DIY portable, useful LiFePO4

[booster]

Originally Posted by markopolo

I don't know the capacity per mV.



I was alerted to this because I needed to enter accurate voltage values for Full, 80%, 60%, 40% & 20% in the BMS software. I was allowing up to 2 hour-ish recovery periods. I also wanted to know what the 50% SOC voltage was. 50% SOC point had an overnight recovery period and the voltage was higher than what I had recorded for 60% SOC.


You see how easy it was to be 10% off on SOC using voltage.


The getting to within 4mV I mentioned might help to avoid a larger error in my example. 13.164V was peak (I think that device reports a bit lower than true voltage). Lose the 4mV and you'd still see 13.16V. Lose 5mV and that's 13.159 and a lot of meters drop the digits they can't display and you'll end up seeing 13.15V. Suddenly you're off by 14mV from actual peak voltage.

Do any of the built in meters and gauges have capability to accumulate and count amp hours or watt hours?

[markopolo]

Prior to this latest testing, I would had said that the BMS counts amp hours. I'm not sure now. I'm thinking it could be tracking watt hours but displaying it as amp hours. Incorrect voltage values entered in the BMS software can result in incorrect remaining capacity being displayed. You choose the voltage for Full and when that is reached the meter resets.

[booster]

Originally Posted by markopolo

Prior to this latest testing, I would had said that the BMS counts amp hours. I'm not sure now. I'm thinking it could be tracking watt hours but displaying it as amp hours. Incorrect voltage values entered in the BMS software can result in incorrect remaining capacity being displayed. You choose the voltage for Full and when that is reached the meter resets.

That is interesting and probably a good thing for SOC accuracy. If it is totalizing from instantaneous watt readings, those voltages would be independent of any later bounce back voltage, I think, so would be measuring the actual energy used out of the battery. IIRC, the Trimetric literature said they measured watt hours and displayed amp hours by calculation, but that may have been some other brand also.

[markopolo]

It appears that voltage, current and temperature are all factored in for SOC reporting.

One of the subsystems of the BQ76920 chip on the BMS is Measurement: BQ76920 Datasheet https://www.ti.com/lit/ds/symlink/bq76920.pdf -

also:

https://e2e.ti.com/blogs_/b/powerhou...or-your-system

The monitor continuously measures cell voltages, temperature and current through the sense resistor and reports this information to the microcontroller

and

perform some basic gas gauging based on voltage, current and temperature information

Data from the BQ76920 chip is passed to the ATmega328 microcontroller on the BMS: ATmega328 Datasheet
http://ww1.microchip.com/downloads/e...S40002061A.pdf

[Winston]

Originally Posted by markopolo

Voltage rebound made it to 4mV of peak voltage after 7.99 hours.

We observed that after a full charge that our pack had not fully stabilized at a resting 100% SoC value even after 83 hours (the point where we terminated the test at 13.77 volts).

This was a huge surprise as all of our subsequent 10/20% discharge segments seemed to stabilize after 8 hours . . . thus our criterion of at least 10 hours of resting. One point, though, we're using a Fluke 177 having only 10mv resolution . . . so our 'stabilization' implies that the voltage has not toggled 0.01 volts within the last several hours.

[Winston]

Originally Posted by booster

The one that sticks in my mind is that we were told that the basic spec from Apple on phone batteries was 700 cycles . . .

It is our understanding that one of the advantages of LiFePO4 (as opposed to the chemistry used in cell phones) is a higher cycle life. Cell battery (chemistry) places more emphasis on weight and energy density at the expense of cycle life.

[markopolo]

The Jiabaida ( JBD ) BMS I used looks to be readily available in fully assembled batteries in Europe in various capacities:

-> https://www.liontron.de/

-> 12V 200AH Lithium Battery RV200-12V Lithium Battery-SOLMAX NEW ENERGY

-> ULTIMATRON | Wohnmobil Untersitz 12V

I particularly like the cases with removable covers.

https://www.youtube.com/watch?v=fQ7x...=youtu.be&t=30

The Liontron, Solmax/Solarfam apps from Google Play Store work with my battery packs. They are view only so you can't change settings.

View only apps on Google Play:

xiaoxiang BMS - https://play.google.com/store/apps/d...m.jiabaida.bms

LIONTRON LX Smart BMS - https://play.google.com/store/apps/d...baida.LIONTRON

Solarfam - https://play.google.com/store/apps/d...iabaida.solmax

The most up to date version of the app ( 3.1.1015 ) with access to change settings is xiaoxiang 3.1.1015.23 and located here:
http://www.jiabaida.com/xiaoxiang.apk

[markopolo]

Coup de fouet is the term used to describe the initial voltage sag when a load is applied to lead acid batteries. That type of sag is shown here: https://www.classbforum.com/forums/f...tml#post104269

I haven't observed it on full discharge tests on these LiFePO4 cells but have seen it when starting a discharge test from various but not all partial state of charge points after a long rest periods. It's quite minimal, a few millivolts at most but interesting because one of the touted advantages of lithium batteries is no Coup de fouet.

It seems to me that here's a chance the coup de fouet effect I've noticed could disappear after one or more full discharge/recharge cycles. The cells that I observed it on have not been through a full cycle. I'll test for it again after a full discharge/recharge cycle.


EDIT: just adding that coup de fouet is readily observed and a concern with lead acid battery banks when high amperage draws are suddenly applied. It's why a UPS or inverter sometimes goes offline rapidly even when there should be enough capacity and capability in the battery for the task. Restarting UPS or inverter often leads to successfully supplying the load. It's a delay in the electrochemical reaction within the battery and once the battery is awakened it performs as expected.

My testing shows it's also observable with low amp loads. The dip is more stretched out with low amp sudden loading.

[JohnnyFry]

It appears that that well constructed battery is not distributed in North America. Too bad, it looks to be a well engineered product.

[markopolo]

I really like that it's configurable and also like that it's repairable / reusable. You can replace the cells or the BMS. Europe does seem more committed to making things repairable instead of disposable.

[markopolo]

Over-voltage protection works.

Screenshot_20200423-150953.png

2020-4-23-15-11-4-EBD-M05.jpg

Pack Two balance after charging looks OK.

Screenshot_20200423-150638.png

I don't see the need to go all the way to 14.6V. 14.4V seems more than adequate for a fast recharge.

[markopolo]

At approximately 16 hours after a full charge (approx 16 hours at rest), Pack Two voltage is 14 volts. The only load on the lithium battery during this time is the BMS.

This attached chart is an extension of the chart in the previous post. I left the EB Tester program running overnight in monitoring mode:

14 volts.jpg

That gives a good idea of what to expect if you decide to fully charge your lithium battery.

Balance is good, only a few millivolts of difference between cell groups.
Screenshot_20200424-063153.png

[booster]

Originally Posted by markopolo

I really like that it's configurable and also like that it's repairable / reusable. You can replace the cells or the BMS. Europe does seem more committed to making things repairable instead of disposable.

We certainly do have a "throw away" culture here compared to other places in the world, at least from what I have seen. It seems that even stuff that could easily be repaired can't be because the parts aren't available or priced so out of line that it makes repair impractical.



There may also be another thing going on here, also, in that it appears that the U.S. consumers are much more likely to go for "magic box" promises on on products, be it new house windows that "will cut your heat bill in half" but don't say how, to sealed up batteries that may or may not have decent components in them with some often exaggerated claims and distorted comparisons. Throw on a "lifetime no hassle warranty" that really has so many disclaimers in it that it would be to call it "useless warranty" and we line up to buy with blind faith.

[GeorgeRa]

Liontron is about $11/Ah vs my Fullriver AGM about $3/Ah by quoted capacity, but real life will likely be even better, perhaps only 2 x cost of AGM and ½ in weight. Liontron with 5 years warranty and 10 years replacement parts is attractive, perhaps attractive enough to get them from EU.

A side note regarding Fuel Cells - I just got advertising from Fisheries Supply about still kicking methanol Fuel Cell EFOY. https://mailchi.mp/fisheriessupply.c...t?e=5adf90d7ae Still expensive fuel - $0.175/Ah, likely to keep contaminants away from a very sensitive fuel cell device.
https://www.imarineusa.com/EFOY150-9...yABEgLQ-vD_BwE

[Winston]

Originally Posted by markopolo

At approximately 16 hours after a full charge (approx 16 hours at rest), Pack Two voltage is 14 volts. The only load on the lithium battery during this time is the BMS.

'Just for grins' we checked our 'plot' ==> ~13.955 volts after 16 hours. Same conditions, just BMS . . . now only question: What are our respective BMS currents?

[markopolo]

Awesome that you compared it to your data.

Looks like our setups have low losses attributable to their BMS. I charged to 14.40V and 0.25A acceptance. Was your cutoff similar?

10% capacity discharge so 90% SOC on this chart: The red vertical line indicates end of discharge. Voltage recovery appears to be rapid but that's not accurate as I forgot to turn monitoring on immediately after the discharge so there was a delay there.

13331V.png

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

[markopolo]

SOC reporting on the App is accurate enough at 80% SOC. 11.6Ah drawn from the 58Ah battery. 58Ah -11.6Ah = 46.4A. The App reports 46.34Ah remaining just before stopping the discharge down to 80% SOC test.

Screen capture attached:
Screenshot_20200425-095808.png

90% = 3.333 vpc
80% = 3.33x vpc (so far, rest period in progress)

Might end up being 1mVPC (4mV Pack) or less difference between 90% SOC and 80% SOC.

[markopolo]

13.331 90%
13.327 80%

80% SOC ended up only 4mV lower than 90% SOC.
------------------------------------------------------

Discharge to 70% SOC done & now the rest period to see what the voltage recovers to begins.

70% SOC 58Ah battery - 17.4Ah removed = 40.6Ah

App shows 68% and 39.98Ah remaining so still good and very useful/usable information.
Screenshot_20200426-090502.png

My 70% 40.6Ah calculation does not account for any self discharge and any parasitic losses over the last three days so the App is likely more accurate than my calculation.

[booster]

Originally Posted by markopolo

13.331 90%
13.327 80%

80% SOC ended up only 4mV lower than 90% SOC.
------------------------------------------------------

Discharge to 70% SOC done & now the rest period to see what the voltage recovers to begins.

70% SOC 58Ah battery - 17.4Ah removed = 40.6Ah

App shows 68% and 39.98Ah remaining so still good and very useful/usable information.
Attachment 9035

My 70% 40.6Ah calculation does not account for any self discharge and any parasitic losses over the last three days so the App is likely more accurate than my calculation.

When all is said and done, will you have enough data to plot a graph of how much the the various parameters change the voltage against each other and particularly voltage? Basically and error stacking comparison and affect on particular items reliable resolution.



As has been mentioned before, for me one of the biggest interests is to see how the whole thing works together in practice to know what is the best way to use and keep track of the system.


One of the biggest would be just how well controlling everything from voltage works. Might it be fine for stopping charging and low voltage cutoff but not as good for keeping track of state of charge because of things like meter error, temp change influences, drop from loads, etc being to large compared to change in voltage across the SOC range?

[markopolo]

No to the first part. I'd need to be much more organized. And, as you noted, so many, too many variables including temperature.

I think a SOC meter is the only way to have at a glance accurate enough SOC info. The App meter is accurate enough so far. Likely more than 99% accurate if factoring in self discharge and BMS usage. Fingers crossed that this continues.

As you mentioned, voltage is useful for some things. Critical actually for low voltage cutoff. Voltage can give you a rough estimate of SOC in broad terms such as knowing that the batteries are above 80% or below 50% but only if you taken the time to chart the voltage during an accurately measured discharge test and you voltmeter has millivolt resolution and the temperature is similar. I don't think you can use someone else's data and apply it to your batteries. Even the same brand batteries and same meters wouldn't guarantee precise info. Meter calibration would be slightly different & the batteries would be aging differently.

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.