An idiot's guide to building a big lithium bank

[AndreHuizing]

Thanks for your extensive reply!Apart from the diving equipment our usage seems quite similar. I’d like to draw approx a max of 8kW AC (220v in our case).
When not using the same voltage for storage as the normal dc circuit (24v) I wonder why 48v. E.g. why not 72v? Challenges of up and down stepping seem the same?
Also wondering about charging this with present alternators (150A mastervolts 24v)?
Seems like I have to “sandwich” the bank in stepping up and stepping down dc/dc convertors?
Btw: I came up with the 64 cell setup as that is about the max that fits in our present (AGM) house battery storage compartment.

 

[twistedtree]

Eg. Balancing full cells when the battery pack is still low and being charged at 200+ amps

Cheers

I'm not following your statement above. Can you say more about it? Is this a way to maintain balance among the cells?



The longer term issue that requires some sort of intervention, either manual or automatic, is re-balancing series cells (or groups of parallel cells). I suspect you know about this, but I'll describe it more for the benefit of others reading the thread. Over time the charge level of the cells drift apart with some at slightly higher charge levels, and others at slightly lower charge levels. The total bank voltage can be fine, but you end up with some cells low and some cells high. If they get too far apart, then when charging you risk over charging the high cells because chargers are only seeing the bank voltage and don't know that one cell is running hot. And when drawing the bank down low, you similarly risk over discharging the low cells. Balancing adjusts the cell's charge levels back in sync with each other.



Drifting out of balance doesn't happen quickly, but it does happen, and eventually you need to do something about it. Charge levels can be adjust manually, automatically. Or I supposed you could ignore it and just wait for cells to get damaged and replace them, or replace the whole pack.

 

[Tator]

follow.

Tator

 

[GoneDiving]

Thanks for your extensive reply!Apart from the diving equipment our usage seems quite similar. I’d like to draw approx a max of 8kW AC (220v in our case).
When not using the same voltage for storage as the normal dc circuit (24v) I wonder why 48v. E.g. why not 72v? Challenges of up and down stepping seem the same?
Also wondering about charging this with present alternators (150A mastervolts 24v)?
Seems like I have to “sandwich” the bank in stepping up and stepping down dc/dc convertors?
Btw: I came up with the 64 cell setup as that is about the max that fits in our present (AGM) house battery storage compartment.

When you say 8kw@220v is this single phase or US style split phase? I have no experience with split phase but 48v 8kw single phase inverters are readily available. I'd be hesitant to run them at 100% load continuously and would consider running 2 in parallel (buy models will parallel kits).

48v is the practical limit of readily available downstream components. Economies of scale give good availability and pricing. 96v is available but it's really only used by industries that order special builds and have deep pockets: eg telcos and military.

Google series parallel relays for info on different house and start voltages. Older trucks commonly had 12v accessories and 24v starters. Two batteries normally ran in parallel but the relay changes this to series to start. It's an option but seems like another point of failure to me but you may be interested. This could facilitate using a 24v alternator. Alternatively, 48v altenators are available up to 300a, both name brands and mass produced Chinese knock offs.

While you do need to have a charging strategy, keep in mind that moving from Lead to Lithium dramatically improves the round trip efficiency: charge, store, discharge all now have less losses. Hence, this reduces the burden. The extra storage capacity also means that recharging can be done when available (solar) or convenient (shore power).

 

[GoneDiving]

I'm not following your statement above. Can you say more about it? Is this a way to maintain balance among the cells?

I was suggesting the opposite: that balancers would struggle to balance a pack where one cell is full at 3.65v but the pack voltage for the whole battery is not high enough to shut down charging eg 3.65, 3.40, 3.40, 3.40 = 13.85v being charged at 100a. Most resort to stopping charge current based on single cell over voltage because they don't have the capacity to divert 10s or 100s of amps between cells. Resistive balancers are commonly rated in milliamps and active balancers 5 or 10 amps only.

A passive approach would be to stop charging at an average 3.40v/cell, realising that this gives a lot of headroom for misbalanced cells and that there is very, very little capacity gained between 3.40 and 3.65v.

You correctly point out that one of the challenges and issues with lithium is that lead self balances in series strings, lithium doesn't.

Cheers.

 

[GoneDiving]

Drifting out of balance doesn't happen quickly, but it does happen, and eventually you need to do something about it. Charge levels can be adjust manually, automatically. Or I supposed you could ignore it and just wait for cells to get damaged and replace them, or replace the whole pack.

Again, you are correct. I don't want to give the impression that any protection device or methodology is set and forget. All need to be monitored and tested occasionally. Perhaps more so for passive systems.

 

[GoneDiving]

For those seeking more general info on Lithium and component testing, I found this a good site:

https://youtube.com/c/OffGridGarageAustralia

A warning that the guy does go a little 'mad scientist' every now and then but he accepts none of the normal conventions and tests everything, sometimes with surprising results.

 

[twistedtree]

I was suggesting the opposite: that balancers would struggle to balance a pack where one cell is full at 3.65v but the pack voltage for the whole battery is not high enough to shut down charging eg 3.65, 3.40, 3.40, 3.40 = 13.85v being charged at 100a. Most resort to stopping charge current based on single cell over voltage because they don't have the capacity to divert 10s or 100s of amps between cells. Resistive balancers are commonly rated in milliamps and active balancers 5 or 10 amps only.



A passive approach would be to stop charging at an average 3.40v/cell, realising that this gives a lot of headroom for misbalanced cells and that there is very, very little capacity gained between 3.40 and 3.65v.



You correctly point out that one of the challenges and issues with lithium is that lead self balances in series strings, lithium doesn't.



Cheers.

Thanks. I see what you are saying.

Your example I think is exactly the sort of average cell voltage approach that you are suggesting, and illustrates the problem therein. The average can be fine, but one or more cells can be running out of spec.

Re balancing, the idea is to keep them from getting as far out of balance as in your example of one at 3.65v and others at 3.40v. If you catch it earlier, only small corrections are required.

For daily charging I think setting the target charger voltages based on an average cell voltage is a good approach, and practically the only way with most chargers. It works great as long as the cells are actually near the average. But eventually one or more will drift, and some action is eventually required. I guess that’s all I’m really saying. Since you are monitoring cell voltages only, and doing so manually, any indicated balancing will have to be undertaken manually as well. It’s no big deal, provided you know to do it, and have a method to carry it out.

 

[GoneDiving]

For daily charging I think setting the target charger voltages based on an average cell voltage is a good approach, and practically the only way with most chargers. It works great as long as the cells are actually near the average. But eventually one or more will drift, and some action is eventually required. I guess that’s all I’m really saying. Since you are monitoring cell voltages only, and doing so manually, any indicated balancing will have to be undertaken manually as well. It’s no big deal, provided you know to do it, and have a method to carry it out.

Yes, I have assumed that the cells purchased are reasonably matched. My example was extreme and I'd hope that the cells' imbalance would be identified before it became that bad. Obviously, if someone is buying the cheapest repacked used cells on AliExpress then any monitoring would have to be very thorough. Especially in the longer run.

We seem to be in furious agreement. ��

 

[Jeff F]

If the cells were well balanced to begin with then the problem is likely a dead cell. On an existing pack a purist would replace all cells. But if it's a new pack then one could just replace the faulty one under warranty.

I think what twistedtree and I have been talking about is managing individual cells on an ongoing basis. You can start with a perfect battery, but it won't stay balanced indefinitely.

I recently threw away an old Prius battery control unit that had 14 voltage taps and provided module voltages, highest/lowest cell, etc to the CAN bus. From there I graphed it on my phone. Maybe I should have kept it

 

[twistedtree]

I think what twistedtree and I have been talking about is managing individual cells on an ongoing basis. You can start with a perfect battery, but it won't stay balanced indefinitely.

Correct. It's entirely possible that we are indeed in furious agreement. I just still don't understand the plan for re-balancing when the time ultimately comes. And that leaves me wondering whether we are fully communicating.

 

[GoneDiving]

Correct. It's entirely possible that we are indeed in furious agreement. I just still don't understand the plan for re-balancing when the time ultimately comes. And that leaves me wondering whether we are fully communicating.

Sorry if I didn't fully answer your comment/question.

Yes, I agree that, even in a well balanced pack, some deviation over time is inevitable. The degree of deviation and the time it takes remains to be seen. Should it become a concern (say, 0.05v or one cell heading >3.40v well before the others) I would just disassemble the pack and to another top balance. Time consuming but not difficult.

Cheers.

 

[Tator]

As I am still trying to wrap my little head around the LiPo4 technology, please indulge me in a perhaps simple question. As an owner in a yet to be installed drop-in system (Lion) do I need to be concerned about balancing the individual prismatic cells in each 12v battery case or does the bms take care of this and do I also need to be concerned about balancing between the 6 12v cases wired in parallel?

Tator

 

[sbman]

As I am still trying to wrap my little head around the LiPo4 technology, please indulge me in a perhaps simple question. As an owner in a yet to be installed drop-in system (Lion) do I need to be concerned about balancing the individual prismatic cells in each 12v battery case or does the bms take care of this and do I also need to be concerned about balancing between the 6 12v cases wired in parallel?

Tator

The drop in units have an internal BMS that will automatically keep the internal cells balanced. If things get out of hand beyond the BMS limits, it will cut the entire drop in unit off from the system and you'll have to make a manual corrective action to fix it at that point, per the manufacturers instructions.

The challenge is that most brands have no indication that the cutoff has occurred, so unless you notice that your parallel pack isn't performing as it should, you can go unaware that a single unit has dropped off line.

 

[GoneDiving]

The challenge is that most brands have no indication that the cutoff has occurred, so unless you notice that your parallel pack isn't performing as it should, you can go unaware that a single unit has dropped off line.

Several also limit/don't recommend at all that their batteries be wired in parallel and/or series. Different brands differ in their requirements so check before buying.

"Can be connected in Parallel
The SuperPack Litium batteries can be connected in parallel to increase capacity. Series connection however is not allowed and will result in damaged batteries if done so."

https://www.outbackmarine.com.au/victron-lithium-superpack-25-6v-50ah-m8

 

[Jeff F]

Gonediving, I think I've found answers to our discussion on balancing.

In post 24 in questions regarding cell monitoring you said

"Yet to be finalised but likely two 8S battery monitor/dischargers. I've used the 6S versions successfully before but want to ensure the 8 individual cell voltages are visible on one screen"

I'm going to guess that the units you're referring to have balancing capabilities. Can you point to an example of one you're considering or have used? Typically these units bleed off high voltage cells at a low current to balance.

I went for a quick tour of eBay and found 24s active balancing units @2a that appear to output data to various networks for under $150. Here's an example
https://www.ebay.com/itm/283822204750

In any case there seems to be no shortage of solutions to my remaining concern about balancing. For me that closes the circle. It all looks doable, and if ordering from China very inexpensive.

I get the need to idiot-proof the 12v drop in units, but I'd be more comfortable with building my own. Thanks for sharing.

 

[GoneDiving]

I'm going to guess that the units you're referring to have balancing capabilities. Can you point to an example of one you're considering or have used? Typically these units bleed off high voltage cells at a low current to balance.

This is the 6S unit that I have been using in other packs:

https://www.ebay.com.au/itm/2-6S-Lithium-Lipo-Battery-Monitor-BB-Ring-Alarm-Low-Voltage-detection-/224423433818?_trksid=p2349624.m46890.l49286

Works well as a convenient readout of cell voltages. It does have balancing capability but is unbelievably slow and gets hot. I melted one balancing a 30ah pack over two days.

This is what I'm considering for the 16S pack under discussion:

https://www.ebay.com.au/itm/G-T-Power-8S-Battery-Voltage-Capacity-Checker-Balance-Discharger-Servo-Tester-/303791797631?_trksid=p2349624.m46890.l49286

https://www.ebay.com.au/itm/ISDT-BG-8S-Smart-Battery-Checker-BattGo-Receiver-Signal-Test-Protocol-Analysis-/152708504865?_trksid=p2349624.m46890.l49286

Note that I haven't put a lot of time into researching them yet and will just use them as cell readouts. I have no intention to use them as balancers.

Cheers

 

[GoneDiving]

Progress Report and hard numbers on balancing and deviation

My stupid power supply charged all cells in parallel to 3.70v CV and 0.00a CC. Yes , I know not good!!

Off charge and resting for 24 hours, the still parallelled cells fell to 3.58v.

When separated, the cells remained at 3.58 +/- 0.00v.

Reassembled as 16S3P (note no parallel connection between cells of the same voltage) and 1.8kw discharge for 40 minutes and voltages fell to 3.34-5v per cell, 53.5v for the pack.

Effectively, all cells are now sitting at what I would consider within their operating range (3.30 to 3.40v) and are all within 0.01v.

Let's see how long it progresses.

The next steps are to assemble the inverters in parallel and test under loads approaching 10kw and for numerous cycles.

Thanks for everyone's comments

 

[Rhinoc]

Great thread and one that I'll re-read in more depth when time permits. I've been lusting after a lithium upgrade for some time but admittedly, don't have the confidence or discretionary time to do the necessary research to go DIY.

Love the vessel also and would love to do a refit of something similar nature as a future live aboard.

Thanks for posting.

Ryan

 

[GoneDiving]

Update:

I did a dummy assembly to allow minimal equipment to get me home after the haul out ie 7.5kw anchor winch, 12v bilge pump and auto pilot. Bench testing showed that, while one inverter could act as a battery charger, the two inverters in parallel will not. When plugged into a regular house GPO socket the house's RCD trips. Tech service confirmed that the inverters need to be installed before a RCD equipped power supply because they have a fixed Safety Earth Neutral bridge. They assured me this will not be a problem for a dockside shore power. This remains to be proven but it's not a big concern for me as I never use shore power now and am unlikely to need it with the larger solar generator set up. I have no idea why one unit by itself could be used as a 48vdc charger.

Otherwise, the units have bench tested well and will be fitted after the haul out.

Photos below. If you are triggered by lack of safety cutouts then look away. As I said, it's a get home only configuration. DC busbars are to the right, AC Out are to the left. The appropriate breakers etc will be added to the permanent install.

The VFD is a 230vac 15kw model to drive the anchor winch. For those that have not used VFDs to convert single phase to 3 phase, double the rating is used to prevent the unit tripping due to over current on the input side. They are a useful way to also prevent inrush current spikes on startup.

Cheers.

 

[Simi 60]

Have spent the last few days down the LFP rabbit hole and stumbles across this thread
Information overload.

 

[Simi 60]

Gone Diving, can I ask who the supplier of these, I assume EVE 280ah batts are and what the damage bill was each landed?

If I went that way I'd be after 3 X 8

 

[GoneDiving]

Gone Diving, can I ask who the supplier of these, I assume EVE 280ah batts are and what the damage bill was each landed?

If I went that way I'd be after 3 X 8

The supplier was OYE Store via AliExpress. Yes they are Eve 280ah. The price was AUD8400 landed for 48pcs.

Note that prices are reported as lower on Alibaba but I've found it's a wash once freight is included.

OYE and AliExpress often have sales so if you are not in a hurry it's worth waiting for these.

To date the cells have only lightly been used but no complaints.

Good luck

PS ensure your purchase includes studs and nuts not M6*10 bolts.

 

[Simi 60]

Thanks

 

[Simi 60]

Well, I hope I don't regret it but placed the order yesterday for more battery than we need but building in extra for the future if deciding to replace gas with induction down the track.

Same brand cells, different supplier, well respected on various LFP forums
Paid the slight premium to get matched and batched cells vs commodity cells.
New version of that battery will be supplied with laser welded studs, larger buss bars and a claimed longer cycle life

Would have liked winstons but they would have us out near $6000 just for the cells for a single 400ah @ 24v and I was not convinced that would be sufficient for our needs

Ended up getting 840ah @ 24v of the EVE 280 + 3x 200amp JBD BMS
making 3 stand alone 24v batteries of 280ah each
If one has issues for any reason we can run on two, or possibly even one if we had to.

Including all freight and taxes this bigger bank with the extras has came in at less than what the small winston bank would cost for cells alone.

 

[GoneDiving]

Well, I hope I don't regret it but placed the order yesterday for more battery than we need...

"I wish I bought a smaller battery" said no-one ever. ?

 

[mcarthur]

Watching carefully Simi - let us know how it goes!

 

[mcarthur]

"I wish I bought a smaller battery" said no-one ever. ?

I've definitely said "why the heck do these weight so much"! [expletive changed to protect the innocent]

 

[Rhinoc]

Well, I hope I don't regret it but placed the order yesterday for more battery than we need but building in extra for the future if deciding to replace gas with induction down the track.

Same brand cells, different supplier, well respected on various LFP forums
Paid the slight premium to get matched and batched cells vs commodity cells.
New version of that battery will be supplied with laser welded studs, larger buss bars and a claimed longer cycle life

Would have liked winstons but they would have us out near $6000 just for the cells for a single 400ah @ 24v and I was not convinced that would be sufficient for our needs

Ended up getting 840ah @ 24v of the EVE 280 + 3x 200amp JBD BMS
making 3 stand alone 24v batteries of 280ah each
If one has issues for any reason we can run on two, or possibly even one if we had to.

Including all freight and taxes this bigger bank with the extras has came in at less than what the small winston bank would cost for cells alone.

Will be really interested to see how this plays out for you Simi. I entertained the idea but just don't have the time to go full DIY. I purchased a 400AH 12v unit locally and am still getting used to it. Takes a while to get your head around things after spending your life watching battery monitors with lead acid. Loving it so far.

Question, some of the reading suggests not to let lithium sit at 100%. What does one do whilst on shore power? I have mine appropriately configured for the lithium which holds them at float 13.5v.

Cheer,

Ryan

 

[GoneDiving]

I've definitely said "why the heck do these weight so much"! [expletive changed to protect the innocent]

All jokes aside, this is another arguement in support of the DIY route. I seriously doubt I could lift a 8D battery into position in my cabin or engine bay let alone the number required to replicate my 43kwh pack. 5kg cells are much more user friendly.