System Design for EVE304Ah 2P4S

[TheAutomotiveEngineer]

I’ve finally received my 8 EVE304Ah A-Grade Cells this week, and I’m thrilled to start designing the battery pack for our van build! I chose these particular cells along with a Multiplus2 (12V/3000W/230V), an MPPT for solar charging, and a DC/DC charger to meet our requirements.

However, since I’m new to building battery packs and 7.5kWh is a lot of energy to handle, I don’t want to do this alone. I’m looking for advice, suggestions, and insights from those who have experience or knowledge in this area. What components would you recommend for seamless integration?

First on the list is deciding the layout. My initial and preferred way is to configure parallel groups of 2 cells and put those in series. This is mainly cost-driven as it decreases the number of components needed to keep the system safe. However, I recognize that two separated banks might have a few advantages in terms of redundancy, though this doesn’t outweigh the cost argument at the moment. With two separated banks (4s2p), I expect to double up on relays, BMS monitors, and current sensors. If there are other compelling arguments for this "alternative" configuration, I’m open to discussing them.

Assuming I stick with the 2P4S configuration, I need to select an appropriate BMS. After some research, I’m considering the 123Electric BMS for these reasons:

• Monitoring of cell current and temperature
• Relatively high balancing current
• Direct cell connection
• Clear documentation and easy-to-reach support
• Integration with Victron (though I’ve seen comments that it might not be fully supported – is this correct?)

However, I noticed that the 123Electric BMS is only top-balancing by dissipating energy instead of shifting it around the pack. I’m unsure if this is a such a bad thing, as an "active" system that moves energy through the pack might not be optimal for a system with extended periods of little to no use (it is a van after all).

Before moving forward, I’d like to clarify these initial questions to ensure a safe and robust system. Any insights, experiences, or recommendations would be greatly appreciated.

 

[OffGridForGood]

First: Welcome to the DIY Solar Forum !
Second: I have built several DIY batteries and my signature block shows some examples. Prior to solar I build DIY E-bike packs from small 18650 cells in 14s 4p and 14s 5p arrangements. I have also built batteries for serveral other mobile thinks like lawn equipment cutters etc.
All to say - I have some experience.

First on the list is deciding the layout. My initial and preferred way is to configure parallel groups of 2 cells and put those in series. This is mainly cost-driven as it decreases the number of components needed to keep the system safe. However, I recognize that two separated banks might have a few advantages in terms of redundancy, though this doesn’t outweigh the cost argument at the moment. With two separated banks (4s2p), I expect to double up on relays, BMS monitors, and current sensors. If there are other compelling arguments for this "alternative" configuration, I’m open to discussing them.

I recommend separate packs of 4s 1p
redundancy
lighter to move install and maintain
one fails, you still have another operating
balancing issues.

Assuming I stick with the 2P4S configuration, I need to select an appropriate BMS. After some research, I’m considering the 123Electric BMS for these reasons:

I am not familiar with 123 electric and can't comment on it.
I use the JK BMS for the DIY packs I build. Available in high amperage models.
The JBD is also well respected.

• Monitoring of cell current and temperature
• Relatively high balancing current
• Direct cell connection
• Clear documentation and easy-to-reach support
• Integration with Victron (though I’ve seen comments that it might not be fully supported – is this correct?)

However, I noticed that the 123Electric BMS is only top-balancing by dissipating energy instead of shifting it around the pack. I’m unsure if this is a such a bad thing, as an "active" system that moves energy through the pack might not be optimal for a system with extended periods of little to no use (it is a van after all).

A well balanced quality set of cells (includes EVE 304's) will not need active balance in my experience, but can be added.

Before moving forward, I’d like to clarify these initial questions to ensure a safe and robust system. Any insights, experiences, or recommendations would be greatly appreciated.

I hope to see more on your project as you make decisions and proceed.

 

[TheAutomotiveEngineer]

I'm looking into the JK BMS but have a hard time navigating through their products and finding proper documentation, even though a lot of people seem to use them and a lot of experience from the public is available I do value some documentation and support service.

I get your point of the dual pack setup regarding the redundancy. Do you have any experience on how the JK BMS would communicate/integrate with a Victron Cerbo GX system?

Furthermore, to the best of my knowledge the Victron system does allow for multiple banks but would prioritize one over the other. If this is the case I assume it will it will always cycle the "main" pack with a higher C-Rate, effectively reducing lifetime albeit close to negligible. Still, this is my assumption, if the experience is different I'm happy to be corrected!

 

[HRTKD]

I have 4s2p connected to a Multiplus, but without any communication between the inverter and two BMS. I haven't seen where two BMS can communicate with the Victron system, but I haven't looked very hard for an example either. For rack mount systems it's different, they're designed to work together.

The discussion about paralleling cells has been going on a long time on this forum. I'm in the camp that says don't do it. Finer granularity of control and reporting, redundancy and lighter weight are some of the reasons to not parallel cells. Paralleling cells does avoid the cost of a second BMS and you may see more success with just one BMS if inverter communication is the goal.

 

[TheAutomotiveEngineer]

Thanks, integration of the complete system is pretty high on the list for me. More control is a solid argument against the parallel option, but I don’t see how weight plays a role here. In the end I will have the same amount of cells in the system, or am I forgetting something here?

In the end I just have to settle for one or the other. At least I get the impression that both options are valid and there is no right or wrong.

For a 2P4S I imagine a power relay and a shunt on the negative pole to suffice before entering at the battery terminals of the MP. I belief load and charge behaviour will be switched on the MP so no need for separate load and charge relays.

How would this work for a 4S2P setup? Does it suffice to double up the power relays and working with a single shunt on the common negative line?

 

[HRTKD]

4s2p has a cable from each battery going to the shunt. Then from the shunt to the common negative bus bar. Everything, including the inverter, connects to the common negative bus bar.

I don't see why you're adding a relay to the circuit.

 

[OffGridForGood]

I'm looking into the JK BMS but have a hard time navigating through their products and finding proper documentation, even though a lot of people seem to use them and a lot of experience from the public is available I do value some documentation and support service.

Check the resources section on the forum likely someone has posted docs.

I get your point of the dual pack setup regarding the redundancy. Do you have any experience on how the JK BMS would communicate/integrate with a Victron Cerbo GX system?

I don't know much about Victron, I use their Smart Shunt and that has been a game-changer for my set up - very accurate and reliable Be Sure to use the Correct cord!

Furthermore, to the best of my knowledge the Victron system does allow for multiple banks but would prioritize one over the other.

No, the packs will all connect to common terminals - batteries of equal capacity and resistance will charge together.

If this is the case I assume it will it will always cycle the "main" pack with a higher C-Rate, effectively reducing lifetime albeit close to negligible. Still, this is my assumption, if the experience is different I'm happy to be corrected!

 

[TheAutomotiveEngineer]

Ok, regarding the BMS option I haven't made up my mind yet. I've been reading up on the differences of common vs separated ports and didn't get a full understanding why to use one or the other. The JK BMS comes along pretty often although I haven't found a way to integrate two battery banks (in case I would go 4S2P) with the Victron system. I've been looking at alternatives such as REC and Batrium, quite like the options they both offer. With the Batrium I found that there is a possibility to use their Watchmon core together with two K9s to control two individual 4S1P banks. In case I would go for the JK BMS appoach, I would prefer a 2P4S configuration due to my concerns of uneven bank usage if separated. Even though @OffGridForGood mentions that this wouldn't be a concern, I'm not completely certain. Here is a draft of how a JK Setup would look like:


As I'm interested in the REC and Batrium modules as well, I'm sketching their schematic layouts as well. For the Batrium I'm drawing up a 4S2P (two individual banks).

In the mean time I'm capacity testing my cells, this is taking quite a bit of time as my equipment (EBC-A40L) allows for 40Amps max. charge/discharge. The first cell I tested looks good and gives out: 306Ah. This is for the case where I charged the battery at 40A to 3.65V with an absorption of 20A and a discharge to 2.5V. I'm now charging them to 3.3V with full (100mA) absorption. Waiting time between charging and discharging has been 1hour (the image shows 30min wait, but I started the cycle test 30min after the charging was finished).

Here is a graph of the first cell:

 

[hwy17]

That's a cool BMS. I encourage you to continue with it if you are up for this battery proliferating into an electrical passion project.

If you would rather get it built and move on with life, a JK or something is a faster, easier, and good enough option and comes with much more community online to ask about each setting. But if you are up for drawing the project out a contactor BMS will force you to learn a lot deeper.

I built with an Orion JR2, it took me 6 months, countless hours, and several multiplications of my initially estimated parts and accessories budget.

I did end up with a seperate charge and load bus design. It is worth it if you can easily separate your chargers and loads. If you can't, there should be a way with this BMS to build a common bus design. Common bus will just mean that when the BMS needs to protect the battery it will disconnect everything. This is ok, as the BMS entering protection should be a rare event and a last resort anyway, only in the event that something has gone seriously wrong.

Don't worry about passive (burn off) balancing, my 304 pack is totally fine with only 150mA of passive balancing current.

 

[TheAutomotiveEngineer]

That's a cool BMS. I encourage you to continue with it if you are up for this battery proliferating into an electrical passion project.

If you would rather get it built and move on with life, a JK or something is a faster, easier, and good enough option and comes with much more community online to ask about each setting. But if you are up for drawing the project out a contactor BMS will force you to learn a lot deeper.

I built with an Orion JR2, it took me 6 months, countless hours, and several multiplications of my initially estimated parts and accessories budget.

I did end up with a seperate charge and load bus design. It is worth it if you can easily separate your chargers and loads. If you can't, there should be a way with this BMS to build a common bus design. Common bus will just mean that when the BMS needs to protect the battery it will disconnect everything. This is ok, as the BMS entering protection should be a rare event and a last resort anyway, only in the event that something has gone seriously wrong.

Don't worry about passive (burn off) balancing, my 304 pack is totally fine with only 150mA of passive balancing current.

Which one are you referring to, the REC or the Batrium approach? My goal is twofold, it should be a robust system which is able to play nice with the loads, chargers and the Victron multiplus2. Second goal for me is to learn and not be ready as quick as possible, if that was the case I would have bought a plug and play battery solution

 

[hwy17]

Which one are you referring to, the REC or the Batrium approach? My goal is twofold, it should be a robust system which is able to play nice with the loads, chargers and the Victron multiplus2. Second goal for me is to learn and not be ready as quick as possible, if that was the case I would have bought a plug and play battery solution

I'm talking about the original idea for the 123, if you're in the EU. I don't have a detailed judgement of it, but it looks like an active product and they work hard at their victron implementation. I don't know, maybe it's a good system or maybe it's not, but my message would be don't be discouraged just cause no one else here is using it.

I do wonder about their encouragement of solid state relays for primary means of BMS disconnect, but maybe they're on to something and those are the future. Contactors have been more common and proven and the 123 can probably use contactors too.

 

[TheAutomotiveEngineer]

The 123 is also still on the list, their solid state relays is anyways not an option as it is for 48V systems, so here I would opt for a more conventional contactor. But I have the same concerns with 123 when I go the dual pack (2 times 4s1p) route, integration seems to be lacking and communication between the different packs isn't officially available. In that case, there are some enthusiast battery merging projects available such as this one: https://github.com/Dr-Gigavolt/dbus-aggregate-batteries . I need to do some more research if these options are actually viable and allow a robust setup.

 

[hwy17]

But I have the same concerns with 123 when I go the dual pack (2 times 4s1p) route

I don't think it is a bad idea to build a 2p4s. This forum is very biased towards multiple packs, but that's also based on the fact that most people here are building with Chinese BMS that are known to fail. In the other lineage of DIY battery builders from the EV scene paralleling cells is a well respected practice. The Orion BMS creator even prefers it to parallel packs, although he did eventually add in support for multiple pack configurations too.