LiFePO4 charging - the overcharging question

[ptourin]

So if I'm understanding you about the chargers, we wish that we had big honking chargers that would put out a rated voltage and LOTS of current, and we'd simply set the charger for X voltage and Y current - but that's not what we've got <g>... So in the world of real chargers:

My charger is rated at 60A, and if my batteries are far down, the charger will provide its rated current but the voltage will be below rated voltage at the beginning because the charger "just can't do it all". As the batteries charge, the battery/charger voltage comes up to the rated voltage (I'm going to ignore wiring losses, etc. and assume that when connected, charger voltage = battery voltage). Once the charging gets to the point where the voltage is up to the rated charger voltage, we're maybe an hour into the charge cycle and we're at the end of the CC phase - now we move to the CV stage, where the voltage stays steady and current starts dropping. This could be anywhere from a short time to a few hours, depending on where we decide to cut off. Does this sound like a fair rough description?

The only thing about your first blue paragraph is that if I set up the monitor as described, I'll definitely move into the CV phase - my charge cutoff point is determined by when the current drops to my setpoint. Right? Is my thinking still on the right planet?

[avanti]

Well, I am far from an expert, but...
I don't think that the cliff in that chart tells us much about the battery. Surely it is the result of an algorithmic decision made by the charger when a target voltage was reached.

At the end of the day, voltage is a kind of pressure. current is a kind of flow, a battery is a kind of closed container, and a charger is a kind of pump. A charger is simply pumping a volume of electrons into the tank at a given pressure. The closer the pressure in the tank gets to the pressure of the pump, the slower the process will proceed. Basic physics suggests that this must be true of Li batteries just like any other. Using this analogy, the only difference between lead and lithium batteries is that latter has a bigger "opening" (i.e., has lower internal resistance), thus allowing higher currents. But, the shape of the curve pretty much has to be the same, and it has to asymptote.

[booster]

In the example they had CC/CV charging so the charger held the 100 amps of input until the voltage triggered transition point was hit. At that point the voltage was surging up at constant current. As soon as the voltage stopped climbing, the current plummeted because there was no more charging. This battery charged at 100 amps from start to finish for all practical purposes. Lead acid slow way down after 80%, with the current dropping quickly.

Here is a similar chart from Elite for their batteries at charge rates up to 3C (1320 amps in our size systems), and they charge at that rate right up to end of charging, as well as at lower rates. If you stop before the big voltage up tail on the profile (about 95%), you can charge at as much (as we can deliver) amps as you want to the end of charge, and also not have the batteries or anything else see elevated voltage, if I am interpreting the chart correctly.

[booster]

Originally Posted by ptourin

So if I'm understanding you about the chargers, we wish that we had big honking chargers that would put out a rated voltage and LOTS of current, and we'd simply set the charger for X voltage and Y current - but that's not what we've got <g>... So in the world of real chargers:

My charger is rated at 60A, and if my batteries are far down, the charger will provide its rated current but the voltage will be below rated voltage at the beginning because the charger "just can't do it all". As the batteries charge, the battery/charger voltage comes up to the rated voltage (I'm going to ignore wiring losses, etc. and assume that when connected, charger voltage = battery voltage). Once the charging gets to the point where the voltage is up to the rated charger voltage, we're maybe an hour into the charge cycle and we're at the end of the CC phase - now we move to the CV stage, where the voltage stays steady and current starts dropping. This could be anywhere from a short time to a few hours, depending on where we decide to cut off. Does this sound like a fair rough description?

The only thing about your first blue paragraph is that if I set up the monitor as described, I'll definitely move into the CV phase - my charge cutoff point is determined by when the current drops to my setpoint. Right? Is my thinking still on the right planet?

Yes your description of how the charger will work is how I also see it. Max current to voltage setpoint, then constant voltage with very fast drop in amps to ending point based on tail amps. I don't think the time to amps would be very long, or need to be. There would also not be any battery harm from stopping very soon. It appears the batteries really don't absorb much once the voltage stops climbing. You would be able tell by how much the drop after resting to evaluate how the time in CV affects the SOC.

[ptourin]

This is going too far! The Victron monitor has datalogging capability, and I've very successfully ignored it up to now - I can see I'm going to get forced to check it out <g>...

[ptourin]

There was just a question asked over on the LTV forum that I don't know how to answer: what constitutes a discharge/charge cycle? The person who asked is thinking about life cycles, of course.

The question is whether 100 cycles of charging from 60% SOC to 100% SOC would be about the same thing as 50 cycles of charging from 20% to 100%, in terms of life cycles - does a partial charge count as 1 full cycle, or as part of a cycle? In other words, does a cycle relate to how much you hit the battery chemistry with charging time, or does it relate to how many times you do a discharge/charge cycle, no matter how deep the discharge?

[Davydd]

That's a good question. I suspect one would have to look up the testing criteria that was done to determine estimated cycles. It certainly hasn't come from field use verification. The only RV field verification report I am aware of is from Technomadia but I don't recall any comment as to cycles over after their 3-1/2 year use.

[ptourin]

Yes, I don't recall ever seeing any discussion of this. It's usually in the opposite direction: "I'll be dead long before I use up 2000 cycles" <gr>...

I remember Technomadia comments about shorter lifetime than expected, but if I remember, that related to very high temperatures or some such thing - it certainly didn't address this particular question.

[booster]

I do recall seeing a depth of discharge vs cycle life for the lithium someplace, and it did similar to the lead acid batteries with longer life at lower discharges. I have no idea where it was, though.

I would think that partial charges would count as cycles, but I wonder is partials in the lower SOC area shorten life more than ones in the higher SOC area.

The life is most likely based on some lab based accelerated testing with no real relation to the RV world of hot/cold/bumps/varying charge rates and types of chargers, etc.

When thinking of the many cycles that lithium would do in theory, I start to wonder if we will get some early results from units with Voltstart and pets. The poster on the Yahoo board with a 200ah Zion says she can run the AC for 90 minutes off of full, then it needs to recharge for 40 minutes to run it 90 minutes again. You could be looking at 6 cycles a day maybe if hot and used all day. The total number of charge cycles could count up pretty quickly that way.

[eric1514]

On the Apple website they define a cycle as anytime the percentages of charging totals 100%. So if you run the battery down to 60% once ( needs 40% to bring to full) and then do it again and again each time recharging to 100% you will have used 1.2 cycles.

[ptourin]

That is, 3 charges of 1/3 capacity = 1 cycle. In your example, 3 charges of 0.4 capacity = 1.2 cycles. That's "option 1" - that the cycle count depends on the load on the battery chemistry, so to speak, and not on how many times you discharge an unknown amount and charge again. I like that very much!! <g>...

[avanti]

Originally Posted by eric1514

On the Apple website they define a cycle as anytime the percentages of charging totals 100%. So if you run the battery down to 60% once ( needs 40% to bring to full) and then do it again and again each time recharging to 100% you will have used 1.2 cycles.

I have no actual data, but I will say that the laws of physics almost guarantee that something like that has to be true. There just isn't going to be some magical threshold that causes the counter to tick up one "charge". Pretty clearly, all the talk of "# of charges" is just a rule of thumb. It has to be some continuous function of amps in and amps out.

BUT, the real question is whether all amps are equal. That is, does an amp in/out of an almost-full battery have the same quantitative effect on lifetime as an amp in/out of an almost-empty battery? On that, I have no intuition.

[eric1514]

Exactly. If you raise the SOC 20% 5 times that's 1 cycle although you might not have ever reached 100%.

[ptourin]

That seemed to me the way it had to work, thinking logically. But I'm always cautious about that sort of thinking - you can think real logically and come to a totally wrong conclusion because of lack of information <g>...

[booster]

Originally Posted by ptourin

That seemed to me the way it had to work, thinking logically. But I'm always cautious about that sort of thinking - you can think real logically and come to a totally wrong conclusion because of lack of information <g>...

I agree, and the lead acid batteries do have some non linearity to them. Most have about 1/2 the number of recharge cycles from 80% DOD to full than they do at 50% DOD, so you get more amp hours of cumulative use at 50%. I have no idea if lithium behaves similarly.

[ptourin]

Avanti's comment: "Are all amps equal?" <g>...
Booster, like you say, it does appear that flooded cells are non-linear - and we have had a lot of discussion for years concerning not discharging too deep if you want your batteries to last, because deep discharge damages the cells more.

I'd have to go back and do a bunch of reading, but isn't your impression that the lith phosphate papers say deep discharge doesn't damage the cell chemistry like with flooded cells? So that there's little non-linearity? The Apple statement from the article that's quoted in Wikipedia is about lith batteries, not batteries in general.

[booster]

Originally Posted by ptourin

Avanti's comment: "Are all amps equal?" <g>...
Booster, like you say, it does appear that flooded cells are non-linear - and we have had a lot of discussion for years concerning not discharging too deep if you want your batteries to last, because deep discharge damages the cells more.

I'd have to go back and do a bunch of reading, but isn't your impression that the lith phosphate papers say deep discharge doesn't damage the cell chemistry like with flooded cells? So that there's little non-linearity? The Apple statement from the article that's quoted in Wikipedia is about lith batteries, not batteries in general.

Yep, the lithium guys have said lots of things about deep discharge, including some that say you can use 100%, which we now know isn't likely, and that you can charge below zero. I wish I had copied the chart I saw, as IIRC, it did show an non linearity similar to lead acid. of course it, too, could be wrong. As has been said "we don't know what we don't know yet", I fear.

The lithium in Apple products is also a different chemistry than the RV lithium is.

[booster]

Found this one on an electric vehicle document so same chemistry. Seems to show non linearity if I am reading it right.

[ptourin]

I'm not sure I understand how to interpret the graph. Why does the voltage curve get lower when the %DOD is more conservative? And I don't understand the end voltages - I'd think that as you test nearer to 100%DOD your end voltage would get lower. This is interesting but confusing. If you say, "I'm going to test at 80% DOD and at 100% DOD, aren't you saying that you're going to test at a certain end voltage to get 80% and at a lower end voltage to get 100%???

[booster]

I don't think the chart is showing the voltage that the batteries are discharged to in each cycle. I do think the voltage they show is what the battery will recharge TO after the number of cycles shown on each charted line to the DOD shown in the lower scale.

After 1000 cycles of 100% discharge the batteries would recharge to 2.4v and be worn out

After 3000 cycles of 92% discharge the batteries would recharge to 2.3v

After 5000 cycles at 80% discharge the batteries would recharge to 2.2v

After 8000 cycles at 73% discharge the batteries would recharge to 2.4v

I think it would have been a much more understandable chart if they took all the lines to a fixed cutoff voltage (worn out) and fixed DOD percentages and then shown what the life in cycles was for those points. They still could have shown the curves to indicate the degradation rate, though, as it is very, very, non linear. They seem to hold performance very well, and then drop very quickly at the end of life period.

It is pretty interesting that the low DOD gives more capacity loss early on, it appears, compared to large DOD, but does last 8 time longer before failure.

The charts we are used to just chart DOD vs cycles based on dropping to 80% of capacity at recharge and do appear to be a lot more easily understood.