DIY portable, useful LiFePO4

[Winston]

Originally Posted by markopolo

I charged to 14.40V and 0.25A acceptance. Was your cutoff similar?

Elite Power Systems specifies 14.2 volt charge. That corresponds to their 'balancing' system that shunts a 1/2 amp resistor across any cell that has attained 3.55 volts (~ 7 ohms). Over a period of several hours, the higher voltage cells (i.e. those >3.55volts) drop in voltage while the undercharged/lower voltage cells continue to charge. Generally all cells will converge on the 3.55 volt number, but sometimes we raise the charging voltage by a tenth or two to force the laggards to 3.55 volts.

Unfortunately we haven't kept records of current vs. time when charging to 14.2 volts. But the current drops pretty low so that the 1/2 amp shunt load accounts for most of the current through any given cell string (meaning, most of the current is flowing through the shunts, not the cells).

[Winston]

Originally Posted by markopolo

13.331V measured using my multimeter after 14 hours at rest. I think that's lower than yours, even lower than your 80%.

Yes, our 90% SoC has been consistent at 13.36 with 80% registering either 13.35 or 13.36 . . . nice and flat . . .

[Winston]

Originally Posted by markopolo

It's looking like 70% SOC will be 7mV (0.007V) or less below 80% SOC voltage so my results are very different than Winston's for example.

Hard to know for certain since we're not members of the Millivolt Club. :~(

[markopolo]

24 hrs passed before I checked Pack Two's 70% SOC voltage and that's longer than the other rest periods so my data becomes less meaningful.

I looked through the logged data and found that peak recovery voltage occurred after 23.61 hours of rest.

Results so far:

13.331V 90% SOC
13.327V 80% SOC
13.325V 70% SOC (after 24 hrs rest)

[JohnnyFry]

Marco

I have some questions about the response of my 150 A/H battery pack.As I mentioned, I am using a Xiaoxiang BMS. I access the BMS using my iPhone and am able to switch the battery on and off. OK, I had the battery fully charged up when I put it to sleep using the app. My other current/voltage monitor indicated 0.0 amps drain. Checking in 48 hours later the BMS app indicated 80% SOC at 13.40 volts. I started the engine (which I do every few days to offset the parasitic drain on the chassis flooded Pb battery. Within about 30 seconds after the DC/DC charger kicked in the SOC returned to 100%. As near as I can determine the only possible drain should be the BMS doing cell balancing and the Bluetooth module. I am not certain whether the BMS is actually giving me the proper SOC reading when I return and find it at 80%. Thoughts?

[markopolo]

It seems to me that a combination of data points are used for SOC readings. Voltages need to be changed from the defaults to match your particular battery.

Here's a screenshot of a page on the iOS app:

voltages.png

If your cells are like Winston's then 80% would be around 3335 mV.
If similar to mine then 80% would be around 3330 mV.

60% 3300 mV Winston's
60% 3295 mV Mine


Edit: just adding 3290mV could be 40% and 3250mV could be 20%. 3450mV could be used for 100% and balancing start could be 3500mV. Your particular batteries and your preferences might lead to slightly different voltage settings.

[markopolo]

Just an update on the resting voltages found:



13.331V 90% SOC
13.327V 80% SOC
13.325V 70% SOC (after 24 hrs rest)
13.174V 60% SOC (after 24 hrs rest)
13.164V 50% SOC (after 24 hrs rest)
13.160V 40% SOC (after 23 hrs rest)

[markopolo]

Very interesting study on LFP capacity recovery: https://iopscience.iop.org/article/1...45-7111/ab7900

The most effective recuperation method was holding the cells at their lower cut-off voltage.

I would never have thought of holding these cells at 2V (vpc) as a method for capacity recovery.

[Winston]

Fascinating. Wonder to what extent the findings of this study carry-over to the larger, non-circular prismatic cells?

Then, we're uncertain of the precise chemistry of our GBS cells . . . they're said to be LiFeMnPO4 . . . with no mention of the anode material (whether it's graphite). Thus, again, we remain uncertain whether we can extrapolate some of the findings of this study to our system.

Then, there's that reference to holding the cells at 2 volts! Really? Our cells reach the discharge knee at 2.8 volts . . . we've assumed that much below that voltage is 'disaster and destruction'.

But still . . . . some eye-openers. First, it doesn't appear that running full 100% SoC charge/discharge cycles is detrimental, at least for the first 2000-3000 cycles (generally thought to be the life expectancy of lithium anyway). There's virtually no 'recoverable capacity' loss with full-cycle discharges and, even after running a very lengthy (3+ weeks) 'recovery' regime for the fractional-discharge examples, full cycle discharging results in only a very small degradation (as compared fractional cycling) until many thousands of full cycles have been reached. We’re uncertain whether we're going to live long enough to worry about this! But even if ‘technically sound’, we’re not going to adopt 100% SoC cycling as we always want to maintain an SoC ‘reserve’ which implies that ordinarily we don’t want to dip below ~30%.

It seems that the best region for fractional discharge is the 'low end' of the SoC range. But, again, wanting to retain a 'reserve', this is not a region in which we want to spend our time. We thought it was 'bad' to operate too close to 100% SoC, but if so, this study doesn't offer any evidence of that. And, unless one intends on spending multiple weeks 'recuperating' their lithium to recovery lost capacity, for us factional users, who want to retain an SoC reserve, cycling in the upper SoC region appears preferable to the 'mid-latitudes'.

Thanks for posting and keeping us 'entertained' for the morning . . .

[markopolo]

In use operation starting from very high SOC seems safe enough. Storing or holding at high SOC might cause more damage. It's not clear yet how much though.

Scroll 90% down this page: https://marinehowto.com/lifepo4-batteries-on-boats/

“The cells lost 11.6% of their confirmed capacity just sitting at 100% SOC”

"The second test we did went for 13 months, under identical testing criteria, and this cell only lost 3.8% of previously verified Ah capacity. While this is quite a bit less capacity loss it still lost capacity."

The cells won't actually sit at 100% for a year when off charging but they would have been at a high SOC.

Until I understand more about this I'll store the batteries at around 40% SOC or maybe less.

[markopolo]

The battery delivered 99% of expected capacity over the 10 day test.

Discharge test results:

13.331V 90% SOC
13.327V 80% SOC
13.325V 70% SOC (after 24 hrs rest)

13.174V 60% SOC (after 24 hrs rest)
13.164V 50% SOC (after 24 hrs rest)
13.160V 40% SOC (after 23 hrs rest)

13.102V 30% SOC (after 21 hrs rest)
12.960V 20% SOC (after 21 hrs rest)
12.818V 10% SOC (after 20 hrs rest)


Charts showing the 10% decrements:


cliff dive.jpg

slightly earlier view.jpg

[rowiebowie]

Marco, I just want to let you know I'm following your thread. If I haven't posted before it's because most of it goes over my head in the level of knowledge and detail. I could never design and build my own power packs.

I can only hope my Renogy package is doing all the balancing etc. that it is supposed to do. So far it seems that way, wth emphasis on "seems".

[Winston]

We see 2.818 volts at 10% . . . which voltage is very close to our 2.8 'discharge/stop test' threshold . . . How 'low' did you take it to reach 99%?

[markopolo]

Originally Posted by rowiebowie

Marco, I just want to let you know I'm following your thread...........

Thanks Rowie.

Originally Posted by Winston

We see 2.818 volts at 10% . . . which voltage is very close to our 2.8 'discharge/stop test' threshold . . . How 'low' did you take it to reach 99%?

It went to the 10.5V low volt cutoff and the BMS took the battery offline so one cell group would have breached 2.625V.

With this pack, there's not much energy below resting 12.8V (it's a cliff dive from there) and no reason for me to go below that other than possible voltage sag from high amp loads. I've since set the low voltage cutoff to 11.6V.

[markopolo]

Charging to 14V with short absorption and balancing:

I charged Pack 1 & 2 while paralleled today. They were below 50%. Charger was set to 14V & 19A. Full current was accepted until just before reaching 14V. Once 14V was reached, current tapered off. The following absorption period permitted balancing to do it's job and by 2A or so tail current, all cells, (both packs) had reached 3.5V. It seems like a gentle & easy way to get a full charge and all cells balanced to within 5mV.

crop.png

charge to 14V current taper.jpg

charging to 14V.jpg

[markopolo]

LFP Pack 3 assembled. I had 12 cells leftover so decided to go ahead build a smaller capacity pack rather than ordering more cells.

Pack 3 4x5 inches.JPG

The capacity of the 4S3P pack is 19Ah & weight is just under 8 lbs. It has the same 150A BMS but I won't try to draw that current from the pack! I might try to run the microwave oven in the van from it. That would be a 5C draw.

I'm very happy to be past the assembly stage! Testing, calibrating, figuring out where to use them etc. is next & more fun.
Pack 1 2 3.JPG
Pack 1 2 3__.JPG


Total Power Capacity of all three would be very near 135Ah so more than 100Ah available if using 80%.

Total Power Capability would be over 300 Amps so more than 3,600 watts.

[markopolo]

Just a project note:

On May 10th I discharged all three packs to under 50% SOC for storage (coulomb counting). That's around 13.16V at rest for these packs.

I checked them all today - just over three weeks later - and they're all still at 13.16V.

I was very pleased to see that. So far, this 150A JBD BMS has met or exceeded my expectations.

There has been much discussion on the forum about some lithium systems very high parasitic losses. The parasitic losses on these packs are so small that they are not noticeable. It's worth noting that the packs were all ON and in a ready to use state the entire time.

[Davydd]

It is my understanding that if batteries are totally disconnected that lithiums have less parasitic loss than AGMs.

[booster]

Originally Posted by markopolo

Just a project note:

On May 10th I discharged all three packs to under 50% SOC for storage (coulomb counting). That's around 13.16V at rest for these packs.

I checked them all today - just over three weeks later - and they're all still at 13.16V.

I was very pleased to see that. So far, this 150A JBD BMS has met or exceeded my expectations.

There has been much discussion on the forum about some lithium systems very high parasitic losses. The parasitic losses on these packs are so small that they are not noticeable. It's worth noting that the packs were all ON and in a ready to use state the entire time.

Was that with BMS connected the whole time? If so, very good.

[markopolo]

Originally Posted by Davydd

It is my understanding that if batteries are totally disconnected that lithiums have less parasitic loss than AGMs.

You might be thinking of self-discharge of the actual LiFePO4 cells that make up the battery. That should be somewhat similar among all brands as it would be a property of the LiFePO4 chemistry variatons.

The parasitic losses I am referring to are caused by phantom loads and in my example it's the load from the BMS itself while it monitors and protects the cells.

That type of phantom or parasitic load will vary greatly depending on the parts used.

Example from the Ecotrek manual:

https://www.roadtrek.com/wp-content/...il-24-2017.pdf

Battery Management System
Each EcoTrek Power Module has its own Battery Management System or BMS. The BMS for each EcoTrek Power Module draws about 60 watts of power. For this reason, it's best to only turn on those EcoTrek Power Modules you need to use.

So you see that you have to turn those type of packs off else they'd soon get depleted. Forget to turn it off and you end up with fully discharged module. Don't forget that when in use 60 watts or 5A or so is used by the BMS so that capacity is lost.

The JBD BMS I referred to uses such a small amount of current while active that you'd need at least millivolt resolution to see the decline after a week of being left on.

Originally Posted by booster

Was that with BMS connected the whole time? If so, very good.

Yes, connected the whole time & Bluetooth module plugged in. Fully available for use and/or monitoring.
-------------------------------------------------

I just started a new test. Pack 3 is paralleled via Schottky diodes to a 13yr old Group 27 wet cell (Walmart brand). I plan to note the voltage of each over the next few days. Once that's done I'll connect a PD charger to the Group 27 wet cell and see if that affects the connected Pack 3 over a few days.


parallel via diodes.jpg