Current Sharing with Large LiFePO4 Packs

[Shale MacGregor]

In response to the main purpose of this post, I do use multiple lifepo4 in parallel with varying capacities and have had fun tracking and logging data on it. I should consolidate at some point and share.

In response to the fusing discussion, I wonder what many people think the purpose of the fuse for their battery packs is. Are you doing it as a failsafe over your DC breaker (if you use one), are you doing it to prevent damage to the wiring? the batteries? the inverter? Would you be better off spending more on a better breaker, or a series of breakers? What part does speed play in your decision based on what you are trying to protect? We know wiring is rated at continuous but can handle higher spikes for shorter times.

I don't personally have answers to these questions yet, so I think the safest thing is to over-protect. This can be more expensive but I always load in more safety when I am not 100% confident why I shouldnt.

 

[Will Prowse]

In response to the main purpose of this post, I do use multiple lifepo4 in parallel with varying capacities and have had fun tracking and logging data on it. I should consolidate at some point and share.

In response to the fusing discussion, I wonder what many people think the purpose of the fuse for their battery packs is. Are you doing it as a failsafe over your DC breaker (if you use one), are you doing it to prevent damage to the wiring? the batteries? the inverter? Would you be better off spending more on a better breaker, or a series of breakers? What part does speed play in your decision based on what you are trying to protect? We know wiring is rated at continuous but can handle higher spikes for shorter times.

I don't personally have answers to these questions yet, so I think the safest thing is to over-protect. This can be more expensive but I always load in more safety when I am not 100% confident why I shouldnt.

If a pack has a shorted cell AND a shorted FET, and is unable to disconnect the pack (very rare), the other packs can feed into the shorted cell. Anytime you have parallel strings, you need to think about the whole circuit. For example, if you have parallel solar panels, you need to match the voltage and provide OCPD for each string. Solar panel cells are large diodes, but there is a limit. If you have a failed solar cell, and multiple strings in parallel, they can feed into the shorted cell. Same issue with batteries.

This is why in large installation packs, each parallel string has over current protection.

If you have two strings of solar panels, you may get by without it. But it really depends. With batteries, each string needs OCPD.

 

[Hedges]

If a pack has a shorted cell AND a shorted FET, and is unable to disconnect the pack (very rare), the other packs can feed into the shorted cell. ...

If you have two strings of solar panels, you may get by without it. But it really depends. With batteries, each string needs OCPD.

In this case, it is excessive reverse current you want to stop.
But I suspect any fuse or breaker capable of carrying rated current of the battery will not open, won't even come near its rating, if a cell fails.

Maybe a circuit breaker with remote trip could be arranged.
That's what I'm thinking could be done for a somewhat similar situation of failed SCC, where PV Voc is available to overcharge battery. (I've got an SCC with relay output to indicate over-voltage, apparently planning for failed FETs.)

 

[Shale MacGregor]

If a pack has a shorted cell AND a shorted FET, and is unable to disconnect the pack (very rare), the other packs can feed into the shorted cell. Anytime you have parallel strings, you need to think about the whole circuit. For example, if you have parallel solar panels, you need to match the voltage and provide OCPD for each string. Solar panel cells are large diodes, but there is a limit. If you have a failed solar cell, and multiple strings in parallel, they can feed into the shorted cell. Same issue with batteries.

This is why in large installation packs, each parallel string has over current protection.

If you have two strings of solar panels, you may get by without it. But it really depends. With batteries, each string needs OCPD.

I will be honest, I hadn't considered that... the more packs you have the more power could be fed through the shorted cell without tripping the other overprotection devices.

Hrm.

 

[mrzed001]

In this case, it is excessive reverse current you want to stop.
But I suspect any fuse or breaker capable of carrying rated current of the battery will not open, won't even come near its rating, if a cell fails.

Maybe a circuit breaker with remote trip could be arranged.
That's what I'm thinking could be done for a somewhat similar situation of failed SCC, where PV Voc is available to overcharge battery. (I've got an SCC with relay output to indicate over-voltage, apparently planning for failed FETs.)

That shorted MPPT and blown battery got us all.
Still can not find a cheap and easy solution for that.
But just posted a possible "theoretical" solution to prevent it.

Why not to use Daly BMS with MPPT controllers

One concern I have with the above "Fail Safe" circuit is that once it energizes the relays, it puts a brief load on the battery which brings the battery voltage back in line, and the relay reconnects the PV panel. It will continue to cycle like this, keeping the battery from overcharging, but I...

diysolarforum.com

In parallel packs if BMS is not disconnecting the shorted cell's pack ... and the short is lower Amp than the fuse ... are we screwed?

 

[Will Prowse]

That shorted MPPT and blown battery got us all.
Still can not find a cheap and easy solution for that.
But just posted a possible "theoretical" solution to prevent it.

Why not to use Daly BMS with MPPT controllers

One concern I have with the above "Fail Safe" circuit is that once it energizes the relays, it puts a brief load on the battery which brings the battery voltage back in line, and the relay reconnects the PV panel. It will continue to cycle like this, keeping the battery from overcharging, but I...

diysolarforum.com

In parallel packs if BMS is not disconnecting the shorted cell's pack ... and the short is lower Amp than the fuse ... are we screwed?

Just for that pack, yes. But it won't hurt the other packs because the other packs will disconnect.

The big danger is using an inappropriate circuit breaker that cannot handle the disconnect current of the entire pack. I isolate groups of batteries with a medium sized circuit breaker. That should help prevent issues.

 

[Hedges]

That shorted MPPT and blown battery got us all.
Still can not find a cheap and easy solution for that.

I think I found a practical solution: circuit breaker suitable to interrupt PV, with remote trip. Midnight Solar has some remote trip PV breakers. Seems to be something like a solenoid ganged to the breaker. If not them, others. If nothing else, gang a 1A breaker and cause it to trip.

And here's a battery sized breaker, with remote trip. One per lithium cell string?
What you need is a suitable signal to trip it. Basically a supervisory/backup BMS. If main BMS fries, a second BMS that detects individual cell voltages severely imbalanced could trip the breaker.

Product - MidNite Solar

MidNite Solar Product Page.

www.midnitesolar.com

 

[mrzed001]

Just for that pack, yes. But it won't hurt the other packs because the other packs will disconnect.

The big danger is using an inappropriate circuit breaker that cannot handle the disconnect current of the entire pack. I isolate groups of batteries with a medium sized circuit breaker. That should help prevent issues.

I am not sure
If one cell goes to SC then a 16s pack becomes a 15s pack
Pack V goes down 3,3V. That is 0,22V/cell.
Andy did test what happens is you put different V cells parallel.

Only when full and empty cells paralleled got he 200+A
So it can happen that less Amp goes into the wrong pack ... discharging all other packs.
(of course cause of SC can melt away)

 

[mrzed001]

I think I found a practical solution: circuit breaker suitable to interrupt PV, with remote trip. Midnight Solar has some remote trip PV breakers. Seems to be something like a solenoid ganged to the breaker. If not them, others. If nothing else, gang a 1A breaker and cause it to trip.

And here's a battery sized breaker, with remote trip. One per lithium cell string?
What you need is a suitable signal to trip it. Basically a supervisory/backup BMS. If main BMS fries, a second BMS that detects individual cell voltages severely imbalanced could trip the breaker.

Product - MidNite Solar

MidNite Solar Product Page.

www.midnitesolar.com

I think a BMS will not detect an MPPT short.
If you direct connect the 400V PV string to the 48V battery I think the battery will pull down the PV voltage to battery level.
Of course PV max Amp will flow into battery but the battery V will only rise with higher SoC.
So when BMS can see the problem that is already too late for the battery.
Or the BMS is already dead. Because when MPPT failed the whole 400V is present on BMS FET's (only after that FET's go into short, connects battery to PV directly, and brings down its volt).
Theoretically


My concept can work with breakers too
Since it has a Volt meter on both side of the PV realy/breaker ... so V1 and V2 nearly identical means breaker is on.

Really easy with a RasPi.
- 2 Volt meter (RasPi can do it, or external V meter)
- 1 out pin to signal to the breakers
- and a little python script that reads the V values, compares, signals.

 

[fopoku2k2]

I've been running two 16s batteries in parallel for almost two years not. Not an issue - contrary to what some commentators on various YT videos say (and that I'm really tired arguing with). Maybe this video will do it

Are they of the same Ah? I also have two 16 batteries that I'm thinking of connecting in parallel but one is 280Ah and the other is 160Ah. Do you think I can hook them up in parallel?

 

[fopoku2k2]

Just checked them out. Looks like they are AC

Gosh those are cheap! I have never used those before. Had no clue they were DC rated. Will need to check those out.

 

[Hedges]

Are they of the same Ah? I also have two 16 batteries that I'm thinking of connecting in parallel but one is 280Ah and the other is 160Ah. Do you think I can hook them up in parallel?

You ought to be able to get away with that, if only looking for more capacity, not higher current. So long as your max current draw never exceeds what lesser of the two packs can provide, shouldn't be a problem.
One might provide more of the current during shallow cycling. If you draw down so far that one would have reached lower knee, the other would definitely be providing current by then. You might eventually observe one reaching end of cycle life before the other. But that should be after thousands of cycles.

If you want higher current draw, may or may not split current optimally. Internal resistance, State of Charge, and cable resistance will make the decision for you. If current exceeds what one BMS approves of it would disconnect.

 

[JJJJ]

Just saw the new video on the server pack battery provided by Will. Very well built and was very impressed by the feature set.

Admittedly they would take a bit of effort to get into place. However, once installed would be OK since they wouldn't require much tinkering.

As I look at server battery designs I still reflect back on initial computer designs. Vendors initially tried to fit everything into one box.

Then over time, networking took place and it became possible to expand capability without packing all the capacity into one box.

Perhaps a vendor will see this and consider a more modular approach. A multi box system with a switchable BMS to handle different capacities might be a consideration. There would be the basic box with the BMS, etc. One could then plug in additional modules to expand capability. Granted DC current has stricter requirements than CAT cable, which restricts distances, but in the future this might be possible. I know the unit sold by Signature Solar tries to do this with their server cases but it still has everything in a single box stacked together.

Again, perhaps the cost of a BMS, and other control hardware is not significant enough to consider this approach. Just thinking outside the box and considering how other technologies have developed over time.

 

[Shale MacGregor]

You ought to be able to get away with that, if only looking for more capacity, not higher current. So long as your max current draw never exceeds what lesser of the two packs can provide, shouldn't be a problem.
One might provide more of the current during shallow cycling. If you draw down so far that one would have reached lower knee, the other would definitely be providing current by then. You might eventually observe one reaching end of cycle life before the other. But that should be after thousands of cycles.

If you want higher current draw, may or may not split current optimally. Internal resistance, State of Charge, and cable resistance will make the decision for you. If current exceeds what one BMS approves of it would disconnect.

Isn't some of that discharge capacity of the other pack then being utilized by the parallel pack that is attempting to equalize? I don't have the math for how that would all be determined.

 

[Hedges]

Isn't some of that discharge capacity of the other pack then being utilized by the parallel pack that is attempting to equalize? I don't have the math for how that would all be determined.

The battery packs have no control or regulation over current, only ability to disconnect. Battery SoC (voltage), internal resistance, and wire resistance are all there is to vary current.
The closer you come to drawing sum of maximum current from all parallel packs, the more likely one detects too much and disconnects.
I think you can get away with it to some extent, but no guarantees.

The way to make it work reliably would be to have CV/CC operation, like bench supplies which can be paralleled. This gets done with the "400V" batteries used by some inverters with multiple battery inputs, such as Sunny Boy Storage. The batteries are actually 48V (or multiples of that?) with a boost converter, i.e. a switching power supply that uses PWM and an inductor. Each an deliver its desired current, and the inverter uses its own switching power supply to draw from each (just like multiple MPPT). This could be done with low voltage 48V batteries but I haven't seen it (except as inverters that can be paralleled, and those I've read about expect a single battery feeding all.)

If a battery can handle 1C discharge, about 300A at 50V, 15kW continuous, should be OK up to there. In the case of 160Ah battery, half that much power. They may (or may not) cycle different amounts and reach 80% capacity end of life at different times. If cycling is using more than one battery's capacity, I think the second battery will begin to handle more of the cycling over time, helping to level the wear.

 

[philatio]

Seems This thread went off on a tangent from the videos ie: diagonal battery connection vs non diagonal battery connection.

I flipped one of the bus bars on my signature solar rack and see no charging/discharging difference. Admittadely I only did 2 charge/discharge cycles. The system I am putting together is not completely installed yet. The voltages stayed equal between all 3 batteries, even though the SOC varies by 1% to 2% between the batteries.

The one flipped bus bar is also creating issues with with cable mounting.

If I used thicker wires, can I keep the bus bar orientation with the connection to the inverter on the same end?

I have 1/0 from batteries to bus bars and 4/0 from bus bars to inverter.

Thanks in advance.

 

[Hedges]

interesting choice of user name

Thicker is better, and I guess it could be OK just on the end.
All depend on how ideal you want performance to be, but I think they'll be happy enough.

 

[philatio]

Prize goes to @Hedges.

Been using this for 7 years and you're the first to catch it.

The old Mikehunt moniker was caught in less than 24hrs

 

[curiouscarbon]

Wiring Unlimited

Wiring Unlimited is a great resource that covers many many common topics and questions relevant to DIYers. It is a free E-book from Victron Written by Margreet Leeftink. Check it out! Download

diysolarforum.com

 

[philatio]

I have read read parts if the wiring unlimited which is why I flipped one of the server rack bus bars.

I'm considering changing it back because of bolt spacing for the positive and negative battery cables to the bus bar dont work very well with the bus bar flipped.