Ok. Thanks for the test tip there @plonkster. I will check it out tonight and report back.
@Phil.g00 Iām getting 2 x BMSās to have one for charge and one for discharge, as per your idea.
Have you given some more thought about how to do that ito settings per BMS and connecting it all up?
Btw, this is how I test directional blocking on the high end. I push the battery past the max voltage. The moment it opens the contactor or starts blocking charge, the inverter voltage spikes up. All the batteries Iāve seen so far reports the internal voltage of the battery via the CAN-bus. So when there is a substantial mismatch, you know itās been disconnected.
With the Multiplus and ESS it is fairly easy to do these tests. I simply instruct the Multi to feed about 50W into the grid and then watch the voltage. If the voltage reported by the Multi drops until it hits the battery voltage and then holds thereā¦ it has directional charge control. If it drops past it and the Multi goes into sustain mode (which it does on a low voltage), I know it doesnāt 
Yes, I have.
Step 1: Establish if you have got a problem (as I described it somewhere above)?
No, then do nothing.
Yes, can you live with it?
Yes, then do nothing.
No, then the simplest solution I can think of only uses 1 BMS.
Butā¦ it only solves the problem of allowing discharge by the load, whilst not allowing charging after a high voltage threshold breach.
It would be premature to discuss this solution.
Maybe discuss it as a hypothetical?
Me and @Gman were also wondering if @Louisvdw can maybe think of ways the Venus driver can be used that when a cell volt hits a set parameter, say 3.58v, to āforceā a discharge ā¦ or dropping the DVCC max amps ā¦
All of this is purely academic, call it ideas for later, āadditional protectionsā if you want, as it is definitely NOT aimed to cater for well-balanced cells/banks.
I had one 250/100 replaced under warranty, Victron is NOT going to do that again.
That would be an ecumenical matter.
I am thinking about just relying on the inverter settings to take out the bank for the low voltage cut-off.
I donāt see this as an issue as this is reliable and used for every other chemistry.
Then using the BMS as is, but with a diode in parallel with the contactor.
With the contactor closed both charging and discharging happens i.e. Normal operation.
A high voltage operation trips the contactor as per before, but it leaves the diode in the circuit.
The diode is orientated to only allow discharging and prevent charging.
This has not been tested, and will probably mess up the pre-charging circuit, which you may or may not want to keep.
Thatās going to break discharge blocking on the other end of the spectrum.
If you do it, use a Schottky diode (or probably a bunch of them). Lower forward drop, and the slightly higher leakage is not going to matter. But as I saidā¦ I suspect itās going to break discharge blocking.
That is what I said.
Discharge blocking at the other end of the spectrum was never a problem that needed a BMS fix in the first place. It is just a functionality that was inherited because BMSās are used in other situations. In our application we will always be using inverters, which to my knowledge are universally pre-equipped with a low voltage cut-off.
I doubt we have to get too fancy, the aim of the diode is just to ensure the loads arenāt dropped.
This means the charging will pick up the load, and when that drops off the high voltage battery issue will be resolved in a very short time Iād think. Then the contactor will close and short out the diode and take all the current.
Regarding the pre-charge circuit:
Well, pre-supposing the pre-charge circuit is actually necessary. It would certainly be far more necessary when that contactor was disconnecting the inverter every time it operated. Which when you use a diode it will no longer be doing.
So in essence the pre-charge circuit was an elastoplast for a problem that you also no longer have.
The same as the seplos design.
1 Like
I picked the same subnet - or rather that is what the router that I first bought came out with 
That was 5 years ago and as every thing else is now set to that subnet I am stuck with it 