Powerwall using 50 AH cells

I have put together a system without the tedium of using 18650 cells which for me was a hell of a lot of work.

I have used 8 LFP or LiFePo4 cells which are in alloy cases in a prismatic form and are readily available for under $1 an amphour plus postage or freight to you from china.

The system has been running well now for 18 months.

The nominal voltage for these is 3.2V and cutoffs are high 3.55V and low of 2V. The BMS is sealed and waterproof.











8 cells gives me 28.4V max and this runs a 24 V Inverter of 500 watts nicely.

I run my 55 inchTV and a light and my security lights at this time to prove the effectiveness and longevity .This approach is rapidly being installed by Boaters and RVers.

With everything turned on there is less than 7 amps at say 27V flowing into the inverter so I am using a quarter of its potential
I am charging it with two normal house panels of 260 watts and can run the TV 24 / 7 if the sun permits . The battery feeds the system for about 10 hours without charging.

Sounds like a "just right" sized system, good job.
What are you doing for fusing, etc?

Fusing? Nothing .
The Inverter is fused for its rating ,the solar controller is fused and protected for its own rating too.
If I put it in a boat I would fuse and battery switch on the negative line with something suitable if I felt a need but I don't.
The battery is capable of outputting a huge current by short circuit but the BMS's maxs protect the battery again on the negative line.
Any more would be overkill and money wasted.

I went 18650 but I have been wondering about LifePO4 / fusing / do same principles apply. For example, if you built a 300ah LiFePO4 battery using 50ah(s) in parallel it would be 6 of them for each 'pack'. Do LifePo4(s) have same general issues as 18650 - e.g. you should fuse each of the 50ah(s) to the busbar for the pack?

>without the tedium of using 18650 cells which for me was a hell of a lot of work.
I must confess that after 5,034 18650 cells for my 1st 780ah @ 48v battery bank - I'm thinking of doing the next one with LifePo4 prismatics as soon as I can find a reliable source.

No that type of cell fusing was not necessary for mine. The BMS connects between each cell and if one cell discharges more than the set current the BMS disconnects the entire pack. If the temp is too high or the voltage too high or low on any cell the BMS just turns the whole pack off .

Systems can be built as you wish , so if you put say 2 off 50AH in parallel, as many do, I would start to think about fusing but perhaps one BMS would still protect each group of parallel cells.

If you put 8 cells in parallel it would not be the way I would go because you can't protect every cell properly.

I would make 8 battery packs of 8 cells each in series and give each one its own BMS. (They are $30 each for 8S),
then I would parallel the 8 packs knowing if one cell goes out then one pack will go out and 7 will remain in service.

Less capacity but full voltage.

tytower said:

Fusing? Nothing .
The Inverter is fused for its rating ,the solar controller is fused and protected for its own rating too.
If I put it in a boat I would fuse and battery switch on the negative line with something suitable if I felt a need but I don't.
The battery is capable of outputting a huge current by short circuit but the BMS's maxs protect the battery again on the negative line.
Any more would be overkill and money wasted.

Click to expand...

At least a fuse or breakerin series with the main battery would be agood idea.
The BMS might not operate as expected if there is a large short current, it's pretty stressful for the MOSFETs...

BMS without fuse is not good enough. A fuse or breaker should be used inline with the system . In this case it should be after the BMS so the BMS isnt disconnected if something trips.

Redpacket said:

At least a fuse or breakerin series with the main battery would be agood idea.
The BMS might not operate as expected if there is a large short current, it's pretty stressful for the MOSFETs...

Click to expand...

As you wish .
That's why I bought the BMS, so how many systems are you going to duplicate for paranoid safety to be good enough ?.
Inverter is fused too and that's all I am using atm.

Daromer - BMS without fuse is not good enough. A fuse or breaker should be used inline with the system . In this case it should be after the BMS so the BMS isnt disconnected if something trips.

Click to expand...

Same for you . Just your personal opinion and I do not agree. The fuse would be too slow and needs 10 times the current for more than one tenth of a second to blow and cut the line .

One of the things about this forum is we share knowledge, give feedback & learn from each other's experiences.
Collectively this adds up to a lot of experience.
We're suggesting accepted basics on the stuff we've all built.
Fuse disconnects/breakers are one of those accepted things - even if just to be able to isolate things to work on your gear!
We get it, your system isn't as complex as some here (not having a shot just saying).
The trouble is the short circuit current of your batteries is many hundreds of amps (at least) & that can easily make for spectacular outcomes with just a slipped tool.

A fuse is not about current limitation. That should be done by the tools connected. Its not My personal opinion. A fuse should Always be used between electronics. You say you have One then you have One and thats fine. But a bms is not a fuse or dead short protection. If you get a dead short a bms potentially can fail in open state. Thats what the fuse are for.

And what do you mean with same for you? I have fuses for each and every device rated att their spec but also have a Main breaker as needed by legal where i live.

I have yet not seen any commercial device without fuse.

OK so i didn't post this to have a discussion on fuses but to demonstrate the practicality now of a reasonably cheap prismatic type alloy cased cell as opposed to a cylindrical cell. This saves the very drawn out process of obtaining 18650 cells ,constructing them into a suitable bank with fusing and the voluminous wiring and connections that can go wrong. The cost too if you are paying $5 a cell for 18650's is a lot less.

I offer my experience in buying and using them and I have found they work well. They are now available on amazon and ebay and other suppliers are importing them.

A note of caution though . Prismatic cells have false cells just like 18650's and weight should be checked.The technology has not changed so discount any statements of "our technology is better".
The cells should be confined lightly on the large flat sides to avoid cell expansion under heavy loads or temperatures but the alloy case does a pretty good job of stopping that anyway.
My preference is to have a BMS on each string of cells so that each cell in the string is watched. When you order make sure the specs are for LFP,LiFePo4 which has a cut off voltage of 3.6V and no more or the cell will be damaged.

The BMS has back to back MOSFETS so current may only flow if both are switched on. If one blows the BMS will not pass
current.
This is a good FB page for knowledge. There is a text file of suggested weights for AH sizes on it.

https://www.facebook.com/groups/271980786862023/

daromer said:

But a bms is not a fuse or dead short protection. If you get a dead short a bms potentially can fail in open state.
And what do you mean with same for you?

Click to expand...

My appologies . I missed the above points .
The same for you meant the same answer was meant for you .
I will have to be careful how I put this next bit ..You being a super moderator and all.

I am intrigued though by how you think a BMS can fail in an open state?
That is how it can fail but still pass a current is what I assume you mean.
I will put a typical BMS circuit here for referral.

There are two back to back Mosfets usually ,each is turned on only if certain conditions defined by the chosen control chip are met. Otherwise it is off .
If both are on current can flow across the mosfets plates. If either one is off no current flows.
In my experience and from some hours of research on the net I can find no mention of a shorted mosfet being able to pass current. If that situation can arise I would like to know about it .

The best I can find so far seems to indicate a blown mosfet would act exactly like a fuse in that there would be no material left between the source and the drain to pass current?

Can you explain why you think that is the case please and perhaps offer some backup info I may study ?
I also attach a diagram of a simplified Mosfet .

tytower said:

I will have to be careful how I put this next bit ..You being a super moderator and all.

Click to expand...

this is of no consequence. We aren't that type of forum where we throw our moderatorship power around. We only act on spammers and people who are trying to purposefully start an argument. I haven't seen either of the above on this thread.
I'm actually intrigued by this thread and have been watching it to get answers myself. Please continue

I also think it's good idea to have a dumb but pretty much fail proof fuse inline with any BMS. For one in case a MOSFET should fail in a closed state, however remote/impossible that may be. But also other components of the BMS could fail and not shut off the MOSFET in the first place. The firmware could hang. Or a dead ant could short the circuit and make it do unintended things. Corrosion, bad soldering, lightling strike (even if not a direct hit), strong radio signal, static discharge... There are hundreds of reasons why it might fail.

We wont kick ur nuts just because we dont agree dont worry.

The bms can "hng" in the same way as an inverter can. Mine blow Up last year in shorted state. The mosfets of some reason was shorted. Same happens with amplifiers and other Gear i have repaired.

Is it worth taking a chance ? I dont think so. I have a Main fuse for everything and There is a reason houses or any other systems do have Main fuses before the electronics.


I dont have any Links to scienimtific pappers and have never thought that should have been needed. I have personally seen it happen. Perhaps not on bms as such but on electronics.

What do a fuse cost? Most likely fraction of what rest of the Gear cost

Yes ... this is a very sensible idea ... main differences between li-Fe and li-ion is that Fe weigh about twice as much for the same storage ... not an issue for static applications ....

Also Fe life cycles are not influenced by how high you charge ... with Li-ion it's paramount to keep top voltage low .. charging to 4.2V gives a life of 400 cycles ... charging to 3.93V gives 3,000 cycles .. Li-fe doesn't care about top voltage and has a naturally high life expectancy of over 2,000 ...

New cells and battery prices continue to drop(slightly) .. around 6WHrs/$ now ... that's equivalent to a price of $1 for a 2200mAHr 18650 cell new including delivery ... ( prices for buying over 100 cells )

This makes recycling less attractive unless you have a good supplier and like pulling things apart.

ozz93666 said:

Yes ... this is a very sensible idea ... main differences between li-Fe and li-ion is that Fe weigh about twice as much for the same storage ... not an issue for static applications ....
Also Fe life cycles are not influenced by how high you charge ... Li-fe doesn't care about top voltage and has a naturally high life expectancy of over 2,000 ...
New cells and battery prices continue to drop(slightly) .. around 6WHrs/$ now ... that's equivalent to a price of $1 for a 2200mAHr 18650 cell new including delivery ... ( prices for buying over 100 cells )
This makes recycling less attractive unless you have a good supplier and like pulling things apart.

Click to expand...

I have to disagree with the top voltage comment I am afraid . In fact the plate voltage should not exceed approximately 3.6 Volts and this is pretty universally stated by all manufacturers I have come across. Exceeding this voltage damages the cell . BMS's suited to LiFePo4 or LFP are set to that top voltage . One important point to store is that the BMS you use is set to LFP top voltage .It is easy to order the wrong one suitable for Lithium ion.

Depth of discharge is the main influence on cycles 80% discharge might give 4000 cycles and 30% will give 12000.(Manufacturers claim)

The charge current also influences cycle life and lesser lasts longer.

On the cost ,its about USD$ 45 for a 50Ah cell and I have been offered 200 AH cells for USD130.
I only have experience with the 50AH cell 3,2V atm so thats 160 watt hrs for USD$45 or 2.80c/Wh if my maths are right.
The larger cells are my pathway forward and that cost is down near 640 watt hrs for USD$130 or 0.49 cents per Wh.
I'm not used to calculating it out in watt hour cost. Somebody will correct that I am sure.

On the weight side a 50AH cell weighs 1.3Kilos . How does that compare to a similar 18650 capacity/ weight?

Agree on the top voltage for LiFePo4, I run my banks at max 3.45 or so per cell = full voltage.

Re failed MOSFETs, I've seen lots of failed short circuit ones, mainly in power supplies. Test a spare one by putting >20V on the gate eg 40V (via 1k resistor), see what happens ;-)
Ironically they're tolerant to voltage spikes across D-S & work like a zener diode mostly there.
Also seen plenty where they are really "open circuit" too eg the body blown apart, literally legs loose.
What happens depends on the energy source when they fail.
There's pics somewhere on this site of blown FETs from inside an inverter.

tytower said:

ozz93666 said:

Yes ... this is a very sensible idea ... main differences between li-Fe and li-ion is that Fe weigh about twice as much for the same storage ... not an issue for static applications ....
Also Fe life cycles are not influenced by how high you charge ... Li-fe doesn't care about top voltage and has a naturally high life expectancy of over 2,000 ...
New cells and battery prices continue to drop(slightly) .. around 6WHrs/$ now ... that's equivalent to a price of $1 for a 2200mAHr 18650 cell new including delivery ... ( prices for buying over 100 cells )
This makes recycling less attractive unless you have a good supplier and like pulling things apart.

Click to expand...

I have to disagree with the top voltage comment I am afraid . In fact the plate voltage should not exceed approximately 3.6 Volts and this is pretty universally stated by all manufacturers I have come across. Exceeding this voltage damages the cell . BMS's suited to LiFePo4 or LFP are set to that top voltage . One important point to store is that the BMS you use is set to LFP top voltage .It is easy to order the wrong one suitable for Lithium ion.

...

Click to expand...

I think you misunderstood that. ozz was not referring to exceeding the maximum voltage, but there being no real gain to under charging LiFe batteries, as can be had with normal Li-Ion batteries. By reducing the full charge voltage of a Li-Ion battery you can drastically increase the number of charge cycles the cell will have.

Over the last 1.5yrs, I bought 5,400 18650 cells in 20 different purchases to build a 780ah @ 48v battery bank. These were mostly used cells at the 85% original capacity / 5yrs old / low cycles. 1/3 were not so good price and 2/3 were good price (I got more knowledgeable over time). Overall I paid between $1.70 - $2.00 per cell. 5400 * $1.85 = $9,990.00

Looking on ebay just now I see things like Lot of 16 pcs 3.2v 86Ah 90Ah LiFePO4 prismatic Lithium Li-ion battery for $1,000. To build a 780ah @ 48v battery I would need 8.7 (8.7 * 90ah = 783ah) sets of 16. So 8.7 * $1,000 = $8700.00

So prices do seem competative BUT.... I have yet to find a source of TRUSTED LifePO4 Prismatics for the ebay price I mention above. In the 18650 world, I had several experiences of complete misrepresentation. For example, an advertised 90ah battery might only have 60ah once you get it.

If you look at a more mainstream site - I see 100ah for $155 + shipping. Say 100ah for $165. To build that 780ah @ 48v battery I would need 16 of these * 8 = 128 of them. 128 * $165 = $21,120.00. That's new of course and 18650's were 85% lower cycle.

I'd love to do LifePO4 as processing 5,000+ 18650 cells took me over a year part-time - but the key thing for me is reliable purchase to know if they are new or 90% etc before I could make the leap. And I'm not about to risk $5000 or $10000 dollars on a single purchase, so it has to be a source or sources where you can get equivalent, reliable prismatic over a year or 2 so that when you space out the purchases you can still get equivalent matched cells... and in prismatic case, since there are *fewer* cells you can't distribute them to mix manufactures as easily as you do with 18650.

There are pros/cons in either direction as far as I can tell