Note that I am filing reports from a phone, so embedding photos in the right place is an exercise in doing things in the reverse order, and typing is mostly correction and correcting the corrections.

Posted by: @vaugi

↑

It could be worth turning the solar DC off though for half an hour so you can get no solar addition intefering.

I think the big switch in the sky is being activated. Give it a few hours more. Mind you the graph I attached in a previous post had a near constant line of power being taken from the batteries.

Posted by: @vaugi

↑

also assume you CT clamp is the right way round

It is the right way around after I pointed this out to the installer last week…

Posted by: @transparent

↑

 

• max voltage, prior to allowing the battery to drop to float voltage. Usually set at 56v-ish
• float voltage at which the battery will be held following a 'charge cycle'. Usually 55v-ish

 

I have float set to 53.0V (3.31V per cell), absorption to 55.2V (3.45 per cell) and equalisation is disabled.

Posted by: @transparent

↑

if you want higher efficiency, then you need to look out for 'transformerless' inverters.

I was seeing pretty good efficiency reported for the dc/ac transformation (200W losses for 5kW being fed to the house). Sunsynk use IGBT devices (doesn’t everyone?). Note that when in Lithium battery mode there was no such continuous draw.

@chickenbig those numbers for absorption/float look sensible…

Posted by: @transparent

↑

So if you want higher efficiency, then you need to look out for 'transformerless' inverters.

The picture I was commenting on showed 1062W from MPPTs, 140W from battery being converted to 1125 on the AC bus, resulting in 286W home load and 846W being fed to the grid. Losses in dc/ac transformation are 77/1125 = 6.8% so not terrible. However I am unsure why the inverter would take power from the batteries for it to be sent to the grid.

Just nipped out to the garage to take some shots of my settings.  5kW version, only connected to "Grid" (no connections to "Load", nor "Gen"), PV attached.

Main power flow Screen:

Lithium Battery page, the Charge and Discharge is really a Global setting on this page.  If you change charge rate on the "Batt charge" page, it overrides this.  Also the BMS overrides and can charge lower when at high SOC.  

I haven't changed anything on this page - as I understood it the right hand side settings don't do anything with Lithium setting (?).  Grid Charge tick allows inverter to charge from grid (@chickenbig, perhaps not being ticked, it also stops your inverter discharging to grid  (i.e. your loads)?

Low battery settings, these are warnings in the inverter, and will override the BMS if higher.  E.g. if shutdown is 20%, it'll stop before BMS own low voltage protection if that setting is lower.

The LiBMS page showing what is reported by the BMS.  Errors will show here if not communicating properly, and it will look different.

The System page - the heart of the control really.  I've got mine set up to charge battery upto 50% during Octopus Intelligent period (actually I miss half hour for reasons transparent mentioned about timing not exactly 23:30, but also I don't need the full slot at 5kW, so have the inverter at 60A charge.)  If battery is already over 50% it will discharge in this timeslot until it reaches 50% at which point it charges to hold at 50%.

All other time slots will discharge the battery until it gets to 20/15%.

On the system 2 page I've got it set to Zero Export so it works to the CT clamp power, only discharging enough to cover load.  If you want this, but allow Solar to be exported, you can click Solar Export.  The other two I believe relate to the Load port (essential loads) not the Load of the house.

(I'm posting this @chickenbig as it might be you want to change to match this to test - it seems like you know what you are doing but thought it worth a try).

The 20W is what the inverter will take from the grid at all times to allow it to control the Zero Export I understand.

Final shot for completeness of the overview page with the figures, this page is useful to check CT clamp is correct on first install.

I might post BMS screenshots for completeness sometime, might be useful to new users.

I have Sunsynk inverter and Seplos Mason with 10E and I’m also slightly perplexed about the comms between the two.

speaking to Sunsynk today I’ve been told that if Lithium is selected in the Battery setup page the Float and Absorption values are not used.

this presents the question; 

What does the Seplos then use to find the stage of Charge?

my view is that once the Overvoltage has been triggered then this is top. It then uses internal shunt to count amps in/out from this point. 

This to me presents the problem that the BMS is always then looking to trigger this voltage to find 100% 

will this settle down and charge to just under the voltage? Time will tell.

there is a lot to be said about using Voltage rather than % however, that said, the easiest way would be just to set the max charge to 95% therefore hopefully never triggering an over voltage.

OK, quick situation update. Having tried "AGM V" I see that approach was somewhat doomed to failure. Voltage is not a good indication of charge level of a LiFePO4 battery whereas it would be a better one for AGM. With "AGM V" enabled I saw batteries getting depleted to 85% while solar was being exported; effectively the battery had a slightly higher voltage than float so it continued to discharge. Maybe using a slightly higher float voltage would have helped, but then one would be fine-tuning a parameter according to the characteristics of the battery and wiring which does not seem like a good idea. In addition the "System Mode 1" settings were not particularly helpful since one could not set the charge.discharge voltage higher than Float.

Perhaps using "AGM %" (which I suppose uses Coulomb counting to measure the state of charge) would have resulted in something better, but instead of that I tried to reconnect the BMS communication cable and try and get the proper way to work.

Giving some context, the wifi connection to the inverter was intermittent until Thursday afternoon (power-line adapter did not work well with Zero Export mode at low rates of solar production, probably due to noise on the mains). Also "System Mode 1" settings kept on getting reverted by the Sunsynk Connect cloud (due to a setting the installers decided to configure). Solving these problems I could upgrade the inverter firmware (more on that later) as well as start to configure the "System Mode 1" properly.

The key was to set SoC to 15% with Grid and Gen unticked. Simple.

I have "Solar Export" ticked in readiness of SEG, along with limiting to 3680W export (*cough* G100 import limiting *cough*).

Battery-wise there is nothing special going on.

So the system now compensates for home consumption and prioritises battery charging over export. One slight niggle is the slight grid draw (~40W) only when charging batteries from solar.

Posted by: @vaugi

↑

If you change charge rate on the "Batt charge" page, it overrides this.

Interesting ... the semantics of these current and voltage limits seem unclear; presumably the BMS values define hard limits but I would have hoped for a single place within the inverter to set the limits inside of which (inset) other settings (grid charge current etc) operate. For me this would be good since I have two batteries connected to the one inverter via 125A fuses and 25mm2 cable; at the moment the BMS tells the inverter that that charge/discharge limits are 200A and I think I set the limits to 100A per battery.

Posted by: @vaugi

↑

The other two I believe relate to the Load port (essential loads) not the Load of the house.

I understood "Priority Load" to be related to the order in which energy is sent. With it ticked then Load (as well as house consumption) is prioritised above batteries and that above grid export. Sunsynk really need to sort out the labels as Load can mean home usage as well as the UPS output.

@vaugi I noticed that your "System Export 2" page does not have a "Fast_Zero_Export_Mode" check-box; are you on the latest firmware (MCU 3384)? Per https://www.sunsynk.org/forum/technical-support/sunsynk-sg01-lp1-not-on-approval-dno-list-for-use-for-export-in-uk you need to be on 3384 (and change the label on the side of your inverter) for it to be ENA approved.

Totally disagree with what you say about Voltage not working to control the SoC. It’s the single most important thing with regards to LiFePO4 cells.

what had you set float and absorption too?

anything under 54.0v is pointless due to this voltage being constant during charge or discharge and there is no way to know if the pack is at 5% or 95%

the only way to do this is set the total absorption voltage at 55.2v (3.45v per cell). This will have the added bonus of active balancing between cells.

if you float the voltage at this same level you can continue to balance for whichever time you want. 

see below result. 55.2v maintained during off peak and this equates to 100% register in both seplos & sunsynk 

Posted by: @chickenbig

↑

Voltage is not a good indication of charge level of a LiFePO4 battery whereas it would be a better one for AGM.

Indeed! ... and all the more so if the inverter/BMS is trying to measure that voltage whilst simultaneously charging the battery at the upper part of its capacity!

If you want to check the SoC, then install a separate colomb-counting meter with a shunt or Hall-effect sensor which covers both/all your batteries.

You can also get the same meter from Amazon UK at a very wide range of price points from sellers who may not appreciate which shunt to send you!
Check what you receive!

Posted by: @chickenbig

↑

I have two batteries connected to the one inverter via 125A fuses and 25mm2 cable; at the moment the BMS tells the inverter that that charge/discharge limits are 200A and I think I set the limits to 100A per battery.

Can you expand that description please?

If you're setting current limits per battery, then you're doing this at each BMS. Correct?

Just think what happens when one BMS detects a current above that figure.
It will switch off the Charge MosFET, leaving the other battery to take all the current being sent from the inverter...
... which means it too switches off.

I'd like to be reassured that the first BMS signals the inverter before switching off its Charge MosFET, and that the current is reduced accordingly.

But does that happen?
It's a logical strategy which, if I was the product manager, would be insisting gets included in the software linking BMS units to an inverter.
However, I remain sceptacle that inverter designers consider such sensible features.

Posted by: @chickenbig

↑

you need to be on 3384 (and change the label on the side of your inverter) for it to be ENA approved.

That's a good point.

However I was informed yesterday that G99/G100 regulations can't be applied retrospectively.
So the legality of this depends on when @vaugi installed that system.

Even so, I would hope that all Members of this Forum would want to keep their installations within the current rules as far as they are are able to.
These regulations are being constantly reviewed and updated by ENA because there's a need to do so.

To help look after or UK energy supply we have a moral duty to be as DNO-friendly as possible.

@transparent I disagree. Voltage is a very good way to measure the SoC of these batteries (at the top end I add). As long as the cells are balanced we know that these will peak around 3.6v / 57.6v 

if we can keep cells balanced we know that the pack will be full at this point.

of course, coulomb counting is perfect when we want to charge to anything less than full but we have to remember that there is an error in this and overtime this will get worse hence we should charge the pack fully based on voltage let it balance for a decent time 

I'm going to keep 'approving' these posts from @sikorsky92 because the points being made are technically challenging 🙂 ... even if somewhat terse!

Yes, voltage is a good way to set what we wish to define as Full within the overall SoC, provided that:

• we make that measurement at the top-end of the voltage range (as @sikorsky92 mentions)
• we do so during a period when the current is low (otherwise it's not a reflection of the actual battery voltage!)
• and when the BMS has had sufficiently long time to apply balancing (but that time will vary according to the specification of that balancing operation)

But... I'm not happy with regularly charging a LiFePO4 battery to such a high point.

a: it will reduce the number-of-cycles and/or lifetime capacity of those cells

b: it's a strategy which requires a level of technical understanding beyond the wider group of potential DIY battery installers. And we want a much greater uptake of what we're discussing here throughout the whole population.

c: it's not a 'natural' way in which we would use batteries that have both solar and grid charging. We would tend to see these batteries used between 30%-80% SoC on most days

d: It's difficult to arrange with the set of controllers presently available. We want to look at the day ahead and evaluate how much solar we are likely to achieve. Then we add enough charge from the grid during the cheap-rate period of the preceding night to ensure we've bought in sufficient for our needs.

If such a controller existed, it would be nice to have one which took the battery to 'Full' once a month in order to recalibrate the SoC measurement.

Posted by: @transparent

↑

Can you expand that description please?

If you're setting current limits per battery, then you're doing this at each BMS. Correct?

Sepois Mason batteries are in master/slave config. Individually I set 100A charge/discharge limits per BMS (IIRC), plus 100A charge limit at the inverter. Each battery has its own 125A fuse before the connections to the bus bars, then a 125A fuse between bus bars and inverter.

Posted by: @transparent

↑

help look after or UK energy supply we have a moral duty to be as DNO-friendly as possible.

Indeed the rules about what happens when under and over frequency are important for predictability of when distributed energy resources drop out. This is why I have taken the time to check all of these settings the installer neglected to check/set. With great power comes great responsibility.

Posted by: @sikorsky92

↑

Voltage is a very good way to measure the SoC of these batteries (at the top end I add).

Although near 100% it is an excellent indicator of SoC of individual cells, voltage is not so useful for SoC in normal working range.

Posted by: @transparent

↑

regularly charging a LiFePO4 battery to such a high point

This is why I was trying for a float voltage of 53.0V which was slightly lower than the natural voltage the battery rests at when at 99% or so. I chose this lower voltage to avoid overcharging the battery.

My conclusion from this is that “AGM V” combined with a lower float voltage is not a good setting for lithium battery energy storage with Sunsynk. Others may differ in their settings; do not let my LiFePO4 choice invalidate yours.

@transparent thank you I’m genuinely here to help.

To give you some background, I’ve supplied over 100 seplos units to the U.K. including various inverter configs to go with these.  I’ve also been here from the beginning using various BMS (JK, Daly etc).

Where the battery suffers degradation is both heat and storage at peak voltage for any length of time.  Where we charge to 100% and then virtually discharge has no significant degradation on the cells.  Especially since at 3.45v per cell which is not technically 100% but pretty close to it and especially since we can balance cells at this level.

are we still speaking about seplos BMS here and a way to get this to work or just prismatic cells as a whole? 

lots of people are now using purely off peak charging to save money these days without solar so this works extremely well. The battery reaches 55.2v about and hour before starting to discharge.  

good discussions