LiFePO4 charging - the overcharging question

ptourin · 10-05-2015, 06:34 PM

I figured I'd start another thread because I still don't understand the overcharging discussion. I think it got discussed in Wincrasher's thread, but that's gotten so big I thought I'd branch off from it.

I'm going to put down a charging scenario - if I get it wrong, flame away <g>...

So let's say I'm down around 50% SOC. I turn my charger on and it puts out lotsa amps - my battery voltage creeps up - 13.1 - 13.2 - 13.3.....
Sooner or later I get to the point in the charging curve where I'm near full charge and the voltage starts rising quickly. Now the charger ends up putting out 14.6 volts and as the battery climbs to meet that voltage, the charge current starts dropping. Finally the battery is up around 14.6v and the charge current drops to around zero. I say, "I'm fully charged now" - I disconnect the charger.

Let's say I don't have any load pulling the battery down. So it sits for some hours, or a day or two (I understand it takes a long time for the lith phosphate batteries to settle) - and the voltage slowly drops down to around 13.4v. I measure that, and I say, "it really worked right - I actually AM fully charged" - and I go set my monitor so that 13.4 is what it considers 100% SOC.

In a perfect battery world, where my battery is perfectly balanced, would I be perfectly safe to leave the charger on? All banks are at exactly 14.6, they're perfectly balanced and no cell draws any current - my charger current is really zero - right?

So now we step away from the perfect world into the real world. If I have one bank that's weak, so that it's not quite up to 14.6v, I assume the charger will try to lift it up to 14.6 - so there's some charger current, and all the banks are seeing it. At this point I start getting vague about what happens if there's no BMS to decide what to do. I assume that the other banks don't just say, "I'm already at 14.6, so I don't need anything" - if they did, we wouldn't need balancing circuitry. What's happening here, and why is it likely to damage the battery? I can't picture yet what's happening because of the imbalance, whether the other cells start going higher than 14.6 in the charger's effort to get the weak cell up. Does anybody have a good enough grasp of this to describe it? I'm of course trying to understand where we get to the point where the battery starts to sustain damage.

gerrym51 · 10-05-2015, 07:52 PM

this might help

LFP Precautions | Home Power Magazine

gerrym51 · 10-05-2015, 07:57 PM

and this

How to charge Lithium Iron Phosphate lithium ion battery packs including packs with high current and High Capacity.

ptourin · 10-05-2015, 08:03 PM

Their first bullet is very different from what I've been getting from SmartBattery - Home Power Mag: "Never overcharge LFP batteries—even a fraction of an hour of overcharging can permanently damage the cells. A battery management system (BMS) is essential to preventing overcharging" - vs SmartBattery: "Don't worry about it" <g>...

In theory - famous last words - I should be safe from overcharging, since my batteries open an internal relay at 15.8 - if I left the charger on 24/7, and it did lift the batteries way over 14.6, the internal relay would sense overvoltage and disconnect. But that's seriously crude. I can program my monitor to connect and disconnect the charger (or turn it on and off), but I need to decide what the connect/disconnect points are (I think I can specify them in either volts or % SOC). Until I understand more about the dynamics of charging, I'm not going to be able to make a good decision. That's why I posted this.

gerrym51 · 10-05-2015, 08:10 PM

roadtrek's ecotrek system has total disconnect switches. also Davydd's ARV.

i suspect this is so stored overcharging and below 32 charging that causes plating can be avoided

ptourin · 10-05-2015, 08:29 PM

Are you talking about manual disconnect switches or relays that sense voltage and automatically disconnect/connect? I have the former, and I'm using them. I want the latter, and don't have them yet - if I want them, I have to do it myself. I'm sure that both the eTrek and Davydd's ARV manage this automatically and very nicely - but I've made the plunge and will have to decide whether I follow my battery supplier's advice and leave the charger connected all the time (except storage and cold weather) and hope they're right. So I'm hoping that Booster of one of the other techies here will have looked into the details of what happens near full charge and can either answer some questions or point me to some articles or other sites that discuss this in detail.

I can do this very crudely. I could, for example wire the system to disconnect my charger whenever I go up to 13.3 volts for more than x minutes, and connect it again whenever I go down to 13.0 volts for more than y minutes. I could do that now, and it'd give me crude protection from overcharging. Every so often (like just before heading out dry camping) I could override this system and let the batteries charge until I see 14.6 volts for 15 minutes, then turn it on again. From what I understand, that'd work just fine. But I'd like to understand it better and come up with something a bit more refined.

gerrym51 · 10-05-2015, 08:37 PM

Quote:

Originally Posted by ptourin

Are you talking about manual disconnect switches or relays that sense voltage and automatically disconnect/connect? I have the former, and I'm using them. I want the latter, and don't have them yet - if I want them, I have to do it myself. I'm sure that both the eTrek and Davydd's ARV manage this automatically and very nicely - but I've made the plunge and will have to decide whether I follow my battery supplier's advice and leave the charger connected all the time (except storage and cold weather) and hope they're right. So I'm hoping that Booster of one of the other techies here will have looked into the details of what happens near full charge and can either answer some questions or point me to some articles or other sites that discuss this in detail.

I can do this very crudely. I could, for example wire the system to disconnect my charger whenever I go up to 13.3 volts for more than x minutes, and connect it again whenever I go down to 13.0 volts for more than y minutes. I could do that now, and it'd give me crude protection from overcharging. Every so often (like just before heading out dry camping) I could override this system and let the batteries charge until I see 14.6 volts for 15 minutes, then turn it on again. From what I understand, that'd work just fine. But I'd like to understand it better and come up with something a bit more refined.

roadtreks are manual disconnect switches-david said his were manual also,the roadtrek BMS will do emergency disconnect

booster · 10-05-2015, 10:18 PM

I really don't know if I can be any help, or not, as I don't feel that I ever found a definitive answer on the charging. Some of what I found is buried in all the multitude of other lithium discussions, including charge profiles, and kind of hard to find.

When I went looking for information some things were generally consistent, with the exception of folks like Smartbattery. Some of the things are:

You can charge very fast, nearly no limit on input current

You should never let lithium get completely discharged as it will hurt them. Generally accepted was to discharge no more than 90%, and because voltage doesn't drop much over most of the range, when it does start to drop, cut off use.

Lithium doesn't like to be full all the time, especially when stored and not being cycled. Very fuzzy on how much damage is done by being full all the time.

Lithium does not want to be on "float" charge once full as it will degrade the batteries. Again fuzzy on just how much degradation happens, and how fast it happens.

Don't charge below freezing (battery temp, not ambient) or bad damage will happen fairly quickly.

Then there is the charge termination question that ptourin is looking at quite thoroughly.

Davydd's setup stops the charging before it gets to full charge, but I don't think we ever really found out how they are doing it.

From what I read on the issue, many talked about going by voltage to stop charging, but they also often wanted it to be no load, rested voltage. This makes it very hard to correlate to what the voltage would be at any given charge rate and charger voltage. I never found a clarification on that, except to allow rest time, which is impractical in the real world.

The best information, I think, came from some tech papers by researchers and battery developers, who weren't selling anything. Their charging profiles and graphs showed the cutoff of charging based on current, just like the return amps that can be used with lead acid. The contention was that if you use current, you can stop charging reliably and consistently at any state of charge you want to leave the battery at, be it 50% 80% or 98%. My guess is that the curve for the current drop is probably pretty steep, but must be usable for this purpose. Of all I saw, this theory seemed to make the most sense, and would also actually be doable in the real world. Chargers exist that will do this charging, just as they are suggesting.

So, what would I do if I were building a lithium system (which I am not at this point because I think too much is unresolved-we are building a big AGM setup)? I would use a constant voltage/constant current charge profile charger, that did a full shutoff of charging based on current to the batteries. As of now, that sounds like the system that would be the most consistent and easiest on the batteries, but it certainly isn't absolutely proven.

ptourin · 10-05-2015, 10:58 PM

That's quite interesting. the Victron monitor will allow me to control its internal relay based on SOC. I define what the monitor thinks 100% SOC is by specifying a voltage and a current - when I reach the specified voltage and the current drops below my specified value - I believe there's a time period I can set, like for 4 minutes - then the monitor says I'm at 100% SOC. I believe it's set now for 13.4v and 4% (4% x 200, so 8 amps).

I can set the internal relay to either engage or disengage at a specified SOC, and I can choose n.o. or n.c. - so I have a good deal of flexibility. So for example, I can use the above setting and tell the relay to disengage at 100% SOC - we know the battery is going to go up way over 13.4 as it approaches 100% SOC, so I know the monitor will really see that setpoint - it should be pretty conservative. Then the monitor waits till the current drops below whatever vaue I pick - and the relay triggers.

Then I can pick whatever I want for a low setpoint - relatively unimportant - start with 40% and start working downwards.

The relay is very small - I think the contacts handle 60V @ 1A. So I'd need a power relay with a 12vdc coil, and use it to interrupt the 120vac into my charger - it's a PD9169AL, and it consumes 1000 watts - more or less 8.3 amps. I have an enclosed 40A DPDT power relay - that should work fine handle switching the AC to the charger if the coil current is low enough. So it shouldn't be hard to breadboard this.

Of course - Jean's going to live in the RV in Salem MA till it gets too cold - then I bring it here to Vermont - and I'm hard against the low temp issues by the time I can wire it and start testing <s>...

One thing I'm not sure of is what the monitor sees if I leave the charger on as the battery approaches 100% SOC, since I've had almost no time with the system. IIRC, the monitor sees the battery climb to match the charger at almost 14.6v. But their definition of 100% SOC is that the monitor has to see BOTH the voltage and current setpoints - so I rise past 13.4 (supposedly the resting voltage at 100%) and watch until the current drops down to - something - I'll have to decide - then the relay switches.

Does this make sense conceptually? Have I missed anything? I drew up a simple schematic, but I'm realizing that I need to think of when I want both relays to be drawing power, so that when I've got shore power and the system is charging, the relays draw power - but if I'm away from shore power, both relays are depowered. Still, it's a simple circuit.

Davydd · 10-05-2015, 11:24 PM

I'm not sold on the idea that fully charge lithium ion batteries are a bad thing. I'm almost always fully charged once on the road. You might be getting anecdotal opinions from battery abusers. My experience with my lithium ion MacBook Pro is interesting if you want anecdotal. I use it almost 100% of the time plugged in and fully charged. It is 5 years old and I unplugged and used it this summer and it still performed as advertised. With my Class B battery pack I don't expect to have it long enough to determine the effects.

booster · 10-05-2015, 11:31 PM

Quote:

Originally Posted by Davydd

I'm not sold on the idea that fully charge lithium ion batteries are a bad thing. I'm almost always fully charged once on the road. You might be getting anecdotal opinions from battery abusers. My experience with my lithium ion MacBook Pro is interesting if you want anecdotal. I use it almost 100% of the time plugged in and fully charged. It is 5 years old and I unplugged and used it this summer and it still performed as advertised. With my Class B battery pack I don't expect to have it long enough to determine the effects.

I think you need to look to see how the chargers for the items are working. All of our phones, computers, gps, etc that have lithium batteries (different chemistry than the RV stuff) have full cutoff chargers, and, sure the readout says 100%, but you have no idea if they stopped the charge at 90% or not. They may well accept longer run in exchange for shorter life, or not. In a full cutoff charger, you can leave it plugged in forever, and it will never see too much charge, because a full cutoff charger does exactly that, fully cuts off.

And remember, you stated in the past that ARV stops the charging short of full charge, maybe 90% IIRC, so technically, you are never fully charged, just charged to the maximum that ARV allowed in the programming.

I don't think we know yet how much life you lose by being fully charged all the time. As I mentioned, that is one of the areas that the data is very squishy.

ptourin · 10-05-2015, 11:33 PM

I'd also like to read whatever sources discuss possible issues with remaining fully charged, or being stored fully charged. Both things are perfectly manageable AFAIK. Especially for you Davydd with all that capacity - it'd be perfectly reasonable to charge to 80 or 90% most of the time, and just charge all the way when you intend to dry camp for a longer period.

I've forgotten the standing rate of discharge in cold weather, but I remember thinking that you could just charge to 60 or 70% and there'd be no problem with storage for 4 months. Oops, another reason not to get solar...

booster · 10-05-2015, 11:36 PM

I dug out one of the charge curves that I had posted on another thread, that shows a current terminating, full cutoff, charge curve.

It shows a constant voltage holding as the current decreases to their 3% full cutoff. I think your Victron would do this pretty easily, as they usually need a time at voltage, followed by a current threshold, which is just what this profile shows.

ptourin · 10-05-2015, 11:52 PM

I remember looking at that chart wherever you posted it. It's interesting - yes, the low current end is perfectly doable. The cell voltages on that chart are right up there!

I'm not quite sure I understand the Stage 4, but I guess it's just saying that as soon as the cell voltage settles after the charger is turned off, and then drops a bit, you can do a topping charge. That's different than what I was thinking - in my scenario I picked a pretty low setpoint for turning the charger on again, but of course it could be 80% or 90% and just bring the charger back on for a short time. It still appears to me that the low end is less problematic than the high end. And I've never heard any discussion of problems with many small charging sessions, only the many discussions about overcharging. Maybe it's starting to be a plan...

gerrym51 · 10-05-2015, 11:56 PM

if overcharging or undercharging,or charging below 32 were not issues then why have complete lithium battery cut offs at all.

My agm does not have it.

and yet zions with ecotrek does.

it's for a reason

ptourin · 10-06-2015, 12:11 AM

No question about undervoltage - if you totally discharge a LiFePO4, it's toast. My internal protective relay opens around 8.5v - that's way down the "knuckle", only short time left from there to total discharge. It opens for overvoltage at 15.8v - again, that's WAAAAYYY higher than I'd ever expect to charge it.

My batteries have no temp protection in the BPS. I believe that the Victron monitor will take a temp input and can be set to respond to under/over temp, but I haven't paid much attention to that yet - figuring out the charging/overvoltage issues has seemed more important, especially since I'm not at this point doing much camping below freezing - I know I can charge the batteries when they're above freezing, then disconnect them and let them sit all winter in safety.

But temp sensing and protection is a topic I'd like to explore - at least enough so I have a basic understanding. As always - can someone tell me or point me to info about how temp sensor leads work? I believe one came with my monitor, though I didn't hook it up. Do you bolt them to a battery terminal, and how can they provide info about the temp of the battery chemistry?

ptourin · 10-07-2015, 07:57 PM

Well bad news about temp protection with the Victron monitor. The temp setpoints can be used to activate the alarm but not to control the internal relay. I suppose the alarm function is nice, but if I really wanted to protect, I'd have to sense battery temp separately. Not what I'd hoped....

gerrym51 · 10-07-2015, 08:09 PM

why not take the isolate method. batteries shut off from any connection to anything

ptourin · 10-07-2015, 08:44 PM

I can easily do that - I'm trying to figure out how to automate some of these functions, especially the ones around preventing overcharging.

ptourin · 10-23-2015, 12:56 AM

I'm posting a question in this thread since it's more a question of mine than Crank's or Eastbay's. Booster, are you here? - hoping you'll chime in on this.

I want to understand your comments about ending the charging cycle based on return amps. I think I understand it conceptually - you mean that you watch your charge current drop as you approach 100% SOC, and when it drops to some chosen point you say "OK, I'm there" and disconnect the charger. Right?

But I've got the usual shunt setup, and my Victron monitor is giving me the difference of charge amps minus load amps.

Now I've got to ramble a bit and try to say what I'm trying to say.

1. I can set tail current on my monitor - Default is 4% but Victron support guys suggested 10% for LiFePO4's, or 20A in my case.

2. The monitor says I'm at 100% SOC when I go over my voltage setpoint (13.4) and my current drops below the tail current setting for over 4 minutes.

3. I can set my internal relay to open (or shut, for that matter - it's a setup param) at 100% SOC - so it cuts the AC feed to the charger. Then I set it to turn the charger back on at some low value - 40% SOC, 30% SOC, whatever I choose.

OK, now I'm to the part I don't understand. If I go outside now and look at the monitor, I'm charging the batteries and I have no loads except the parasitics like the CO detector - a while ago it was 45A, then 35A - if I go out later this evening it'll keep getting lower. Sounds great - I can set the monitor to open the relay and turn off the charger, and I'm home free.

BUT - what happens if I turn on all the lights and set the fridge to 12vdc, and start drawing about 10 amps of load? The monitor is going to read the charger current minus the 10 amps, right? So it'll read 100% SOC early, won't it - voltage setpoint plus current drop below a setpoint. So any time I have a load, the monitor will see less positive current and open the relay early, won't it? Am I missing something here?

You may also like: