Testing a Shorai LiFePO4 on solar!

[PNjunction]

PN just like lead acid batteries, especially AGM, for LiPo or any of the lithium batteries is to measure and keep track of the Internal Resistance from the start. Most of the better Smart Chargers made for Lithium will do that for you and I assume yours does?

That's what Optimate-lithium claims, although there is no user-accessible output on internal resistance measurements. Along with other sanity checks, it looks for dropped / shorted cells etc during the charge process and periodically stops and tests them for voltage retention and I suppose IR before issuing a pass. I liked it for the fact that it had more than just voltage smarts, but also looked at internal IR.

I have an IR checker for PB, (ac waveform type, not a load test) and I'm NOT going to use that for any IR measurements. Not sure the LiFepo4 will take kindly to that type of IR checker.

Thanks for the tips - although I have adjusted the voltages downward for Shorai's LiFepo4, instead of the higher Li/cobalt values. The Shorai charger has a convenient store mode which discharges (or brings up if need be) a pack to about 80-90 percent and then shuts off and monitors.

To simulate a typical solar charge controller sitting at 14.4v for a few hours waiting for the sun to drop, I used a gel-specific battery tender that rises to only 14.4v max, (which surprised me since I expected 14.2v max) but doesn't seem to drop to float for about 6 hours, and no short-term damage seems evident, although I'm sure that isn't good for cycle life on a daily basis. An agm-specific battery tender will rise to 14.6v and does drop to float quickly, but I see no need to hit 14.6v daily.

Either one of my Morningstar or Xantrex C12 controllers do an ok job of charging, although I'm going to look into one that just stops once they reach the target voltage. Maybe the Genasun's do that, I'll have to check.

Right now I am the one doing the charge cutoff since my stuff is portable and I keep an eye on terminal voltages. I'm really enjoying the flat charge/discharge curve, and slowly losing my obsession over obtaining 100% charge which is not really necessary. Quite convenient for solar, but it doesn't relieve one from actually planning their power-budget to do it right. And at this point with LiFepo4, that's expen$$ive.

[Sunking]

PN there is a lot of chatter about the complexity of charging Lithium batteries ,ut in fact it is a very easy algorithm. Problem is non of the solar chargers are capable of doing it.

Basically it is a Precision Float or constant voltage charge with current limit and taper. For example a 3 Cell LiPo uses 14.4 volts and you limit current to 1C or less. As the state of charge or the battery approaches to equalize with the charger voltage, charge current starts to taper off or go down What is important is that the charge terminates once the charge current tapers down to 3 to 5% of the battery C AH rating. Hopefully by now you see Solar Chargers are not capable of limiting current or terminating a charge based on a voltage. That is the catch.

Only way to really limit current on most solar charge controllers is by using the panel wattage design. So for example if you need a 10 amp maximum current and have a 3 cell LiPo using a PWM controller you are looking for a 170 to 180 watt panel. However you are SOL disconnecting or terminating the charge when the charge current tapers off to 50 to 80 ma.

One thing I thought of but have not tried it is playing around with the Bulk, Absorb, and Float voltages on a controller that allows you to program the voltages. Set Bulk/Absorb to 14.4 volts, and float to say 13 volts. Once the battery reaches 14.4 volts the voltage is folded back to 13 thus shutting off the charge. Problem is what would th echarge current be once it reaches 14.4 volts. If still at 1C or near that the battery would not be fully charged.

Manufactures can easily make a solar controller for lithium, it is just a matter of demand and for now there is not any because Lithium prices are not low enough yet to justify commercial use for solar applications. Hang in there.

[PNjunction]

Yep - I'm looking for complexity when it really is very simple compared to Pb.

My Xantrex C12 controller will only sit in absorb for 1 hour before going to float no matter what, so that's easy to set to just 14.4v absorb, and then fall back to float, which I can adjust to more or less emulate a charge shutoff. So with the Xantrex, I'm guaranteed only 1 hour max of 14.4v (assuming my solar insolation actually allow me to reach that point). Still, I'd prefer an immediate shutoff.

Thing is, I know I've introduced too many variables to come to any conclusions, other than to just use the dang things within their specs and see.

One problem I see with solar is the dire need for an LVC cutoff for those that don't monitor closely. Let's say one accidentally heavily discharges their LiFepo4 bank beyond 80% DOD, or loses a cell, and then with the sun rapidly rising, hammers their low-soc bank, rather than allowing it to come up gently before reaching say 12.8v. Not good.

I'm hanging in there, but trust me, I'm not throwing away my TPPL agm's.

[Mike90250]

.... and then with the sun rapidly rising, hammers their low-soc bank, rather than allowing it to come up gently before reaching say 12.8v. ...

sun rapidly rising HA, I have about a 2 hour window from 20W PV input, to 1,000w in the mornings. By then (if I can keep the wife away from the electric toaster) the batteries have been getting a taste of a charge, and they should not be so "flat", by the time the sun input really gets cranking.

[PNjunction]

I just realized I got all caught up in bench-racing.

From a neighoring pacific-rim bike forum, (caution - adults only) I found out from the manufacturer that my Shorai cells are indeed well balanced from the factory as designed from a 10-90% DOD verified with my own tests. Check. Well above and below that (I took them to 14.75v (testing only, NOT recommended) for about half an hour today), and they are STILL balanced for the most part, although there is an observable gap now operating well into the edge-case knees. One minute on a simple Hyperion EOS pack sentry quickly balanced them to 0.005v gap. The store mode in the BMS01 charger did much the same.

Essentially, I have NO qualms about two-terminal charging these things, and operating them within a sane 10-80% DOD, A balancing charger is nice to have and I use either of the Optimate or Shorai. Under normal use most of the time, I'll be using my simple pwm solar charge controllers set to 14.2 or 14.4 and call it a day and perhaps check on balance every few months or so. Keeps me from wearing out the balance port, and is really unnecessary for daily use since these cells are matched at the factory. This whole balance thing is under control.

I also have to remember that my beloved Odyssey AGM's don't come with ANY nanny-protection bms, and for longest life just treat them right within spec! You don't want to under or overcharge these babies either, so I'm used to taking quality care of my batteries. Crucially, the Shorai's pass the over-the-counter *physical* availability test so nerds like myself can swing voltmeter leads over them before taking them home.

The biggest problem with the Shorai's is that they are now making me look at building my own 4S high-capacity packs with CALB lifepo4's. Another day perhaps.

[PNjunction]

Morningstar Prostar 15 notes

Interesting - With Fluke 87V attached to the Shorai terminals, watched the prostar with metering differ much more than it did with Pb. I'm not faulting morningstar for my use of a chemistry that the controller wasn't designed for however.

During bulk, my Shorai was reading 13.8v, whereas the Prostar was reading 14.2v. It wasn't until the start of absorb did the Shorai terminal voltage meet the prostar display voltage, although obviously it's never going to be as accurate as the Fluke. But when set to flooded, it DID eventually settle at 14.3v with very little current flowing. It never dropped below 14.3v since I wasn't patient enough for the Prostar to figure things out. I don't want it sitting fully charged at high voltage longer than necessary.

TEMP COMP! It also occured to me that when set for 14.4v "flooded" on the prostar, if taken into colder weather, temp comp will actually raise it above that. Soooo, I'd have to think that for a solar charge controller, perhaps 14.0 to 14.2v (much like a real gel setting) would be appropriate but for different chemistry reasons of course.

Ok, so yeah, backyard fun can be accomplished with lead-based solar charge controllers if you set your voltages appropriately, and take into account temp-comp voltage swings if your controller has that feature. I'd definitely doublecheck with an accurate meter what the REAL voltage setpoint is if you are using el-cheapo stuff.

In a fixed / serious installation, I'd definitely want a controller specifically designed for LiFepo4.

[inetdog]

Interesting - With Fluke 87V attached to the Shorai terminals, watched the prostar with metering differ much more than it did with Pb. I'm not faulting morningstar for my use of a chemistry that the controller wasn't designed for however.

During bulk, my Shorai was reading 13.8v, whereas the Prostar was reading 14.2v. It wasn't until the start of absorb did the Shorai terminal voltage meet the prostar display voltage, although obviously it's never going to be as accurate as the Fluke. But when set to flooded, it DID eventually settle at 14.3v with very little current flowing. It never dropped below 14.3v since I wasn't patient enough for the Prostar to figure things out. I don't want it sitting fully charged at high voltage longer than necessary.

TEMP COMP! It also occured to me that when set for 14.4v "flooded" on the prostar, if taken into colder weather, temp comp will actually raise it above that. Soooo, I'd have to think that for a solar charge controller, perhaps 14.0 to 14.2v (much like a real gel setting) would be appropriate but for different chemistry reasons of course.

Ok, so yeah, backyard fun can be accomplished with lead-based solar charge controllers if you set your voltages appropriately, and take into account temp-comp voltage swings if your controller has that feature. I'd definitely doublecheck with an accurate meter what the REAL voltage setpoint is if you are using el-cheapo stuff.

In a fixed / serious installation, I'd definitely want a controller specifically designed for LiFepo4.

Some CCs do display the "temperature corrected equivalent" voltage instead of the actual voltage. I guess they think it will save you the trouble of doing your own conversion to see what the actual SOC of the battery is or what stage the CC is in.

Also, some CCs with remote voltage sensing only null out the drop in one of the two battery leads (hiss, boo.)

[PNjunction]

Still a starter battery!

Heh, just because I can doesn't mean I should.

At the end of the day, these Shorai batteries are intended to be *starter* batteries, and are definitely the least bang for your buck when you do continual low-current discharges. I knew that before going in, and in fact there is a sticker on the bottom showing the watt-hour and Ah rating. Typically about 1/3 of the nebulous "Pbeq" used for starting comparisons. For bikers, that means listening to tunes for a long session while the bike is off is nearly a guaranteed recipe for quick complete discharge. This applies to ALL typical LiFepo4 powersports batteries.

While it is amazing that what I hold in my hands can turn over a bike, even a car, or perhaps even the backup diesel emergency genset in my office building, at the other end of the scale, it may only provide a handful of low-current charges to my tablet. Doesn't matter if it is prismatic, pouches, or cylindrical cells designed for high-rate discharge.

RESPECT - while all batteries pose a danger if abused, while the LiFepo4 is certainly much safer than LiCoo2 (older "cobalt" designs), it still has enough instantaneous power even at a very low SOC, that only a FOOL would randomly connect these without using wiring harness fusing despite their tiny size and weight. Note that fear is different from respect, and fear will get you into trouble. Respect these things by using / wiring in common safety measures.

[PNjunction]

Parallel is a bad idea

Learned an interesting lesson that basically applies to all chemistries, not just lifepo4.

I have two LFX-14's that on their own, maintain balance just fine over repeated standard cv charging cycling especially in my non-starter application.

However, when I parallelled them together to get more capacity, at the end of a cycling run and pulling them apart, one of them was totally out of balance with itself by about 200mv! Not good.

What the battery gurus said was true - when building a bank of 2-3v cells, go parallel first to build up capacity, THEN go series to get your voltage to make balance easier.

In the Shorai's case I thought about what I did by putting two internally series-connected cells in parallel with another series-connected battery and it became apparent that the cells on each end are now freaks having "plates" if you will, twice as large on one side than all the others. No wonder balance became an issue.

Even the Shorai charger balked at rebalancing this wide of a spread, so I used my Hyperion EOS pack sentry to discharge them all together. Maybe the "store" mode of the Shorai charger would have done the same when the charge mode balked. Either way, the LFX-14's are back to being single again, and so far, no major balance issues.

Even if I go back to lead, I'll be certain to parallel cells first to build up capacity, and then do the series connections to get the voltage.