LiFePO4 charging - the overcharging question

[booster]

I have had a lot of similar questions as I continue to read about the lithiums and their care and feeding.

There seem to be some things that are a bit contradictory related to how they charge. We always hear that lithiums will accept as much current as you can throw at them, but I don't ever recall seeing a current vs charge voltage graph. The charging profiles I have seen seem to indicate the batteries can be charged with a charger that is set to the battery voltage that will give you get the state of charge you want. I think 100% charge is about 13.8v and 13.4v is the desired cutoff at 90% full as 100% is normally considered not good for the batteries. Lots of the profiles I have seen for vehicle stuff charged at the finish voltage they wanted. But almost all the RV stuff is recommending 14.6v to charge with, which is way above the finish voltage you would want to get for a 90% (or even 100%) state of charge. To compensate for that they put in a control system to stop the charging when the battery voltage gets to 13.4v or so, which assumes the batteries are able to pull down the charger to that voltage because it can't output enough to hold the voltage. My question is would the batteries charge as fast at 13.4v (same as finished voltage) as they would at 14.6v shutting off at 13.4v at the batteries. My guess is they probably would because that is what is happening anyway. So just why have the extra voltage on the charger which gives damage risk if it doesn't shut off when it should.

There probably is a reason for the 14.6v, but I haven't been able to find any explanation anywhere that makes sense. The only conceivable thing I can think of if that they set the voltage much higher than needed to overcome wiring voltage drops, so that they be sure they are getting enough voltage at the batteries to control charging accurately. Of course, the way around that is to have the voltage reference on the charge sources coming from the batteries themselves.

I had found some charge profiles a while ago that showed the CC/CV charging. I wish I had copied them, as I don't recall where they were. It didn't matter what charge voltage they were going to at the batteries (state of charge). The current would stay high until the CV step kicked in and then it would drop very quickly. This same thing happened on all finishing voltages so also for all states of charge. Those profiles would also never have the charger putting out a higher voltage than the finish charge voltage you want for you state of charge.

Perhaps someone can shed some light on it, but it doesn't make much sense to me.

 

[ptourin]

Are you thinking of this graph? Constant Current Constant Voltage (CCCV) Charging
It lays it out very clearly so it's helpful conceptually. 4.2V charging voltage - that'd be real high for a 4S battery...

Edit: I just found out where that graph came from:
http://pdfserv.maximintegrated.com/en/an/AN4169.pdf

Also (I'm on vacation, have time to cruise the web <g>...), I just ran into several pages from January 2016 by PowerStream that are interesting -

http://www.powerstream.com/LLLF.htm (charging discussion)
http://www.powerstream.com/lithium-phosphate-charge-voltage.htm (charge tests)

They cover a lot of the charging issues we've been tossing about. Their discussion puts nominal cell voltage at 3.2 and peak voltage at 3.65 (there's your 14.6 voltage for CCCV chargers!), and their discussion of conventional charging switches from CC to CV when the voltage rises to 3.65. They say that LiFePO4's have a large overcharge tolerance and can be safely overcharged up to 4.2 - that's way above the charging algorithms we're discussing.

Their charge tests show that reasonable charging starts at 3.3 and starts approaching 100% SOC at 3.4-3.5. They test up to 4.2.

They talk about ~3 hours to charge via conventional charging - 1 hr for the CC phase and another 2 hrs for the CV phase. They point out that you can force charge to 95% via CC with no damage. It's all interesting reading.

 

[booster]

That isn't the one that I was thinking about, as it shows charger voltage. In the one in the link, it shows charge voltage of 14.6v, which in our world is essentially impossible, I think, as the the charge sources don't have enough capacity to hold the voltage that high. Perhaps Davydd or someone else with lithium can fill us in on if they have ever seen 14.6v while the batteries are still charging, if they normally stop at 90% full. I would think that they would have to look at whatever controls they have to tell if they are still charging or not, unless everything is full shutoff on the charging. This is the major part of the paradox I was referring too. If you are stopping charging at 90% full, which seems to be in the 13.8-14.0v range depending on who you believe, you would never see any voltage higher than that unless you had a huge charger that could generate more amps than the battery could accept. Even at only 5C acceptance, which is tiny for lithium, it appears, Davydd's 800ah would accept 4000amps. What I have seen no information on is how much voltage it takes to get the batteries to accept the amps when you don't go to the 100% full battery voltage where the acceptance current rapidly drops off. This is what was different in the graphs I mentioned, as there was no charger voltage shown, only the charge/battery voltage, which climbed as the battery charged to whatever SOC they wanted and then flatlined (I think they were using 3.45v/cell or 13.8v for a 12v system). To me, that indicated they were using a 13.8v CC/CV charger. That was also the impression I got on many of the electric vehicle sites.

I don't know what to make of the Powerstream stuff. They are huge into fast charging on most battery types, I think, but the charts have to be for small batteries if the charger can hold voltage. They are also taking them all to the 3.65v/cell (or higher) or 14.6v so 100% full, which we are hearing isn't great for the batteries. My consideration was for if you were putting together a system that you wanted to stop charging at 90%, which seems to be getting pretty common. In that case, it might be much easier to use a CC/CV at 13.8v and not have to worry about a separate full cutoff at full system, as the charger would shut off at 13.8v or 90% full. Many of the chargers will restart again when the voltage drops, so it would be just like the separate control system in that way.

The 14.6v would certainly be because that is what it takes to get to 100% full, if that voltage were settable, folks should be able to dial right in on their desired max SOC, right at the charger (maybe). This would certainly make the "drop in" switch to lithium a whole lot easier.

 

[Davydd]

I have taken lots of pictures of my Silverleaf screen over the past year and went through them. It seems the individual battery cells are generally around 3.3 to 3.4 volts when seemingly fully charged. Each cell varies in a small range. The total average for the 16 cells for 12v readout on the screen is usually 13.3 to 13.4 volts and rises up to about 13.6 when being charged. The average voltage drops as the SOC goes down when parked and not being charged.

When we first got our ARV it would charge up to 99-100% SOC but the alternator amp input from over 200 amps would gradually slow down to 0 like stepping on a brake starting at about 90% SOC when driving. There is a steady draw from the inverter and other things so it would bounce in that it would charge and not charge to maintain 99%. ARV reprogrammed it last fall and now it charges up to 99-100% and charging stops until the SOC drops to 90%. Then charging starts again whether it is solar, alternator or shore power.

In reading some of this discussion I wonder if 99-100% SOC readout is truly that. I wonder if there is not a safety factor at the top end similar to what ARV built in cutting off the bottom 20% SOC. I understand their measures they implement to to prevent battery depletion with a total battery shutdown cutoff at 20% SOC which does not depend on Autogen startup to prevent. Autogen can be programmed at any SOC to start to prevent ever getting to 20%.

420w solar in February under ideal conditions seems to charge at a rate of about 10-14 amps. The sun angle makes it almost negligible and disappointing, IMO, compared to the overwhelming second alternator input. At home I am always plugged in and parked under trees so I haven't as yet determined performance of solar in the summer. I mentioned before I could at home plug into shore power and run my engine and get an input of over 300 amp input charging simultaneously. There is a 4/0 wire from alternator to battery and inverter to battery. That is a nearly 1/2" dia. stranded copper wire.

Those are observations. I don't know how all this works and interfaces with Silverleaf other than I know the alternator works with a Balmar regulator and the battery bank is controlled by Elite Power Solutions BMS and there is computer programming involved in those systems.

 

[booster]

I have taken lots of pictures of my Silverleaf screen over the past year and went through them. It seems the individual battery cells are generally around 3.3 to 3.4 volts when seemingly fully charged. Each cell varies in a small range. The total average for the 16 cells for 12v readout on the screen is usually 13.3 to 13.4 volts and rises up to about 13.6 when being charged. The average voltage drops as the SOC goes down when parked and not being charged.

Thanks, davydd, that is exactly the information I was looking for. It does seem to indicate, as it appeared from some of my reading, that unless you want to go to 100% state of charge, having a charger set at higher than about 13.8v is mute as it will never get to the constant voltage stage do to battery acceptance rates. 13.8v would probably be about 90% SOC so the 13.6 davydd sees would probably be slightly less than that, which makes sense.

This fills in another blank in the DIY drop in lithium possibility. It totally removes some of the controls, if you chose to, I think. If your chargers are CC/CV set at 13.6-13.8v, you will stop charging at the SOC you want. Just set the CV timer to a very low time. It also takes more worry about using the engine alternator to charge out of the picture. It doesn't really matter if your alternator is running at 13.8 or 14.8v when the batteries are being charged, as they will drag it down to whatever they are at anyway. For those of us that wouldn't have a centralized control system, all you would need to do is watch a voltmeter on the batteries and turn off the separator when you got to your 13.8v if that's what you want.

The BMS would then just need an high and low voltage emergency protection, not charge control. A low temp cutout in the BMS would also be good.

 

[Davydd]

I looked at all my photos and I have plenty as I was trying to document what was going on. I have seen individual cells up around 3.65v and that would be 14.6v with four in 12v series. If it got higher then the alert message would be pack not normal and action would be taken. This photo is about as typical as I see. This was last December when I was monitoring how often the battery heater came on. I was plugged into shore power at the time so had to be somewhere above 90% SOC. The "Heater" button with the wings filled with the green color means the heater is set to come on. When the flame comes on to the right it means the 10a 12v electric heating pad is actually heating. I found it did not take much heat to keep the batteries warm in below freezing temperatures as the heat was off way more than on. Also, it is programmed to keep the cells above 41 degrees F and not just freezing. Above 41 I understand is an optimum range. At the other end, BTW, when it was 96 in Palm Springs the battery temperature was about 84. Day/night temperature fluctuations and time lag is the reason. This is something I won't be worrying about as long as I continue to pursue following the weather and not using air conditioning in my ventures.

 

[booster]

When you are on shore power, does your display show if the charging sources are connected, or not? It would be very interesting to see what the cell voltages look like at the disconnect point.

 

[ptourin]

Thanks as always guys - this last few posts have been very interesting. It's appearing to me that even in my very simple system, the Victron monitor gives me lots of flexibility for charge control. My charging voltages will be high by the nature of the system - up to ~14.2 off the alternator and 14.6 off the charger. The Victron will provide several things:

SOC definition via voltage, tail current and a time value - I can define 100% SOC as reaching 13.4V and dropping to 10% tail current for 4 minutes, for example - that seems like a conservative config. Having the Tail Current param allows the use of simple CCCV chargers like the Progressive Dynamics unit that was recommended to me.

Now - I can trigger the monitor control relay using just SOC or both SOC and voltage. Let's suppose I use both. Their logic is interesting. Assume I'm going to use both: when *EITHER* SOC *OR* voltage drop below their setpoints, the relay will close and allow charging; when *BOTH* SOC *AND* voltage go over their setpoints, the relay will open and end charging. The former seems not to take much thought - maybe we start charging any time we're below 50% or 12.5 - whatever...

The latter takes a bit more thought. We could simply say that any time we reach 90% SOC, or 95%, or 100%, we stop charging - that'd be simple. We could also try to fine tune it - perhaps we might say stop charging at 100% SOC *AND* 14.1V (or 14.2, or really stretch a point and use 14.5 <g>...) - you see what I'm getting at - lots of programming flexibility, but it does assume that you have some vague idea what to program. That's what spurred this last set of questions - my RV is out of storage next week and I'll be setting this up - and a friend has already done a similar setup and is in the config and testing phase - so we've been talking about configuration.

The oddest thing about the Victron is that they sell a battery temp monitor and you can use it to set off an alarm or trigger the control relay - I wanted to use it to prevent charging when the temp gets down near freezing - but the logic apparently doesn't work. As an example, it might be: if SOC drops below 50% OR voltage drops below 12.5 OR temp goes over 40 degrees F, close the relay and allow charging - this works fine. BUT if SOC goes to 100% AND voltage goes up to 14.1 AND temp goes down to 34 degrees, open the relay. OOPS - NO GOOD! - if the temp is AND'ed, I'll never stop charging until I drive way way north <g>...

It's hard for me to believe that the monitor really works this way, though that's what one of Victron North America's tech guys told me. I emailed him with a detailed description of the above and he said he'd check with Engineering - haven't heard back yet. But if I can also do temp protection, that little monitor is flexible indeed for a simple battery system.

 

[Davydd]

When you are on shore power, does your display show if the charging sources are connected, or not? It would be very interesting to see what the cell voltages look like at the disconnect point.

The Silverleaf Home Screen gives you the the house battery and chassis battery voltage and what status you are such as Shore Power connected or Inverter connected for power.

The Silverleaf DC Power screen gives you a constantly updating snap shot of what is going on.

The average battery temperature

The state of charge of the battery pack in %

The voltage of the battery pack

How many amps are discharging from the battery or charging the battery (+ or -)

Whether the charger is enabled or disabled. You can turn the inverter on or off at the AC Power touch screen.

That Li-ion Battery Status screen is accessible tapping on a battery icon on the DC Power screen or via the Features button which gives all kinds of sub screens of information and Silverleaf customization. At first it is all kind of overwhelming.

I have an iPhone app called RVCair that can mirror this information. I often use it at my dash just to see what the charge rate is and how fast the alternator will recharge the batteries while driving. It is connected to my internal wifi as is the Silverleaf monitor.

 

[gerrym51]

are you guys saying that to guarantee protection at both ends of the range

for charging they never go to full charge?

 

[booster]

If you are referring to keeping the batteries between less than full and more than empty, yes we are. Most of what I have seen lately repeats what we have heard for a long time on low end, with not going below about 10 or 20% SOC. Lots of places are also starting to hedge on the top end also. I kind of started by saying the lithiums don't like to be full all the time, and has kind of morphed into stopping charging at 90-95% SOC. More and more are also now saying no float or continuous voltage on them once they are fully charged to whatever level you choose. I am sure the rules for lithium will continue to evolve for some time to come.

 

[ptourin]

That brings me into a question about the definition of SOC <s>... It seems practical to say I've reached 100% SOC when voltage climbs to 13.4, and tail current drops to ~C/10 - that's the resting charged voltage, so it'd be reasonable to start from there in tracking discharge under load. BUT - my understanding is that if I really stop charging there, the batteries won't remain there after they settle. So maybe I bring them up to 14.0 and then stop the charging. And at rest they'll drift down to... something... assuming that I'm right that they pick up a surface charge during the charging cycle and lose it a few hours or a half day after. That's a piece I don't understand yet - I assume that if I bring the batteries to 14.6 and stop charging when current goes to C/20, the batteries will settle to around 13.4V. Suppose that instead, I stop at 13.6 and C/20 instead? Will the batteries settle to a lower voltage than 13.4?

So I still find myself not sure of how I want to configure for charging - even how I define 100% SOC isn't absolutely clear to me. At the moment, I think i want to set it as above, because when I load the batteries I want the monitor to think 100% until the voltage starts to drop below 13.4, the resting full charge voltage. But I'd think I'd want to charge to some higher voltage since I know the batteries will settle - if we end up thinking it's not good to push them too high on each charge cycle, maybe not 14.6 and very low current, but maybe 14.0 or thereabouts.

 

[booster]

That is a very similar thought path as the one that got me thinking of why the high charge voltage is chosen when they still wanted you to stop charging based on voltage, and/or amperage.

Going past and guessing a settling point just sounds pretty imprecise and unrepeatable to me. I seems like if you went into CV at the same voltage you wanted after settling, and held there until you hit the tail amps that are correct for you batteries, you would be much more accurate and consistent because you wouldn't have 1.3v of over voltage surface charge (I think). I wish I had a lithium battery to play with on the bench, as the Blue Sea charger I took out of the Roadtrek, and now use as a bench charger, would allow all of these theories to be tested pretty quickly. I wonder how much a 20ah battery would cost?

 

[booster]

I found this on the Cruisers Forum. I have no idea if this guy is knowledgeable or not, but he is addressing what we have been talking about.

LYP/LFP cells need the charge voltage to be limited, some say high cell voltage while charging won't harm but the Carnegie professor says different and I tend to believe him. With LYP/LFP cells there is no need for a boost/bulk charge rate that drops back to and absorption rate, they just want to be charged to 3.4v per cell and left at that. It takes a little higher voltage to get the 3.4v in there but not much, I use 3.45v per cell and that works out to 13.8v for a 12v nom. battery. Set the regulator to 13.8v and it will charge the cells as fast as they will take it till they are full, no need to stop at a certain amp acceptance rate like is required at higher cell voltages.
When the cells are full the terminal voltage will be 13.6v or a little higher and it will stay there till the batteries are discharged. The terminal voltage will remain higher than 13.v till the batteries are virtually completely flat.

 

[ptourin]

About your first post: Yes, it'd be handy to simply define SOC on my monitor and disconnect charging at 90%, 95% or 100% - whatever I choose. But I have that 14.6V CCCV charger - can I still do this? I suppose I can, based on tail current. I can define 100% SOC as when voltage climbs to 13.4, tail current drops to ~C/10 and 4 minutes go by. So my charger comes on and goes to CC - the voltage climbs to 14.6 but the charger stays on because current is high. It goes to CV and the current starts to drop - when it gets down to 20A the monitor shuts off the charger. That's pretty simple - I like it.

As for testing - you can buy Shorei motorcycle batteries for $150 - $200 that are 12V and 14 or 21 Ah. I'm sure there are cheaper choices, I just happen to know about these because of my vintage and racing bike friends. If you've got the time to test, I'll contribute to the battery cost - I'd love to know a bunch of these things and I don't have the time to set up a test bed. Just let me know <g>...

About your second post: I can't say much about his post because I still don't understand how charging voltage and charging time are related. I emailed the PowerStream guys and asked if they had any info on this or could point me to some - we'll see if they're talkers or not. I'd say that his last paragraph is a simplification. I'd still expect his batteries to settle a bit from 13.6 down to maybe 13.4. In general, yeah his batteries are getting down there when they go down to 13.0. But I popped a big load on mine and saw them go down momentarily to the mid-12's. It was just a short-term effect, but it surprised me, as I wasn't all that much discharged.

And while I'm at it - this may be a real bonehead question, but I keep seeing distinctions between charge voltage and battery voltage. Do you have any comments on this? I keep thinking that if I connect my 14.6 charger when my battery is at 12.8, it won't matter if I measure at the battery or the charger - I'll see some voltage between the two - charger will be pulled down and battery will be pulled up. Is there another whole concept that I'm missing here?

 

[avanti]

I have always assumed that "battery voltage" refers to the voltage that the battery shows when the charger is not providing current. No?

Since you mentioned it, I've always wondered how chargers measure this. Do they briefly stop providing current in order to measure the battery voltage? If so, one ought to be able to see this by watching the voltage carefully.

 

[gregmchugh]

You can get a 20ah 12v pack with the balancer board for $160 from Elite, I am sure there are cheaper options...

Elite Power Solutions

 

[booster]

Interesting it won't post with the comments inside the quotes, says the post is "too short" so this statement makes it long enough

About your first post: Yes, it'd be handy to simply define SOC on my monitor and disconnect charging at 90%, 95% or 100% - whatever I choose. But I have that 14.6V CCCV charger - can I still do this? I suppose I can, based on tail current. I can define 100% SOC as when voltage climbs to 13.4, tail current drops to ~C/10 and 4 minutes go by. So my charger comes on and goes to CC - the voltage climbs to 14.6 but the charger stays on because current is high. It goes to CV and the current starts to drop - when it gets down to 20A the monitor shuts off the charger. That's pretty simple - I like it.

I think that procedure would work fine. It probably would be more accurate than the systems that just kick out on voltage. Once you knew what amps gave you the right "settled" voltage on the batteries, I would it expect it to repeat very well. As you say, with the non adjustable charger, the voltage will continue to climb past whatever you set, but you will be able to control how much climb you get by setting the tail current higher to reduce the amount of climb, or lower to increase it. The further you stay away from the point of the charge curve where the voltage climbs quickly and current drops quickly, the less voltage rise you will have, also. Going out at 13.4v would leave you quite a ways away from the point, as it seems to really take off at about 13.8v in most profiles I have seen. The guy on the cruiser forum liked 13.8v cutoff to settle at 13.6v.

As for testing - you can buy Shorei motorcycle batteries for $150 - $200 that are 12V and 14 or 21 Ah. I'm sure there are cheaper choices, I just happen to know about these because of my vintage and racing bike friends. If you've got the time to test, I'll contribute to the battery cost - I'd love to know a bunch of these things and I don't have the time to set up a test bed. Just let me know <g>...

About your second post: I can't say much about his post because I still don't understand how charging voltage and charging time are related. I emailed the PowerStream guys and asked if they had any info on this or could point me to some - we'll see if they're talkers or not. I'd say that his last paragraph is a simplification. I'd still expect his batteries to settle a bit from 13.6 down to maybe 13.4. In general, yeah his batteries are getting down there when they go down to 13.0. But I popped a big load on mine and saw them go down momentarily to the mid-12's. It was just a short-term effect, but it surprised me, as I wasn't all that much discharged.

And while I'm at it - this may be a real bonehead question, but I keep seeing distinctions between charge voltage and battery voltage. Do you have any comments on this? I keep thinking that if I connect my 14.6 charger when my battery is at 12.8, it won't matter if I measure at the battery or the charger - I'll see some voltage between the two - charger will be pulled down and battery will be pulled up. Is there another whole concept that I'm missing here? This is the part that gets back to my original question of what really happens. Theoretically, if the charger has enough output available to not lose voltage, the batteries should charge faster. But with lithium, it appears the batteries will accept so many charge amps, even at lower voltages, that it is essentially impossible to have a charger big enough to hold its setpoint voltage. At that point the charger output will the same as the battery voltage plus whatever voltage drop there is in the wiring to the batteries (which will change as the amps change during charging). This would be why you would want to measure voltage at the battery. If you are charging at 200+ amps off the alternator, you can have significant voltage drop even in big wires. The battery and charger voltage will bothclimb as the batteries charge. Most seem to say that the lithium batteries don't get "pulled up" much during charging do to their very high acceptance rate which seems to be confirmed by the Cruisers Forum guy that says he sees only .2v drop with the charger removed. I don't think you are missing anything.

 

[avanti]

Interesting it won't post with the comments inside the quotes, says the post is "too short" so this statement makes it long enough

The trick is to carefully edit the "QUOTE" and the "/QUOTE" tags in order to intermingle your comments with the old post. Just don't screw up the nesting.

Theoretically, if the charger has enough output available to not lose voltage, the batteries should charge faster. But with lithium, it appears the batteries will accept so many charge amps, even at lower voltages, that it is essentially impossible to have a charger big enough to hold its setpoint voltage.

Yes, but doesn't the "accept so many charge amps..." statement only apply while the battery has a low state of charge? I assume that Li batteries are like any other battery and lose their ability to accept high currents as they approach 100% SOC. So, once we get to the end-game (where all of this matters, presumably), the battery can no longer sink all those amps and the charger will gain the ability to control voltage. Does this make sense?

 

[booster]

Yes, but doesn't the "accept so many charge amps..." statement only apply while the battery has a low state of charge? I assume that Li batteries are like any other battery and lose their ability to accept high currents as they approach 100% SOC. So, once we get to the end-game (where all of this matters, presumably), the battery can no longer sink all those amps and the charger will gain the ability to control voltage. Does this make sense?

That is the big question that I haven't been able to find any really good data on. Most times all you get are "they accept huge charge amps up to about 95%" or other jargon, but I have yet to find a %C charge rate vs SOC graph. I have seen folks talking about 3/5C charge rates, though, so if those continue to 95% no moderate size bank would ever have enough charger on it to not be pulled down in voltage. If you do stop you lithium charge short of full at 90-95%, the end game parameters don't matter any more, like they do in topping lead acid batteries, so that is one of the big benefits of what we have been contemplating.

Here is a charge profile for a 200ah battery charged at 100 amps that shows the profile they saw, with slowly rising voltage and constant 100 amps available. At about 95% the voltage soars and current crashes, so the question remaining is how much current would be required to have the charge voltage run higher than battery voltage. In our system, if it were lithium, we couldn't even get to .5C in the example, so it would definitely apply to us (we can get a bit over .4C on the alternator when it is hot).