge controller the voltage drops down to below 58 and never gets back up to 58.65 during the 2 hrs. Why? My understanding is absorption should keep the set absorption voltage and slowly lower the current.
I would like to see the voltage stay nearer to set 58.65v during absorption so that my balancer can kick in (set to start balancing at 3.44V). Switching from Optimised to Keep batteries charged fixes things but why should I have to do that?
Switching to “Keep batteries Charged” - perfect voltage except now Solar is at 0W. I dont want to have to use “Keep batteries Charged” in the first place.
If there are loads it will use the battery to feed them not keep the battery full.
If you need the battery to stay charged then you have to use keep batteries charged.
I don’t want to muddy the waters with my comment on the no solar when on keep batteries charged. I simply want to point out that the voltage is corrected to 58.65 the moment I switch to this mode. My issue is that optimised mode never gets back up to over 58V once absorption hits. Plenty of sunlight and low loads during this time.
This may be the issue - Is the BMV correctly calibrated for the bank (types of cells) would be my question.
I assume you could also use the BMV to manage the MPPT charging instead of just letting the MPPT assume things from the voltage?
Is the MPPT being managed by ESS as the settings on the MPPT will be overruled if this is the case.
I would enable tail current and see if this doesn’t affect the “charged detection time”
It could be that as soon as the Charged voltage gets to 58.65 the system thinks it needs to float (no tail current) and drops immediately. I not clued up with this so may wait for some “guru’s” to comment as well.
Edit: What are your BMS settings (since I assume you set up the JK yourself?)
Mr @TheTerribleTriplet can also advise here… The issue could be there…
Well, first thing is you must turn on DVCC. Don’t worry that you don’t have battery comms, it is still a good idea. A lot of newer stuff becomes available.
Then, when in “Keep charged”, it will tell the solar charger to charge at 0.4V higher than the Multiplus. So this will push the batteries slightly higher when the loads are low, but at 0.4V over 16 cells… it’s not going to matter. Then the Multi will feed power to the loads as long as the voltage is higher than what it sees as the charge voltage (which is 0.4V lower), ie, it will use that “overvoltage” margin to power the loads. The battery will then be kept full, while only the excess is used to power loads. That’s how the whole thing works.
But in this case, I feel something else is wrong, because…
So a current limit is kicking in or the solar charger is artificially lowering the charge voltage for some reason? That’s very weird. I thought of temperature compensation, but then you need a battery temperature sensor somewhere, and then the Multi would follow the same lower voltage. So that still does not explain it.
Only other things I can think of is 1) PV voltage isn’t high enough, it needs to be 5V higher than the battery voltage, 2) Solar charger is running hot and pulling back, or 3) solar charger is latching on to the wrong power point, partial shading might be involved, or 4) solar charger is broken.
If you are adept at a unix shell, and you root the GX, you can run dbus-spy and look at the .solarcharger service. Look for the path /MppOperationMode. This will tell you if the solar charger is tracking a maximum power point (2), or running limited (1). If it throttles back while reporting that the operation mode is 2, then it is probably tracking a local minima in the PV curve, and it might be partial shading.
Actually just turned OFF DVCC to see if the old stuff works better for me. Only reason DVCC was ON was I use the Limit Charge Current to 90A but I don’t really need that.
Generally my PV is 70V+ so that should be fine.
Doubt it - in an AC controlled room.
No shading whatsoever.
Maybe?
I can try this. Is rooting essential? No other way to get that number?
Are there any combination of settings in an ESS system where the primary Voltage does not come from the installed smart shunt? I want to be sure the system is definitely using the shunt V over the reading taken by the Multi or Mppt.
You can get it via mqtt as well. Or via modbus, register 791. Look under Settings → Services → Modbus TCP etc etc to enable that service, and to get a list of devices and their unit addresses (since multiple solar chargers are possible, you need to fetch register 791 from the right unit id).
There are two reasons why you get different voltage measurements. The obvious one is calibration, there will always be a small difference between devices. The second is voltage drops on the cable.
During charging, you would expect the Solar side to measure higher than the battery side, since the voltage drop is always higher upstream if you catch my meaning. So this is probably calibration.
In an ESS system, the solar chargers are always synchronised with the Multi. The voltage measured by the Multi is shared with the solar chargers, and they calculate a delta to account for any calibration differences.
If you also turn on SVS (shared voltage sense) in the DVCC settings, then the Multi will also be synchronised with the battery monitor, and all your voltages should then agree. That might be something to try. But if this works, it still says there is a rather gross voltage difference of as much as 700mV. I would not say that I’ve never seen that before, but it is exceedingly rare.
In any case, I think you’re looking in the right place by asking about voltage. Maybe charging actually stops at 58.65V, as measured by the solar charger, but at that time the shunt actually measures lower, so it looks (to the untrained eye) as if it stops prematurely.
Still strange that it does go up to above 58V before mysteriously drooping down. My theory does not account for that.
Under system setup, you can explicitly select the Multi as the battery service. Then it will use the voltage, as well as the SOC from the Multi instead of the SmartShunt. This is not advised for long term operation (because the SOC will be less accurate this way), but as a test it’s fine.
With DVCC off, all devices use their own voltage measurement (in an ESS system, however, the charge voltage, that which the solarcharger aims for, comes from the Multi, not from internal settings).
With DVCC on, SVS off, the solar chargers use the reading from the Multi, but everything else uses their own measurement. This is usually the most stable configuration, you just have to accept that the displayed battery voltage might not be quite on the dot what you asked for.
With SVS on, the Multi synchronises with the shunt, and since the solarchargers also syncs with the multi that means everything agrees.
In all cases, the voltage you see on the user interface comes from the shunt. What voltage your other devices use to make their decisions depends on the settings.
You can, from the GUI, go into each device and look what they measure…
