• Have you tried out dark mode?! Scroll to the bottom of any page to find a sun or moon icon to turn dark mode on or off!

diy solar

diy solar

Calculation of parallel string battery currents

Anyone else have thoughts on my post. Appreciate the excellent info here.
I haven't seen Solarod posting since maybe March,
I assume your 12 100Ah packs are in two racks of 6 each? and if you have the racks, these will have bus bars for connection of each pack.
From his earlier comments, there will be no 'ideal' way to connect your main 4/0 feeders, close best option may be between 1/2 on pos and between 5/6 on neg (or vise versa) but this would be better with a four pack set up. There will be some imbalance with six batteries.
I would just try it and then monitor the packs for a while, after some cycling up and down the SOC range. See how it does.
I suppose if it doesn't work well, you can change your 12 batteries to 3 sets of 4.
In my own set up, where I have the ackward use of 2 100Ah packs plus 3 280 Ah packs, connecting the two 100's up as a pair (as if a 200 Ah pack) has worked out pretty well, and now that the bus bars have only four connections, I was able to shift the main pos and main neg to the 1/3 2/3 locations Solarod noted for a four battery set up on bus bars and this has been far better than where I started.
 
I've done some calculations on a 5 battery parallel string and here are my results.

I'm using a value of 1 milliohm for the battery IR, and .1 milliohm for the link resistances which I think is more representative of what one gets with busbars rather than cable links. The load current is 100 amps. I'll reference this image:

batt11-png.91674


The red and black load connections are the standard diagonal connection, and that connection gives me the following theoretical battery currents in amps:

Example 26
23.5
18.2
16.6
18.2
23.5

Some have suggested moving the load connection toward the middle of the string. If the black load connection is moved to the point on the busbar adjacent to the negative terminal of battery 2, and the red load connection is moved to the point on the busbar adjacent to positive terminal of battery 4, we get these theoretical currents in amps:

Example 27
18.3
21.8
19.8
21.8
18.3

This is a substantial improvement. In some of the earlier hookups it was possible to get a perfect balance by connecting the load to the right place on the busbars between the last and next to last batteries in the string. Of course, I have tried to do that for this 5 battery string by varying the connection point. It isn't possible to get perfect balance, but it is possible to get improvement. I had to consider how to measure the improvement, and what seems reasonable is to get the least variation in the battery currents. I had my calculations include the standard deviation of the battery currents, and tried all the connections between the end batteries and the next-to-the-end batteries in 5% increments. For this 5 battery string, the connection which is 66% of the way from the last battery to the next battery inward gives the minimum variance in the battery currents. This would be a connection like the yellow/blue connection shown in the image above, at the 66% point. Here is the result of the 66% connection theoretical battery currents in amps:

Example 28
20.0
20.6
18.7
20.6
20.0

I haven't given up on the possibility of a perfect balance for this string, but no joy so far. :(
I guess this as good as we can get on the SOK racks.

Will try this at some point and see how it goes.

I can tell you your calculations are correct just based upon what I see with battery draw on the other different scenarios listed.
 
Last edited:
I haven't seen Solarod posting since maybe March,
I assume your 12 100Ah packs are in two racks of 6 each? and if you have the racks, these will have bus bars for connection of each pack.
From his earlier comments, there will be no 'ideal' way to connect your main 4/0 feeders, close best option may be between 1/2 on pos and between 5/6 on neg (or vise versa) but this would be better with a four pack set up. There will be some imbalance with six batteries.
I would just try it and then monitor the packs for a while, after some cycling up and down the SOC range. See how it does.
I suppose if it doesn't work well, you can change your 12 batteries to 3 sets of 4.
In my own set up, where I have the ackward use of 2 100Ah packs plus 3 280 Ah packs, connecting the two 100's up as a pair (as if a 200 Ah pack) has worked out pretty well, and now that the bus bars have only four connections, I was able to shift the main pos and main neg to the 1/3 2/3 locations Solarod noted for a four battery set up on bus bars and this has been far better than where I started.
Thanks for replying, yes two racks of 6. Looking back at post 57 he does cover banks of 6 and I plan to do it that way by drilling my own holes in the busbars for precise balancing. Would be interesting to see @Will Prowse do a geeky video on this subject. Tricky part may be controlling variables for a good test.
 
Last edited:
Hi,

I have this setup connected as such:
Four Battery Packs
I use 50mm² wires to interconnect them as shown.
Two Deye Hybrid Inverters
Deye1 has ~10% more Solar Panels than Deye2.
I use 35mm² to connect Inverters to Battery
Inverter Cables are 2meters, except for Deye1's Negative which is ~2.5m (Deye1 is ~0.5m farther away)
1724450198758.png

From my observation:
During Charging (Day), Battery3 and Battery2 receives more current.
During Discharging (Night), Battery0 and Battery1 delivers more current.
There is little current flow in the stripe wires between Battery2 and Battery1.

Should I connect the inverters exactly on the same posts?
Need help.

Thanks!
 
 
Hi,

I have this setup connected as such:
Four Battery Packs
I use 50mm² wires to interconnect them as shown.
Two Deye Hybrid Inverters
Deye1 has ~10% more Solar Panels than Deye2.
I use 35mm² to connect Inverters to Battery
Inverter Cables are 2meters, except for Deye1's Negative which is ~2.5m (Deye1 is ~0.5m farther away)
View attachment 238729
If you follow the link @Danke posted you will see the discussion retired EE 'Solarod' explained not only how to balance multiple batteries better, he also explained why. In PM he assisted me with balancing 6 packs - two 304Ah, two 280Ah and two 100Ah and I am eternally grateful.
From the thread linked you will see he recommends connecting between your 0 and 1 on neg and between 2 and 3 on pos for the best overall balance of four batteries. I found clamping a ring terminal to the bus bar as an easy way of playing with small changes of location until things worked the way I wanted, then drilling the bus bar for a bolt.
 
It has been done, but I have no idea how to locate the particular thread where that analysis was carried out.

I ran into this exact same problem recently, when trying to top balance thirty lithium cells. With some very long aluminium busbars and diagonal connections, the voltage drop along the busbars still created differences in cell voltages along the string of cells. Only a millivolt or two, but the difference was there, and continued until the current became effectively zero.

But the short answer is, if the parallel busbar part of the circuit has a low enough impedance with respect to the relative current, the diagonal connection method does work pretty well in practice.
Another solution is to have equal length conductors to each individual cell on both the positive and negative sides.
The equal length wires kind of fan out from each cell to a single large connection point on each side.
That may arguably be a preferred method where the voltages are low and the current very high.
 
Last edited:
@ragnakore, post #4 of this thread covers the exact situation you're trying to fix. It's probably not the way the inverters are connected on different posts. It's how they're connected to the outer batteries.

While you're reworking the connections, make sure that ALL of the connections are clean and tight. During charge or discharge, check the voltage of each battery at it's terminals. Don't check on the lugs, it has to be on the terminals. Checking on the lug could mask a bad connection.
 
As I said, it's probably not the inverters. Focus on the connection from the system to the battery.

If you really think it's the inverters then put in a set of common busbars. Run the battery cables to the common busbars, then cables from the common busbars to the inverters.
 
It takes common busbars and specific placement of the uplink cables to get near perfect balance. Also a YR1035 meter to check the impedance of batteries, cables, connections, etc.

If you haven't actually read this whole thread, you need to.


And shouldn't this be a new thread verse adding on to the very useful one about parallel batteries?
 
Last edited:
Hi,after reading the thread i still have some mouses in my head and that’s..
I am planning to have 3 banks from 2p16s lifepo4 304Ah(608Ah) connected to a tinned copper busbar and from there going to a t-class or HRC fuse with 2 95 mm(2/0 AWG) copper cables but now i need to find out where i can put my cables on to the busbar going to fuse,if i see this thread it’s not that easy…or do i overlook something
All battery’s have same manufacturer,age and gonna be balanced before connecting them
Thanks for help advice and recommendations
 
Hi,after reading the thread i still have some mouses in my head and that’s..
I am planning to have 3 banks from 2p16s lifepo4 304Ah(608Ah) connected to a tinned copper busbar and from there going to a t-class or HRC fuse with 2 95 mm(2/0 AWG) copper cables but now i need to find out where i can put my cables on to the busbar going to fuse,if i see this thread it’s not that easy…or do i overlook something
All battery’s have same manufacturer,age and gonna be balanced before connecting them
Thanks for help advice and recommendations

Post #14 addresses three batteries in a bank. 1/3rd of the way towards the middle battery.
 
Thanks for replying, yes two racks of 6. Looking back at post 57 he does cover banks of 6 and I plan to do it that way by drilling my own holes in the busbars for precise balancing. Would be interesting to see @Will Prowse do a geeky video on this subject. Tricky part may be controlling variables for a good test.

@offgridfarmgineer What did you end up doing? I'm in the process of integrating a second rack of batteries, and I'm curious what your experience was.
 
Does anyone want to take a guess at the best configuration for 10 batteries with bus bars?

I ran for a month or so with both connected to one end. Once I was able to inspect the cell dayta I did see imbalance, and have now reconfigured to "diagonal", where I am seeing much closer balances.

However I am still seeing more current throughout 1&2 and less through 9&10, which I think is indicative of some wear from when they were initially misconfigured.

Initially: red: 1 black: 1
Now: red: 10 black : 1

I am wondering if I would get better performance from: red 7 black :3.
 
Does anyone want to take a guess at the best configuration for 10 batteries with bus bars?

I ran for a month or so with both connected to one end. Once I was able to inspect the cell dayta I did see imbalance, and have now reconfigured to "diagonal", where I am seeing much closer balances.

However I am still seeing more current throughout 1&2 and less through 9&10, which I think is indicative of some wear from when they were initially misconfigured.

Initially: red: 1 black: 1
Now: red: 10 black : 1

I am wondering if I would get better performance from: red 7 black :3.
Why not just try it out and see?
I did some adjustments to where a ring terminal connected to a main copper bus bar - I used a visegrip locking pliers to hold the ring terminal in test locations and let it run up and down the SOC for a few days - as a way to test a new location without actually drilling any new holes until I was happy with the result.

As to ten battery packs or twenty in balance... Here is my attempt...
 
Does anyone want to take a guess at the best configuration for 10 batteries with bus bars?

I ran for a month or so with both connected to one end. Once I was able to inspect the cell dayta I did see imbalance, and have now reconfigured to "diagonal", where I am seeing much closer balances.

However I am still seeing more current throughout 1&2 and less through 9&10, which I think is indicative of some wear from when they were initially misconfigured.

Initially: red: 1 black: 1
Now: red: 10 black : 1

I am wondering if I would get better performance from: red 7 black :3.
I had a similar nightmare trying to initially balance thirty lithium cells with very long and very crappy improvised busbars.
Diagonal connection should have worked in theory, but once everything becomes seriously unbalanced, its difficult to get it to recover quickly.
It probably will recover and settle down, but it might take a great many charge and discharge cycles to slowly get back.
I don't think you can really force it, just be very patient and keep an eye on progress.
 
Does anyone want to take a guess at the best configuration for 10 batteries with bus bars?

I ran for a month or so with both connected to one end. Once I was able to inspect the cell dayta I did see imbalance, and have now reconfigured to "diagonal", where I am seeing much closer balances.

However I am still seeing more current throughout 1&2 and less through 9&10, which I think is indicative of some wear from when they were initially misconfigured.

Initially: red: 1 black: 1
Now: red: 10 black : 1

I am wondering if I would get better performance from: red 7 black :3.
This should help: https://diysolarforum.com/threads/calculation-of-parallel-string-currents-addendum.94178/
 
I had a similar nightmare trying to initially balance thirty lithium cells with very long and very crappy improvised busbars.
Diagonal connection should have worked in theory, but once everything becomes seriously unbalanced, its difficult to get it to recover quickly.
It probably will recover and settle down, but it might take a great many charge and discharge cycles to slowly get back.
I don't think you can really force it, just be very patient and keep an eye on progress.
Break up the string of thirty into 5 strings of 6. Use the optimum connection for a string of 6
shown here: https://diysolarforum.com/threads/calculation-of-parallel-string-currents-addendum.94178/

Just clamp the output cables temporarily to the busbars at the optimum locations rather than drilling holes.

The strings of 6 should individually charge their cells to 3.65 volts per cell fairly quickly.
 
Last edited:
Wow. I read this thread early on but did not realize how long it got.
It is amazing how very small differences in resistance can make a huge difference in the current.

A few observations:
* The analysis is phenomenal! It puts some numbers behind what many of us have been saying for a while. I have done much cruder analysis to convince myself on a lot of this, but nothing near as detailed as this. THANK YOU.
* The minor resistance differences pointed out may be beyond our ability to control consistently. Very slight differences in crimps studs etc. will throw off the balance. Can we reasonably expect to have milliohm consistency between multiple crimps and connections? Can we even expect that level of consistency between one BMS and the next (even if they are both the same make & model)?
* The analysis is a static analysis and does not take into account the dynamic nature of LiFePo4 cells. I suspect that the cell behavior in the flat part of the charge/discharge curves are different than in the low and high end of the curves. This would be very difficult to analyze. It would need to be simulated, but that would require an accurate model of a LiFePO4 Cell...and that could be different between cell types, size and manufacturer.
* As shown in all of the examples the imbalance gets worse as the current goes up. It may look really bad at 400A, but for most home systems it is going to be a *lot* lower than that 99.9% of the time.
* Since we are talking about balancing the wear on the batteries, perhaps the best test is to run them under your typical load for a month and compare the total KWH of charge or discharge through each battery. This will take into account *your* typical use patterns and the dynamic nature of the cells.
* Something that @Will Prowse points out probably applies here: The calendar life aging of the cells may get you before anything else.

My bottom line: the analysis shows that we must pay attention to balancing the batteries, but I fully agree with some of the posts that advised not to get carried away with things. Do whatever is practical and call it good.
 
The cabling resistance matching has been overly beaten to death.

It has an effect but is by no means the last word on current sharing variance between parallel batteries.

Other big effects.

-Terminal lug clamping connection resistance. Oxide coated aluminum battery terminals. Poor torquing. Terminal surface to surface compression force distribution.

- Crimped cable lugs. If you crimp your own cables you better have a milli-ohm meter like YR1035+ to check them and have computed what the cables resistance should be based on wire gauge-length copper resistance. Very easy to have 2x-4x of resistance for shorter cables due to lug crimp contact resistance. Shorter cables are what you will have if parallelling batteries together near batteries. If you buy pre-made crimped cables it is a good idea to check them with milli-ohm meter against their gauge-length resistance expectation. YR1035+ is well spent $35 for test meter.

-BMS MOSFET Rds_On resistance.
This factor is somewhat self-compensating between parallel batteries, each with their own BMS. As BMS MOSFET's have greater current flow, the MOSFET's get hotter, which causes their Rds_On resistance to increase, which adds resistance to that battery path, which lowers its parallel current contribution, which reduces the variance in current sharing between parallel batteries.

-Breakers. Breakers with short circuit electro-magnet solenoid quick trip generally have greater series resistance due to series wire coil for the electro-magnetic trip solenoid.

-Overpotential voltage variance in battery. Manufacturing tolerances cause overpotential voltage variance between cell batches.

-Local temp difference between batteries. Overpotential is very dependent on cell temperature. Battery mounted on outside wall versus battery on second row, inside heated room. Rack mounted batteries, top of rack is often warmer than bottom of rack.

Example below is two 12v LFP self-contained batteries with a minor amount (51mv vs 56mv) of overpotential voltage slump at about 50A. All cell ohmic resistance, as you would measure with a 1 kHz battery impedance meter, BMS and cabling resistance are identical. Only the overpotential is minorly different between batteries.,
Batteries in parallel matching.png

Little trick.

You will get better parallel battery current sharing if you parallel connect batteries near inverter with separate smaller gauge cabling going to each battery versus parallel connecting near batteries with one larger gauge cable going to inverter.

By having multiple cable runs of smaller gauge cables on each separate parallel battery, you will have better current sharing balance due to the ballast resistance of the separate cabling on each battery and still have the same net cable losses of one large gauge cable with batteries parallel connection close to batteries.
Battery parallel at inverter.png
 
Last edited:

diy solar

diy solar
Back
Top