Visit our Popular Forums

OceanSeaSpray · 22-04-2016, 02:57

Quote:

Originally Posted by Anjin

I'm not really concerned about reaching "full" charge. My experience charging these packs is that when the pack reaches 13.8 volts, assuming cells are reasonably well top balanced, terminating the charge will cause the cells to drop back to around 3.35 or 3.4 volts per cell. At 3.3 volts per cell we are well into the "knee" above which there is no substantive charge - they will continue to accept current and the voltage may climb, but you've already got most of what you're going to get in capacity. See the attachment that shows the discharge curves for my CALB cells. And even if I am missing the top 2-5%, that's ok with me if it means my (rather expensive) cells stay out of the high charging "knee". There aren't a lot of certainties yet with LiFePO4 but there seems to be a growing consensus that cell life gets shortened by spending unnecessary time in that knee.
- Art

This is incorrect, if you charge at fairly high current and don't absorb, there is very substantial capacity left behind, especially with a termination voltage as low as 13.8V. Terminating without absorbing was the first thing I did and the myth didn't stand to the test. It does work much better if the C-rate is low.
Also, if the matter is charging, you need to use a charge curve, not a discharge one. They are different. It also needs to match the charge rate you are using.

And 3.3V is nowhere near the knee. The voltage is no absolute indication of the knee approaching, because the higher the current, the higher up the knee moves as well. As I wrote earlier, I have charged cells at 3.6V and they still weren't anywhere near the knee. The knee is where you are about to run out of additional capacity.

OceanSeaSpray · 22-04-2016, 03:16

Quote:

Originally Posted by Anjin

Thanks for clarifying where that recommendation came from. I knew that Rolf had experienced some difficulties but I also know that several GBS banks based on less than 200AH cells also suffered some substantial capacity losses in a few short years. I also know that Maine Sail has had a 400AH cell 4S battery that has suffered virtually no capacity loss in over 5 years use and more than 500 cycles. Not a lot of consistent data to build on.
- Art

What happened to Rolf and loss of capacity are two completely unrelated matters. Rolf ended up with cells that have enormous leakage and self-discharge, i.e. they are shorting internally. This all points out to mechanical damage.

Unless I am mistaken, Maine Sail operates one stationary test bank (the older one) and another more recent on his boat, but it is also near-stationary compared to what happens on passage. Those are not relevant data points when it comes to mechanical robustness.
This being said, the Winston 400Ah cells are long and thin and they might be at an advantage mechanically compared to many others that are more square.

TurninTurtle · 22-04-2016, 03:18

Got very lucky... and was GIVEN four LiFeMgPo4, 40 ah 12.8v (nominal) batteries.
Outdated medical equipment batteries that had never been installed.

Unfortunately these things are extremely expensive to buy.
Valence U27-12XP 12.8V Lithium Ion Battery type LiFeMgPO4

4 years later, these are still performing at rated capacity and current delivery. (That's 4 years AFTER labeled expiration...)

I'm using them in a small solar power system with 120 watts of panels for charging large RC electric power models. My system can supply ALL of the charging for everyone in the club.
Kind of a waste of the capability of these things...

If you know someone in medical equipment maintenance, ask about batteries they are going to discard.

Maine Sail · 22-04-2016, 03:47

Quote:

Originally Posted by OceanSeaSpray

This is incorrect, if you charge at fairly high current and don't absorb, there is very substantial capacity left behind, especially with a termination voltage as low as 13.8V. Terminating without absorbing was the first thing I did and the myth didn't stand to the test. It does work much better if the C-rate is low.
Also, if the matter is charging, you need to use a charge curve, not a discharge one. They are different. It also needs to match the charge rate you are using.

And 3.3V is nowhere near the knee. The voltage is no absolute indication of the knee approaching, because the higher the current, the higher up the knee moves as well. As I wrote earlier, I have charged cells at 3.6V and they still weren't anywhere near the knee. The knee is where you are about to run out of additional capacity.

Bingo! Just like with LA's charge current plays a role. This is why most prismatic cell manufacturers recommend 0.3C, or 30% of Ah capacity, as the "optimal" charge current or about 120A for a 400Ah bank. Sure they can "take" more current but most suggest the optimal as 0.3C..

If you let an LFP bank absorb at 13.8V to 14.0V it will get as fully as you'd need. I can still take over 400Ah out of my 400Ah bank (at a 30A constant current discharge and the average vessel load is below 10A) by allowing them to taper to less than 10A at 13.8V.

OceanSeaSpray · 22-04-2016, 04:07

Quote:

Originally Posted by evm1024

Way back a few pages (post 5057) I made a comment about the voltages on my Winston 700 ah bank being some 60 mV different from the published voltage and that I "should" do some testing. Well, I did do some testing, And here is what I found.

I charged my bank up to what I thought was 100% (but now realize was only about 80% as explained in post 5127). I have not ever run a full capacity test on the bank and generally cycle around 60% to 40%.

Starting the test I charged the bank up until it reached 13.750 volts at around 30 amps. Then letting the bank rest for a day I measured a pack voltage of 13.358 volts. The cells were (3.340, 3.339, 3.338, 3.341)

Then I would run a space heater drawing 100 amps or so for the length of time that it took to drop the bank by 70 AH (as measured on a victron BMV-602s). Then let the bank rest 24 hours measure the voltage again then draw another 70 AH.

0 AH 13.358v (3.340, 3.339, 3.338, 3.341) Assumed 100% SOC
-70 AH 13.257v (3.317, 3.308, 3.310, 3.321)
-140 AH 13.202v (3.302, 3.299, 3.300, 3.301)
-210 AH 13.185v (3.296, 3.295, 3.295, 3.295)
-270 AH 13.152v (3.290 3.284, 3.283, 3.291)
-350 AH 13.036v (3.262, 3.253, 3.254, 3.266)
-420 AH 12.890v (3.226, 3.214, 3.216, 3.232)
-490 AH 12.090v (3.090, 2.937, 2.924, 3.143) Terminate test

Based on post 5127 and on fitting the data above to the open circuit voltage curves I would say that I was only charged up to around 80% and that the curves I've seen fit the published curves closely in the 20% to 60% SOC range with my (used) cells rising faster above 60% SOC.

I did note that the charge voltage did rise significantly at higher charge rates (100 a) as I got to the 80% SOC area. This of course defines the transisition point from CC to CV.

I do not have a power supply that can do more than 30 amps in CV so I was manually adjusting my inverter/charger to mimic a CV charger. I did this with a 3.5v max cell voltage. It there a better cell voltage to use for the CV range? Cell #3 was at 3.475v while cells #1 and #4 were at 3.408v. This was at 30 amps or so.

Regards!

Good on you. That is a lot of time and effort.

At 12.09V you would have been very close to the bottom, so 490Ah drawn out of 700Ah nominal capacity would suggest that they might only ever have reached ~70% SOC. If that much, because the 700Ah rating would normally be for a higher discharge rate than C/7 and I would have expected the cells to outperform the 700Ah figure.

You perfectly illustrated that they need absorption to charge properly and low-ish voltages are not effective. 30A was only 0.04C, but 13.75V is just too low.
I had carried out a number of charge termination experiments initially and I now absorb at 14.0V whenever a full charge cycle gets triggered. This is then followed by a long, slow deep cycle that gets extended by trying to hold 13.4V whenever solar power is available.

Thanks for posting the data. I will plot it to look at it more closely again later.

Maine Sail · 22-04-2016, 04:12

Quote:

Originally Posted by Anjin

I also know that Maine Sail has had a 400AH cell 4S battery that has suffered virtually no capacity loss in over 5 years use and more than 500 cycles. Not a lot of consistent data to build on.

- Art

Art,

I am in excess of 765 cycles now and still exceeding 400Ah (at a .075C discharge rate limited by my CC dischargers ability) but keep in mind that my cells have;

--Never been floated, they get charged, then discharged
--Only absorbed to a net 8A - 10A at 13.8V -14.0V
--Not charged above 14.0V unless for testing purposes (I now have a few other banks for that)
--Max charge rate at approx .3C
--Stored at 50% SOC when not being used or cycled
--Stored in 45-60F temps when not being used or cycled
--Only very rarely exceeded 80F
--Highest voltage they have ever seen was 3.8VPC while top balancing initially.

On my last few capacity tests I ran five 100% DOD tests back to back and noted that I gained nearly 14.5 Ah's of capacity so there seems to be some truth to what T1 was suggesting a few pages back. I actually captured one of them with an intervalometer, but I need to figure out how to convert it into a .mov etc..

I suspect I am not seeing as much gain with the back to back 100% DOD cycles because I am normally discharging to 80% DOD & then recharging at .23 - .3C but with five back to back Ah capacity tests at .075C I did see an increase of about 14.5Ah's before it stabilized.

Also keep in mind that my bank is an n=1.... Cells vary and every boater will treat them differently. I have customers with over 12 years on GEL batteries and others who murder them in 2 years.

I am treating my LFP bank as I suspect they do well with in a PSOC environment, and it seems to be working, but an n=1 really tells us little in the whole scheme. The frustrating part is that very few of us are actually capacity testing, with repeatable equipment, to know what is actually happening with our LFP banks. Until more fractional C data in a PSOC use situation comes out, it's learn it as we go.

OceanSeaSpray · 22-04-2016, 04:33

Quote:

Originally Posted by Maine Sail

Bingo! Just like with LA's charge current plays a role.

It does indeed... high current makes the battery voltage read higher than it really is in terms of state of charge (because of the internal resistance) and by the time the system gets into CV-mode (limiting the voltage), the SOC is not as high and more capacity needs to be absorbed.

Maine Sail · 22-04-2016, 05:22

Quote:

Originally Posted by OceanSeaSpray

It does indeed... high current makes the battery voltage read higher than it really is in terms of state of charge (because of the internal resistance) and by the time the system gets into CV-mode (limiting the voltage), the SOC is not as high and more capacity needs to be absorbed.

If you stick around .3C the taper is pretty quick, hours and hours faster than LA...... I have also noted that the rate at which you discharged the battery has some effect on how the battery performs during charging. I have not sat down to analyze it yet but at frac C discharge rates the battery voltage seems to rebound a bit easier than the same charge rate after a higher discharge rate......

Anjin · 22-04-2016, 08:09

Quote:

Originally Posted by evm1024

I charged my bank up to what I thought was 100% (but now realize was only about 80% as explained in post 5127). I have not ever run a full capacity test on the bank and generally cycle around 60% to 40%.

Starting the test I charged the bank up until it reached 13.750 volts at around 30 amps. Then letting the bank rest for a day I measured a pack voltage of 13.358 volts. The cells were (3.340, 3.339, 3.338, 3.341)

I'm surprised at the amount of voltage "settling" you experienced and it's clearly something I should test for when my configuration is fully assembled. Based on my experience with other packs I expected it to drop from 13.8 to around 13.5 while yours dropped all the way to 13.35 - and that last .15 volts is pretty significant. I'll post my results here in a few weeks.

Quote:

Originally Posted by evm1024

Then I would run a space heater drawing 100 amps or so for the length of time that it took to drop the bank by 70 AH (as measured on a victron BMV-602s). Then let the bank rest 24 hours measure the voltage again then draw another 70 AH.

0 AH 13.358v (3.340, 3.339, 3.338, 3.341) Assumed 100% SOC
-70 AH 13.257v (3.317, 3.308, 3.310, 3.321)
-140 AH 13.202v (3.302, 3.299, 3.300, 3.301)
-210 AH 13.185v (3.296, 3.295, 3.295, 3.295)
-270 AH 13.152v (3.290 3.284, 3.283, 3.291)
-350 AH 13.036v (3.262, 3.253, 3.254, 3.266)
-420 AH 12.890v (3.226, 3.214, 3.216, 3.232)
-490 AH 12.090v (3.090, 2.937, 2.924, 3.143) Terminate test

Based on post 5127 and on fitting the data above to the open circuit voltage curves I would say that I was only charged up to around 80% and that the curves I've seen fit the published curves closely in the 20% to 60% SOC range with my (used) cells rising faster above 60% SOC.

This is incredibly valuable test data - thank you for sharing it! Yes, it looks like you consumed 490AH and probably had some left so that seems like it would be close to 80%. That could be from lack of charge or just a reduced capacity from used cells - I'm not sure how you would tell unless you had an earlier baseline. In any event I would not be unhappy with cycling my pack only up to 80% initially until I full understand it's behavior. Again, thanks for sharing - I will try to return the favor.

Quote:

Originally Posted by evm1024

I did note that the charge voltage did rise significantly at higher charge rates (100 a) as I got to the 80% SOC area. This of course defines the transition point from CC to CV.

Yes, although the rapidly rising charge voltage is usually an indication that you are entering the charging "knee." My goal is to find a charging voltage that will allow the charging sources (primarily solar and genset driven inverter/charger combinations - alternator charging is pretty minimal) to shift into CV mode a bit before entering that knee.

Quote:

Originally Posted by evm1024

I do not have a power supply that can do more than 30 amps in CV so I was manually adjusting my inverter/charger to mimic a CV charger. I did this with a 3.5v max cell voltage. It there a better cell voltage to use for the CV range? Cell #3 was at 3.475v while cells #1 and #4 were at 3.408v. This was at 30 amps or so.

3.5 cell voltage is charging at 14 volts for the pack which seems pretty reasonable. I have a bench supply rated at 0-15V and 0-60A so I'll try some experimentation on my specific cells and let you know what I see. Again probably about 2 weeks away from starting that testing.

Thanks -

- Art

CarstenWL · 22-04-2016, 08:36

Quote:

Originally Posted by evm1024 View Post
I charged my bank up to what I thought was 100% (but now realize was only about 80% as explained in post 5127). I have not ever run a full capacity test on the bank and generally cycle around 60% to 40%.

Starting the test I charged the bank up until it reached 13.750 volts at around 30 amps. Then letting the bank rest for a day I measured a pack voltage of 13.358 volts. The cells were (3.340, 3.339, 3.338, 3.341)

Quote:

I'm surprised at the amount of voltage "settling" you experienced and it's clearly something I should test for when my configuration is fully assembled. Based on my experience with other packs I expected it to drop from 13.8 to around 13.5 while yours dropped all the way to 13.35 - and that last .15 volts is pretty significant. I'll post my results here in a few weeks.

I did the same on my 4x 400Ah Winston cells a few days ago when I was testing my Victron charger.

I charged the bank until the current went down to 5 amps at 13.80 volt.

Then letting the bank rest for a day I measured a pack voltage of 13.41 volts. The cells were 3.353, 3.352, 3.353 and 3.354V

Than I took 12A for 11 hours from the bank.
Pack voltage was 13,12V than, the cells were at 3.288, 3.276, 3.284, 3.273V

Letting the bank rest for the night (12 hours) without any load I measured a pack voltage of 13.29 volts. The cells were at 3.327, 3.327, 3.311 and 3.327V

9 days later still without load pack voltage is 13.20V with the cells at 3.300, 3.301, 3.299, 3.300V

FWIW ...

Carsten

Anjin · 22-04-2016, 08:45

Quote:

Originally Posted by Maine Sail

Art,

I am in excess of 765 cycles now and still exceeding 400Ah (at a .075C discharge rate limited by my CC dischargers ability) but keep in mind that my cells have;

--Never been floated, they get charged, then discharged
--Only absorbed to a net 8A - 10A at 13.8V -14.0V
--Not charged above 14.0V unless for testing purposes (I now have a few other banks for that)
--Max charge rate at approx .3C
--Stored at 50% SOC when not being used or cycled
--Stored in 45-60F temps when not being used or cycled
--Only very rarely exceeded 80F
--Highest voltage they have ever seen was 3.8VPC while top balancing initially.

On my last few capacity tests I ran five 100% DOD tests back to back and noted that I gained nearly 14.5 Ah's of capacity so there seems to be some truth to what T1 was suggesting a few pages back.

Yes, I've read your posts and your "musings" very carefully and would like to try to largely mimic your approach (although my boat will be in the Bahamas so the ambient temps will be a bit higher). I'll be experimenting with the charger settings and the cutoff voltages to see how much absorption I can get in an automated way and still stay out of the higher voltage parts of the curve for any extended period.

Quote:

Originally Posted by Maine Sail

If you stick around .3C the taper is pretty quick, hours and hours faster than LA...... I have also noted that the rate at which you discharged the battery has some effect on how the battery performs during charging. I have not sat down to analyze it yet but at frac C discharge rates the battery voltage seems to rebound a bit easier than the same charge rate after a higher discharge rate.

Unfortunately with my "Rolf sized" bank (1440AH) I would be hard pressed to reach .3C. I will come close when running the genset but will be lucky to get more than .1C from solar. My thinking (for now) is that I may want to settle for 70-80% charge rather than letting it sit at higher charge voltages for an extended period of time. One of the many good things about LiFePO4 is that you don't hurt them by not topping them off - it's something I still have trouble adjusting to.

Thanks!

- Art

evm1024 · 22-04-2016, 09:17

Quote:

Originally Posted by OceanSeaSpray

Good on you. That is a lot of time and effort.

At 12.09V you would have been very close to the bottom, so 490Ah drawn out of 700Ah nominal capacity would suggest that they might only ever have reached ~70% SOC. If that much, because the 700Ah rating would normally be for a higher discharge rate than C/7 and I would have expected the cells to outperform the 700Ah figure.

You perfectly illustrated that they need absorption to charge properly and low-ish voltages are not effective. 30A was only 0.04C, but 13.75V is just too low.
I had carried out a number of charge termination experiments initially and I now absorb at 14.0V whenever a full charge cycle gets triggered. This is then followed by a long, slow deep cycle that gets extended by trying to hold 13.4V whenever solar power is available.

Thanks for posting the data. I will plot it to look at it more closely again later.

The cells are sitting at 70% or 80% SOC right now. No real discharge after recharging after the test. I knew I was near the lower knee and was watching closely and a bit edgy. And recharged them after the 24 hour setting period.

I think that I will go ahead and set my bench supply to 14.0 volts for an absorption phase. I do need to get shorter/heaver cables (currently #8 or #6 wire but 6' long) as that I do see a voltage drop.

My guessing that the bank was charged to 80% before the test was due to curve fitting. It appeared that the results fit the OV discharge curve best at 80% but looking at the lower knee position gave 70% as well. Who knows at this point.

If I assume that they are at 80% then my bench supply at 14v and 30 amps would take a minimum of 4.6 hours to reach 100% (put the last 20% or 140 AH in) but of course it will not be constant current at 30 amps so it could take many more hours to reach less than 10 amps at 14v.

With 70% SOC we could be looking at 7 hours minimum....Perhaps overnight.

As was noted it is a bit of effort to set aside 2 hours every night for 10 days to conduct this test - and no using the boat in the test period! But I may give it a go again after recharging.

I do think that I might like to make a charge curve. However I don't have a good way to measure the AH into the bank. No wait - I could reverse the leads on the BMV-602S shunt. Then rather then reading AH discharge it would read AH charge. I can't think of a problem with that. The internal ground reference in the BMV602s will be reversed but I would have to see the circuit to give a guess as to what it will do oddly. Anyone hook one up backwards? If nothing else I could power it off an isolated battery and just let it do the differential on the shunt.

But the charge curve will have to wait. I really do not want to be 20 days without the boat.

--Ethan

Anjin · 22-04-2016, 09:31

Quote:

Originally Posted by CarstenWL

I charged the bank until the current went down to 5 amps at 13.80 volt.

Then letting the bank rest for a day I measured a pack voltage of 13.41 volts. The cells were 3.353, 3.352, 3.353 and 3.354V

Than I took 12A for 11 hours from the bank.
Pack voltage was 13,12V than, the cells were at 3.288, 3.276, 3.284, 3.273V

Letting the bank rest for the night (12 hours) without any load I measured a pack voltage of 13.29 volts. The cells were at 3.327, 3.327, 3.311 and 3.327V

9 days later still without load pack voltage is 13.20V with the cells at 3.300, 3.301, 3.299, 3.300V

FWIW ...

Carsten

So your settling voltage was similar to Ethan's. Great data to have. Thanks!

It also looks you went from a nearly "full" charge to about two thirds charge (400AH-131AH) while losing ~.07 volts (~.053 resting volts) from the cells. If the curve is as linear as they illustrate I'd say you have a long way to go (at least the other two thirds of charge) before you hit the discharge knee. Will you continue the test to see?

Thanks!

- Art