LiFePO4 Batteries: Discussion Thread for Those Using Them as House Banks

[ebaugh]

Terry,

That article did deal with various lithium chemistries, but I assumed the conclusions he drew applied to all those mentioned and LFP was one of them. Of course the voltages are different for each variation.

GBS only rates their batteries to 1500 cycles, yet I attached a purported test of a 40 AH cell showing 2000 cycles to 100% discharge. At the end, it only holds about 80% of what it did when new. I read the Winston cycle life charts differently than you as "or" not "and".

The conclusion I drew after studying everything I could find is all the prismatic LFPs are basically the same, LIFEPO4, LIFEYPO4 or LIFEMNPO4. There is nothing demonstrated by any firm data that differentiates the Y or MN from the base chemistry. Other than perhaps low temp operation for the MN variant. If you've seen something, I'd love to read it.

But there is a consistent theme over time where the operating ranges were lowered. Especially Thundersky which started out at 4.25 and now the Sinopoly spec is 3.8V for what I think is the same battery. Even my GBS cells are made in a plant that started out part of Thundersky. GBS lowered their specs at some point from 3.8 to 3.6V. I believe the initial lab research used 4+ volts, but operationally it's been determined pointless.

And as you point out, they are fully charged well before those numbers anyway. At 3.5V, mine are at least 96-97% charged if 3.6 is the end. This is based on watching 48 cells individually under initial charge. At a 10A charge rate, the meter just starts clicking off at 3.5V and in about 4-5 minutes your done.

So I think subjectively, less is more, in regards to cycle life. Especially since we don't have fine control over our chargers until the marine market adds logic for lithium.

If I followed your attached file right, you were showing a series of cells from top of charge, loaded, then unloaded and the recovery voltages. If so, there is a typo in the second white box. How much capacity was drained in each cycle?

I already have the LVC alarm you suggested. It's my inverter. I have it set to drop out before the BMS and alarm. It's not the entire load, but probably represents 90 percent at anchor and 70 percent underway.

On the HVC side, I have multiple charging sources coming via different paths. I could use the SS relay for the alternators since their max practical charge rate is probably just under 100 A. But my normal charge rate on generator power starts at 160 amps and runs 150 plus until it finally starts a taper to zero at around 85% SOC. The taper effect is really caused by the voltage drop in the charging cables as the charger sees the finish voltage before the batteries do.

I tested the HVC function during installation and observed the cutoff. It supposedly has a timer and will reconnect after a period of time has elapsed, but I did not test to see how long that was. But while I want the protection from accidental overcharge, I'm really satisfied with normal charges to less than the 3.6V cutoff of the BMS.

Would you suggest a one time observed charge to 4.0V at this point? How long is it safe to hold a 3.6V charge to give all 12 cells time to even out? Both of these are to attempt a better balance.

Thanks, Bob

[T1 Terry]

Yeah, I saw my typos there are a few :lol: The discharge rate is 100amps and for 12.5 mins each time so 83 Ah were drawn from the 90Ah battery, you can see it starting to sag towards the end of the last cycle, it’s coffee percolator powered by an inverter.
Sinopoly have achieved some great results in regards to better high discharge current and voltage sag, excellent for the EV market they are designed for, not much of anything for us though. Winston seem to have extended cycle life so as the technology develops there is a greater difference between manufacturers approaches to what they consider important so real differences are appearing between the products. Where factories were producing for every one in the early days now manufacturers have taken control of certain factories to produce only for them and the chemical make up and closeness of tolerances are changing all the time, thankfully for the better. As you said earlier, it will be years before we know if the claimed cycle life is real or not but I know my 720ah pack gets some real punishment running the household stuff via an inverter as the only charging is solar, when it rains for 4 days straight the batteries get dragged way down, I've had the inverter drop out at 10.5v a few times and over the 720Ah recorded on the Plasmatronic controller/monitor, yet they have bounced back every time so I'm very happy with the Winston cells.
As far as charging for cell balance, I have not carried out any testing of the chemistry that GBS use, it must be different as the claimed cycle life is less than half that of the Winston cells, but again, it's unknown yet if their claims are fact or fiction. As the manufacturer http://michellegbs.en.hisupplier.com/product-547941-Lithium-Ion-Battery.html recommends a max charge voltage of 3.8v I would not charge beyond that but at a 0.1C rate till the 3.8v was reached, then let the cells sit for 1hr and see if the voltage remains above 3.5v rested, if any cell pack doesn't hold 3.5v repeat the process. I was quite surprised just how many more Ah of charge went into the Winston cells after they reached the initial 3.5v per cell but once they had been fully charged they seem to accept charge readily all the way to 95% and stay as a pack but until then they jump around a lot, some times high, some times low compared to the others in the series group and taking up to a mth or two of daily cycling to settle down.
As far as using the solid state relays, they can be linked to handle higher current so multiple units acting as one could handle the generator charger, by raising the charger output voltage and allowing the BMS to cut charging when a cell reaches 3.6v you could reduce the charging time as the full 160 amps would continue till the BMS opened the relays. Once the cells have had that initial full charge I think you will find the charge acceptance will remain high up until around 95%, the last 5% can either be ignored or topped up with solar, 5% of 1200Ah is still 60Ah so a reasonable amount of capacity to not be utilising.
As you have such a big battery pack utilising the whole capacity will reduce the number of cycles needed per week so the actual calendar cell life will be longer and that's the bit that is important not how many cycle but how many years the batteries last. Still haven't heard of any one wearing a set out yet, murdering cells yes but not having them die of old age

T1 Terry

[ebaugh]

Terry,

It's all marketing...Chinese style I think? I went to the actual Chinese GB Systems website to double check the data there and found the same 3.8V max you did. Elite, the US distributer is using 3.6 now, but that may be because they only market "packs" of 4 cells. But low and behold, the manufacturer webpage has been changed in the last month or so to show a cycle life of 3000, versus the 1500 it used to show. Same picture and part number. Wonder what changed?

I think properly charged and discharged, the cells will last as long as the weakest chemical component lasts. Perhaps the electrolyte? And in such a fashion where ultimate life may well depend more for most users on calendar life, then cycle life. But this has to be proven, and I agree not all cells are equal. What I don't know is who really has the best Mfg process and Quality assurance since the market is so dynamic right now. I bought GBS because I could get them quickly, not because I thought they were better than any of the others. It would be nice to see some independent comparative lab testing of random samples from Winston, Sinopoly, Calb and GBS to see what might pop up after 500 to 1000 cycles. It would tell us a lot, but I think I'm dreaming.

Anyway...I will be in a marina in Aruba in a couple of weeks. I will bypass the BMS and try your suggestion to 3.8 V and see what happens and report back.

We have an all electric boat...including the stove...so we charge when we cook, usually twice a day. We need 1500 watts of solar, better refrigeration, and a propane stove to really get with the program. But if we had all that, I'd be less inclined to go lithium.

We burn through about 600 amp hours a day in two 200-400 amp hour cycles. This is why we are so focused on cycle life.....for now...but that number would fall to maybe 200-300 amp hours a day on one cycle if we ever get to the improvements mentioned above. But still, the admiral requires both a large freezer and large refrigerator, and the captain requires an ice maker. So I don't think solar would be able to replace what's lost at night and keep things going during the day regardless. And don't forget the washer, 2 computers, vacuum cleaner etc....hey we're a powerboat....

We need to see at least 5 years of mostly on the anchor cruising ahead to make the payback worthwhile. The savings are .5 to 1 gallon of diesel per day at anchor. At 200 days a year, that's still only $700-1000 year versus at least 5K, maybe 10K in improvements to get there with a new Bimini frame to house the solar and the carpentry work to replace the refrigeration that is harder to estimate.

Thanks for your help...Bob

[T1 Terry]

What size inverter do you have Bob? Could you increase the inverter capacity to suply all your electricity needs including the stove? Possibly te ice maker could be on standby while the stove is in use and the refridgeration put through an early cycle as well the also in standby mode till the stove is off again. then you could reduce the gen run time by 50% or more.
Reading through the GSB specs it would appear they are still developing their product so the improved cycle life could be the result of further testing. The 0.2C discharge rate to gain the 100Ah rated capacity is a surpising one, all the lithium ferrous cells are 0.5C, could be the different chemical make up. I'm sure they will improve that figure as well given time.

T1 Terry

[ebaugh]

Terry,

That .2C rating surprised me too. I thought it was higher, but that one I can't swear about it changing.

The stove has 3 stovetop elements at about 15A each at 120V, plus a two oven element at at 15A each, your choice of broiler or bake only one time. No practical way to cover even one stove top and one oven element without a huge inverter. That's 3000 watts, plus the other loads. I believe the existing DC inverter wiring even at 4/0 US sizing is only good for 150 amps or so at the 5/16 inch terminals. They get pretty warm as it is... You need to go to 24 to 48V inverters to get more watts...IMHO.

I have two inverter chargers, one 2K watts and one 2.5K watts, but different makes so they don't parallel. I get one or the other at a time for AC loads. But they do parallel for charging.

Best Bob

[deckofficer]

Just bought (8) 100 a/hr cells from Balqon - Advanced Transportation Solution for $880. Best price I've found so far, and only a 3 week lead time.

[Cotemar]

deckofficer,

How do you plan on wiring your new (8) 100 a/hr cells?
A picture or diagram would be great.

I do not see connectors sold on the Balgon web site Balqon - Advanced Transportation Solution .

Will you be making the connectors yourself?

[ebaugh]

Originally Posted by deckofficer

Just bought (8) 100 a/hr cells from Balqon - Advanced Transportation Solution for $880. Best price I've found so far, and only a 3 week lead time.

Good luck with your project!

[ebaugh]

Originally Posted by Cotemar

deckofficer,

How do you plan on wiring your new (8) 100 a/hr cells?
A picture or diagram would be great.

I do not see connectors sold on the Balgon web site Balqon - Advanced Transportation Solution .

Will you be making the connectors yourself?

I looked at the Balqon site. They have some great deals on 700 and 1000 amp hour 12V packages. Using what looks like standard Winston 700 and 1000 amp hour cells connectors included. The price range when I was shopping ranged from $1.10 to 1.25 per amp hour for 3.2V cells. The 1000 amp hour 12 V package works out to 86 cents per amp hour at the cell level. This pricing is essentially the same as good AGM batteries, before you factor in the usable capacity. Compared to a golf cart bank of equal usable capacity, its more expensive, still almost twice the cost.

Look for the HIQAP series.

Wish this was around when I bought in....

[Cotemar]

deckofficer,

Wouldn't you need (9) 100 a/hr cells to make a 3x3 -300Ahr - 12 volt stack?

[deckofficer]

It is for (2) 12.8 volt, 100 a/hr batteries to be used in an Ocean Torque kayak. One battery outbound for hopefully 35 nm (80% DOD @ 30 w/hr/nm), 2nd battery to return.

Also plan to use with a 9 lb, 1500 watt inverter to power an electric chain saw.

I'm still looking for my series bar connectors, as Balqon doesn't supply them, anyone have a source?

If these were for my house bank, I would size for capacity needed and have a single series string, as I'm not a fan of series/parallel.

[Cotemar]

deckofficer,

Series buss bar connectors are pretty easy to make.
1) Cut copper pipe to length needed.
2) Fatten copper pipe with a hammer.
3) Drill two holes where needed.
4) remove all burrs with a file

[deckofficer]

I just talked to Mario at Balqon, and he will have Steve band the (4) cells together, install the connector bars and hardware for both. Turn key package, which is what I like.

Cotemar, I know I could roll my own, thanks for the input. I'm not a fan of drilling on the drill press, small metal strips, so called the company and got the above results. Steve will call me with the cost, but to insure a tight banding, perfect current carrying bars, and the correct hardware, at the price they sold me the cells, there is plenty of headroom for his quote.

I'm glad I put my order in first, then started sharing the link on this and other EV sites. They had (32) of the 100 a/hr cells in stock when I placed the order 24 hours ago, now they have (4).

Since they build their trucks using the 1,000 a/hr cells, they have 346 of them. The popular for us using for the house bank 400 a/hr, only one left, out of stock for the 300 a/hr, but have 49 of the 260 a/hr. 101 of the 90 a/hr are in stock, but all other sizes other than the above mentioned, very low or out of stock.

Mario is a fellow EV-head, knowledgeable and fun to talk with. At $1.10 per a/hr, in the long run much cheaper than any other battery. There is only 13 million tons of lithium worldwide, with each electric car needing 1/4 ton, enough for 52 million cars and hopefully by then we will have a new chemistry that is even more energy dense at maybe a lower price.

[hellosailor]

Bob, I'm not understanding the math.

Four cells in each battery, each cell "Operating Voltage:2.8V~4.0V" according to Balqon. So, each of your batteries will be ranging from 11.2 volts to 16 volts? And you will relly be running 16-volt batteries, normally in a state of partial discharge?

Or is your electric kayak system designed for 16-volt batteries?

[deckofficer]

Originally Posted by hellosailor

Bob, I'm not understanding the math.

Four cells in each battery, each cell "Operating Voltage:2.8V~4.0V" according to Balqon. So, each of your batteries will be ranging from 11.2 volts to 16 volts? And you will relly be running 16-volt batteries, normally in a state of partial discharge?

Or is your electric kayak system designed for 16-volt batteries?

2.8 volts is fully discharged, 4.0 volts equalizing so not run that high on charge very often. More than 90% of its use, holds close to 3.2 volts, so 4 in series is 12.8 volts. Besides, electric motors can be run at higher voltages as long as you do not exceed the design rpm. At my 20 amp draw, that would be less than the 0.5C rate discharge, and as you can see from the graph that battery cell will be 3.2 volts down to 3.1 volts for 80% DOD, or on the 12 volt pack 12.8 volts to 12.4 volts.


Makes ya drool, doesn't it?


One of my humble cells