Yes, there have been a few posted research papers here on the forum.
Summary
absolute min 2.5
Absolute max 3.65
Most bms use range 2.8/9 to 3.55v.
Some allow a short period overshoot to 3.65 and this allows a short absorption time to make sure full charge and balancing happens.
Its a pity this thread has been hijacked from “how do I set the charge levels used by the MPPT in a typical victron battery system” to “how to charge a LiFePO4 battery”.
However to properly answer this I have taken a look at the literature as well as considering general electrochemical basics.
Like most/all battery systems the LiFePO4 battery has three zones. Overdischarged (generally taken as below 2.5V/cell) below which irreversible physical damage occurs to the cell. Overcharge (generally considered as over 2.65V/cell - although many disagree and prefer lower) where irreversible physical damage occurs to the cell. Useable zone, somewhere between these where life is maximised, and irreversible damage is slow. In a constant current regime this shows up as (at least at the upper end) a sharp knee in the voltage). Note the sudden rise in what had been a steadish voltage. We have now moved off the linear part of the charge curve and moved to the non-linear part. IMHO this is not a good place to be, even if damage is low it would be best avoided if not needed. The second voltage spike a few hours later is the MPPT also doing a full balancing, completely unnecessary and can only harm the battery, if even only a bit. I can control the daily grid charge but NOT the MPPT one.
I note that under dynamic ESS a full balancing charge is only done weekly (despite the huge battery charge-discharge cycles) as they specify, a regimen I totally concur with.
There is a reason why LiFePO4 cells are uniformly stated to be best stores at 50% charge, and that 100% charge is best avoided for storage.
The very first reply was how to stop it reaching absorption and therefore balancing. As simply as it could have been answered, however from your side there is no feedback as to whether you have tried it and confirmed is working or not working.
As you can see its set to 55V, but this does not work.
Hence the thread continuing. You have virtually my complete setup visible so you can check if I have already done it or not.
Are you sure that is the actual battery voltage or is that the ve bus voltage because the two can be very different when feedback is enabled and power is being processed.
277 is the quattro address is why i ask.
I imagine at 57 or 58v the ve bus bms would have triggered a trip or disconnect at that voltage as it is out of bounds.
Smart shunt show 53v on other pictures. Etc.
Obviously i only have second hand information here and can’t zoom into the graphs like you can. But it just appears as if the information mix up here is something to consider.
I asked you before, what is ‘feedback’ and where is it set?
All the DC voltages are almost identical (95MM^2 and short cables) and simply overlay on the graph, ditto battery/system voltages. The battery voltage at balancing is 56.9 (typically) as set in the quattro. Hmm, I note its slightly higher when balancing from MPPT even though the current is less, but not really significantly so. Both are way over 55V as set in DVCC.
Smart shunt shows 53V?? Where under what conditions, charging, discharging or balancing (excluding massive discharge caused by dynamic ESS). Not a figure I recognise. Its 53V discharging, 54 float, and 57V balancing.
PS I am using the pictures displayed here, not directly off my system, that would be confusing. What you see is what I see.
The only reason why the AI says this is that it is not good for an LFP battery to be stored for longer time at full charge. Store it at a SOC between 40 and 60%.
Nobody is charging his LFP battery to a real SOC of 100%. We all use algorithms that set the SOC to 100% at i.e. 55.2V (3.45 V per cell) instead of 58.4 V (= 3.65 V per cell). At 55.2 V the battery ist probably at at real SOC of 98%.
It is absolutely no problem to charge a LFP battery fully every day. But even when you are afraid to do so: How many cycles does you battery count per year? Mine counts about 150 cycles per year, and I doubt very much that many count mor than 200 per year. Do you really expect that your battery will get 40 years old? Who cares if it get’s only 30 years?
And here is what the specs say: “The cells have a lifespan (up to 80% capacity) of 8000 cycles when using a pressure of 280-320kg and charging at 0.5C (0.05C termination current up to 3.65V) 0.5C discharging (up to 2.5V)”
Charging from 2.5V to 3.65V is charging from absolute 0% SOC to absolute 100%.
Apologies i assumed you were trying some kind of feedback based on this statement.
Does this mean you aren’t exporting any solar?
This was the screenshot i was thinking of
For about 9 months of the year two balancings/day (say) that’s 500/year. The victron suggests (from memory) 3000 80% cycles BUT I do not believe this in practice, for one thing there is a continual micro- charge/discharge with a cycle time of perhaps a minute. Anyway, why wear your cells out for no reason??
Note balancing is at 3.55V/cell or 56.8V in a standard Victron system. There is argument for dropping this but one may fall foul of the internal, non settable, balancing system.
Once again an answer that doesn’t answer the actual question I pose, which should be simple. I cannot imagine victron have not considered this, particularly as dynamic ESS aims for one balancing a week.
Yes, I am exporting solar. At present 20-30kWh/day.
Sorry but I cannot see any enclosure of the particular image you mean.
The cells don’t wear out when you fully charge them every single day. Your cells will show less voltage drift when balanced at every cycle and this will increase the health of your battery.
And once again: since your question is based on wrong assumptions I will answer with my different opinion. It is not to convince you, but to prevent others to make the same wrong asumptions as you.
Ok so you have a Ve.bus BMS V2 then this applies:
the presence of a VE.Bus BMS V2 does not control the charge voltage of the solar chargers, Inverter RS, Multi RS or a Multi.
• In an ESS system, the Multi controls the charging voltage of the solar chargers, Inverter RS and Multi RS using the configuration made with VE.Configure or VictronConnect. In other words: The charge algorithm must be configured in the Multi.
• In a non-ESS (off-grid) system, the solar chargers, Inverter RS, Multi RS and Multi follow their own internal charge algorithm. Here, all devices must be set to the appropriate lithium charge algorithm.
For this to work DVCC must be enabled!
Feedback is the CVL value, as far as I’m aware of.
BAnders, I agree, this is how the system seems to work. Its a good way, one ring to rule them all, and in its day quite forward thinking BUT I cannot control balancing if I have an MPPT as well.
It does need a small tweak and that is to have a settable delay between balancing. Heavily thrashed batteries may need rather regular balancing whilst lightly used ones may need far less. At the very least it should be possible to set balancing using a variant of the ESS Scheduled charge levels to include sources other than just the grid or to actually set balancing time intervals. As it stands I can easily control balancing this way from the grid, but the MPPT is uncontrolled.
Which is where I came in waaay up the thread!
Bizarrely the cerbo could read most victron battery cells via bluetooth assuming its reasonably close (which will often be the case) and decide when a cell is deviating from the rest sufficiently to require balancing. This is relevant for victron batteries, which have a very basic equalisation system. This would be very cool and very easy to implement.
Another possibility is active balancing being a retrofittable option to replace the existing one.
So, I think you are actually saying that there is no way, which is a strange gap in the usual victron setup system. Normally rather all-encompassing even for trivial settings.
Victron batteries feature a highly unusual configuration; nevertheless, they remain subject to the same laws of physics as all others. In the case of LFP cells, specifically, cell drift only becomes detectable above approximately 3.4 V per cell. Were balancing to commence at a lower cell voltage, it would frequently prove highly counterproductive.
However, at 3.4 V, an LFP cell is already charged to approximately 95% capacity. This implies that it is futile to attempt to balance the cells of an LFP battery that is less than fully charged.
Hallo Funktioniert eigentlich ganz einfach du brauchst nur die Ladespanung begrenzen im 48 V System habe ich auf 54,5 V eingestellt, (ca 80-90%) Also: Einstellungen - System Setup - Ladekontrolle - Ladespannung der Batterie begrenzen - maximale Ladespannung (54,0 - 54,5V sind rund 80%) Maximale Ladespannung sind bei mir 56,8V Kannst du auch bei Laderegler (Solar) einstellen. alle paar Wochen oder bei bevorstehenden Schlechtwetter einmal hochstellen. So bleiben deine Ladung der Accus auf die 80% Während der Nacht entlade ich dann auf 40% bis 60% (Sommer - Winter) im Winter immer auf maximal belassen. Mein Ladebereich der Accus liegt bei mir immer über das ganze Jahr zwischen 40 - 80%. Grüße Roman
This does not seem to work in an all-victron system, quattro, rs-450/100, BMS-2 and victron battery. See graphs above all used with enclosed settings, Vcb = 55V.
It would seem that the quattro BMS over-rules this.
OK I have found the details on the smartli internal balancer. It would appear I am wrong and there is some internal active balancing, which is (to be honest) a relief. Judging from the PCB its not that powerful, the spec sheet suggests 1.6A (=~80W). Reading the manual it would appear that cells are bypassed/actively balanced below 3.3V (but probably at quite low power).
Balancing is complete when Vcell~3.55V (the famous 56.8) AND current gas dropped “even further below” 1.5A (presumably double that for me as I have two strings).
NOW if the charging is controlled by the Quattro and its using VOLTAGE and NOT current during the balancing phase and with the battery generally already very close to balanced before the start of another cycle then a very wasteful (in energy) and probably incrementally damaging and unnecessary balancing is BAD. It is never going to be good.
Enclosed the solar rebalancing cycle. I have reduced this to 1hr and its clearly fully balanced in 15 minutes (down to 0.4A per string). The rest of the balancing does nothing but waste power and damage the battery (if only minutely). Now we all know why the Quattro on its own cannot be used to sensibly measure battery charge current, only the smartshunt can. BUT the proper way to balance is as per the manual, which is 56.8V until charging current drops below some figure (probably about 0.5A/string.
So ideally victron should change the Quattro/Multi to perform a full balancing only after a set period has passed or perhaps (where a smartshint is present) when the current has dropped below a set figure. Below I have cut and pasted the section from the manual and I note it suggests monthly or weekly, not twice daily!
==========Here relevant manual entry============
8.3. Cell balancing
Why is cell balancing needed
Though carefully selected during the production process, the cells in the battery are not 100% identical. Therefore, when cycled,
some cells will be charged or discharged earlier than the other cells. These differences will increase over time if the cells are not
regularly balanced.
When fully charged, the current through a lithium cell is almost zero. Lagging cells will not be charged further unless they receive
“help” with this from cell-balancing electronics.
How does cell balancing work
The battery has built-in “active” and “passive” cell balancing. This ensures that all cells will be balanced. Each cell voltage is
monitored, and if required, energy will be moved from the cell(s) with the highest voltage to the cells with a lower voltage. This
process will continue until all cell voltages are within 0.01V of each other.
When does cell balancing take place
“Active” cell balancing begins when the first cell reaches 3.3V or less for severely unbalanced batteries.
“Passive” cell balancing starts when the cell voltages are 3.50V. This can happen only during the absorption charge stage, as
during this stage, the charge voltage (14.2V or 28.4V) is high enough to for the cell voltages to also be sufficiently high to allow
smaller cell differences to be corrected.
The cell balancing process is nearing completion when all cells have reached a voltage of 3.55V and the charge current has
dropped below 1.5A. Balancing is complete when the charge current has dropped even further.
How to ensure that the battery remains balanced
A 2-hour fixed absorption period is recommended for lithium batteries so that there is enough time for cell balancing to take place.
It is important to regularly fully charge the battery. This so the battery spends enough time in the absorption stage. A full charge
once a month should be sufficient. However, there are some applications where the cells will become unbalanced quicker than
usual. This is the case when the system is used more intensively or if the battery bank consists of multiple batteries in series. To
ensure a well-balanced battery, a weekly full charge is required for:
• Systems with a battery bank that contains batteries that are connected in series.
• Systems that are charged/discharged every day or a few times per week.• Systems that have high discharge currents.
• Systems that have short charge periods or low charge voltages.
It is not possible to speed up the cell-balancing process
Please note that a higher charge voltage will not speed up the cell balancing process. The cells are charged by current and not by
voltage. Feeding current into a cell will cause the voltage to increase over time, but this is a fixed process. Applying more voltage
will not speed this process up. In addition to this, the balancing speed is determined by the maximum current rating (1.8A) of the
active and the passive balancing circuits.
================end of manual entry
When it is an active balancer then 1.6V is not so bad, but since balancing is done for single cells this is just 5.4 W.
No, I don’t believe this, since cell drift normally occurs only above 3.4V. Balancing at 3.3V and below will often be counterproductive. In many cases an active balancer will then uncharge a cell which indeed has a lower SOC then that cell, that will be charged. When one cell is at 3.30V and another is at 3.31V nobody can tell which one needs more charge current than the other.
The standard value used by most BMSs is an 1 hour absorption period, and in most cases even 15 minutes will be enough. A more sophisticated approach is that of the Victron Smartshunt, which doesn’t decide by time but by the dropping current at fixed charge voltage. In other words: when an LFP cell gets close to 100% SOC it’s internal resistance will rise, and when you have a fixed charge voltage then according to Ohm’s law the current will fall. In the Smartshunt’s settings the user can set the threshold, most users probably use 2% of the nominal capacity. For a 280 Ah battery this is a charge current of 5.6 A. The charge voltage mostly is set to 55.2V for a 16s battery, this is 3.45V per cell.
Odd, I have a smartshunt and can find no way to use it for charging/balancing. Please explain how I do that in my system.
You know very well that the Smartshunt doesn’t balance your cells. And I did not say that it does.
Just take a look at the smartshunt settings. The first dialog field is the capacity (in my case 280 Ah), and the second is “Charged voltage”. For my 16s battery I did set it to 55.2V. And a few steps down in the dialog you will find the “Tail current” setting, I have this on 2% (which means a current of 5.6A = 2% of the nominal capacity). And just below there is the “Charged detection time” setting. I have this on 3 minutes.
Balancing is done by your battery’s BMS, or by any other device. My BMS starts balancing when the first cell reaches 3.42 V. I takes about 15 minutes after the battery reached 55.2V (= 3.45V per cell) to bring the charge current down below 5.6A, and 3 minutes later the Smartshunt will set the SOC to 100%.
The BMS however continues absorption charging for another 45 minutes, which definitely is not necessary.
