Optimum LFP Battery Cabling/Wiring Attachment Points

[Dave McCampbell]

Hi Dacian,

I am considering increasing our LFP capacity from 4S2P to 4S3P, ie 4 sets in series of 2 cells in parallel to 4 sets of 3 cells in parallel.  I am interested in finding the best place to attach Cat 6 monitoring/balancing wiring and the main system supply cables to the parallel cell blocks.  

I want to get the best connection point for balancing and monitoring the cell block voltages.   Also, I want to get the best attachment point for the main system supply cables going out to the shunts in order to draw as evenly as possible from all the 12 cells.

My thought is to attach the Cat 6 wiring to the center cell of each cell block rather than on an end cell of each block.  For the system supply attach the + and - cabling to the shunts to the center cell at each end of the battery.   Is this optimum or is there some better attachment scheme?

TIA  Dave

[Dacian Todea (electrodacus)]

Dave,

Yes center cell will be a good choice but if cell bussbars are connected symmetrically you should have the exact same voltage at any point.

I recently answered a question to someone that has a 8s3p regarding how to arrange the cells and how to connect them so current splits equally. I do not think he will mind posting the photo here.

The connection between positive terminal should be made out of 4 short but equal length wires connecting the 3 terminals to the battery shunt.

The green bus bars will see the battery current divided by 3 while the blue bus bars will normally not see any current at all at most a few mA so all points on the blue bus bars should have the same voltage but middle cell is symmetrical so it is a good connection point.

[Axel]

Dacian,

Since the blue bars don't experience much current can they be a smaller gauge then? What about their length in relation to the green wires?

Looking at your EVE battery build you also have symmetrical flexible copper busbars, two "blue bars" - one across the positives and the other one across the negatives. Is that necessary?

How does one size those "green" wires? According to max load/charge current and then divided by 4 in case of a 8SP4 battery?

Thanks,

Axel

[Dacian Todea (electrodacus)]

Axel,

The blue wires will not have any current flow trough them in normal conditions but if there is a bad contact or one of the green bars is missing then there will be some current flow.

No it is not necessary to have the blue bars on both positive and negative but since I got those included with the cells I used all of them as I had no other use for them.

Yes for 8S4P in normal conditions the current will split in to 4 but since it is possible one connection is bad I will use the total current divided by 3 to allow for such fault cases and also in that case the blue will be designed for same current as green.

Same will be valid for yellow.

So say you protect that EVE 8S4P battery with a 400A circuit breaker or 400A fuse then each green and yellow bus bars will experience 100A max in normal operation and in case of a single fault it will be 133A thus you design the bus bars to handle 133A

So 16mm^2 flexible bus bars will be more than sufficient in this example but if you expect long periods at 400A you can use thicker bus-bars.

My Battery 8S2P uses 16mm^2 flexible silicone wires as bus bars but my max total current is 125A (breaker rating) and since I only have two cells in parallel if one connection is bad then all current will flow trough a single 16mm^2 bus bar.

The included braided bus bars are complete overkill at 35mm^2 or maybe even 50mm^2 (I do not remember exactly).

The important part to pay attention to is contact resistance between bus bar and cell terminal as that can be significantly higher than bus resistance and usually that is the main problem.

[Axel]

Thank you for the clarifications, Dacian.

For the cell bus bars I would like to use 1AWG marine cable with ring terminals (I have a good quality hydraulic crimper).
The problem I have is that my cells came with 10mm M6 studs. The locknut is 5mm, the ring terminal is 2.1mm thick (0.083"). So I can stack two but three will be pushing it, plus I read that's a not really good practice for ring terminals.

Would something like this aluminum "mini bus bar" work? I will use it only on the battery terminal that needs three connections. I'm thinking this way I would have better contact with the ring terminals with 3 separate stainless steel bolts and it's not too expensive to use a service like sendcutsend.com to fabricate that (USD3.45/piece). Thoughts, comments?

[Dacian Todea (electrodacus)]

This 280Ah - 300Ah cells are not designed for more than 0.5C so less than 150A per cell charge or discharge limit.

Not sure where will you need 3 connections to a single terminal. And stacking two will not be an issue.

Aluminum will create problems and oxidize the terminals are plated aluminium or copper depending on terminal not just aluminium

There are plate aluminium lugs for 2 or even 4 connection points that are plated like say the one in the image below

All you need is aluminium or copper conductors and ferules.

[Axel]

I have three connections on the middle cells, but I guess I was overthinking it. I could do a lug for the blue wires and ring terminals for green and yellow, and would check periodically on those set screws since it's a RV, not a static application.

In a previous post you were recommending for the positive battery connections to be connected directly to the shunt. I have 4 wires with ring terminals. Is it okay to connect them to a nickel-plated copper bar which is connected to the shunt like in the drawing below? What would Dacian do? :)

[Dacian Todea (electrodacus)]

What will the max continues battery current be ? What is the fuse or breaker value ?

You want the resistance between the 4 positive terminals and shunt be equal and that will not be the case with the shunt be on the side of the buss bar.

Having the shunt connected in the middle will be better still not equal resistance that is why I recommend equal length wires all connected directly to shunt.

Something like this connected to shunt and then 4 equal length wires connected to it. There will be no issues with vibrations if you use flexible wires and torque to spec.

[Axel]

Max continuous current could  be reaching 300A for an hour in a cooking scenario, with the AC on, water heater kicking on and other always-on DC loads. I doubt we would ever get there especially with solar production. RV life is also slow life where one does things in series not in parallel, and we would be mindful of not inducing stresses on the system. In any case, we do have the power consumers and wouldn't want to limit the system. So rounding up to 320A which gives a  400A breaker.

Point taken on the shunt-bus bar connection.

Will those lugs provide enough surface contact? Inserting a hexagon or square into a circle? Not being facetious, just want to learn since I haven't used ferules for thicker cables.

Also apologies to Dave for kinda hijacking his thread, but hopefully this is still relevant to the original post.

[Dacian Todea (electrodacus)]

I use similar lugs on my battery.  They connect the cells in series using two 16mm^2 (closer to 5AWG but they are marked with the more common 6AWG) flexible silicone wires super flexible I think around 3000 strands and ferules that are directly crimped by the lugs.

The screws in those lugs are half sphere at the bottom and that will make an indent in the ferule shaping the ferule on the shape of the lug so as much contact surface as possible and ultra low contact resistance.

My battery is 8s2p and just 125A breaker. I frequently charge with 100 to 110A for hours with zero issue all connections are basically at room temperature.

[Marinepower]

I know this slightly off topic, but i have a similar parallel / series arrangement (4P, 4S), however, due to space constraints (see picture below) I could not optimize the current sharing as you show above.     Cells are 100ah each and have very robust, copper M10 terminals allowing for good cell-to-cell connections.

BTW - picture below was taken during install, so the electrical connections are yet made in this image (just a "dry-fit").

See drawing below for the wiring configuration.

I decided to put my sense wires on the last cell of each parallel 4-cell pack, as I figure the 1st and 4th cells in the packs will be the first cells to get to HVC or LVC.

While I know this is not an ideal current sharing lay out, my loads and charging C-rates are pretty modest and I assume the deviation in cell SOC and voltage would be pretty small?  I only charge to about 3.5 v per cell and so far (according to the SBMS0) cells pack have stayed within 0.08 mv of each other.  It's actually a bit alarming as i have never seen the SMBS0 start balancing this battery since i have installed it.  I have been running it intermittently for about 2 years now ( as its on a boat) without  any problem.     However, I know I am only seeing pack-voltage data from the SBMS0 and not individual cell voltages.

Not sure if I should be concerned that I'm not recording any balancing from the SBMS) (inbalances are being hidden) or happy that the 4-packs are keeping such good balance?

MP

[sailingharry]

Dacian has made a point of the need to keep the resistances between the cells identical (same length wires, etc).  This is conventional wisdom, and repeated very often in many forums.  The fundamental theory is easy to understand.  However, I have strong doubts about the practical implication (has anyone ever measured a real-world voltage difference across a 6" bus bar?)  I have never seen a single study of the impact of using, for instance, the 4-cell bus bar with shunt at one end shown up thread.  I suspect that the result might be loss of 1-2Ah at the end of discharging, if the end is reached when discharge rates are above .2C.  At anything below .1C (with a cabling system designed to support .5C), I suspect the actual impact is below measurement limitations.

I'm not looking for an explanation of differing discharge rates due to differing cable resistance.  I'm looking for someone (or some company/organization) who has run actual tests of various cabling systems, and produced quantifiable differences.  If Andy or Will have done something, I'd love to see it.  I'm very skeptical.  This is, in my opinion, an example of pursuing perfection for perfection's sake. 

[Dacian Todea (electrodacus)]

MP,

If your charge discharge rate is very low it is irrelevant how you connect the cells as the difference in current flowing trough the 4 parallel cells will be super small.

Also the way you have them connected is by far not the worse and current will be shared fairly equally.

The reason you do not see cell balancing most of the time is because the charge rate is likely very low. Well below the standard of around 0.2C that people use.

Those cells are also very low internal resistance compared to typical cells people use like the EVE cells and that also helps on top of the low charge discharge rate.

 

Also a typical 280Ah 300Ah EVE cell has a single terminal using an M6 bolt while the equivalent for your battery is 3xM10 terminals as the batteries are just designed for different type of applications.

Of course you paid a premium for your cells.

[Marinepower]

Thanks Dacian for your review of the cell wiring.  Yes, my C-rate is fairly low.  I charge at around 0.15C and discharge only intermittently up to 0.2C.  (Victron 1200VA inverter).

The cells are not that premium.  The are made by a company called "Frey Energy". and often sold in the NA market under the brand "Fortune".  Will Prouse (DIY Solar) did a review on the Fortune cells a while back.  The big terminals, plastic covers and metal frames make them very convenient for mobile applications.

Here is a pic of the installed unit (located under a small nav-seat).

MP

[Dacian Todea (electrodacus)]

By premium battery I was referring at the cost per unit of stored energy not the quality of the cells. And cells are designed for higher discharge rate than EVE cells.

So yes there will be less cell imbalance with this cells even at higher charge discharge rate but you pay a premium for that.

3 of this cells will likely cost more than a single 280 to 300Ah EVE cell.

[Marinepower]

Thanks Dacian for the clarification of your observation.

For a small system like mine (400ah, 12v) cost of the ancillaries (fuses, relays, wire, lugs, bus bars) and just the amount of labor time spent begins to make the cost of the cell less relevant .  

For eg. I spent a lot of time and money getting my old EVE 280 cells compressed and fixed to the boat.  

These cells were about $60 USD per 100ah.  So about 2.5x the price of an EVE 280.  Likely could be had cheaper now....

[Dave McCampbell]

Modern EVE Grade A 280 ahr matched cells from any one of 3 Docan Power's warehouses run about $55 US now.  Less than half what they were about 4 years ago.  Our new cells just arrived here in Malaysia in about 10 days from China very well packed.  

[Marinepower]

Wow, that is cheap.  I bough EVE 280ah  4 years ago for 125USD each.

[Dave McCampbell]

We paid about $150 each from a different company 4 years ago.  These are the latest technology and with matched internal resistances and capacities.  If you are in the US shipping from their US warehouse is a bargain.  Hard to pass up.