My Lithium Upgrade

[rowiebowie]

Originally Posted by @Michael

The discharge curves that I've seen show a quick drop to approx 13.3v then a gradual, near linear taper down to 12.5 or so. See the 10A discharge curve here:

https://battlebornbatteries.com/comp...teries-series/

I don't think that one can necessarily associate 13.3v with 80% SOC.



Renogy would be the best source for this.

Assuming passive top balance, my thinking is that if the cells are already balanced, then the time needed to balance would be minimal and the battery voltage would rise from 14.0 to 14.6 very quickly.

BTW - the recent YouTube video where Will interviews the president of Battleborn is informative. Seems like much of the conventional wisdom related to LiFePo4's isn't as is often presented on the Internet.

Thanks, Michael. The charging profile in the Battleborn charts do show a very steep (nearly vertical) rise between the high 13's and 14.4 volts. This is similar to what I experienced. But wow, did mine occur within a couple of minutes.

[booster]

Originally Posted by rowiebowie

Thanks, Michael. The charging profile in the Battleborn charts do show a very steep (nearly vertical) rise between the high 13's and 14.4 volts. This is similar to what I experienced. But wow, did mine occur within a couple of minutes.

That is what we would expect, to see the voltage climb very quickly, but what is suspicious in your case is that the current is dropping as fast as the voltage is rising like is shown on most of the lithium charge profiles. Unfortunately, the Battleborn profile doesn't show current, so that wouldn't be obvious from their literature.

[rowiebowie]

Originally Posted by booster

I think the test would be to run the same test at higher and lower SOC starting points from this test, and see if and where the voltage changes.

Originally Posted by markopolo

Can you force the charger to have another go at it? Hopefully you'll see current tapering. After that you could set the monitor to 100%. It probably won't be actual 100% but much closer than what it shows now.

Renogy should be able to tell you what voltage balancing kicks in.

I definitely agree I need run more charge cycles from different SOC starting points. It will not be hard to do, but what will be hard is observing the voltage and charge changes.

I had pleasant 70 degree weather to sit and observe my recent test, but really don't know if I want to devote that time again. Plus, I was just lucky to be paying attention when the final voltage sure happened.

To the question of forcing a recharge, it certainly does not happen every time I plug in. Observations from our trip were that the SOC needs to be near 70% (possibly just below 80%). I really wasn't paying close attention other than it also sometimes needed a load (microwave) to trigger the charge cycle. So could charging parameters also include voltage readings?

Since the SOC monitor is just that, a monitor, the charging SOC is determined in the Inverter/Charger. It's display does not show a numerical % but rather a battery symbol with bars. 4 bars fills the battery symbol and represents full charge. But there is no way to know within +/- 25% since 4 bars could be just about ready to change to 3. The charger must have an internal monitor that is more precise than that, right?

I want to add that the system worked great and we are very satisfied with it. Further testing is just for inquiring minds and to know we are seeing actual capacity numbers. Well that, and to ensure we are treating the batteries well in order to extract the most capacity and life cycles from them.

I'll contact Renogy after a few more discharge/charge cycles.

I appreciate all the input.

[booster]

What you are seeing may actually not be all that far off, it this more zoomed in charge profile is correct as it shows more offset before the current drops compared to many others.





This profile shows the current not dropping until the voltage hit setpoint plateau, which is more like you are seeing.


The 4 light monitor is probably just triggering off voltage, without knowing load so could be way off. Your battery monitor, if setup properly and running on a shunt will be much more accurate for SOC.



In this case the charger is possibly just holding current steady and the 3.4v is what it can hold, but it is still odd that is was so steady voltage until the end if it wasn't being controlled.

[markopolo]

It does seem like the batteries were really close to the point where current would drop dramatically. They would have been very close to 100% with the 14.6V set-point. But why only 13.3V in the morning assuming no loads overnight? I would have expected 13.4V or higher.

Maybe just meter inaccuracy?

[Winston]

Originally Posted by rowiebowie

You could almost blink and miss it.

Your chart looks normal to us - - it is what we'd expect for a 50 ampere Constant Current charger. The "blink" interval you referenced, again, seems normal . . . the voltage will rise very sharply if you attempt to force large amounts of current into a fully charged lithium battery - - this being the so-called 'knee' region.

We question the use of a Constant Current charging profile, particularly in this region (nearing or fully charged) of the charge cycle. We think Constant Voltage is the better approach.

Indeed, many (most?) balancing systems (at least those using the larger prismatic cells) won't balance at such high currents. These systems operate on the principle of 'shunting' current past cells that have reached a predetermined voltage. In the case of our GBS 100ah cells, this current is just 0.5 amperes. We know other systems shunt more, but, still, the amount is just an ampere or so.

When applying an appropriate Constant Voltage (in our case, 14.2 volts), as the cell reaches full charge (i.e. the point when that shunt kicks-in), the cell current has dropped dramatically, often to less than 1 ampere (thru each 100ah cell string). It is easy to see how balancing works under these conditions . . . as most of the current through the fully charged cell is shunted while the laggard cells are receiving the full current.

Now consider the situation you describe. If we were to 'force' 50 amperes through our pack (of 5 parallel strings), this would constitute 10 amperes thru each string. Shunting a mere 0.5 ampere past a fully charged cell leaves nearly all of the current, 9.5 amperes, still flowing through that cell. These cells would rise to dangerously over-charged voltages long before the laggard cells had a chance, at just 5% more current, to catch up.

We agree with all the comments that caution, 'do not talk about voltage' of a lithium cell/battery in the context of a cell under charge or load. The reason is this: Our measurements reveal that under a sustained load (of 10 amperes), the terminal voltage of the lithium pack can be as much as 0.1 volts below its otherwise resting voltage. But in lithium, a mere 0.01 volt difference can represent more than a 10% difference is SoC. See the attached table of lithium Resting Voltages vs. SoC, in particular, note that the 80% and 90% SoC resting voltages are either the ‘same’ or within 0.01 volts of one-another. Thus, the comparatively 'huge' swings in terminal voltage of a lithium battery under charge or load overwhelms and masks those 1/100th volt 'deltas' necessary to see SoC charge levels. (Incidentally, as we've found that it takes nearly 10 hours for the cell voltage to ‘settle-in’ after a discharge segment, our resting voltages are taken at least 10 hours after completely disconnecting the pack from charging, discharging [including metering ‘overhead’], devices.)

While we’ve been getting a very repeatable 13.36 volts for 90% SoC, our 100% SoC numbers seem to wander so we don’t know what the resting voltage of a fully charged lithium pack is. And a further problem with your “Morning After” measurement of 13.3 volts is, first, you’re missing that critical 4th digit - - is it 13.30 or 13.39 volts? Further, if your pack remains attached to any of your charging/metering or other devices, there remains some residual ‘overhead’ drain which, while small, could impact the integrity of that next day reading.

In an attempt to answer you questions . . . the sudden rise in pack voltage is a strong indication (assuming balance) that you’ve reach full charge. Not knowing your battery balancing system, we hesitate to comment on the sufficiency of this ‘Wonder 3 Minute Treatment” - - our gut feeling is that this does not guaranty balancing. But the good news is . . . evidence abounds that frequent balancing is not required.

[rowiebowie]

Originally Posted by markopolo

It does seem like the batteries were really close to the point where current would drop dramatically. They would have been very close to 100% with the 14.6V set-point. But why only 13.3V in the morning assuming no loads overnight? I would have expected 13.4V or higher.

Maybe just meter inaccuracy?

Originally Posted by Winston

Your chart looks normal to us - -

While we’ve been getting a very repeatable 13.36 volts for 90% SoC, our 100% SoC numbers seem to wander so we don’t know what the resting voltage of a fully charged lithium pack is. And a further problem with your “Morning After” measurement of 13.3 volts is, first, you’re missing that critical 4th digit - - is it 13.30 or 13.39 volts? Further, if your pack remains attached to any of your charging/metering or other devices, there remains some residual ‘overhead’ drain which, while small, could impact the integrity of that next day reading.

I agree. I would like to see 13.4v to ensure 100% charge.

It is disappointing that the Renogy SOC monitor doesn't go to hundredths when showing voltage, because it does when displaying amps. I just did a check 42 hrs. after charging ended.

SOC Monitor = 13.3v
Multimeter = 13.36v
Cheap Amazon 12v plug in monitor = 13.36v
Victron Bluetooth Solar Controller (solar off) = 13.34v

This max .02 voltage difference is consistent with what I've found in the past.

[Winston]

Well, for our GBS cells, that represents 80-90% SoC. Maybe it's time to play the fun game of measuring battery capacity . . . get yourself an adjustable resistor (10 amp load) and an alarm clock . . . you could run 4 x 5 hour discharges (followed, each time, by a 10 hour rest period . . . we know you said that you don't want "to devote that much time" but, heck, aren't you a little curious? Then, you'll have a benchmark should you desire to verify how your pack is doing in the future.

[rowiebowie]

Originally Posted by Winston

you could run 4 x 5 hour discharges (followed, each time, by a 10 hour rest period . . . we know you said that you don't want "to devote that much time" but, heck, aren't you a little curious? Then, you'll have a benchmark should you desire to verify how your pack is doing in the future.

I'm retired. I ain't got that much time.

[markopolo]

Winston's opinions, methods and approach to care of these batteries seems like good advice to me.

-------------------------------

I thought the term I needed to describe why slowing at least the finish is a better approach was intercalation but I not finding support for that usage right now. Perhaps it's the process of intercalation. I've read through a lot of stuff and was left with the impression that charge process (ability of the battery to convert charge energy) slows as the LiFePO4 battery nears being fully charged.

Scroll all the way down to the last graph on this page: https://lygte-info.dk/review/Review%...dule%20UK.html

Shown here for convenience:



That tiny charger charges to 3.6V VPC (think 14.4V for our purpose). Some would say that will result in a fully charged battery. It doesn't at the presumed rate of at least 0.3C. You can see it in the graph.

It appears to restart charging at around 14V (converted for our purpose). That's too soon for anything other than items that need to always be near 100% charged. You can see that the restarts begin immediately and so often that it creates a solid red block. You can also see that the delay between restarts grows as the battery actually gets full. That's the really interesting part.

IMO, that graph shows why slowing the charge process as Winston does results in a more fully charged LiFePO4 battery. He doesn't specifically use a CC phase or an absorption phase because the more gentle approach of charging to a specific voltage at a utilizable rate results in a full or near full battery.

[booster]

I wonder, looking at Marko's graph, if the restart oscillation that we we is actually what is being called top balancing, but isn't just done at a very low current like is described in some of the places. Apparently, the current reduction is needed because many/most of the batteries aren't really absorbing much energy, just the ones that need the time to get full. In the on/off restart they may just be using some sort of capacitance in the system or surface charge to basically pulse charge to reduce the current. As the lowest batteries get balanced, the stored pulse takes longer to get used because fewer batteries are needing it. Most articles on balancing and memory talk about the gained capacity you get from balancing or counteracting memory, but usually just say that they do it a very reduced current, which may be all that this is and it just an easy and inexpensive way to do it. I am not positive, but I think this is how smallish chargers like the Ctek do their desulphate and mini equalize cycles on AGM batteries also.

[markopolo]

The results would be the same or very similar. Either process provides the time needed for slower cells to catch up.

Winston's setup might allow balancing of individual cells (I can't remember). I think Rowie's balancing would keep series layers in check/balanced. If balancing occurs at 14.6V for example then voltages at the layers would be 3.65V / 7.3V / 10.95V / 14.6V. The parallel connections would/should get each cell to 3.65V in this example.

[booster]

Wouldn't the concern in rowiebowie's setup be that it was happening when the cells were still accepting large amount of current? Top balancing seems to always be mentioned as being done once the majority of cells have reduced their need for current, with only the low cells taking the much reduced current supplied, if I read what has been said correctly.


Perhaps rowie's setup is actually doing a mid range balance of some sort? I have never really seen much of a detailed explanation of the details of mid level so don't know. It appears from his stabilized voltage after semi rest that his battery was not full to 3.65v/c.



At least in lead acid batteries, they tout the benefits of pulse voltage to control current as being the best way to deliver otherwise potentially damaging voltage to counteract sulphation without adverse effects like heat or water loss, so maybe similar stuff goes on with lithium cells.

[Winston]

Originally Posted by markopolo

Winston's setup might allow balancing of individual cells (I can't remember).

We discovered that the cell balancing in our system is an interesting/curious process. Charging voltage is 14.2 volts, individual cell 1/2 ampere shunts engage at 3.55 volts/cell. 3.55 x 4 = 14.2. Hmmm. So how does this work . . . if two of the cells reach 3.55 first, then exceed 3.55, don't we end up with the possibility of an unequal voltage split of, say: 3.52, 3.51, 3.58, 3.59 = 14.2? And the answer yes, initially. But what happens after an hour or more is that the total string current drops very low . . . (after all, if the string of cells, added together, equals the charger voltage, the current should, in theory, fall to zero). At some point the current drops below 1/2 amp and the shunts on the 3.58 and 3.59 volt cells actually start to discharge these cells . . . their voltages drop . . . allowing a margin for the non-shunting cells to charge toward their 3.55 volt goal.

So we think we know how our balancing works, but due to its low current, it can take a period of hours. Yet, while balanced, can we be certain these cells are fully charged? As we're in the 'knee' region and drawing very little current, we're willing to accept this as full charge.

Originally Posted by booster

Wouldn't the concern in rowiebowie's setup be that it was happening when the cells were still accepting large amount of current?

As our yet to be reported attempt at lithium battery assassination reveals, there is nothing that automatically forces the charging current of lithium cells to drop as they reach full charge. We had no difficulty pumping 100+ amperes into our fully charged pack! We wish we had rowiebowie's Constant Current charger back in our AGM days (to keep the charge current high, longer), but we're reluctant to go that route with lithiums.

[avanti]

Originally Posted by Winston

We had no difficulty pumping 100+ amperes into our fully charged pack!

Where was all this energy going? That current represents a lot of heat.

[Winston]

Originally Posted by avanti

where was all this energy going? That current represents a lot of heat.

tba :~)

[markopolo]

Note: I'm just making some guesses about what Renogy does in that battery assuming that it's like other drop-ins and also based on the BMS I bought recently. It has 5 small gauge wires. They get attached to the 4 voltage points & one positive connection. Only the negative side heavy gauge goes through the BMS. Nominal -12V in and -12v out. The positive side heavy gauge has a direct connection from the battery to the load or charger.

You can see the small wires in both of the teardown videos shown earlier in this topic. One of the teardown videos was a Battle Born battery.

[Tonycamco]

Have 2 relion lithium battery’s. Rb 75 150 amps total run micro wave or toaster or hot water maker not at the same time works great have a blue sky pro remote
Monitor tells me every thing about the battery’s and 240 watts of solar blue sky controller 3000i works like a Swiss watch not cheap but great never look back!

[rowiebowie]

Originally Posted by Tonycamco

Have 2 relion lithium battery’s. Rb 75 150 amps total run micro wave or toaster or hot water maker not at the same time works great have a blue sky pro remote
Monitor tells me every thing about the battery’s and 240 watts of solar blue sky controller 3000i works like a Swiss watch not cheap but great never look back!

Congrats Tony.

As you said, Lithium is nice, but it's not cheap. In addition to running appliances, it's nice to check my van after weeks of sitting (no solar) and the batteries are sitting at their same charge level. That wasn't true with my old lead-acids.

[booster]

Originally Posted by Tonycamco

Have 2 relion lithium battery’s. Rb 75 150 amps total run micro wave or toaster or hot water maker not at the same time works great have a blue sky pro remote
Monitor tells me every thing about the battery’s and 240 watts of solar blue sky controller 3000i works like a Swiss watch not cheap but great never look back!

What charge profile are you running on the Blue Sky on lithium like yours?