Parallel BMS's

[fxykty]

Originally Posted by CaptainRivet

The contactor BMS can all do that BUT neither a TAO or Orion are meant switch off an MPPT or shorepower charger at Overvoltage in normal operation
The MPPT as well as shorecharger have their parameter for absorption and float which are below the cutoff voltage of the BMS and the BMS only intervenes when the regulation fails.
With electrodacus it's different and the absorption voltage is higher the end of charge value of BMS and if that's reach BMS shut only the MPPT off. If BMS fails to do so the MPPt does its regulation absorption/float and if that fails also the electrodacus cuts with Overvoltage the whole charge devices.
That is the different philosophy they operate.
Sure i can force and setup a TAO the way electrodacus works, it's flexible enough to do so but that's not how it's designed.

This is nonsense, but it also reveals that there are in reality different ways to control charging. I still don’t understand what Electrodacus does that is so special.

Regarding our TAO BMSs, we can define up to twenty events within the BMS. Each of these events can, when triggered by whatever condition we’ve defined, raise an alarm and/or power one or more of 6 output relays. So for stopping charging during normal operation we have an event that triggers when cell voltage reaches 3.550V. Triggering the event displays a light on our monitoring panel and sends power (closes the circuit) to our stop charge relay (we have it set by jumper to be normally open). Powering the stop charge relay turns off all 7 of our charge sources via either their remote on/off port (VE MPPT chargers (150/100 and 250/60) and Quattro) or separate relays (cutting the ignition wire on our Balmar external regulators and opening the contactor on the wind generator output). This is all ‘normal’ operation.

But before any of that happens our TAO BMS via CANBUS assumes control of all the VE chargers that are connected to the Cerbo: the BMS is the external control and tells the Cerbo when to allow charging at full current, when to restrict charging to limited current (low or high temperature, higher battery voltage, high cell voltage differential), and when to stop charging. Typically the CANBUS settings trigger ahead of the BMS relays. We have defined the settings in each charge source to work as we want in case the BMS fails or is taken offline, but during normal operation the BMS is the boss. BTW, we do have DVCC enabled on the Cerb; this is what enables the BMS to take control of all equipment connected to the Cerbo. Still normal BMS operation.

In the case of over voltage or some other emergency situation defined by either voltage, temperature, or state of charge, a BMS event will trigger that disconnects the battery. Still normal operation, but in this case should never happen due to events that will trigger before the emergency situation develops.

[fxykty]

Originally Posted by Dockhead

OK, so in this case, the BMS is not needed at all for controlling the charge sources -- is that right?


So the Cerbo is used for this?


So what then exactly do the BMS's tell the charge sources?


And is Wakespeed the only alternator regulator which is easily integrated with the BMS's and/or Cerbo? Wakespeed speaks Canbus, but perhaps that's not needed? Just modulate the field wire somehow?

Belts and suspenders and adjustable drawstring is (in priority order, with normal everyday operation at the bottom):
1) setting up each charge source so that it operates as you wish independently of anything else. This generally means defining and modifying the Lithium charge profile within the device’s settings environment.
2) setting up your BMS events and wiring relays and/or contactors to each charge source so that if the BMS triggers a stop charge event, each charge source will turn off charging, or have charge output disconnected from the battery.
3) Connect your BMS via CANBUS to all equipment that can use it (such as Cerbo and Wakespeed regulators) and use CANBUS settings in the BMS to control charging parameters. In this case the Cerbo is used to distribute the CANBUS parameters to all VE equipment such as MPPT controllers and Multiplus/Quattro and whatever else you’ve got. Any equipment that can’t be controlled directly or indirectly by CANBUS will need to rely on the hard-wired controls in item 2.
3 alternate) Connect your BMS via CANBUS to all equipment that can use it (such as Cerbo and Wakespeed regulators) and use the BMS only to provide the most accurate battery information. In this case the Cerbo is used to define and control the CANBUS charging parameters that are distributed to all CANBUS and VE-Direct connected devices such as MPPT controllers and Multiplus/Quattro and whatever else you’ve got. Any equipment that can’t be controlled directly or indirectly by CANBUS will need to rely on the hard-wired controls in item 2. This option may require programming the Venus OS (Cerbo operating system) as not everything is in settings.

[Dockhead]

Thanks for the education, guys.


So let me get this straight:


LP can actually be charged on the same basic principle we charge lead -- charge sources regulate VOLTAGE, turn on at a certain voltage and turn off at a certain other voltage.


So in daily operation, Plan A is to have charge sources regulate themselves according to programmed charge profiles. Did I get that right?


Plan B is that the BMS commands charge sources to stop, based on one cell getting overcharged, or the whole pack if for some reason the charge sources didn't stop charging. I guess that's a core function of BMS -- right? Because it sees individual cell voltage and unlike everything else in the system, the BMS can know if there's a problem INSIDE the pack.


Plan C is none of this works for some reason and either BMS or Cerbo throws the pack offline via the contactor.


Then secondary functions like propagating charging profile over CANBUS.



Is this all correct?


So if this is how it works, it begs a couple of questions:


1. Most important communication is the "stop charging" command. This goes out only by CANBUS -- right?

[fxykty]

Originally Posted by Dockhead

Thanks for the education, guys.


So let me get this straight:


LP can actually be charged on the same basic principle we charge lead -- charge sources regulate VOLTAGE, turn on at a certain voltage and turn off at a certain other voltage.


So in daily operation, Plan A is to have charge sources regulate themselves according to programmed charge profiles. Did I get that right?


Plan B is that the BMS commands charge sources to stop, based on one cell getting overcharged, or the whole pack if for some reason the charge sources didn't stop charging. I guess that's a core function of BMS -- right? Because it sees individual cell voltage and unlike everything else in the system, the BMS can know if there's a problem INSIDE the pack.


Plan C is none of this works for some reason and either BMS or Cerbo throws the pack offline via the contactor.


Then secondary functions like propagating charging profile over CANBUS.



Is this all correct?


So if this is how it works, it begs a couple of questions:


1. Most important communication is the "stop charging" command. This goes out only by CANBUS -- right?

Not quite in my opinion, but getting there :-).

Regarding LFP charging: yes, define charging voltage and use battery voltage to turn off charging. For 100% charging use tail current at constant charging voltage to turn off charging - generally current down to 2.5% of C will indicate 100% SOC. However, with LFP, unlike LA, you want to turn off charging once the battery is full, so no absorption charging and float can be off or at a low enough voltage that it keeps the battery at 50% SOC (say 13.1V).

Plan A is the fallback plan for when the BMS goes offline - the chargers will operate as they should with no BMS telling them what to do. Very rarely should your system operate like this (though it is preferred by some who want a simple system).

Plan B with communications via CANBUS to the Cerbo and alternators to fine tune their charging. Regarding BMS communications, stop charge can be sent via relay outputs and via CANBUS. In addition to stop charge, CANBUS commands can specify a variety of charge voltages based on cell and battery conditions. This is your normal operating regime.

Plan C should never happen as long as A and B are good. Certainly the BMS should trigger the disconnect. Have the Cerbo do that if it loses BMS connectivity, though that will require additional wiring.

[CaptainRivet]

Originally Posted by Dockhead

OK, so in this case, the BMS is not needed at all for controlling the charge sources -- is that right?


So the Cerbo is used for this?


So what then exactly do the BMS's tell the charge sources?


And is Wakespeed the only alternator regulator which is easily integrated with the BMS's and/or Cerbo? Wakespeed speaks Canbus, but perhaps that's not needed? Just modulate the field wire somehow?

The first general decision you have to do is:
1) BMS is running completely independent and have in maximum one way communication to cerbo or home assistant or any other smart boat steering which optimize use inbetween the end.
2) The cerbo or Home assistant is the spider in the web steering all with a bus and BMS is kinda last resort that influenced in its decision by cerbo/HA via two way communication

After 20 years having both and having both on my cat, I returned my cat 6month ago back to 1) with one way communication from BMS to cerbo and now getting home assistant as spider which cerbo feeds with data.
No 2 needs a fully canbus communication capable BMS, the JK inverter BMS can do this to 95%

The general standard operation should be:
1) each charge source has its own set of parameters for absorption and float and works independently shutting itself off or into float when needed. This includes the alternator internally or externally regulated. The inverter as biggest load have their own internal low voltage warning and low voltage cut off. BMS only interfere independently when BMS measures that parameters are exceeded or an event happens that is a danger to the battery bank.
That's maximum KISS and a stableworking and save system.

2) next level up is charge sources should be connected which each other to synchronize. The biggest benefit here is if you have multiple solar controllers in your system that these sync. The rest of charge sources has tiny advantages in sync and in my experience are only worth doing if it comes for "free" with or you have an exceptional case. Thats the case with victron BT sync so always choose the smart devices and the BT connect is fully sufficent for doing that sync wirelessly without routing cables between your MPPTs, also multi/Quattro with the BT dongle and Battery protects to deliver reference voltage.
A Bus system does it perfect but I never saw big enough differences to BT sync when using bus and DVCC (what victron uses to interfere and steer the different connected device based on the reference voltages DVCC receives) to justify the complexity and additional issues these "perfection" brings plus costs this brings (eg requiring an expensive wakespeed instead a Balmar 618).
Again here the BMS only intervenes and eg cutting charge when the BMS measures that the device exceeds its parameter, so something is wrong.

3) next level up is one way communication from BMS to install. This means the BMS sends all its information of cells/SoC... to the cerbo or home assistant and cerbo/HA syncs charge and also optimizes the loads to achieve best SOC and charge efficency away/outside from the parameter edges/limits of BMS. This includes also switching off an alternator at 13.6V so that it cannot cause trouble in the knee and its very inefficient anyhow to absorb or float a battery, solar is much better in doing that anyway. The BMS still acts fully independently on what it measures and only intervenes when parameter exceeded limits/thresholds set.
1 till 3 has a single failure point, the BMS so it's highly recommend to have a 2nd independent device that only does emergency cutoff if the whole systems exceeds its parameters, a BMV 702 or 712 is a simole but very effective = perfect device for that.

4) the final evolution is having everything interconnected with a bus and a central brain which can be a cerbo or home assistant or a BMS that has external steering to manage all that besides is security/protection functions.
That's perfection IF everything works flawlessly together but the additional complexity introduces a big amount of additional failure points that need then additional measures to cover them=more that can go wrong.


Regarding wakespeed or the Zeus these alternator regulators or better computer regulated alternators makes in my opnion only sense if you need and profit strongly from its additonal features which for sure 80% of boats don't need that. Typical cases are big heavy duty alternators on small auxiliary engines where the alt takes so much power away that during propulsion nearly nothing is avaliable for propulsion so that needs seperate regulation. The wakespeed eg allows for limition based on RPMs so you strongly derate the alt in lower RPMS you need for maneuver or propulsion and only run it full output when in the upper range but derate again at close to/full WOT as there you need all you have for propulsion. And even then you don't need a bus delivering you all the info, both can do this independently. but sure profit from a bus that steers that expentional charge device I your specific installation.
But for the average cruiser boat with a bigger alternator on their engine a standard external regulator that charges based on absorption and float parameter and that you can derate with a switch to an adjustable level for eg low rev motorsailing is more then sufficent and a wakespeed would add cost and complexity you cannot profit from.

[s/v Jedi]

Back to the basics:

All LiFePO4 batteries must be charged with an CC/CV charge algorithm at a specific charge rate, like 0.5C or 0.2C as specified by the manufacturer. When the charged voltage is reached, it is to be considered the termination voltage, i.e. charging stops at that point because it is complete.
This is similar to the bulk phase of LA charging.

For house banks on boats this never works because we can’t do 0.5C. I have often posted on how to adjust charging based on your specific charger and setup so won’t repeat that here but it entails that you first establish accurate SOC data, then adjust charging to complete when full.

A BMS is a device that protects the battery. If you want a hardened system that can be kept operational wherever on the planet you are, then it is crucial that you do not use it for anything more than this function. It doesn’t need to control your charger because you can do that as described above which works with every BMS, not just a specific one. When your system needs a specific battery or BMS or charger or regulator then you have been locked into a commercial trap.

The basic Balmar regulators are way good enough for charging LiFePO4, just configure them correctly and stop charging when SOC reaches a point just below 100%. When you have accurate SOC data this is child’s play using a BMV or Cerbo with programmable relay contacts to switch the regulator ignition wire.

That’s all, you don’t need magic devices.

[Dockhead]

Originally Posted by s/v Jedi

Back to the basics:

All LiFePO4 batteries must be charged with an CC/CV charge algorithm at a specific charge rate, like 0.5C or 0.2C as specified by the manufacturer. When the charged voltage is reached, it is to be considered the termination voltage, i.e. charging stops at that point because it is complete.
This is similar to the bulk phase of LA charging.

For house banks on boats this never works because we can’t do 0.5C. I have often posted on how to adjust charging based on your specific charger and setup so won’t repeat that here but it entails that you first establish accurate SOC data, then adjust charging to complete when full.

A BMS is a device that protects the battery. If you want a hardened system that can be kept operational wherever on the planet you are, then it is crucial that you do not use it for anything more than this function. It doesn’t need to control your charger because you can do that as described above which works with every BMS, not just a specific one. When your system needs a specific battery or BMS or charger or regulator then you have been locked into a commercial trap.

The basic Balmar regulators are way good enough for charging LiFePO4, just configure them correctly and stop charging when SOC reaches a point just below 100%. When you have accurate SOC data this is child’s play using a BMV or Cerbo with programmable relay contacts to switch the regulator ignition wire.

That’s all, you don’t need magic devices.

Thanks; very valuable.

So -- you stop alternator charging via Cerbo or BMV just cutting the field wire via the relay? Based on SOC? Not voltage? And you rely on the Balmar to stop charging based on voltage, right?

So -- the BMS only controls HVC and LVC? And doesn't talk to the alternator regulator?

And last question (I promise) -- this is different with your Multiplus, you do have the BMS talking to that? Or not?

It might be helpful to have diagrams of what elements of the system have authority over what actions -- stop charging, alarm, HVC, LVC, etc. Different people will have different diagrams; it would be interesting to compare.

[s/v Jedi]

Originally Posted by Dockhead

Thanks; very valuable.

So -- you stop alternator charging via Cerbo or BMV just cutting the field wire via the relay? Based on SOC? Not voltage? And you rely on the Balmar to stop charging based on voltage, right?

So -- the BMS only controls HVC and LVC? And doesn't talk to the alternator regulator?

And last question (I promise) -- this is different with your Multiplus, you do have the BMS talking to that? Or not?

It might be helpful to have diagrams of what elements of the system have authority over what actions -- stop charging, alarm, HVC, LVC, etc. Different people will have different diagrams; it would be interesting to compare.

First it is important to configure the regulator correctly for charging LiFePO4. You basically copy the settings of the Multiplus which you figured out using the SOC method I explained.

This should be good enough but because of the risk of blowing out the rectifier diodes on the alternator during a HVC, we can stop just before finishing the charge using rely contacts inside a BMV or Cerbo.

You don’t switch the field wire… you switch the ignition wire. Read the Balmar manual which explains this in detail. If whatever you switch carries more current than the relay contacts are rated for, then simply add a relay close to the regulator and drive that from the BMV/Cerbo.

[Dockhead]

Looks to me like some are controlling charges sources purely by voltage, and others are doing tail charge or SOC measurements and using that.

So I'm still having some difficulty seeing what the best (optimal balance between effectiveness, reliability, simplicity) algorithm is for this.

Maybe something like this:

1. BMS's control HVC and LVC for each battery separately. This is a freestanding system, last ditch protection of the batteries if everything else fails, which does not use any networking functions.

2. What information can the BMS communicate to Cerbo and/or Multiplus, which is useful? I understand we don't want JK's SOC measurements which are not accurate, right? Do we let the BMS inform Cerbo or Multiplus that there is a rogue cell with too high voltage, for example? So that charging is stopped prior to some other reason to stop charging (reached the target SOC, reached voltage)?

3. We let Cerbo calculate the SOC of the the whole pack, based on a Smart Shunt placed AFTER the bus bar which combines the negative leads of each battery, like in Nick's schematic.

4. So we let Cerbo control the alternator, right? The regulator is programmed to start and stop charging according to voltage, but we let Cerbo preempt the regulators "stop charge" command by cutting the ignition wire if Cerbo judges 100% SOC (or whatever has been programmed) has been reached? Right? That would be in case of a non-CANBUS regulator like the Balmar. Otherwise, in case of Wakespeed etc., this command is given over CANBUS, right?

5. Multiplus also has a self-managed charge profile based on voltage, right? But we also allow Cerbo to tell Multiplus via CANBUS to stop charging, if Cerbo judges that 100% (or whatever has been programmed) has been reached?

Please rip this apart if I've got something wrong.

Some further questions:

1. Alternator protection. We have relays on both Cerbo and BMS's which can be programmed and used for different stuff. I think (if I understand the principles right) that we will not blow the diodes in the alternator if only the ignition or field is cut at least a couple of seconds before the contactor is disconnected. Right? So maybe it would be good to let the BMS's do this -- one relay (on each BMS) is programmed to cut the ignition wire at a slightly lower voltage, than the relay which disconnects the HVC contactor. Or -- if voltage isn't right, since a rapid voltage spike would activate both relays almost simultaneously -- the second relay acts with a couple of seconds of delay after the first one? If we're using a CANBUS regulator, then this would be completely separate from the high voltage or full SOC "stop charging!" signal from the Cerbo, which we would program to happen at a slightly lower voltage, so redundant.

2. What does the Cerbo communicate to the Multiplus, which makes it behave any different than its own regulation does? Is it a signal the 100% (or whatever) SOC has been reached, so preempts the voltage-determined decision to stop charging? How does this work?

3. What if anything do we want the BMS telling the rest of the system? And to do what? For example, do we want the BMS to tell Cerbo to stop the charge sources charging based on some abnormal condition (one rogue cell for example)? Or what exactly? What commands from where, preempt what? Or do we just have the BMS controlling HVC and LVC and perhaps cutting alternator ignition wire, and nothing else.

I hope this conversation is benefiting others, and I'm grateful to the experts for their patience.

[Dockhead]

I would have asked fewer dumb questions, perhaps, if I had read this first:


https://marinehowto.com/lifepo4-batteries-on-boats/


;sheepishgrin:


Among other topics discussed in this thread, Rod talks about alternator protection. He recommends the Sterling alternator protector (and says the Balmar one is better), but also goes into strategies for shutting down the alternator before a HVC occurs. Interestingly, he states that the field, not ignition wire should be cut, as an ignition wire cut doesn't shut down the alternator quickly enough. Balmar say cut ignition plus power supply to the regulator.

[CaptainRivet]

You are overthinking things.

To 1) With JK BMS you use master/slave so one BMS is master and care about all together. Each battery does its own HVC/LCV.
It communicates SOC and cell voltages also which cerbo aggregates to one battery he is using for steering like a bat monitor.

I suggest you download the manual of the cerbo to read what it does and what options you have.

To your 2) Then you need to decide if you wanna give cerbo the spider in the web by activating DVCC function or not. There is no I want it partly, only yes or No.
DVCC means in short cerbo is taking the voltage from BMS and communicates this as reference to all connected devices AND sync/control the charge, it tells eg all MPPT go to float+deliver 0 and one MPPT delivers the float current/voltage, same with Multi....and also HVC and LVC is controlled by DVCC and not the BMS anymore.
Due to manual you simply shut DVCC on and it works...nice theories as they are so many things that are not documented

To 3) cerbo takes SOC from the source you define as master and with DVCC on that's the master BMS. You can connect BMV as seperate battery monitor but it won't take the SOC from it.

To 4) that depends on your decision in 2) DVCC yes or No. If yes the DVCC has the control via the communication interface used CAN/VE.Direct....
The cerbo has 2 relay ports you can switch manually on/off. You can programm them with MQTT to do task eg shut on/off a seperate active balancer or at a specific voltage of battery shut on a shore charger. You can add more relays. These are not steered by DVCC. So you can use one of these relay and program with MQTT that relay cuts the alternator at the beginning of the knee/when a cell reaches 3.4V.
5) same with Multi, with DVCC on the cerbo runs the show=Multi
Cerbo is not a wish concert. Home assistant is much more flexible.

As describe in my other statement.
A) if you want BMS in control you cannot use DVCC.
B) charge sync can also be done independent via BT connect and the reference voltage comes from the BMV as well as SOC. Define in cerbo as bat monitor that's then can then be used so cerbo also gets the SOC and reference voltage from BMV. DVCC and BT connect does not work in parallel, stands nowhere in viction documents as well as dozen other stuff that's needed that DVCC works correct, if it does you don't know unless it doesn't.

To your questions
1) Alternator if DVCC on the cerbo steers it IF it has CAN connection which wakespeed has and balmar no.
If you have a JK BMS it's mosfet and it simply cuts the chargebus if 100% SOC reached, it doesn't and cannot steer an alternator. If you take a contactor BMS that can steer the alternator directly.
What options you have depends on your previous decisions.
So with balmar you can simply set the charge parameter like in all other devices, so that in absorption you don't reach HVC. If one battery has a runner the other is still online.
You then can use the relays steering in the cerbo and program with MQTT cell voltage 13.6V or SOC value is reached that a relay cut ignition and or field wire (depends also on alternator used and cutting both is safest). Alternative the BMV can do this cut off with its relay output or even a voltage sensing relay that cuts at 13.6V or 28.3V. That avoids the alternator is running when bank gets into knee and a runner/imbalance can cause a disconnect. If a spike cuts the alternator too early it will be automatically re-engaged as voltage is lower and relay closes and out alternator back on.
And for the emergency cut off you have the balmar surge protector. An emergency comes sudden an you cannot switch the alternator before off as it doesn't know and a delay if eg an MPPT sends panel voltage which can be above 100V will kill the BMS then before it can perform its HVC.

2. Read cerbo manual. If DVCC on it communicates eg the reference voltage which overseers the multi measures on his input connection. With DVCC on the cerbo sends via communication to shut on or off and oversteers the direct load and charge disconnect coming from BMS....

3. There is no BMS telling rest of system if DVCC on cerbo steers everything that has a connection and understands victron protocol. For what has no communication As written before cerbo has relays that are independently programmable via MQTT so you can switch on off when a or multiple condition are met the cerbo/Venus OS has the data for eg temperature or SOC or....
Eg Switch on a water heater when SOC (from BMS or BMV or....) is 85% and solar production over 400W and switch off at 70% SOC.
Or eg. switch of alternator when 13.6V/27.2V or 98% SOC is reached.
If DVCC on the Cerbo runs the show. If off BMS has controll. then BMS does HCV and LCV as well as emergency disconnect and cerbo or BMV or VSR cut alternator when beginning knee is reached so no disconnect under load can happen unless emergency disconnect because a device failed...

As jedi wrote before if you wanna have a harden system that kept operational wherever you are then the BMS has control and does nothing else then HVC and LVC plus emergency shutoff.
Additional then you can deliver with one way communication the cell voltages, total voltages and SOC to smart system (one way!) with then uses this information to optimize load or charge eg shutting the alternator off when it reaches the knee so mitigating the risk it gets damaged in a HVC. Or running the waterheater above 90% SOc as dump load. The smart system like cerbo or home assistant can then issue warnings when close to BMS disconnect conditions like too hot, voltage to low or high or.... endless variation BUT ALL of that doesn't influence or modify the behavior of the BMS at all doing its protection tasks. But you can prevent with this smart systems that BMS needs to act and damage happening by cutting alternator earlier or you get a warning low voltage so you can switch off manually or automatically steered by smart system loads so the LCV doesn't happen.

[Dockhead]

OK, thanks very much for this.


DVCC = "Distributed Voltage and Current Control"

"Enabling DVCC changes a GX device from a passive monitor into an active controller."

"Managed CAN-bus batteries: In systems with a managed CAN-bus BMS battery connected, the GX device receives a Charge Voltage Limit (CVL), Charge Current Limit (CCL) and Discharge Current Limit (DCL) from that battery and relays that to the connected inverter/chargers, solar chargers and Orion XS. These then disable their internal charging algorithms and simply do what they're told by the battery."

OK, I think I understand this -- it means that Cerbo takes over charge control from all the charge sources, and manages them according to what it gets from the BMS's.

I think both you and Nick said you ended up preferring the charge sources to manage themselves. So, DVCC OFF. I can see that this could be simpler and more reliable.

In that case, then no particular reason for a CANBUS alternator regulator, right?

So then we use the Cerbo to cut off the alternator in specified situations, and we use the BMS's to manage the HVC and LVC.

But what about the Multiplus? Do we leave that self-managing or does it take some orders from Cerbo? Nick was talking about turning the Multiplus down based on tail charge current or SOC. How does that work?

You say Cerbo uses SOC from the BMS's -- there is no option for it to do its own based on Smart Shunt data?

[s/v Jedi]

1. Exactly

2. Nothing, use a BMV for each battery instead so you gain full control over its configuration and can program the relay for whatever you wish.

3. Yes, I add a SmartShunt for the whole bank but mainly because the Cerbo doesn’t allow defining a sum of multiple BMV’s, so it’s a workaround by brute forcing the solution.

4. The regulator must be a smart type that controls itself. The reason I recommend to make it stop charging early is the risk of blowing diodes and the reason to use a relay programmed based on SOC is just one parameter that can be used. You can also use voltage (after monitoring your bank you have a dataset that shows which voltage is which SOC for the part in the upper and lower knees) or even midpoint monitoring when you have access to the midpoint of packs of cells or when using two 12V batteries in series for 24V etc.

Normally the Balmar regulator wants it ignition wire to be disconnected to stop charging as this is the primary on/off method for it. What they state in the manual is that for HVC control you can switch B+ instead (not both) to make the switch off as certain and as quick as possible. But I recommend to stay away from HVC, you don’t want any of that, so simply stop charging from the alternator at 98% or 95%. In that case there is no danger/hurry. Of course there will be boats that only have the alternator as charge source but this is so far from a solid system that I simply don’t make any recommendations for such a setup (must have solar or generator)

I recommend against CAN bus controlled alternator regulators because they are a niche product so there isn’t enough choice and it is overpriced and probably not tested for reliability and longevity. In 5-10 years when they all have it at prices like that of a Balmar controller today and they are all compatible with everything then sure.

5. No, Cerbo should not control Multiplus charging. This is a thing designed by people wanting to lock customers into their brand. They want their batteries to tell the Cerbo to tell the Multiplus what to do. The Multiplus already knows what to do. When there is a problem with the battery it will be protected by its BMS; it doesn’t need a Cerbo or Multiplus to help with that and they simply continue with batteries that didn’t fault, because a HVC means a cell blew up or some other disaster happened.

Originally Posted by Dockhead

Looks to me like some are controlling charges sources purely by voltage, and others are doing tail charge or SOC measurements and using that.

So I'm still having some difficulty seeing what the best (optimal balance between effectiveness, reliability, simplicity) algorithm is for this.

Maybe something like this:

1. BMS's control HVC and LVC for each battery separately. This is a freestanding system, last ditch protection of the batteries if everything else fails, which does not use any networking functions.

2. What information can the BMS communicate to Cerbo and/or Multiplus, which is useful? I understand we don't want JK's SOC measurements which are not accurate, right? Do we let the BMS inform Cerbo or Multiplus that there is a rogue cell with too high voltage, for example? So that charging is stopped prior to some other reason to stop charging (reached the target SOC, reached voltage)?

3. We let Cerbo calculate the SOC of the the whole pack, based on a Smart Shunt placed AFTER the bus bar which combines the negative leads of each battery, like in Nick's schematic.

4. So we let Cerbo control the alternator, right? The regulator is programmed to start and stop charging according to voltage, but we let Cerbo preempt the regulators "stop charge" command by cutting the ignition wire if Cerbo judges 100% SOC (or whatever has been programmed) has been reached? Right? That would be in case of a non-CANBUS regulator like the Balmar. Otherwise, in case of Wakespeed etc., this command is given over CANBUS, right?

5. Multiplus also has a self-managed charge profile based on voltage, right? But we also allow Cerbo to tell Multiplus via CANBUS to stop charging, if Cerbo judges that 100% (or whatever has been programmed) has been reached?

Please rip this apart if I've got something wrong.

Some further questions:

1. Alternator protection. We have relays on both Cerbo and BMS's which can be programmed and used for different stuff. I think (if I understand the principles right) that we will not blow the diodes in the alternator if only the ignition or field is cut at least a couple of seconds before the contactor is disconnected. Right? So maybe it would be good to let the BMS's do this -- one relay (on each BMS) is programmed to cut the ignition wire at a slightly lower voltage, than the relay which disconnects the HVC contactor. Or -- if voltage isn't right, since a rapid voltage spike would activate both relays almost simultaneously -- the second relay acts with a couple of seconds of delay after the first one? If we're using a CANBUS regulator, then this would be completely separate from the high voltage or full SOC "stop charging!" signal from the Cerbo, which we would program to happen at a slightly lower voltage, so redundant.

2. What does the Cerbo communicate to the Multiplus, which makes it behave any different than its own regulation does? Is it a signal the 100% (or whatever) SOC has been reached, so preempts the voltage-determined decision to stop charging? How does this work?

3. What if anything do we want the BMS telling the rest of the system? And to do what? For example, do we want the BMS to tell Cerbo to stop the charge sources charging based on some abnormal condition (one rogue cell for example)? Or what exactly? What commands from where, preempt what? Or do we just have the BMS controlling HVC and LVC and perhaps cutting alternator ignition wire, and nothing else.

I hope this conversation is benefiting others, and I'm grateful to the experts for their patience.

[s/v Jedi]

Originally Posted by Dockhead

Some further questions:

1. Alternator protection. We have relays on both Cerbo and BMS's which can be programmed and used for different stuff. I think (if I understand the principles right) that we will not blow the diodes in the alternator if only the ignition or field is cut at least a couple of seconds before the contactor is disconnected. Right? So maybe it would be good to let the BMS's do this -- one relay (on each BMS) is programmed to cut the ignition wire at a slightly lower voltage, than the relay which disconnects the HVC contactor. Or -- if voltage isn't right, since a rapid voltage spike would activate both relays almost simultaneously -- the second relay acts with a couple of seconds of delay after the first one? If we're using a CANBUS regulator, then this would be completely separate from the high voltage or full SOC "stop charging!" signal from the Cerbo, which we would program to happen at a slightly lower voltage, so redundant.

2. What does the Cerbo communicate to the Multiplus, which makes it behave any different than its own regulation does? Is it a signal the 100% (or whatever) SOC has been reached, so preempts the voltage-determined decision to stop charging? How does this work?

3. What if anything do we want the BMS telling the rest of the system? And to do what? For example, do we want the BMS to tell Cerbo to stop the charge sources charging based on some abnormal condition (one rogue cell for example)? Or what exactly? What commands from where, preempt what? Or do we just have the BMS controlling HVC and LVC and perhaps cutting alternator ignition wire, and nothing else.

I hope this conversation is benefiting others, and I'm grateful to the experts for their patience.

1. The reason for not disconnecting the field wire is that this can damage the regulator. Yes this is a rather small current (say 7A) but also the driver in the regulator is sized accordingly. Disconnecting field wires is done using manual switches or the safety switch inside a main battery switch. Some even use a rheostat in this line to do manual control. This is all from the time before we had smart regulators.

You want to stop charging from the alternator well before any HVC caused by overcharging happens so there is no timing issue. Trying to charge to the last Coulomb from the alternator is fools wisdom.

2. Again, saty away from this. What other brand inverter/charger could you use when the Multiplus fails, the brand is bankrupt or sold out and you must conform to such communication requirements? And when everything works well, what is gained by it? (Answer: absolutely nothing)

3. Again, nothing. The B in BMS stands for Battery. Some want it to steer the boat but it’s just to manage the battery it is in.

A word about alternator protection devices. I do recommend those so let me explain why and for what: these are a last line of defense for when things go very wrong. They are hopefully never used.

So normally the alternator is switched off long before a HVC can happen because we only charge to 99% or less. I like 95-98% better than 99%. But things can go wrong, let’s say a cell fails and shorts and the battery goes HVC. Now you have only one battery left online which has to do double the power of normal and this happens exactly when you decided to use microwave, induction and A/C all simultaneously even though you know this requires both batteries online but as they always are (were) complacency has sneaked in. The battery left can’t sustain the load and it’s BMS does a cutoff based on overload and this leads to an unexpected total loss of power and hopefully the alternator protector does its job.

[CaptainRivet]

Originally Posted by Dockhead

OK, thanks very much for this.


DVCC = "Distributed Voltage and Current Control"

"Enabling DVCC changes a GX device from a passive monitor into an active controller."

"Managed CAN-bus batteries: In systems with a managed CAN-bus BMS battery connected, the GX device receives a Charge Voltage Limit (CVL), Charge Current Limit (CCL) and Discharge Current Limit (DCL) from that battery and relays that to the connected inverter/chargers, solar chargers and Orion XS. These then disable their internal charging algorithms and simply do what they're told by the battery."

OK, I think I understand this -- it means that Cerbo takes over charge control from all the charge sources, and manages them according to what it gets from the BMS's.

I think both you and Nick said you ended up preferring the charge sources to manage themselves. So, DVCC OFF. I can see that this could be simpler and more reliable.

In that case, then no particular reason for a CANBUS alternator regulator, right?

So then we use the Cerbo to cut off the alternator in specified situations, and we use the BMS's to manage the HVC and LVC.

But what about the Multiplus? Do we leave that self-managing or does it take some orders from Cerbo? Nick was talking about turning the Multiplus down based on tail charge current or SOC. How does that work?

You say Cerbo uses SOC from the BMS's -- there is no option for it to do its own based on Smart Shunt data?

Now you got it right.
Yes you don't use the managed can bus battery.
And DVCC has so much dependencies, as I understood Jedi uses DVCC (??) and I don't and do the charge sync via BT connect.

No need for a canbus alternator. To cut the alternator earlier use cerbo relay steering or BMV or VSR. If your alternator is an important charge source to you and you want it's data, the best is to use a dedicated BMV 702 or 712 that a)cuts with its relay output at beginning of the knee and b) reports via Ve-direct into cerbo the charge current and voltage, you just set this BMV as meter and give it the name alternator meter. And you use one of the cerbos temp measure input and put a temp sensor on the alternator. Like this you have all data in the cerbo for warnings and automation like putting a fan on for the engine room....

Yes you can use the BMV that is your last resort Device connect via VE.DIRECT to cerbo and define it as main bat monitor. This will deliver SOC and battery voltage to cerbo and you can use it for steering plus this delivers also via BT the reference voltage for BT connect charge sync. That's the way I have it as electrodacus BMS has no communication to victron.

Multi has its own internal charge algorithm and parameter so you simply use a lower or higher tail current in the parameter of the multi then eg in your MPPTs. There are additional ways best Jedi explains.
I use the multi 98% as inverter and 1.9% the honda as backup charge where the target is to get the battery in the bulk charge phase full as quickly as I can, don't care about the knee as there I basically never use it as honda is shut off before. As electrodacus BMS as contactor can steer individual charge sources I simply have absorption in multi set higher then end if charge value in BMS so multi runs in bulk mode till electrodacus cuts it off so bank is charged to approx 95% but in the shortest time possible. The rest is job of solar.
But for a dock boat with a berth and frequent shore power you use different parameters for the multi then your MPPTs to optimize the perfect charge