LiFePO4 reference diagram

[sailingharry]

Okay, the alternator should not be connected directly to the bus in case of cumulative short circuits that are large enough to overheat the alternator and start a fire, but not large enough to blow the fuse.

How will the BMS know about this in order to disconnect?

And, unless you go with a combined bus instead of a charge bus and a discharge bus, a partial short in the alternator will lead to a low voltage cut out, not a high voltage, and so will not be disconnected. I guess HVC and LVC pre-alarms could open the respective buses (and shut off devices like regulators) but an actual event would open both buses?

With regards to the alternator dump, I think you were mixing the parts of my reply. In the lead acid world, a common suggestion is to have the alternator connected on the battery side of the battery switch, so that an inadvertent battery switching (most commonly on an a/b switch) doesn't disconnect the alternator. Alternator dump on a properly designed lithium bank is easy - kill the regulator before the disconnect.

[Baronkrak]

Originally Posted by s/v Jedi

When a BMS supports both a charge and a load bus then it has control over both of them either via MOSFET’s or latching solenoids like the Blue Sea RBS switches. In case of a LVC, it will cut off the load bus but not the charge bus and this is fully ABYC compliant because it is under BMS control.

Originally Posted by s/v Jedi

I gues that would be my diagram? If so then you’re correct, I have one combined charge+load bus. Most (if not all) boats have that, which is why some posting on the forum don’t understand the dual bus setup.

The reason for only one bus is that it just works great. We never have any hvc nor lvc events because our chargers are working perfectly and we rarely ever get below 50% SOC.

So even with my perfectionist engineer point of view, I also believe in keeping everything as simple as possible where that has proven to work as good as a more complex setup.

I program my BMS myself, so it’s only a couple lines of code and a couple extra I/O lines. But it’s also an additional bus bar.

So you have a combined charge+load bus.
Your drawing shows only one latching solenoids switch when two are required. One to Load Disconnect, the other to Charge Disconnect.
Can you comment about the discrepancy?

[goboatingnow]

Originally Posted by Baronkrak

So you have a combined charge+load bus.

Your drawing shows only one latching solenoids switch when two are required. One to Load Disconnect, the other to Charge Disconnect.

Can you comment about the discrepancy?

Given the bms in a safety event , ie LVC , HVC, overtemp , over current must isolate the Li battery completely as per ABYC/ISO , and current safety thinking anyway , I fail to see this issue about dual busses , the control of the alternator field coil can be done by the bms, in advance of any safety disconnect

In my opinion after any such a safety trigger , no automatic reconnection of loads or charge sources should occur, until manual inspection of the Li batteries had occurred and then a controlled manual override can be initiated.

[s/v Jedi]

Originally Posted by Baronkrak

So you have a combined charge+load bus.
Your drawing shows only one latching solenoids switch when two are required. One to Load Disconnect, the other to Charge Disconnect.
Can you comment about the discrepancy?

You are misguided; a single bus is the most used option, especially on boats and it is compliant with ABYC and ISO.

[Baronkrak]

Originally Posted by goboatingnow

Given the bms in a safety event , ie LVC , HVC, overtemp , over current must isolate the Li battery completely as per ABYC/ISO , and current safety thinking anyway , I fail to see this issue about dual busses , the control of the alternator field coil can be done by the bms, in advance of any safety disconnect

In my opinion after any such a safety trigger , no automatic reconnection of loads or charge sources should occur, until manual inspection of the Li batteries had occurred and then a controlled manual override can be initiated.

You are not paying attention to Nick statement.

Originally Posted by s/v Jedi

In case of a LVC, it will cut off the load bus but not the charge bus and this is fully ABYC compliant because it is under BMS control.

[Baronkrak]

Originally Posted by s/v Jedi

You are misguided; a single bus is the most used option, especially on boats and it is compliant with ABYC and ISO.

You are not paying attention to your own statement

Originally Posted by s/v Jedi

In case of a LVC, it will cut off the load bus but not the charge bus and this is fully ABYC compliant because it is under BMS control.

[goboatingnow]

Originally Posted by Baronkrak

You are not paying attention to Nick statement.

And Nick is wrong in my opinion. ( if are suggesting he’s saying charge sources remain connected after a LVE/LVC. I don’t actually believe he’s said this. )

ABYC


13.7.7 No electrical connections should be made directly to a lithium ion battery that would bypass a BMS or the
protection relays.


hence both load and charge busses must be disconnected.

Common sense safety would dictate that a safety bms disconnect should isolate the battery. That’s the whole point. It’s a “
Safety “ disconnect. It should never happen in normal operation.

On a bms action disconnect everything is disconnected and. no automatic reconnects until manual I inspection has occurred in my opinion . In particular LVC should not be designed as a charge start point. This should happen way before LVC.

Load and charge buses should be disconnected. Whether this is one relay or two is irelevant.

[fxykty]

Originally Posted by goboatingnow

And Nick is wrong in my opinion. ( if are suggesting he’s saying charge sources remain connected after a LVE/LVC. I don’t actually believe he’s said this. )

ABYC


13.7.7 No electrical connections should be made directly to a lithium ion battery that would bypass a BMS or the
protection relays.


hence both load and charge busses must be disconnected.

Common sense safety would dictate that a safety bms disconnect should isolate the battery. That’s the whole point. It’s a “
Safety “ disconnect. It should never happen in normal operation.

On a bms action disconnect everything is disconnected and. no automatic reconnects until manual I inspection has occurred in my opinion . In particular LVC should not be designed as a charge start point. This should happen way before LVC.

Load and charge buses should be disconnected. Whether this is one relay or two is irelevant.

I’m not sure that you have that right. We have two positive bus - one for charge sources and one for loads (and inverter/charger). We have independent contactors between the battery and each bus on independent BMS relays. One disconnects the load bus on LVC and the other disconnects the charge bus on HVC. With either bus disconnected there is still a BMS relay controlling a contactor for the other bus. There are no electrical connections between the battery terminals and the contactors (other than alternator regulators’ voltage sense). In the case of temperature or voltage differential shutdown triggers both relays are activated to cut off both bus.

I believe our system is compliant with the ABYC statement above.

Not that we’ve ever seen an HVC or LVC event in the 1.5 years our system has been running full time live aboard.

[goboatingnow]

Originally Posted by fxykty

I’m not sure that you have that right. We have two positive bus - one for charge sources and one for loads (and inverter/charger). We have independent contactors between the battery and each bus on independent BMS relays. One disconnects the load bus on LVC and the other disconnects the charge bus on HVC. With either bus disconnected there is still a BMS relay controlling a contactor for the other bus. There are no electrical connections between the battery terminals and the contactors (other than alternator regulators’ voltage sense). In the case of temperature or voltage differential shutdown triggers both relays are activated to cut off both bus.

I believe our system is compliant with the ABYC statement above.

Not that we’ve ever seen an HVC or LVC event in the 1.5 years our system has been running full time live aboard.

I do not believe what you are describing is ABYC compliant. The specification does not go into the detail you suggest.

Hence you are “ reading between the lines “

What TE-13 says is quite general


“ Redundant Protection - A BMS should be equipped with HVC and LVC actions in response to an HVE or LVE when the programmed functions in the charging sources, inverters, or inverter/chargers, etc fail to do so.
13.9.3 13.9.4
A BMS should monitor cell voltage to determine if an HVE or LVE is imminent.
HVE/HVC/LVE/LVC - A BMS should protect the lithium ion battery cells in response to an HVE and an LVE.
“

Hence I would argue on no account should a charge source be left connected on an LVE.

it may be very dangerous to reconnect a charge source to a very low cell voltage.

Hence the control of charge sources should occur before HVE or LVE

Hence I see no purpose to a separate load and charge bus as both should be disconnected to comply with 13.9.4.

Safety first thinking should also ensure that manual intervention should occur before loads or charge sources are reconnected so appropriate manual overrides should be installed ( including bms overrides ) even though such overrides void ABYC compliance per se.

[Baronkrak]

Originally Posted by goboatingnow

And Nick is wrong in my opinion.

And so must be Victron?

Load Disconnect output
The Load output is normally high and becomes free floating in case of imminent cell under voltage.

Charge disconnect output
The Charger output is normally high and becomes free floating in case of imminent cell over voltage or over temperature.

Under voltage alarm (cell voltage < 2.8V)
Over voltage alarm (cell voltage > 3.7V)
Over temperature (T > 50°C)
April 10 2017

[s/v Jedi]

Originally Posted by Baronkrak

You are not paying attention to your own statement

No, you are wrong. I am stating that both options are fully ABYC compliant. If you do not agree then describe which part does not comply with which text in the guidelines.

[s/v Jedi]

Originally Posted by fxykty

I’m not sure that you have that right. We have two positive bus - one for charge sources and one for loads (and inverter/charger). We have independent contactors between the battery and each bus on independent BMS relays. One disconnects the load bus on LVC and the other disconnects the charge bus on HVC. With either bus disconnected there is still a BMS relay controlling a contactor for the other bus. There are no electrical connections between the battery terminals and the contactors (other than alternator regulators’ voltage sense). In the case of temperature or voltage differential shutdown triggers both relays are activated to cut off both bus.

I believe our system is compliant with the ABYC statement above.

Not that we’ve ever seen an HVC or LVC event in the 1.5 years our system has been running full time live aboard.

GoBoating recently has an interest in countering my recommendations blindly, so rest assured that your installation is not only fully compliant but it is the gold standard that my diagram and my installation doesn’t even achieve.

[s/v Jedi]

So for the readers, my explanation of how this works:

In case of an LVC (low voltage cut-off) one or more cells have gone so low in SOC (state of charge) that their voltage has reached the point where the BMS must react to protect the battery. To be compliant with ABYC, the following must be implemented:

1. Prior to the LVC, a warning signal “LVC IMMINENT” must be given. This signal can sound an alarm but also trigger relays which then shut off loads like an inverter, a dc-dc converter etc. Doing so may even prevent the LVC from happening at all.

2. When LVC triggers, the battery must be disconnected from the loads so that it can not be discharged any further. This can be done in two ways: the battery negative may go through a bank of MOSFET’s inside the BMS, which function as a large switch, or the second option: the BMS controls an external switch, like a latching solenoid. In my diagrams I show a Blue Sea Systems RBS (remote battery switch) which can be operated manually like every regular battery switch, as well as remotely by a trigger signal. They come with a Contura switch that you can mount elsewhere in the boat, but some BMS’s can also control the RBS. Make sure that the RBS is compatible with the way the BMS controls it, because there are two different options available (pulsed or continuous signal).

Now we get the HVC, the high voltage cutoff. This occurs when one or more cells have reached a voltage so high that the BMS must disconnect the battery from the charge sources that cause this. The following two points must be implemented to be compliant:

1. Like for LVC, prior to this happening, a “HVC IMMINENT” warning signal must be raised. Again an audible alarm must go but also relays may be triggered to stop charge sources like the alternator, mppt controllers, dc-dc converters or AC powered chargers. When implemented correctly, this would prevent the actual HVC from happening.

2. When HVC triggers, all charge sources must be disconnected from the battery. This is what burns out the alternator diodes and can cause damage to other charge sources as well. It is done the same way as for LVC description above.

Now we get to the disagreements; there are two versions of BMS’s: one has MOSFET’s or controls for a charge bus as well as discharge bus and another version combines both into a single charge/discharge bus. One party here claims that only a combined bus is compliant, a second party claims that only separate busses is compliant and then there’s others incl. me who claim both options are compliant.

The text in the recommendation is clear: no connections directly to the battery, outside of BMS control are allowed. But this is the case for both options. When the BMS triggers an LVC, the discharge bus is disconnected, while the BMS allows the charge bus to be connected, as it hopes a charge source may clear the LVC by recharging the battery. This charge is fully under BMS control and the BMS may decide to disconnect the charge bus at any time it deems charging poses a danger to the battery.

The option with separate charge and discharge busses is superior, but it also requires a more expensive BMS and two RBS’s with two busbars which all adds significant costs to the installation, so this is normally only done when high availability is desired. I chose to not put this in my diagram because when the warning signals are used to control chargers and discharge, the likelihood of a HVC/LVC is very low.

[goboatingnow]

My issue is not the separate load and charge buses. My issue is not removing charge sources on an LVE. ABYC charges the bms with protecting the batteries.

If you allow the charge sources to remain while the cell falls there is no way the charge source should automatically be connected. The cell may be damaged and needs inspection , the cell may need a very low C prequalifying charge to bring the cell to 20 %

This is the kernel of my objection , LVE is a critical safety event. The battery safety ( under ABYC 13.9.4) is the sole focus of the bms.

Low Soc should trigger charge sources either by the charge sources or the bms way before LVE.

As to blindly challenging anyone. This is my domain expertise. I designed charging systems fit industrial batteries etc. I’m comfortable in my interpretation.

Notably on dirsolarforum a group has worked up a marine Li reference design. It does not require or consider separate charge and load relays

As o said it’s not the two buses that’s the kernel , I would contend that automatic charge source connection at ( or below LVE ) is not ABYC compliant. The bms cannot be sure what caused the LVE and therefore must protect the battery first

[s/v Jedi]

Originally Posted by goboatingnow

My issue is not the separate load and charge buses. My issue is not removing charge sources on an LVE. ABYC charges the bms with protecting the batteries.

If you allow the charge sources to remain while the cell falls there is no way the charge source should automatically be connected. The cell may be damaged and needs inspection , the cell may need a very low C prequalifying charge to bring the cell to 20 %

This is the kernel of my objection , LVE is a critical safety event. The battery safety ( under ABYC 13.9.4) is the sole focus of the bms.

Low Soc should trigger charge sources either by the charge sources or the bms way before LVE.

As to blindly challenging anyone. This is my domain expertise. I designed charging systems fit industrial batteries etc. I’m comfortable in my interpretation.

Notably on dirsolarforum a group has worked up a marine Li reference design. It does not require or consider separate charge and load relays

As o said it’s not the two buses that’s the kernel , I would contend that automatic charge source connection at ( or below LVE ) is not ABYC compliant. The bms cannot be sure what caused the LVE and therefore must protect the battery first

No, primarily the BMS and secondarily the charge source decides if it is safe to charge the battery. If a cell is low, it indicates that the battery must be charged and possibly balanced. If it is too cold to charge or if the cell overheats then the BMS can trigger HVC to protect the battery as designed.

This is how it works, how the gold standard has been designed and implemented a million times; it simply is how it is and not subject to opinion. I even believe the separate charge and discharge busses predates the single combined bus.