Mixing Lishen 272ah & Eve 280ah LiFePO4 cells in a 3p8s configuration

[DavidS]

Question: is it advisable to create a battery bank made up of (2) 3p batteries of Eve 280ah LiFePO4 cells and (6) 3p batteries of 272ah Lishen LiFePO4 cells.  See attached configuration image.

Details: I currently have an 8s configuration of Eve 280ah cells managed by an SBMS0.  I'd like to expand this to a 3p8s configuration.  I understand that, due to intentional choices in the way the SBMS0 works, it cannot communicate with another SBMS0 in any meaningful way.  I also understand that it is expected that the battery managed by an SBMS0 present as a 3s-8s config.

 

Currently, it is not possible to easily obtain 280ah Eve LiFePO4 cells due to recent changes in the marketplace.  Lishen 272ah LiFePO4 cells are now common and relatively inexpensive.  Their dimensions are similar enough that intermingling them should be approachable from a physical perspective.  Two of the current Eve 280ah LiFePO4 cells are lower performing (but in spec).  In the proposed configuration, those 2 cells would be scrapped.

Thanks!

[Dacian Todea]

They may be the same cells or very similar anyway so yes there should be no problem to combine then. The pack will have the same capacity as the lowest group of 3 parallel cells.

Parallel cells are in no way different from a single cell.

[Dave McCampbell]

David.  I would be interested to know what has changed in the market so that EVE cells are not easily obtainable?  Also, what is wrong with your two EVE cells?  I understand that they do have a warrantee replacement provision.

We have just done that with our RJ 271 8S2P bank.  We had one unusable RJ cell due to excessive self discharge and replaced it and its mate with 2, easily obtained here in the Philippines, Lishen 272 cells.  The RJ warrantee replacement cell won't get to us for another month.  It has taken a bit of manual balancing to get the Lishen cells balanced with the RJ cells so that they are all within 15mv now at charge termination.  And it appears from our Node Red charting that the charge curves are similar, but not a perfect match.  They and EVE cells all have exactly the same dimensions and external appearance, however the connectors are not of the same quality as those from RJ and there is no warrantee with the Lishen cells.  UTube testing and comparisons of these cells are very good and the price is right.

[DavidS]

I'm not 100% sure what's changed.  I believe that BMW may have announced an agreement with EVE regarding using their cells (of some type) in BMW vehicles.  That could be rumor.  Regardless, when I last looked, most places seemed to be out of EVE 280ah cells and Michael Caro over on diysolarforum.com couldn't get them for a decent price anymore.  WRT my 2 "lesser" cells, they're probably ok, they are just consistently the lowest voltage cells out of the 8 pack at the end of a charge cycle.

[Jhon]

Basen had a problem getting enough Lishen cells, but had plenty of EVE cells.  They advised me not to mix them and on Will's forum every post advised against mixing them as well.  I would have thought you should be able to mix them, but I have no experience and Dacian seems to think you can as well. 

I am getting one of the Lishen cells replaced, yesterday my pack cut out at 38% as that one cell reached 2.5V well before the others. They were not totally  unresponsive, just took a month before they finally send one.  They seem nice and I have a feeling they will get better, if not people will have to look for other suppliers.

[DavidS]

Most of the posts advised against mixing on diysolarforum.com except in specific cases. I posted on this specific scenario there and the advise aligns with Dacian’s advice here. It shouldn’t be an issue. 

[Dacian Todea]

It is not ideal to mix cells but this are similar enough that it will not be a huge problem. They will all be top balanced so total usable battery capacity will be as low as the lowest group of parallel cells.

Those are 272 and 280Ah cells so about 3% difference in cell capacity and I do not expect much difference in internal impedance of the cells that is within tolerances of some lower quality cells that will be all the same model.

Ideally if they where the exact same physical size you will have groups made of two 272Ah cells in parallel with one 280Ah cell but not sure what is wrong with the two cells you have.  It seems strange you will have cells that are defective so maybe you just misinterpreted your test results.

[Dave McCampbell]

Jhon.  Did your Lishen cells come with a warrantee and QR code documentation?  If so for how long?  Also, do you know if the problem cell had a lower capacity, or if there was some other problem what it was?

[Dave McCampbell]

David.   If just a problem with their voltage being lower at charge termination, is your balancer not effective?   If not have you tried adding fractional amps to them with a bench power supply until they get closer in balance with the others at charge termination?  This has worked for us when we had the problem cell and recently getting the two new Lishens in balance with the 6 original RJ cells.  Our small adjustable load also helps when we need to subtract fractional amps.  The overall cell delta is currently less than 20mv at charge termination, and we can get it closer in the future.   But for now we are watching to see how stable that is.

We can now track cell charge curves with Sherry's Node Red display.  Using this we can see the different charge curves for our Lishen and RJ cells.  With these cells the problem using any balancer while charging seems to be that the voltage curves intertwine on the way up.   So a balancer will sometimes be taking mamps from a cell that is high on the way up but low at charge termination.   So it doesn't seem to be very effective at reducing the voltage delta at charge termination.  We were using default SBMS0 balancer settings for that.

Does someone have an explanation for how a balancer can be effective with these types of cells?   Are we missing something here.?   

[DavidS]

I tend to agree with the analysis that my interpretation of the results is wrong :).  I'll be running a battery capacity test soon, so we'll see where they bottom out, then we'll recharge to see where we sit.

The SBMSO brings them back into alignment nicely as they settle after charging.  I guess one cell will always get to charge termination value first.  More experimentation!

[Jhon]

"Did your Lishen cells come with a warrantee and QR code documentation?  If so for how long?  Also, do you know if the problem cell had a lower capacity, or if there was some other problem what it was?"

Cells were sold to me as 280Ah Cells, and came with the wrong lables as they were Lishen 272AH cells. But the difference is so small I didn't worry too much, just kept quiet (them knowing you know and staying silent is an effective bargaining chip) Terminals were clean as was the QR code.  This one cell had a solid blue vent cover, while the other 7 had a blue ring around them.  After top balancing, I got 274Ah out of the pack with another cell than this one getting to low first.   Vendor offered a five dollar discount per cell on a subsequent order because I got only 274AH instead of the 280AH.  Batteries are still a good deal so I let it slide.

Now a month later, that cell with the blue vent cover is losing capacity fast.  Last Friday, the pack cut out at 38% 4 weeks before it was 24% Top balancing does not help and at 99% charge the cell delta is 12mV.  (I am new to this so I tried a few things to see if I had done something wrong.)  A month ago I talked to the vendor and they said they would send a replacement cell, but it took them about a month to actually ship it.

I also ordered another 8 cells from them at the same time because I thought they had to match the manufacturer, (there is a cheaper vendor now, but I am not sure if they are reliable, Will and others are going to test...)  I did that after reading about half a dozen posts saying I shouldn't mix them.  I am stubborn, but some of the posts seem to be from people who, unlike me, know what they are talking about.  So even though I see no reason why they can't be mixed, it is too expensive for my budget to experiment.

Now I have no idea why that cell is misbehaving, I checked the temperature, the connections etc  The box it was in had been dropped during shipping, but I did not see any external damage on that cell (I did see some slight damage on another one) and the only difference is the colour of the vent cover.

I payed a week ago, they have not shipped them yet and after shipping, I will have to wait a couple of months for the shipment to arrive, it will give me time to build the pergola (with 8 more solar panels) "she who must be obeyed" wants.   I did not expect perfect batteries from China, but I did not expect one cell to fail this badly.  I am sure that with a little gentle persuasion, those Chinese vendors could well become decent suppliers.

[Dacian Todea]

Jhon,

What do you use to charge the battery ? What is the charge rate ? Did you changed anything in advanced parameter settings ? If charge rate is very high or you have bad connections between cells charging may terminate quite a bit earlier than real 100%

If your current shunts are not properly set and current measurement is wrong you will also get this sort of bad results as SOC is calculated based on measured battery current.

A photo of the first monitoring screen when battery is fully discharged can also help determine if a cells was at fault or bad current measurement.

[Don Fukushima]

One option is to buy a few spares just for cases like this. Testing is easier as you can swap in a spare and keep your full pack performing well. You'll have time to diagnose the issue with the poor cells/modules without the stress of the whole battery pack not being 100%

[Jhon]

What do you use to charge the battery ? What is the charge rate ?

I use Solar during the day, which can peak at 56A, (yesterday for a brief moment I got 27A out of one set of two 300W panels in parallel.  But generally they charge at 35A  from four 300W panels.  At night, depending if it is going to be cloudy the next day, I charge with the multiplus at about 40A max (I set it to draw 10A from the grid as North American sockets can't always cope with the full 15A for long.)

Did you changed anything in advanced parameter settings ?

No.

If charge rate is very high or you have bad connections between cells charging may terminate quite a bit earlier than real 100%

I checked the temperature on all of the connections, if one was a little bit less, I would have expected a slight increase.

If your current shunts are not properly set and current measurement is wrong you will also get this sort of bad results as SOC is calculated based on measured battery current.

Shunts seem to be calibrated, the charge and discharge rates seems to match the clamp meter.

A photo of the first monitoring screen when battery is fully discharged can also help determine if a cells was at fault or bad current measurement.

Attached a screenshot of the battery 10 minutes after the BMS disconnected the load (currently only the multiplus), I can't remember if the battery was being charged with the multiplus or from solar at the time, my guess is the multiplus as my wife  and son need it to power their laptops and internet.  The other load at the time connected to the multiplus was a small (800W) heater, a large fridge and a large freezer.

I have to check the temperature readings, for now they are "good enough"

[Jhon]

"One option is to buy a few spares just for cases like this. Testing is easier as you can swap in a spare and keep your full pack performing well. You'll have time to diagnose the issue with the poor cells/modules without the stress of the whole battery pack not being 100%"

That is good advise Don, problem is that my budget is tight and I don't have enough experience with these cells from China to justify getting spares.  This whole project is about trying to learn more and even the bad cell is a learning moment.  If I was doing this commercially, I would go a different route and buy local to reduce the maintenance headaches.  If people wanted my help for free, I would be happy to assist with the installation of whatever system they bought.

[Dacian Todea]

Your pack is super imbalanced due to cell 6 that is the one empty. All other cells may still have the remaining 38% SOC if not a bit more.

Not sure how that happened ? Is that cell 6 from a different group. You will need to manually charge that cell 6 as it is missing about 100Ah to be equal with the others so that will take forever with SBMS cell balancing that is designed to keep the cells in balance once they start relatively balanced.

[Electrons4me]

Hello da...@svsoggypaws.com

Could you possibly let me know what "Sherry's Node Red display" is? I'm quite interested about the comment you made about tracking cell charging with this device.

Thanks

E4M

[Dave McCampbell]

E4M.  I will pass this on to Sherry who can answer your questions better than I.  In the meantime have a look at this link on our website:

http://www.svsoggypaws.com/electrodacus.htm

See these two sections below that:

Soggy Paws Notes on Electrodacus SBMS

Electrodacus MQTT & Node-Red Getting Started Guide

Dave

[Dave McCampbell]

Jhon.  Did your Lishen cells come with a warrantee and if so for how long?  

Did they have a QR code sticker and other stickers indicating capacity and internal resistance that would indicate they may be grade A cells?  Finally, did they come with terminal studs or bolts and connectors.

The Lishen cells we bought from an internet store in Manila had a 1 week warrantee, no QR code or other stickers and had installed terminal studs and connectors.

[Acercanto]

As a more general question, say I have 200ah cells and want to expand my capacity. Assuming it will work physically with the connections and all, is there any contraindications for adding, say, a couple of 50ah cells in parallel? So each hypothetical parallel group would be 200ah+50ah.

Acercanto

[Jhon]

"Your pack is super imbalanced due to cell 6 that is the one empty. All other cells may still have the remaining 38% SOC if not a bit more.

Not sure how that happened ? Is that cell 6 from a different group. You will need to manually charge that cell 6 as it is missing about 100Ah to be equal with the others so that will take forever with SBMS cell balancing that is designed to keep the cells in balance once they start relatively balanced."

Yeah, the cell seems to have a lover capacity, maybe something broke inside during shipping and only showed up after a month of use. At 100% charge that cell is close to the rest of the pack, so I don't know what else to do but replace the cell Dacian.

Thanks for the node red link David, I bookmarked it as well.  I like the display on the SBMS0 but I am also interested in doing some "decision making" with a raspberry pi  next winter, and unifying the display with the multiplus, when I have more time to play.

"Jhon.  Did your Lishen cells come with a warrantee and if so for how long?"

No idea to be honest E4M, I got mine from alibaba and they were recommended in Wills forum, but they are sending a replacement.  I had the battery for about a month before I contacted them asking for a replacement,

"Did they have a QR code sticker and other stickers indicating capacity and internal resistance that would indicate they may be grade A cells?  Finally, did they come with terminal studs or bolts and connectors."

I can't interpret the QR code Dave, I am still very much a noob, but they did look like brand new cells.  I did not trust the sticker as they had 280AH written on them and they are Lishen cells, which come as far as I am aware at 272AH.  No idea if they are grade A cells, they sold them as such, but I have no way of telling.

The Lishen cells we bought from an internet store in Manila had a 1 week warrantee, no QR code or other stickers and had installed terminal studs and connectors.

Mine came with screws and buss bars, the new batch will come with grub screws and hopefully an extra set of them to replace the others.  I don't know how a week's guarantee would even work, it took me a week to balance them the first time, because I had no idea what I was doing.

[Jhon]

"As a more general question, say I have 200ah cells and want to expand my capacity. Assuming it will work physically with the connections and all, is there any contraindications for adding, say, a couple of 50ah cells in parallel? So each hypothetical parallel group would be 200ah+50ah."

On Will's forum, they are saying that you can't, Acercanto, my supplier told me I can't parallel an EVE cell with Lishen, but I have no idea why that would be the case.  There are also people who say it can be done. 

[Soggy Paws]

E4M,

I just uploaded an updated copy of my PDF file to our web page about Electrodacus.

http://svsoggypaws.com/electrodacus.htm

This is v.002 of my guide on trying to get going with Mosquitto and Node-Red

[Dacian Todea]

Acercanto,

You will not find 50Ah cells that will be the same dimensions with 200Ah cells but in the hypothetical case where something like that existed then yes you will have been able to parallel a 200Ah cell with a 50Ah cell to have a 250Ah equivalent cells.

All this prismatic cells require compression so when building a pack the cells will need to have the same size and structure. For example plastic cells have ridges and so you can not compress those with the aluminium case one even if there where exactly same size as pressure on the aluminium will not be uniform due to those ridges on the plastic one.

I already explained this somewhere in the forum but here is a cut prismatic LiFePO4 cells to see how it is build inside.  Say that yellow Winston cell is a 160Ah cell then you can say internally is build out of two parallel 80Ah cells or hard to say from the photo but can also be considered as made of 32 parallel 5Ah cells (hard to number from the photo) but you get the idea. Adding another 160Ah in parallel or even a 80Ah in parallel will not be any different than having a single larger 240Ah or 320Ah cell. Of course the cells need to have the same chemistry you can not parallel a LiFePO4 with a LiCoO2 cell. 

[Dave McCampbell]

Thanks Jhon for that info.  Intact QR, capacity and testing info stickers plus price are all good indicators of grade A cells, but no guarantee.  Will P's recommendation and no issues replacement are all plusses for those cells and the seller.  We will see how our probably grade B cells do.  Hopefully, one will perform properly.   The one week warrantee is just enough time to get them, see that there is no physical damage, and check to see if there are any obvious energy issues.  Our warrantee replacement will be a month getting to us.  Good to know Will's Alibaba seller is reliable.  

[eksine]

unless your diagram is generic it looks like you have a single cable connection to make the series connection between the 8 sets of 3 parallel banks. and not only that but you have each cable at the very end of each parallel bank so that  means all 3 cells have to flow through that single battery in order to complete the circuit. I would at the very least move that cable so it's at the center of each parallel bank so that no single cell has to hop thru 2 cells just to complete your circuit. actually it would be better if you had all 3 cells connected to the 3 cells they are series connected to, meaning cables for each cell. hopefully that sort of made sense.

I also wouldn't mix lishen and eve if you can help it because the specs for internal resistance are different, but they are very similar so it shouldn't hurt the entire bank hardly at all unless your eve cells are really old compared to the new lishen cells, unless those are old too. I would measure internal resistance , that's the best way to tell if these will work well together, if they are say 0.15 ohms and 1.0 ohms, etc, I'd be worried  

[eksine]

just to add I will never trust will's recommendations again. I bought lishen cells based off his "$433 are these the cheapest cells" video and it was the worst 3 month nightmare buying experience of my life. spend a fortune buying and testing the cells for capacity. they only pulled about 245-255 AH each and internal resistance seems extremely high, I mean extremely high and the terminals were heavily scratched, plastic flaps on top peeled back, one had swelling. aliexpress refused and said work it our with seller, seller refused to help and just responded with generic nonsense dribble. both asked for "photograpic proof" although they seemed to just ignore them. in the end thank god visa finally did a chargeback although they require a return, I told them there is none, seller refused to even communicate with me. 

the only positive that came out of it is I now have advanced knowledge of how to build resistor discharge banks and I now know how to discharge a single lifepo4 cell using any size load I want. I can do a 70-80 amp discharge using 0.5 ohm resistors and be done in 3 hours vs 14 hours + before. I also discovered duo chargers that can do 70 amp charge and discharge with whatever end voltage I want. all with relay control so I do not need to worry about over discharging. shoot I even learned about 250 amp relays and solid state relays, so good for that, but terrible for the nightmare I went thru. never ever will I listen to will's recomendations, he always uses the lowest quality products, never have I seen him use Samlex, etc or any high quality brands, other than battleborn, which are overly expensive

[Dave McCampbell]

Bnphoto.  You should write Will directly to let him and others know of your problem with that seller so others don't have the same problem in the future.  As I recall looking at Will's several cell seller recommendations on his website, there were several buyers that had similar problems you did and reported them on the seller evaluations.  Did you see those?  Also, can you give us a link to the 70amp duo chargers that charge and discharge using one unit?

[eksine]

If you're talking about leaving a review for the product review on the Aliexpress website I did, I did read a whole slew of bad reviews similar to mine with either swollen cells or other major issues. The thing is mine is on there and those bad reviews appeared all together so that means they had a very large batch of bad cells. The fact the seller chooses to ignore the customers when they complain means I will never buy from them again and since Will has listed these cells as his top recommendation I would definitely not listen to anything he recommends. I think he's blocked me anyways because he doesn't reply to my comments any more. The charger is the Junsi iCharger 4010 duo, the best one they have although it is a few years old. Their older duo chargers like the 306 and 308 have known failure issues so avoid those. An alternative is the icharger S6 which is 40 amps. several things to note is that the 4010 has a sequence of connecting the wires up otherwise the balance circuit can get damaged, basically connect the charger up to the cells first than power on the power supply. There is a "hack" with RC chargers where you can modify used server power supplies by HP or Dell and get 12.3 volts - 62.5 amps , 750 watts or 75 amps 900 watts, Don't quote me on the exact numbers but that should be close. on 220v power they go to 1200 watts but since I don't use that I don't remember the amps. you can parallel them for 24.6v or 36.9v, etc but you must perform a DC float/ DC ground isolation mod with plastic M3 screws and washers. These server supplies cost $15 or $20 each and look near new. the 4010  requires specific power, you cannot use the server power supply hack, because if 1 power supply fails it will permanently damage the charger. The one I use and was recommended is " Meanwell AC-DC Single Output RSP-2000-48 2016W 48V 42A Switching Power Supply" I got it on a closeout sale for $225, both the charger and meanwell cost $350 each. Expensive I know and maybe controversial since many regard RC chargers a joke, but this is the only charger which supplies I think something like 1-50v (or maybe it's 1-48v DC) and up to 70 amps in synchronous mode where you combine the 2 channels bridged. each channel is 40 amps. yes I know 40+40=80, they derate it in dual mode. There is no other charger that has allowed me to set any voltage I want but also set any amperage I want and will shut off. It will do a capacity test however for 70 amps you need to instead use a partially charged bank of cells so it can dump the load into that as well as provide power for the charger. the learning curve is kinda steep, you just need to know there is 1 setting that is vague but important, it basically tells the charger to wait until the current drops to 10% of the 70 amps in order to get every last ounce of power you can if you're going to do a capacity test. I guess that setting isn't too important since that's only a tiny fraction of the capacity. Progressive RC is where I bought the 4010, but I also bought the 3 year warranty for $50 , this company is different, any time it breaks, they do the repairs in house because Junsi is overseas and Progressive RC is in america. I think they do repairs even if you bought it from somewhere else but they will charge you. A programmable charger for lifepo4, that's the advice I'm giving you. I have also tested the P30 from ISDT and do not recommend that brand.
 as for reliability this is the only dual charger with reviews that suggest they are a new design unlike anything before them however should they fail the fix is pull the cover off and replace fuses, mosfets, etc, anything burnt just like in any amplifier. 

[eksine]

[eksine]

I know it's a bad picture but it's the only one I had, you will encounter connectors you have never heard before if you go this route like EC connectors which require a special CNC milled metal punch $6 from progressive RC (the only place that makes and sells this tool) or 4mm barrel connectors and of coarse XT60 and XT90 connectors. or you can just hack off the connectors and do what you want, you can even solder your own wires on the circuit board as shown in the other pictures

[Dave McCampbell]

Eksine.  Thanks for all that information and the pics.  I thought you were describing a dual bench power supply/adjustable load for manually testing and adjusting cell balance.  Re the cells I would just make sure you have documented your problems with this seller as many places as possible.  There are several other lithium forums you might consider sending a short note to.