I don’t know if this document has been posted here yet, but it seems tailor made for local conditions and having a Sunsynk 8kW I’ll surely give it some serious attention - although written for the Sunsynk the basics are applicable to all similar setups; enjoy!
SystemModeSetup(1).pdf (1011.3 KB)
CREDIT WHERE IT’S DUE…
2 Likes
The part I figured out (luckily very early in my solar journey) is that you don’t drive your car around the block and go fill it up every time afterwards.
So what if the SOC never reaches 100% everyday? In my case this past month I have never been able to do that and SOC never gets past 85%. The only reason I believe why this could be important is if you have problematic cells that goes out of balance.
If all is well in that regard, happy to continue like this until the sun is able to take care of things again which should be within 2 months. If not an absolute inclement weather or load shedding emergency, I absolutely refuse to “top up” batteries using the grid just for the sake of it.
What I have seen though is that once the SOC has reached the highest it could (lets assume this 85%) and then a high load is switched on, the SOC will drop with about 8 -10% whilst it is on and then revert back to 85% if only on for a short while - kettle boiling. Could this be a cell balancing thing?
Nope, that is what happens with BMS SOC, volts drop, SOC drops.
Load goes off, volts jump back, SOC increases.
That is why it is a good idea to have a coulomb counter to help calculate what goes out of the battery/gets put back.
Thing is, the large drop in SOC only happens if the batteries did not get full everyday.
In other words for couple of days in a row never reaching 100%.
So the GX will only know then what exactly is in the battery once it is fully charged and calibrated itself?
If starting from 100% and going down, SOC percentage as VRM displays it gradually tinkers down in 1% increments.
That seems weird. What batteries are these again?
BMSes sometimes have voltage waypoints, where they reset to a better estimate based on the voltage.
To make an extreme example, There is no way a cell that is above 3.35V is below 50% SOC. So a BMS may well adjust its estimate upwards based on such a voltage. Similarly there is no way a cell at 3.2V is above 50%, so the same would happen in reverse.
This sounds like the coulomb counter causes an initial 8% to 10% drop, and then afterwards as the voltage rebounds, the BMS resets to a higher SOC.
It suggests that in the long run, your battery is probably below the estimated SOC, and you will see severe SOC drift over time.
Rather be safe. Make sure it gets a full charge once a week. The one chemical signal you cannot argue with is a cell voltage of over 3.45V. That cell is full or close to full. Anything else is a guess.
GX is not a battery monitor. It is either the BMS, the inverter itself OR a BMV/Smartshunt, the source of the SOC.
What plonk said. Follow that advice and get it to 100% SOC once a week.
Freedom Won batteries, 4 of them.
To get them full at least once a week is not an issue, I am mostly able to do that over a weekend. But point taken.
Is such big SOC drift uncommon for batteries less than 2 years old?
OK, I sense there might be some confusion here. I am not talking about SOC drift where different cells/modules in a battery end up with different SOCs. Rest easy.
I’m talking about a battery that thinks it is at 85%, but in reality it is only at 55% (for example). IE the SOC estimate has drifted away from the real SOC.
This has nothing to do with the age of the battery, but with how it is used. Small errors always creep in over time, and the only time those errors get “reset” is when you charge to 100%.
Let me tell you of an actual support case I had. The customer had a demo system in his showroom. This system had 4 x BYD B-Box modules and a Multiplus installed. The customer wanted to know why the system sometimes (without any instruction) suddenly discharges the batteries really quickly.
Inspecting the system showed the SOC dropping off a cliff, but there was no corresponding power or current draw to account for this. After digging around some more, I found out that these particular batteries cannot really sense accurately below 1 ampere, and since there were 4 modules this meant that below 4A (or about 200W), it could not sense any energy leaving the battery. Since this was a demo system with very little load… the battery would slowly discharge without the BMS noticing anything. So the SOC would remain at 100% for days on end, and then suddenly drop off to zero as one of the cells registered a low voltage.
This is of course an extreme example of SOC drift, but that’s the basics of it. Any inaccuracy in current measurement results in either an over- or underestimation of how much energy moved into- or out of the battery, and these small errors add up over time.
When combined with a battery that has waypoints – the FW Lite series does, I’m unsure about the Pace-BMS – the battery sometimes corrects its SOC (or attempts to) based on cell voltage measurements. This could, under some perfect condition, actually make things worse, by repeatedly pulling the estimate back up to 85% (based on cell voltage) while it has actually drifted below that.
The more time passes since the last 100% charge, the more SOC drift there will be. That’s just how it is 
Thank you Isaak, your explanation makes sense to me.
Yes I believe the Freedoms have a Pace BMS.
When there are days without good sun, I use those times to use Eskom to “balance” the batt again.
Cause Eskom power is more stable compared to MPPT’s and loads on/off and clouds.
I set it to Keep Charged, drop the amps to like 10a, and let the batts take their time.
Just a thought.
I’ve just been using BatteryLife and it’s been working amazingly for me especially since the tumble dryer was added with one or two more things increasing consumption compared to last winter.
Back to the document. The principle highlighted in the document is don’t charge from the grid if you have a chance to charge from solar later.
Another principle I think may be relevant if you are chasing efficiency is to not let the battery discharge if you need to charge it from the grid later. If you will charge it from the grid then you may as well use the grid now. My thinking is that there are losses going to and from the battery.
So if due to loadshedding or something you are going to need 40% SOC charge tomorrow morning at 6am then don’t let your SOC fall below 40% over night then because you will only be able to charge it from the grid. May as well directly use the grid before 6am to avoid discharging the battery below 40%. Because sending that through the inverter twice is way less efficient use of kWh.
Spot on!
From AC to DC to AC has losses yes.
Use the panels.
as grid per kW is still cheaper than battery we must use solar and grid as far as possible - especially now that it seems to be available; for how long is anybody’s guess - I even have heard that the extra solar in SA is in fact necessitating ESKOM to curb production at stages… a new kind of ‘load shedding’ as there is no energy storage available on their side…
EDIT
my argument is based on total capital outlay for a system, of which batteries are the most expensive - and remember that lifepo4, although having sometimes 10 year guarantees, have not been tested over that period - so the ‘lifetime’ as with many other things boils down to the lifetime of the owner 
1 Like
I disagree, if you have a battery and it’s charged (using solar in particular) I would use it before using the grid. That’s cheaper because the marginal cost of that is basically 0.
Even if you don’t have a battery buying a battery would still probably be cheaper in the long run as you get to use your free solar at night basically.
This is the basis of all solar/battery grid tied systems.
Batteries cost when you buy them and they only have limited number of cycles. So the cost isn’t for nothing…
Well, if you are going to bring in capital costs then the whole calc is more complicated yes. But the point remains, we buy batteries because we want to be able to use our free solar later when the sun isn’t shining. This is generally more cost effective than using the grid.
So when you are optimising your inverter settings (the topic of this thread) one would generally use your batteries until they discharge to the minimum level you set, before you start using the grid.
It’s probably not so much capital costs, but rather an eroding of the value your capital bought.
Car analogy time… cause that is how we roll The TCO (total cost of ownership) is composed of:
- Devaluation
- Opportunity cost (money you lose because it is not earning interest elsewhere, or because you took a loan)
- Running costs (maintenance, fuel)
- Insurance
Then what you are supposed to do, is add all of that up, and divide it by the distance you drive… and then give up on the whole deal and call an Uber instead (let someone else take that hit!).
Or… as we all do, since these things are decided as much by emotion as otherwise… you ignore all that and do what makes you happy.
(Similarly, while a battery is nice to have, I think you can probably have 80% of the benefit and far lower cost per kWh by not having a battery… but what’s the fun in that).