Lithium Batteries (for the rest of us)

[familycruisers]

I'm east coast btw...

[Singleprop]

Lets do this on PM to avoid cluttering the thread.

[poiu]

Originally Posted by T1 Terry

Li cells are fully charged at 3.4v, anything more is over charged and continually damages the cell every time it happens.

T1 Terry

I've been doing this. I have 8 700Ah Winstons. The current doesn't stop going in though. The Mastervolt chargers are set to 3.45v float charge - constant volts. Volts keep rising and current drops to about 30A as charge progresses, then more or less stays at that level for hours. How long I don't yet know as I manually turn off after 2 hours as a maximum, but I am a bit concerned they will overcharge if left permanently exposed to high volts like this - is this ok to do?

[goboatingnow]

OK, sure, the Li charger can be simple. Now, offset that simplicity with the need for a BMS that can double the cost of the batteries themselves and...ooops, simplicity just went out the window. Mind the broken glass.

A BMS for Li Ferrous can be as simple as a simple $20 voltage monitor. Nothing more complex is actually needed

The discussion from some in this forum , strikes me as a phrase I once heard
" Yes its all very well that it works in practice , but does it work in theory!."

The fact is the cells work the way people like me, Terry, Lagoon etc say they work , thats what happens in practice.

OceanSailor you strike me as an EE with experience but not in Li Ferrous and you are mixing up issues with other Li topologies. AGMs suck for boats BTW in real life.

dave

[OceanSeaSpray]

Originally Posted by goboatingnow

A BMS for Li Ferrous can be as simple as a simple $20 voltage monitor. Nothing more complex is actually needed"

At $20, the BMS is YOU basically. You will be disconnecting the loads on LV and stopping charge on HV. Well... this is if you happen to be around at the time.

Here is an IC that terminates the charge on residual current - as it should - for you. The ones that push all the way to the over-voltage limit at 3.8-3.9V/cell and pull the pin there, leaving everything nice and toasty are for $15 cells in little portable devices.
I wouldn't do that with my house bank.

The thing is, when a technology is forgiving, it gets abused. Does it mean that it should be abused? Maybe not. Personally I am after a long service life and a setup that I can almost forget, won't fail, screw up or damage itself. So I am not going to abuse it. And it won't cost much either, I can already see that now.

[goboatingnow]

Here is an IC that terminates the charge on residual current - as it should - for you. The ones that push all the way to the over-voltage limit at 3.8-3.9V/cell and pull the pin there, leaving everything nice and toasty are for $15 cells in little portable devices.
I wouldn't do that with my house bank.

The thing is, when a technology is forgiving, it gets abused. Does it mean that it should be abused? Maybe not. Personally I am after a long service life and a setup that I can almost forget, won't fail, screw up or damage itself. So I am not going to abuse it. And it won't cost much either, I can already see that now.

Noone is disagreeing with that statement as a general rule. I woudl agree, given the costs of current Li tech, its makes sense to do cell monitoring and LVC cutoff and as you yourself said. it doesnt cost much.

Ive designed in dozens of small Li charge system into my designs over the years and I have a GSM monitor with LI charging and backup on my CAD system at the moment.

Note that charge termination by sensing current is fine in standalone applications. In applications with load sharing, such as a boat, its application has to be carefully applied as the charge cycle can go on for too long. If you map charge current against terminal voltage you will see that there is a increasing rapid dv/dt as you approach the Knee. This in my experience is a better method of termination.

IN fact for very conservative use and to avoid charging mini cycles in a boat, very conservative stop voltage ( 3.5-3.65) are best used. IN these cases in excess of 10% of the capacity remains and hence termination by charge current in itself is not a good idea.

IN my designs once I get a specific voltage I tend to stop as I cant be sure exactly what the battery receives. In some designs Ive looked at direct battery current sensing as a way around this so as to enable higher charge densities. Bit in large prismatics it makes no sense to try and get the last bit out of the batteries.

Remember also not to float charge Li. when the charge cycle completes the charging should disconnect.


what I was trying to say in general , is that Li technology y, especially Li ferrous can actually take quite significant abuse. Its doesnt need to be surrounded by all sorts of monitoring devices persay. Thats not to say that it shouldn't be

dave

[noelex 77]

Originally Posted by T1 Terry

This would indicate the batteries will cycle for at least 10 yrs http://en.wikipedia.org/wiki/Opportunity_(rover)#2013

I am not sure quoting the life of litium batteries on another planet is very indicative of boat owners are likely to experience.

The sort of temperatures experienced on Mars are much colder. Lithium battery lifespan is very dependent on temperature, with hot temperatures leading to a much shorter life, but I guess only NASA knows what happens at -150 C.

I plan my crusing destinations around slightly warmer areas

[goboatingnow]

I think life cycle is a very open question in regards LI , especially large prismatic Li Ferrous cells at the moment. I know Ive been surprised sometimes at the severe reduction in cycle life on certain LIs. This is especially true where some part of the Li envelope was pushed ( especially attempting 100% charges in small Li installations )

Its going to take time to build the data.

Dave

[goboatingnow]

BTW teh way OceanSeaSpray, Just to elaborate on charge termination.

If you charge Li at about upto .25C, then voltage termination is a fine method and tends to be very safe. This would be upto 100 amps on a 400 AH bank , probably not untypical for a boat. However if you engage in higher C charging, then you must also do di/dt charge termination detection. At higher C charging the battery spends longer in the "constant voltage tail" and hence can be undercharged if only voltage is used.

(again wasnt arguing with you , was merely making the original point that Li can be charged simply using voltage cut-off as an indicator, ie you can do it from a bench power supply)

Like everything, its the devil in the detail.

[ebaugh]

Originally Posted by poiu

I've been doing this. I have 8 700Ah Winstons. The current doesn't stop going in though. The Mastervolt chargers are set to 3.45v float charge - constant volts. Volts keep rising and current drops to about 30A as charge progresses, then more or less stays at that level for hours. How long I don't yet know as I manually turn off after 2 hours as a maximum, but I am a bit concerned they will overcharge if left permanently exposed to high volts like this - is this ok to do?

Are you saying that at 3.45V, the charger stays at 30A for two hours, with only a small voltage rise? I would not expect that. And it's probably a too high for float, I would drop the float voltage to 3.35 or less.

Here is some data from new Forzatec LFP cells from a capacity test cycle from 3.65-2.0-3.65V. These cells are 97.7% charged if you stop charging at 3.4V. At 3.45 they are 99% charged.

This was testing at .3C, 30A on a 100Ah cell. Only CC with no CV stage. However, go back in the next day, and continue the charge, there is a very very trivial capacity remaining. Charging to 3.65 CV terminating at .01C takes less than 1Ah of charge. The charge numbers started at around 2.8V, that was the recovered voltage after resting 1 hour after the discharge to 2.0V.

[poiu]

Originally Posted by ebaugh

Are you saying that at 3.45V, the charger stays at 30A for two hours, with only a small voltage rise? I would not expect that. And it's probably a too high for float, I would drop the float voltage to 3.35 or less.

Here is some data from new Forzatec LFP cells from a capacity test cycle from 3.65-2.0-3.65V. These cells are 97.7% charged if you stop charging at 3.4V. At 3.45 they are 99% charged.

This was testing at .3C, 30A on a 100Ah cell. Only CC with no CV stage. However, go back in the next day, and continue the charge, there is a very very trivial capacity remaining. Charging to 3.65 CV terminating at .01C takes less than 1Ah of charge. The charge numbers started at around 2.8V, that was the recovered voltage after resting 1 hour after the discharge to 2.0V.

Thanks for this information, it is very useful and nothing like what I see, though I wonder how different the numbers would be with CV charging. I really need to set up the logger to monitor it like this. I note your charge voltage limit. Is there any information source on CV charging? I can't find anything on the battery maker's web sites - they all provide information expecting a more sophisticated charging regime.

The charging cycle with my float voltage set charger goes something like this: 3.1v prior to charging then steady voltage rise at full charger output of 200A, tapering off to say 120A at 3.35v then dropping steadily to a low of 30A just before I switch off and the last period of about two hours the amps only drop a little and the voltage only rises a little.

It may be the current is reduced by the charger as according to the Mastervolt manual output will reduced at higher operating temperatures. They don't say by how much though. The charger cases are quite warm.

[poiu]

I just looked at my notes. After disconnecting the charger the voltage settled at 3.43v.

[ebaugh]

Originally Posted by poiu

Thanks for this information, it is very useful and nothing like what I see, though I wonder how different the numbers would be with CV charging. I really need to set up the logger to monitor it like this. I note your charge voltage limit. Is there any information source on CV charging? I can't find anything on the battery maker's web sites - they all provide information expecting a more sophisticated charging regime.

The charging cycle with my float voltage set charger goes something like this: 3.1v prior to charging then steady voltage rise at full charger output of 200A, tapering off to say 120A at 3.35v then dropping steadily to a low of 30A just before I switch off and the last period of about two hours the amps only drop a little and the voltage only rises a little.

It may be the current is reduced by the charger as according to the Mastervolt manual output will reduced at higher operating temperatures. They don't say by how much though. The charger cases are quite warm.

Remember this data is for 100 Ah cells, most of the data at .3C. If you have a 24V bank, this should be like charging at your 200A rate. But if your bank is 12V, it's more like a 400A rate and the data won't line up. Think my quick math is right...

Anyway, both my GBS cells and the Forzatec ones see the voltage skyrocket above 3.45 if .3C is held. It's under 2 minutes at 30A from 3.45 to 3.6V.

Some of your current tapering is likely because the charger sees a different bank voltage (voltage drop) at the end of the battery cables than you see on a voltmeter attached to the cells. I see this on both my chargers. That may explain the 120A at 3.35V.

It may explain the entire situation if the cells stay below 3.45 or so, since the cells will accept full rate easily to that voltage, but it's getting limited by the charger adjusting slowly as the voltage drop across the cables equalizes.

I've configured my chargers to go flat out "bulk" to 14.2V, then switch to float, skipping "acceptance". The physical LFP bank never gets there, it gets to about 3.4-3.45V, or 13.6-13.8V then drops off.

[ebaugh]

Originally Posted by poiu

I just looked at my notes. After disconnecting the charger the voltage settled at 3.43v.

This value means little without a time at rest, and then only if there is no load.

We measured some brand new cells out of the box after arriving ocean freight at 3.4V. They were shipped at full charge!

One other thought....if you have a "load" 12V refrigeration, watermaker etc, the charger may be powering that at the end of the charge.

[OceanSeaSpray]

Originally Posted by goboatingnow

BTW teh way OceanSeaSpray, Just to elaborate on charge termination.

If you charge Li at about upto .25C, then voltage termination is a fine method and tends to be very safe. This would be upto 100 amps on a 400 AH bank , probably not untypical for a boat. However if you engage in higher C charging, then you must also do di/dt charge termination detection. At higher C charging the battery spends longer in the "constant voltage tail" and hence can be undercharged if only voltage is used.

There is nothing to elaborate much about, dv/dt or di/dt. Bulk charge always terminates when a voltage target is reached and if bulk is enough, there is nothing else to do. You just get a higher charge if you take it slowly and you will get a nearly full charge if you take it very slowly.

What you get with alternators on boats is bulk-charging only. I can't see any simple, practical way of performing a full charge this way without building a special regulator and use a current shunt. And there are no good reasons for wanting to do it, except that it could do current limiting.

When charging starts, the alternator outputs everything it can (which varies with RPM and later temperature) and the voltage starts rising. Once it reaches, say 14.4V (modern alternator), it holds there for a little while, not very long, current starts dropping off, and then voltage is stepped down to 13.8V typically, and stays there. As soon as this happens with LiFePO4 cells, charge has been terminated for all intents and purposes.
It just happens that 14.4V translates into 3.6V/cell, target charging voltage for most LiFePO4, and by the time the LiFePO4s reach 3.6V/cell they are 90% charged or so already, especially if charge current was reasonable. This is also why, if there is temperature compensation, it should be eliminated. In cold weather, it can increase that figure of 14.4V to 15V and beyond and conversely it gets lower in a hot environment.

Things get more muddled once you start splitting into banks. Lagoon4Us is splitting and sensing his SLA starting bank. The behaviour of the system will now depend a lot on the state of charge of the SLAs.
If they are full (usual), there will be greater voltage drop over the splitter (diodes or MOSFETs) on the path of the heavier current (to the lithiums) and the SLAs will always read higher. This will cause early current drop-off and then charge termination for the lithiums, before they reach the regulated voltage. You can leave quite a lot on the table this way. Low SLAs would result in a better charge for the lithiums.
Sensing the lithiums instead would give higher current for longer to the house bank, but the starting bank would tend to go over-voltage while the lithiums are charging.

The problem has been around since blocking diodes and other voltage switches came on the scene, short of having one regulator per bank, something is always unhappy. Because an alternator is basically a regulated source, sharing it between banks always leads to over- and under-charging issues.
Keeping everything as simple as possible, the best would be dedicating a suitably sized stock standard alternator to the LiFePO4s, the meaning of suitably sized being not too large so it doesn't start charging at 1C.

When it comes to trickle charging at the top, I also think it is a complete no-go. Alternators won't do it even running for days, the problem is with solar systems etc. I have the feeling that there could be more lithium banks ruined on boats sitting around doing nothing than on the ones with someone living aboard.
A fully charged LiFePO4 bank left alone should sit around 13.3V. If it is enough to cause the regulator to cut in, or if it starts by charging to the cut-off limit every morning...