8th Annual Lithium Report

[Winston-ClassB]

Each year we measure the capacity of our 500ah lithium pack to help document the "real" life-expectancy of lithium batteries used in the RV environment. Some authors have referred to the use of batteries in an RV as "fractional use" - - this term reflecting the reality that RV'ers rarely fully discharge (cycle) their batteries and, at least in our case, we maintain our pack at fairly high charge levels as we never know when circumstances may require access to our full capacity. Further, we do not 'power-down' our system. During those periods of travel idleness (a/k/a winter), we often find ourselves 'repairing' or adding new things to our DIY Promaster CamperVan - - so it has never seemed necessary or appropriate to power it down. And while there are many 'threads' discussing how we should treat our lithium batteries for maximum life - - the rules 'distilled' from these technical treatises often conflict with the realities of RV life - - thus, 'best lithium practices' yields to the realities of our RV lifestyle.

Our lithium pack consists of 20 - 100ah Chinese GBS cells arranged in a 4 high, 5 wide matrix. When installed, this pack netted 515 ah of capacity. We lost an additional 40 ah of capacity this 8th year - - down from 420 ah after 7 years, to its current 380 ahs. While 380 ah capacity likely remains sufficient for our needs, this is not the whole story.

Arguably the greatest feature of lithium batteries is their ability to draw literally hundreds of amperes current from the pack with the pack voltage hardly sagging. We can pop our induction stove onto "high" - - drawing 150 amps off the pack - - with none of the other electronics (computers, file server, router) flinching. This is the consequence of lithium's very low internal resistance.

Not so much so after 8 years. The sagging is reminiscent of lead-acid batteries and, on one occasion, the sagging voltage caused a momentary glitch in our inverter and the involuntary re-booting of our file server. So, this season, not due to too few amp-hours, but the existencd of "internal resistance" - - we are seriously looking into pack replacement

 

[booster]

Hi Winston.

As always a very useful data point for all of us to take note of. I always wondered about where in the degradation process high amp delivery started to be a problem.

Where are your cell voltages and balance at this point?

It also seems to follow the life others have seen when using good, but practical methods.

It is looking like 6-8 years of moderate use may be the norm for lithium. About the same as good AGMs, but with the current pricing nearly a wash, just need to select which benefits and detriments suit your needs best.

I will be watching to see where you go next, if you replace the pack. The lower prices also give more options, I think.

 

[Winston-ClassB]

Hi Winston.

As always a very useful data point for all of us to take note of. I always wondered about where in the degradation process high amp delivery started to be a problem.

Where are your cell voltages and balance at this point?

It also seems to follow the life others have seen when using good, but practical methods.

It is looking like 6-8 years of moderate use may be the norm for lithium. About the same as good AGMs, but with the current pricing nearly a wash, just need to select which benefits and detriments suit your needs best.

I will be watching to see where you go next, if you replace the pack. The lower prices also give more options, I think.

The difficulty with replacing what we have with anything other than GBS cells is our BMS system which is completely external to the cells and provides more granular information (e.g. the voltage and temperature of each cell) than many of the 'drop-in' solutions. And our present system (with new batteries) literally allows charging and discharging at any current we desire . . . charge and discharge rates in excess of 500 amperes is possible (well, if we had the ability to generate or consume such high currents). Finally, we have such a "pretty panel of gauges" - - some of which would 'go dead' if we abandoned GBS. :~(

As far as cell voltage tracking goes - - when new, these cells tracked within 0.01 volts (the resolution of our BMS system and external Fluke meter). Cell balancing is performed by charging at 14.2 volts (3.55 volts/cell). When any cell exceeds 3.55 volts - - a 1/2 ampere shunt is enabled across that cell which effectively terminates charging of that cell while the laggard cells in that chain 'catch-up'. In reality, some cells will overshoot the 3.55 volt target. Early on we pondered: "If one cell is sitting at 3.6 volts, and with the total charging voltage set to just 14.2 volts (3.55/cell), how can a laggard cell ever reach the 3.55 volt level?" The answer appears to be that - - at this charging level - - the current in each string is less than 1/2 ampere . . . thus, the shunt across any cell above 3.55 volts actually acts to discharge the 'over-charged' cells causing them to drop back to 3.55 volts while the laggard cells continue to charge to 3.55 volts.

All this said, this is no longer the case with our 8 year old lithiums. After many hours of charging at the balance/equalization voltage of 14.2 volts . . . some laggard cells have not, and will not, ever rise to the 3.55 volt level. Indeed, we have one problematic cell that does not rise higher than 3.35 volts - - a full 0.2 volts below most of the others.

So we experimented. We forced this laggard cell to 3.55 volts (and above) by placing a charger on this one cell. At 10-15 amps, the cell 'conceded' and rose, as noted, to and above 3.55 volts. We weren't surprised, however, when this cell sank back to the 3.35 volt level when our forced charge was removed.

What we've found interesting the past couple of years is that these laggard cells "catch-up" to the good cells during the discharge process. Catch-up?? Maybe we should have said that the good cells "catch-up" (drop down to) the voltage of the laggard cells early during the discharge cycle with all 20 cells, even after 8 years, tending to track lower and lower within that original 0.01 volt range.

At least until this year, we have concluded that whatever causes a laggard cell not to reach the balance voltage of 3.55 volts, this is a "top-end" effect only and that the capacity of the laggard cell(s) remains about the same as the normal cells. This year, however, our laggard 3.35 volt cell was the first to reach the lower cut-off (2.8 volts) - - by 0.15+ volts ahead of the others - - causing the termination of our capacity test.

Hope we answered your question or, at least, you found this interesting.

Winston

 

[booster]

You bet I find it interesting and informational, great data.

You may, or may not, find interesting about the balancing of our 4 cell strings X 3 for 618ah.

When I had AGM batteries, I was meticulous about cable length matching, but still got different charge and discharge rates between the two strinsgs.

I checked the litium strings an found the same thing even after top balancing the cells individually.

To get the charge and discharge rates to match I selectively put .02" stainless washers under the 6 cable connections to the 3 batteries and changed the order of cables that were stacked on a connection.

The batteries passive balance at 14.4v and we never charge to there, so thy never have been balanced since I put them in.

The cells will get up to about .030mv difference when charging or discharging, but as soon as the are at low current again, thy stay at under .003mv.

We also have them connected all winter and often used during van work.

I was quite surprised how well they stayed in balance.

 

[Winston-ClassB]

You bet I find it interesting and informational, great data.

You may, or may not, find interesting about the balancing of our 4 cell strings X 3 for 618ah.

When I had AGM batteries, I was meticulous about cable length matching, but still got different charge and discharge rates between the two strinsgs.

I checked the litium strings an found the same thing even after top balancing the cells individually.

To get the charge and discharge rates to match I selectively put .02" stainless washers under the 6 cable connections to the 3 batteries and changed the order of cables that were stacked on a connection.

The batteries passive balance at 14.4v and we never charge to there, so thy never have been balanced since I put them in.

The cells will get up to about .030mv difference when charging or discharging, but as soon as the are at low current again, thy stay at under .003mv.

We also have them connected all winter and often used during van work.

I was quite surprised how well they stayed in balance.

Your comments on balancing are reminiscent of the lithium guru we studied back in 2016 - - can't remember his name, but he was in the marine business and wrote much on the topic of lithium. And we remember him commenting that after the initial balance, he never balanced again.

And speaking of lithium which - - as we haven't had to think about it in 8 years, we haven't - - what/which are currently considered the top contenders for RV lithium installations. We're still inclined to replace our GBS cells so we don't have to reconfigure anything, but want to consider all options.

PS: Thinking you meant "volts", above, rather than millivolts. We know you have that 'extra digit' on your multimeter so you can measure .003 volts, but if you've refined your accuracy to .003mV, we're really impressed.

 

[booster]

OOOOPs, definitely a voltage glitch here

I think I have heard and read it done wrong so many times it starts to seem correct....kind of like politics.

Even the the 5 digit meters that read to .001 can be a pain to use sometimes as thy tend to be slower than 4 digit ones to settle on readings.

The SOK batteries report cell voltages to .001v and are more accurate than I expected when compared to the 5 digit meter so I can check plenty well with them on bluetooth. I would need to open the cases to use the meter.

The per battery amps from the BMSs is not accurate, but they do seem to compare to each other well enough to balance bat to bat current for matching. I confirm the matching by carefully using the clamp on meter which is easy to mess up the consistency in crowded wiring areas.

I would be inclined to just replace the cells, unless you are looking for a science project.

 

[BikerMonkey]

The difficulty with replacing what we have with anything other than GBS cells is our BMS system which is completely external to the cells and provides more granular information (e.g. the voltage and temperature of each cell) than many of the 'drop-in' solutions. And our present system (with new batteries) literally allows charging and discharging at any current we desire . . . charge and discharge rates in excess of 500 amperes is possible (well, if we had the ability to generate or consume such high currents). Finally, we have such a "pretty panel of gauges" - - some of which would 'go dead' if we abandoned GBS. :~(

As far as cell voltage tracking goes - - when new, these cells tracked within 0.01 volts (the resolution of our BMS system and external Fluke meter). Cell balancing is performed by charging at 14.2 volts (3.55 volts/cell). When any cell exceeds 3.55 volts - - a 1/2 ampere shunt is enabled across that cell which effectively terminates charging of that cell while the laggard cells in that chain 'catch-up'. In reality, some cells will overshoot the 3.55 volt target. Early on we pondered: "If one cell is sitting at 3.6 volts, and with the total charging voltage set to just 14.2 volts (3.55/cell), how can a laggard cell ever reach the 3.55 volt level?" The answer appears to be that - - at this charging level - - the current in each string is less than 1/2 ampere . . . thus, the shunt across any cell above 3.55 volts actually acts to discharge the 'over-charged' cells causing them to drop back to 3.55 volts while the laggard cells continue to charge to 3.55 volts.

All this said, this is no longer the case with our 8 year old lithiums. After many hours of charging at the balance/equalization voltage of 14.2 volts . . . some laggard cells have not, and will not, ever rise to the 3.55 volt level. Indeed, we have one problematic cell that does not rise higher than 3.35 volts - - a full 0.2 volts below most of the others.

So we experimented. We forced this laggard cell to 3.55 volts (and above) by placing a charger on this one cell. At 10-15 amps, the cell 'conceded' and rose, as noted, to and above 3.55 volts. We weren't surprised, however, when this cell sank back to the 3.35 volt level when our forced charge was removed.

What we've found interesting the past couple of years is that these laggard cells "catch-up" to the good cells during the discharge process. Catch-up?? Maybe we should have said that the good cells "catch-up" (drop down to) the voltage of the laggard cells early during the discharge cycle with all 20 cells, even after 8 years, tending to track lower and lower within that original 0.01 volt range.

At least until this year, we have concluded that whatever causes a laggard cell not to reach the balance voltage of 3.55 volts, this is a "top-end" effect only and that the capacity of the laggard cell(s) remains about the same as the normal cells. This year, however, our laggard 3.35 volt cell was the first to reach the lower cut-off (2.8 volts) - - by 0.15+ volts ahead of the others - - causing the termination of our capacity test.

Hope we answered your question or, at least, you found this interesting.

Winston

Totally off topic, sorry. But my first job in the industry was a technical writer for the Fluke Corp, our motto was “If it works, it’s a Fluke”

Cheers!
Lee

 

[booster]

Totally off topic, sorry. But my first job in the industry was a technical writer for the Fluke Corp, our motto was “If it works, it’s a Fluke”

Cheers!
Lee

Not anymore from what I have seen the last decade after being a lifelong Fluke fan.

I got a 110 series meter thinking it was a real Fluke. Not so. Chinese and with the same guts as the the rest of the the dtuff on Amazon, and it worked poorly from day one. Also got their higher end IR temp gun that had the worst battery compartment I have ever had. Loose batteries that caused leakage and destroyed it in a couple of years just like most others of them based on the reviews. No help from Fluke on either of them.

I now have 2 Bryman meters with similar performance to Fluke but under half the price. 4 digit daily us one and a 5 digit for the picky stuff. For other not super accurate stuff, I have been using Kline.

What are you using these Winston?

 

[BikerMonkey]

Not anymore from what I have seen the last decade after being a lifelong Fluke fan.

I got a 110 series meter thinking it was a real Fluke. Not so. Chinese and with the same guts as the the rest of the the dtuff on Amazon, and it worked poorly from day one. Also got their higher end IR temp gun that had the worst battery compartment I have ever had. Loose batteries that caused leakage and destroyed it in a couple of years just like most others of them based on the reviews. No help from Fluke on either of them.

I now have 2 Bryman meters with similar performance to Fluke but under half the price. 4 digit daily us one and a 5 digit for the picky stuff. For other not super accurate stuff, I have been using Kline.

What are you using these Winston?

No, no it was a joke. When I left Fluke in the 70’s I was given a multimeter, not only does it work flawlessly but it has the same 6v battery in it.