Have Lithium Ion batteries come of age for off-grid systems?

karrak said:....................................I wonder what sort of protection the modern commercial MPPT controllers give one, especially the 150V and higher units. The thought of 150Vdc + at the output terminals of the MPPT controller does not appeal to me.
Simon

They should be robust enough to not short the PV input to the battery output, there's a transformer and switching FET's in the middle, and without it functioning well, nothing would get through, unless a catastrophic event happened. (full of conductive slag from a lightning strike)
But they still need voltage surge protection, especially on the PV input side.  Either Delta or Midnight (or both) Surge Protection.

BB. said:  I am not sure that 150 volt and less charge controllers use transformer isolation. Typical buck switcher is a "switch" (FET/Transistor) between the the array and an inductor (energy storage device)--The inductor (more or less) connects to the battery + terminal.    -Bill

Years of mil and space design, and I forget about the shortcuts !

Interesting article.

Until they can get the cost equal to or less than my FLAs' with the same reliability and capacity demonstrated over one battery lifecycle, I wouldn't risk making the technology change on a battery bank as large as ours $$$.

Dave's point about the importance of same manufacturer networked communicable inverter, cc, ags and bridge components.

Looking at say 8 years until li-ion's are economically mainstream? Sooner?

Agree Marc.

Jimmyaz - Ok, let's review the basics from a single-cell standpoint then.  You can refine it from there and do the math for your desired series connected voltages.  I'll divvy this up into several messages to keep things from going off the rails.

First - know the differences between your lithium chemistries:

LFP, aka LiFeP04 is the *only* lithium chemistry that is a "nominal 3.2v" cell.  Don't charge with voltages any higher than 3.6v.

ANY OTHER lithium chemistry, such as those from the Tesla pack or those commonly found in laptops which are NOT lfp, are "nominal 3.6v" cells.  Don't charge any higher than 4.2v per cell.

Keep your lithium chemistry voltages in mind!!  LFP = 3.2v nominal, 3.6v max.  If you don't keep this in mind, you'll overcharge and burn up an LFP, and merely undercharge a non-lfp cell.

LFP has discharge rates of about 1C maximum.  3C in short bursts.  Very suitable for our solar housebanks, which when sized properly for reasonable discharge rates, are about .05C or lower max!  Basically LFP cells laugh at our application, being tickled by very slow charge and discharge rates, provided of course one is making a housebank of suitable capacity.

Non-lfp cells, can usually take up to about 10C recharge!  Suitable for vehicular charging / regenerative braking and the like.  Sounds great right?  Not really - for us.

Using / charging Tesla packs has already been discussed elsewhere, such as on the DIYELECTRICCAR forum:

http://www.diyelectriccar.com/forums/forumdisplay.php/batteries-and-charging-35.html

BUT, the problem is that while yes, you CAN use these, as long as your voltages are correct for this NON-LFP chemistry, is the fact that under solar housebank use, which never needs a 10C discharge / recharge capability (not to mention USED cells in most instances), only means you'll be paying for a high-rate current capability you'll *never use*.

Thus, LFP with it's lower nominal voltage per cell (makes it easier to construct multi-cell batteries to your desired voltage), along with it's already high-enough discharge / recharge current capability, makes more financial and DIY simplicity for a solar housebank.

TBC..

Full charge vs 90% charge:

An initial FULL charge is needed upon receipt to make sure that all the materials inside are truly active.  If you don't do this, even if charged somewhat from the factory, you can create inactive areas, and thus "hot spots" can form.

NORMAL USE:  don't always charge to full!

We've seen that a full charge is an absorb voltage anywhere in the window from 3.45 to 3.60v, but it is the end-amps, or absorb current that determines the full state of charge.  So voltage is not a good determination of full charge, but end-amps is - provided you are between the 3.45 to 3.6v window.

Thus, to charge to say 90%, don't allow for a full absorb.  It's as simple as that.  You can merely cut off charge when a voltage is reached.  Or you can allow the cell to reach full voltage, but stop absorb half-way.  Heck, you can even stop charge well before 3.45v, but of course your overall capacity is reduced.  You choose your stopping point - based on the knowledge of 3.45 to 3.6v and an end-amp current of .05C is the generic value for a full charge.

The holy grail voltage:  3.5v

A lot of newcomers to LFP are in search of the holy-grail value for the CV voltage, thinking that voltage is the determination of full charge values.  But now you know that it can vary between 3.45 and 3.6v, with the absorb current being the REAL value.

However, why do we see all over the place to keep your overall CV voltage low, like 3.45 to 3.5v max??

It is because we are dealing with series-connected cells, each with their own slight differences in overall capacity and internal resistance, and we don't want our neighboring cells to go sky-high when doing a two-terminal (pack level) charge.

Assuming we bought from reputable sources, new and not used trash, the cell capacities and internal resistance should be close enough for use to merely keep our CV voltage in the low-to-mid area of the 3.45 to 3.6v spec.

When you choose an overal low-to-mid CV voltage spec, the slight differences in cell characteristics will not be enough to drive one cell over 3.6v.

Thus, provided you have done an INITIAL individual cell charge prior to placing high quality / matched cells together, a CV charge of 3.5v max gives you a little leeway.

Ie, if you have a 4S set of cells to make a 12v battery, have done your initial individual charge, for normal use, set your CV to no higher than 3.5/cell, or 14v overall.

Because you set your CV voltage no higher than mid level, (3.5v per cell) during charge you may see something like this at the end when you reach the canonical .05C current drop in absorb:

Cell 1: 3.48v
Cell 2: 3.43v
Cell 3: 3.52v  << You still are under 3.6 and have a bit of headroom.  This is also the lowest capacity / highest resistance cell.
Cell 4: 3.49v

This is perfectly fine!

How low can I go - my LVD or low-voltage disconnect setting ....

Generally, you don't want to take any LFP cell below 2.8v resting which equates to a 100% discharge for the most part.

In practice, that means using a bit more conservative values under load.

Generally, about a 90% discharge under the typical housebank currents from an appropriately sized battery means not taking the cell below 3.1v under load.  3.2v is a nice conservative value.

For the 12v battery example above, that means an overall pack level LVD of 12.8v  (3.2v * 4) under load.  This is assuming you have quality cells that are reasonably matched for capacity and internal resistance, AND have had an initial full charge upon receipt.

HOWEVER - if you want to refine it somewhat, the better thing to do is this:

1) Discharge your bank at the normal level.  Watch the voltage of the cells.
2) When the **FIRST** cell reaches 3.1v, measure the overall voltage of the pack.
3) Set your programmable LVD to this overall pack voltage.

EX:
Cell 1: 3.18v
Cell 2: 3.20v
Cell 3: 3.10v   <<<  This is the trigger cell being the first to reach 3.10.
Cell 4: 3.12v

Add up the voltages.  3.18 + 3.20 + 3.10 + 3.12v = 12.6v LVD.

At this end of the discharge curve, there is not much left.  Choose your level of conservative values.  You could have watched for the first to be 3.2v for instance.

Keep your application in mind.  We are different from an EV, drawing huge currents just to get back up the driveway, where this simplistic LVD could lead to a cell-killing reversal.  This is far different from the "bottom balancing" technique so often discussed for EV's.   Because our current draw is very minimal compared to an appropriately sized bank, we have this luxury of time for the LVD to trip properly without allowing for a cell reversal using "pack level" techniques.

The other reminder is that we are assuming you are starting out with high-quality cells from a reputable vendor.  Anything else, like used trash cells from the dump or waste recycler requires a whole different strategy.

Obsession with top-level charge voltages:

We've kind of tried to steer clear of balancing issues to get to the meat of the basics.  So I'll keep this somewhat short.

A lot of DIY guys go nuts trying to keep their cell voltages during the end of charge at *exactly* the same voltage.  They freak out if there is a .001v difference between the cells.

In our usage, a difference of 0.1 (zero-point-one, aka 100mv) or less is acceptable to most manufacturers.  Of course, don't exceed 3.6v on any cell.  Normally quality cells, that have had a proper initial charge to full, end up much less, like 0.07v (zero-point-zero-7, or 70mv) difference or better.  AND it all depends on which part of the final charge curve you measure them at!

They are playing a detrimental game trying to make all their cell voltages line up exactly when reaching the CV pack voltage, which can NEVER be in their natural state due to slight cell differences in overall capacity and internal resistance.  Once again, we are assuming that each cell has been individually charged to full upon receipt, not a hodge-podge of individual cell SOC values slapped on a charger.

Oh, you can play this game, but TIME is going to hurt you in the end.  They never think about TIME.  What do I mean?

Let's put the bleeder-boards on the 12v battery as discussed above.  Assume something like 100ah cells, with only 500ma bleeder boards.

The user gets them from the distributor, and each cell as received in reality is thus:

Cell 1: 40% charged
Cell 2: 50% charged
Cell 3: 80% charged
Cell 4: 58% charged

"I'll just slap my bleeder boards on this thing and charge away".  Guess what happens:

Cell 3 reaches the high voltage limit of 3.6v and stays there waiting/bleeding forever for cell 1 to catch up.  Could take DAYS.  TIME is coming after you now.

After several days of bleeding, during normal use, all these cells are programmed to do the "lets make all the voltages line up exactly" dance.  Again, this takes TIME.  That is, time spent sitting at fully charged waiting for the others and the blinky-light dance begins.

Basically what I'm saying, is that for a simplistic LFP bank, made from *quality* components, you don't need to do this dance every cycle.  A difference of no more than .07v (70mv) between each cell during charge can be quite acceptable.  Of course, less is better, BUT take into account that your cells have slightly differing capacity and internal resistances which lead to their natural SOC voltage, which may differ a bit.  As long as you don't exceed 3.6v, you should be fine.

Any difference larger than 70 to 100mv means something else is wrong:
1) Cells were not individually charged to their full capacity initially.  (3.45 to 3.6v CC/CV w/end amps of .05C)
2) You were sold a low capacity / high resistance cell.
3) Your wiring infrastructure is poor - corroded/oxidized terminal connectors, loose wiring, OVER-torqued bolts etc.

The question to ask is that will my cells have a longer life if I spend TIME each cycle doing the blinky-light dance trying to make the voltages exactly the same, or will small and reasonable differentials in voltage, with an overall lower amount of TIME reaching my full-charge limit, be better in the long run?

It is the latter method that actually results in the longest lifetime.  But, there are a lot of blinky-light boards to sell to the unknowledgeable DIY'er to feel-good in the short term.  And of course to those who never want to deal with a proper *initial* charge to each cell upon receipt.

This is akin to a guy making a 48v bank of 12v batteries of unknown characteristics, slapping them all together, and wondering why one battery is gassing upon first charge.

So yeah - would I give an LFP diy bank to my Grandma?  Heck no.  To experienced battery guys on this forum?  Sure!

EV'ers and benchtop guys are in a different application than us, and have different needs.  So when discussing LFP, keep your solar-application in mind, otherwise you can end up in the weeds / tangents of discussion.

Ah, ok kind of a different arena, one-off scenarios.  Sure, anything could be done.

In the old days, it was ripping apart DeWalt tool packs for the A123 LFP cylindricals.  Then of course you also have crashed/salvage Nissan Leaf batteries that were the rage for awhile.  Now salvage Tesla stuff.

The keyword here is "salvage", where safety / performance / maintenance is unknown.  Keep your wallet shut, because the "salvage" guys purposely target the benchtop hobbiest, who don't know what a prismatic LFP, along with the intended application is all about.

Somebody else here might have more info for you - though to be sure see the EV forums for additional notes.