diy solar

diy solar

"Smart Balancing" - A better way to maintain cell balance.

the_colorist

"Move over... let me fix it" Installer/Engineer
Joined
May 1, 2020
Messages
389
Location
USA - Working in Central America
As an installer/engineer, I'm usually explaining this once a week or so (sometimes more) so I felt I would write up a quick post on it. Perhaps it has been covered elsewhere (I know I've written about it a few times) but I didn't bother to dig around.

I'm not going to get technical here, this is a very high-level post. Intended for the benefit of everyone. There are a lot of guys here who will instantly pick up on what I'm talking about and help to flesh things out for anyone coming along with questions. Don't get me wrong, I'm glad to help with explanations but my time these days is unfortunately extremely limited.

I get asked regularly: What is the benefit of some of the more expensive BMS units (Orion, REC, etc) over the cheaper ones?

Well, there are a lot of advantages but I'm only going to mention one in this post.

BALANCING

At first, this may seem odd as it doesn't appear that those units on average have any higher current balancing capabilities or really anything else in the "specs" that would lead one to believe that they are more advanced in any way, there is however one extremely important difference that sets them apart.

COMMUNICATION

What I'm referring to here is the CAN bus communication link between the BMS and the system (Victron/SMA/Growatt etc) that is charging/discharging the bank.

I find in talking with people/clients that this specific feature seems like a black box and even appears to possibly be unnecessary/optional but in fact it serves a very, very useful purpose with regards to balancing.

There has been plenty of discussion of active balancing vs passive balancing and how passive balancing doesn't appear to do anything in some cases and yet we see balancing readouts on BMS units such as Orion that look like this (@cinergi I hope you don't mind I stole this):

1622685229883.png

So what's the diff here? Does passive balancing work or not? Do I care about communication or do I not?

Here is how it works:

During the charging phase, the BMS monitors the per-cell voltage among many other things (nothing strange here). At the same time, it uses the per-cell information to make calculations about the most optimal charging current/voltage for the bank and then it uses the communication link to send the following commands/information to the system the battery is powering. It varies some across the market but here are the basic commands.

Bear in mind these commands are sent every few hundred milliseconds, sometimes even faster. And the commands (Max Charging Current for example) will vary based on the state of the cells. Lines in italics are information messages.

1. Max Charging Voltage (Pack Full Charge Voltage Target)
2. Max Charging Current (DC charge current limitation)
3. Max Discharging Current (DC discharge current limitation)
4. State-of-Charge (SOC value)
5. State-of-Heath (SOH value)
6. Battery Pack Voltage (Total pack voltage as read by the BMS)
7. Battery Pack Current (Pack charge/discharging current as read by the BMS)

8. Alarms (cell over-voltage etc)

During the charge phase, as the pack voltage rises, there is usually a target voltage per cell when the passive balancing resistors come on.

Given enough time, even the lowest current balancers could balance a bank the amount of time it would take would be ridiculous depending on the size of the bank. Assuming a properly sized BMS/balancer vs bank capacity, it could actually achieve it in a reasonable time. The key is in communication.

As the first cell begins to reach the balance start voltage, the balancer kicks on. Nothing strange here.

As the first cell reaches the per-cell full charge target voltage (3.45VPC etc) however, this is where the magic happens.

The BMS uses the communication link to instruct the inverters/charge controllers/etc to begin to dynamically back down their charging current to the point where:

The balancer is able to keep up with burning off the power from the cells that already at the per-cell full charge voltage target (preventing a rise to cell over-voltage) while the low voltage cells are still charging.

The BMS at this point could be requesting a pack charging current of 15.6A, 10A, 2A or even 0.5A. It all depends on how many cells are at or near their full charge target voltage and how many are still needing to charge.

This phase at the end of every charge cycle will last as long as it takes (or as short as it takes) to get all of the cells to reach the full charge voltage target (per cell).

This could be 3.45VPC etc. It will depend on the targets programming into the BMS.

On the Orion, there is an extra special feature.

Victron explains it like this:

Adaptive Maximum Charge Voltage Set-point – the value transmitted varies from 54.8V to 55.8V depending on the state of balancing on the particular ****. Typically a well-balanced **** battery will request a maximum charge voltage of between 55.0V and 55.8V. This advanced feature is only available from ***** and allows superior system control and optimal battery management. The GX-device uses this set-point to control the real time operating or target voltage of the inverter/charger devices and the MPPTs."

That comment is associated with a specific brand of commercial battery banks that use Orion BMS units internally. I'll let you figure out who it is.

REC explains their charging algorithm like this:

************

The communication between the REC BMS and the Victron CCGX is established through the CAN bus. All the parameters that control the charging/discharging behavior are calculated by the BMS and transmitted to the CCGX unit in every measurement cycle.

The charging current is controlled by the Maximum charging current parameter. It’s calculated as Charging Coefficient ('C','H','A',’C') x Battery capacity.

The parameter has an upper limit which is defined as Maximum Charging Current per Device ('M','A','X',’C') x Number of Devices ('S','I','S',’N').

When any cell reaches the voltage interval between Balance Voltages Start and Balance Voltage End, the charging current starts to ramp down to 1.1A x Number of Devices until the last cell rises to the End of Charge Voltage. At that point, the Maximum charging voltage is set to Number of cells x (End of Charge Voltage per cell –0.5 x end of charge hysteresis per cell) and the charger is disabled also via the BMS I/O interface.

End of Charge, SOC hysteresis, and End of charging cell voltage hysteresis prevent unwanted switching. SOC is calibrated to 100 % and Power LED lights ON 100 %. Charger turn-off can also be caused by some of the system's errors (See System Errors indication chapter). SOC is calibrated to 96% at the 0.502 x value between Balance Voltages Start and Balance Voltage End.

************

So active balancing? Does it solve the issue?

Well, we using them with "Smart BMS" units with the type of communication/algorithms mentioned above and they do help. They aren't always needed BUT with A- cells that have varying IR and are powering systems with large loads with high startup currents pulling the cells out of balance, I feel they can be a big plus.

In the future I feel "Smart Active Balancing" in a single BMS unit would be even better but for now, communication backed by a great charging algorithm and appropriate cell voltage targets will get us where we want to go IMHO. There are a couple units on the market that I'm aware of with "Smart Active Balancing" such as Autarctech's BMS units and REC's Active BMS but unfortunately one is over $1K last I checked and the other is 4S so not appropriate for budget 15/16s banks. I'm sure we'll get there though. Even some commercial companies are realizing the usefulness of active balancing. Tesvolt APU is a great example of this.

That said, nothing helps to get things rolling in the right direction like a great top-balance...
 

Attachments

  • UserManual_REC_Victron_BMS_Protocol.pdf
    1.8 MB · Views: 35
That method of balancing is kinda what I did with my Chargery BMS with passive 1.2A balancing and and adjustable power supply.... but I had to be the "communication link."
So far I have only needed to do this once and the normal balancing the BMS does has kept things in balance. I can't afford a gazillion $ of equipment to do this for me. If things get out of balance I will use the power supply again.
 

For what it's worth, I do not have my Orion changing the target charge/pack voltage (it's fixed at 55.2v), and I can get the above balance to occur on my "Grade A-" cells, even in a 540Ah bank in a 2p16s configuration (with cells varying from 270Ah to 281Ah). My personal conclusion is that standard passive balancing works quite well with the right (AKA quality) BMS, even without CANBUS. The one huge advantage of connecting/controlling via CANBUS is that the BMS can instruct the Victron ecosystem to slow down or even stop the charging if any of the cells get too high, despite the target charge voltage. Without this, your granularity is limited to the entire pack voltage and you risk the BMS cutting off the battery entirely. In my case, CANBUS is a safety control, sort of like the contactor or MOSFETs of other BMS without the downside of disconnecting your battery; the contactor is a secondary safety cutoff. I have yet to see the primary (CANBUS) or secondary (Contactor) controls kick in for my battery bank -- again, meaning that passive balancing is working just fine.
 
Good write up. I have been using a REC BMS for many years, the ability to slow charging when temperature is high or balancing is required is a priority for me.
 
Great information.

Probably less important to those of us underpaneled folks that charge at 0.1 or 0.2C but more critical when you generate enough current to overdrive the passive balancing I would imagine.
 
This is true, cost is many times a factor. This is why we've been working on getting a lower-cost BMS that we can program with the logic we're looking for. We don't have one yet but I feel we are close possibly.
Have you done any investigation of the Chargery BMS paired with their power supply / charger?

I have read that the Chargery BMS will control the power supply but haven't found anyone who has actually used this combination.
 
Great information.

Probably less important to those of us underpaneled folks that charge at 0.1 or 0.2C but more critical when you generate enough current to overdrive the passive balancing I would imagine.

Unless your maximum charge current is less than your balancing current, no. What needs to happen is that your first stage charge voltage is set appropriately so that you never drive any single cell too high (that's why initial top balance is important). Then either at this voltage or at a slightly reduced absorb voltage (Second stage), the balancer does it's work. 0 amps will come into the bank unless there are cells out of balance, in which case the current coming into the bank should be roughly equal to the maximum balancing current that the BMS can perform.
If your battery bank gets SO out of balance between charges that a cell will go too high, then a CANBUS solution is a great way to deal with that situation -- assuming you can hold the bulk/absorb voltage long enough for the balancer to do its job.
 
Hi, did you know where we can find the canbus protocol detail for limiting current or change target voltage from these corporate charger?
Only one day canbus will fail, and only god can know what's happens
 
Last edited:
If the canbus signal fails on my system, the batteries disconnect.

The canbus protocol as used by SMA and Victron is readily available.
 
If the canbus signal fails on my system, the batteries disconnect.

The canbus protocol as used by SMA and Victron is readily available.
if alternator are loading massive amp to the battery then there is a drama , this single point of faillure is crazy
 
There is no single point of failure.

CANBUS signal fail - battery disconnect.
Over voltage at cell or pack level - battery disconnect.
Excess current - battery disconnect.

Doesn’t seem crazy to me. What CANBUS system failures have you observed?
 
Canbus transceiver in my car, canbus in the sailing boat on nmea2000 systems because lightening near from the boat.
 
I am in agreement with nonoceb that their is danger in relying on communications for "critical" functions. Some may have seen where I have stated this on other posts.

I would define critical as anything that will cause damage to the battery or other equipment if it doesn't work properly .... or create a safety concern.

This can be done safely, however, but the coding must be done with meticulous care to set up a communication "heartbeat check" with actions taken to keep things safe if communications fails.
 
I agree, but the CANBUS isn’t the last line of defence. There must be millions of CANBUS monitored battery systems around, i’ve seen CANBUS failure, but haven’t heard of one that caused cell damage.

Based on failures that i have seen, CANBUS is well down the list of things that will cause my power to go off - and not even on the list of things that will damage my battery.

I’d suggest a rethink of your system design if your system can cause cell damage due to CANBUS failure.
 
I agree .... if all that is being done is monitoring .... but some of the description is describing full control.
 
I've tested this on a REC ABMS, with Victron Cerbo, Multiplus, and SmartSolar 150/100 with VE.CAN. The BMS goes to the BMS.CAN on the Cerbo, and the VE.CAN goes to the SmartSolar, and VE.Bus to the Multi.

The main contactor is bypassed for these tests. DVCC is on, with the REC providing SOC info as well as SCS, SVS, and STS.

Test 1:

System at 85% SOC and under a 50A discharge load via the Multiplus. Disconnect CANBUS. Within 30 seconds the Multi drops the loads as no discharge current limit being provided to it.

Test 2:

System at 70% SOC and being charged by the Multi at 110A to the batteries. Disconnect CANBU. Within 30 seconds the Multi ramps down the charge rate to zero.

Test 3:

Sartsolar putting out 20A to the battery to charge. Disconnect CANBUS. Same situation - Smartsolar goes into off state.

Test 4:

Under charge from Multi. Use heat gun on batt temp sensor to raise temp above 55C. Multi stops charging. 55C is a programmed set point.

Test 5:

Same as 4 but this time put battery sensor onto a frozen ice pack. Again, once it reports 9C the Multi stops charging. This is a programmed set point.

The only place I would have issues:

- B2B charger not being under CANBUS control.
- DC loads behind a Smart Battery protect is not under CANBUS control.

Both of the above two are dealt with using the output IO signals to get the desired behavior.
 
I've tested this on a REC ABMS, with Victron Cerbo, Multiplus, and SmartSolar 150/100 with VE.CAN. The BMS goes to the BMS.CAN on the Cerbo, and the VE.CAN goes to the SmartSolar, and VE.Bus to the Multi.
thank you for your feedback, so you can confirm the main contactor never open on canbus loss?
I will install a recbms16s next week and your information is really appreciated, i will just add WS500 alternator controller on your setup!
Because i have seen 20000$ damage (autopilot and every 24v smart devices fryed) on sailing boat just because a not smart enough alternator controller, not properly configured for a good talk with bms. it's was mastervolt but same topology, software incompatibility lol.
Damage is from very inductive charger, typically alternator winding, not from high frequency switching charger.
 
Hi,
i lookingfor some information to setup rec bms + gateway victron cerbo gx + victron skylla charger in smart way to slowdown charge or even stop charge when higher cell > 3.65V.
On the gateway everything is link with VE.can or VE.bus (inverter phoenix), recbms is link with BMS.can and nothing happen before the BMS decide to release lifepo4 to protect overcharge.
Every device is well listed on gateway, i tried lifepo4 mode and canbus mode on skylla charger...
It's not smart situation for the moment :poop: and hell is near when you need to top balancing 24V 1400Ah winston lifepo4 bling bling cell with a poor resistive balancing!
 
As the first cell reaches the per-cell full charge target voltage (3.45VPC etc) however, this is where the magic happens.

The BMS uses the communication link to instruct the inverters/charge controllers/etc to begin to dynamically back down their charging current to the point where:

The balancer is able to keep up with burning off the power from the cells that already at the per-cell full charge voltage target (preventing a rise to cell over-voltage) while the low voltage cells are still charging.

The BMS at this point could be requesting a pack charging current of 15.6A, 10A, 2A or even 0.5A. It all depends on how many cells are at or near their full charge target voltage and how many are still needing to charge.

This phase at the end of every charge cycle will last as long as it takes (or as short as it takes) to get all of the cells to reach the full charge voltage target (per cell).

This could be 3.45VPC etc. It will depend on the targets programming into the BMS.

Thanks for this explanation. Sounds great. I bought a BMS (Heltec #1026) with ports for CANBUS and RS485 (aka MODBUS?) and a Sol-Ark 12k with support for CANBUS/MODBUS, and I hope to eventually get the inter-communication working, although the documentation seems lacking. In the meantime I hope that the Heltec's integrated 2A active balancing keeps things in order.

My biggest question is what if there are multiple batteries with separate BMS? It seems like this communication wouldn't be very effective as one BMS might request one charging current while another BMS requests another. Or perhaps the charger would just accept all requests and then default to the lowest requested current? That would undercharge some packs while one receives final full charge, but then presumably that pack's BMS would "disconnect" to prevent overcharge and other packs could charge at the next requested current? Complicated and could add some time to for all battery cells to reach full charge, but ultimately workable? Or better off skipping the CANBUS altogether and let active balancer deal with any issues?

Also, what's the difference between CANBUS and RS485/MODBUS? Are they just competing communication standards that mainly accomplish the same thing with different syntax, so I just need to pick one to use on both ends? Any good reasons to pick one over the other?

FWIW, Sol-Ark support said multiple separate batteries should work fine with CANBUS (although they did caveat that only Orion is fully tested/supported), but they didn't get into any technical detail about how the intercommunication would work.

My question:
I'm planning to install a Sol-Ark 12k outdoor unit and will build or buy a compatible 48V battery array. Can you provide any information on which battery BMS support CanBus/RS485 and are compatible with Sol-Ark? Can the Sol-Ark 12k communicate with multiple BMS in this way if there are several batteries attached in parallel, each with their own BMS? If so, is there an upper limit on how many BMS are supported? For example, I'm currently planning to assemble 64 LiFePO4 cells into four separate 48V batteries with four dedicated BMS. Is that viable? Please advise.

Sol-Ark support response:
This shouldn't be an issue. If you go with Orion and use CanBus this should be fine. You can contact Orion for additional information on their protocols. You will likely use address 02 when doing your settings.
 
Last edited:
Back
Top