Does the BMS have to be connected to the inverter?

[DrPhil]

I'm preparing to connect a 20kwh LiFePo4 pack (when it arrives) to a Solis hybrid inverter.

I've been advised by Solis support that I have 2 options on connecting them (unofficially that is. Officially they don't support DIY batteries).

1 - connect the batteries using the PylonTech option in the Solis menu. Use a Can cable to connect the BMS to the Solis and it should (but not guaranteed) communicate OK.

2 - connect them using the default Lead Acid setting on the inverter, and don't bother connecting the Can cable. The battery parameters can be entered on the Solis and it will then be able to estimate the SOC.

However a friend of mine who built his own battery has told me that there is no point in having a BMS, if it isn't communicating with the inverter because the BMS's only function is telling the inverter to stop when it's supposed to.

So who is right?

 

[John Frum]

The BMS's job is to protect the battery.
I don't know anything about this solis inverter but it sounds like it speaks one of the management protocols and the bms can use that protocol to turn the inverter off in order to protect itself.
You say "can cable" so I guess it speaks canbus.
If you link to the inverter in question we can learn more about it and hopefully give you better advice.
What is the dc voltage of the Solis inverter?
Also what is its continuous rate watts?
Is it single phase, split phase or other?

 

[DrPhil]

Thanks @smoothJoey

It is a single phase 5kw hybrid inverter, RHI-5K-48ES

Yes sorry I mean canbus. I've spent so long talking to the dudes on AliExpress who keep referring to it only as Can and Can to usb cables...

Apologies I'm very new to this so may show my ignorance.

I get that the BMS's job is to protect the batteries. So if they're speaking over canbus then the BMS will stop the inverter from charging.

But without that communication, can't the BMS shut down the battery and therefore the inverter will not be able to charge it so it will export instead?

 

[DrPhil]

https://www.google.co.uk/url?sa=t&source=web&rct=j&url=https://static.solartricity.ie/wp-content/uploads/2018/04/Solis-Hybrid-Inverter-Manual.pdf&ved=2ahUKEwiJ6b2UpszvAhXyXRUIHU8xBW0QFjAAegQIAxAC&usg=AOvVaw1EaINgeZcBq3MhfNDdZkyL

 

[DrPhil]

I've spoken to a few people who have LiFePo4 batteries connected to a Sofar battery inverter without the BMS communicating directly with the inverter and they've had no issues.

 

[John Frum]

Thanks @smoothJoey

It is a single phase 5kw hybrid inverter, RHI-5K-48ES

Yes sorry I mean canbus. I've spent so long talking to the dudes on AliExpress who keep referring to it only as Can and Can to usb cables...

Apologies I'm very new to this so may show my ignorance.

I get that the BMS's job is to protect the batteries. So if they're speaking over canbus then the BMS will stop the inverter from charging.

But without that communication, can't the BMS shut down the battery and therefore the inverter will not be able to charge it so it will export instead?

I didn't look at the doco you linked below.
but assuming the inverter will work without seeing another canbus hub and/or endpoint yes the bms can disconnect itself from the system for self preservation.
5000 ac watts * 1.5 low frequency inverter fudge factor / 48 volts = ~156 dc amps
156 dc amps / .8 fuse headroom = 195 fuse amps.
You should be looking for a quality 48 volt bms rated for 200 amps.
I don't know of any that I can recommend at that rating.
The workarounds are to use a contactor or solid state relay in conjunction with the bms.
That looks like top shelf kit.
Perhaps someone can recommend a bms that speaks canbus.

 

[DrPhil]

Thanks

I thought that the 5kw inverter had a 3kw limit for battery charge/discharge? The 5kw is the PV input limit.

Again I'm risking revealing my ignorance but I thought that the battery would be charging at 62.5a and the 16s 100a Overkill BMS would be sufficient.

Daly have a 16s 100a BMS also that speaks canbus, but it only comes with a Can to USB cable and the Daly folk were vague about whether it would have a matching protocol.

 

[John Frum]

Thanks

I thought that the 5kw inverter was 3kw for battery charge/discharge rather than 5kw?

Again I'm risking revealing my ignorance but I thought that the battery would be charging at 62.5a and the 16s 100a Overkill BMS would be sufficient.

Daly have a 16s 100a BMS also that speaks canbus, but it only comes with a Can to USB cable and the Daly folk were vague about whether it would have a matching protocol.

I didn't look at the doco much but I did see the word hybrid which I assume means grid tie, confirm?
Assuming you are correct 3000 ac watts * 1.5 low frequency fudge / 48 volts low cutoff = ~94 dc amps
~94 dc amps / .8 fuse headrooom = ~117 dc amps.
That is pushing it for a 100 amp overkill BMS.
Overkill is the one I recommend usually recommend BTW.

 

[John Frum]

I'm tempted to say go for it.
The Overkill BMSs have earned a good reputation.
The 48 volt one is quite new though.

 

[GXMnow]

I am running a Schneider inverter and a DIY battery bank. There is no communication between the BMS and the inverter, and it all works just fine.

The inverter and charging source should be programmed to not exceed the safe limits of the battery bank. It will stop charging at your full setting, and cut off inverting at your low battery cut off setting. The BMS will monitor the individual cells, and only cut off the current from the charger if any cells go to high, or cut off the load (inverter) if any cells run down too low. It is best if the inverter shuts off first and the BMS never has to cut off, but it is there if the pack goes out of balance and a rogue cell is going too high or too low, even though the entire pack voltage is still safe. The better balanced and matched your cells are, the closer you can run your charging and inverting to the voltage limits. Leave a little room at both ends and it will be fine. The BMS will only step in if the pack gets out of balance enough for you to need to deal with it. My cells are NCM so the voltage has a fairly linear slope and works great with only voltage control in the inverter. LFP cells are much flatter through most of their discharge curve, so it does not have as good of control on state of charge. If (when) I go to LFP cells, I might also go with a Battery Monitor that will calculate battery state of charge that the inverter can understand. If running on voltage control only, you would do need to leave a little more room as the cell getting close to full will start to increase it's voltage very quickly at the end of charge, and the low cell will drop voltage quickly at end of discharge. The worst case is one cell going rogue in a 16S bank. Full charge is 3.65 volts x 16 = 58.4 volts. But if one cell is topping out while the rest are still in the flat curve at just 3.4 volts, a single cell could actually be over 7 volts. 3.4 x 15 = 51 volts. 58.4 - 51 = 7.4 volts for the 16th cell. For this reason, I would not set the absorb voltage on a 16S LFP pack any higher than 51 + 3.65 = 54.65 volts. 54 volts would be even better. If all of the cells are all balanced, that is 3.4 volts per cell, which is 99% charged anyways. Then have the BMS cut off in a cell hits 3.65 and you are good to go on the charge side. The the low end side, give it even more room. Have the BMS cut off at 2.5 volts per cell (or higher if the manufacturer has a low limit above 2.5) but then set the inverter cut off at 2.9 x 16 = 46.4 volts or higher so even a fair bit of cell imbalance won't cause the BMS to trip. But even at this level, with 15 of the cells still at 3.1 volts, one rogue cell could be down to 0.1 volts without shutting off the inverter. Obviously, this should trip out the BMS long before this happens. This is why true State of Charge counting on LFP becomes important. You will need to monitor your cells for a few cycles to ensure your balance holds down low. Ideally, you want to have a system that leaves 20% in the battery to make them last longer. One weak cell could cause that cell to also age faster. Having good balance and matching between cells makes it all work so much better.

 

[DrPhil]

Assuming you are correct 3000 ac watts * 1.5 low frequency fudge / 48 volts low cutoff = ~94 dc amps
~94 dc amps / .8 fuse headrooom = ~117 dc amps.
That is pushing it for a 100 amp overkill BMS.
Overkill is the one I recommend usually recommend BTW.

Well I don't need to charge/discharge at full force, I can limit it a little to keep it on track for the 100a BMS?

 

[John Frum]

Well I don't need to charge/discharge at full force, I can limit it a little to keep it on track for the 100a BMS?

Humans are fallible I prefer to engineer them out of the loop .

 

[DrPhil]

Ps I forgot to mention, the batteries are CALB 3.2v 200Ah LiFePo4.

32 cells, to be joined in sets of 2 parallel and then 16s.

 

[DerpsyDoodler]

Humans are fallible I prefer to engineer them out of the loop .

That’s a set and forget setting, right?

 

[DerpsyDoodler]

Ps I forgot to mention, the batteries are CALB 3.2v 200Ah LiFePo4.

32 cells, to be joined in sets of 2 parallel and then 16s.

If you decide to have an Oprah moment, I would love 32 calb cells, or event 16. Just sayin...

 

[John Frum]

Ps I forgot to mention, the batteries are CALB 3.2v 200Ah LiFePo4.

32 cells, to be joined in sets of 2 parallel and then 16s.

Make them 16s2p which is 2 discrette batteries each with their own bms and Bo's yer uncle.

 

[DrPhil]

Make them 16s2p which is 2 discrette batteries each with their own bms and Bo's yer uncle.

I thought about that, which would be fine if I'm not canbus connecting.

How do the 2 separate batteries connect into 1 inverter though?

 

[John Frum]

I thought about that, which would be fine if I'm not canbus connecting.

How do the 2 separate batteries connect into 1 inverter though?

The batteries and the inverter are paralleled via busbars.
For best results put the inverter between the batteries to keep the resistance as well matched as possible.
Also the battery cables should be the same length.

 

[the_colorist]

I feel I should step in and clarify a few points with regards to BMS comms.

LFP and lithium cells in general have extremely low resistance and they will practically take whatever current you throw at them. This, however, isn't great and there is an optimum charging current curve, particularly at the end of the charge cycle. Typically the primary purpose of the BMS is to handle the charging parameters in realtime by sending max current and voltage parameters every few hundred milliseconds to the inverter. By doing this, it's able to optimize the charging curve and taper the charging current at the end of the charging cycle.

There are 2 reasons to artificially taper the curve at the end. The first one is to settle the chemistry gently so that the cell retains the most amount of capacity for as long as possible. Otherwise, the internal resistance grows over time and the cell capacity diminishes.

The second reason is to keep from overloading the internal cell balancers. So as the first cell reaches the target voltage, the BMS sends instructions to the inverter to back down the charging current to an amperage that will keep that cell and subsequent cells from exceeding the target while not overloading the balancers burning off (passive) or shunting power (active) other cells.

Instructing the BMS to power-down is actually the last resort. For discharge for example, prior to that, it begins sending commands lowering the usable discharge current until the max discharge current reaches 0A. Then the inverter shuts down.

Overall it's much easier on the cells and helps them to retain capacity for as long as possible.

 

[the_colorist]

On the flipside, it is definitely possible to use a BMS without comms to great success. That explanation was just to give a more full overview of the purpose of BMS comms.