LiFePO4 Batteries: Discussion Thread for Those Using Them as House Banks

[Lagoon4us]

Thought you'd like it, it gets worse up to 4.2v!

lithium & solar power LiFePO4, FAQ: Initial charging information The above is...

[daddle]

What do "we" plan to do about the charging below 0 degrees C issue? If "we" are unfortunate enough to select the wrong latitude on the chart plotter?

My information implies the damage may be permanent and dangerous because of ... Okay forgot the fancy word ... Something.

[Lagoon4us]

Daddle i just don't reckon that's cruising, apologies to the folk in cold climes, as Aussies's it's just not in our plans.......

[OceanPlanet]

Originally Posted by T1 Terry

[/COLOR]
[/I] Where did that come from? It's completely against all I've learnt from cell manufacturers and testing so far.
As far as the 4v per cell, this is known as a conditioning charge, not suitable for traction batteries where high current demains are required as it will result is serious voltage drop at high discharge rates, but it makes the cells last a lot longer as it forms an even coating on the cathode and resists dendrite build up as it stops isolated attachment, there just isn't enough materials in the plates to build an even layer of dendrite large enough to short through the separator to the anode. The coating is a result of partial breakdown of the electrolyte but it traps lithium ions within the coating, this loss of free lithium ions from the electrolyte is part of the reason for voltage drop at high discharge rates, the part is because the ion can't escape from the intercalation bonds of the graphite cathode as fast as the lithium ferrous anode can absorb them, resulting in a reduction of ion flow, the driver that creates current and therefore voltage..... is anyone still awake out there :lol: you really don't want to know the rest of it do you
It's one of those very fine line things, the slightest loss of concentration and irreversible cell damage will result because the voltage will go over 4v in the blink of an eye, you certainly don't want to be doing it on a regular basis.
This is where the basis for the "staying in the middle of the charging curve" theory comes from, 4v is the top of the curve, the bottom is somewhere south of 2.4v, so 2.8v and 3.6v is safely inside this envelope, it fits well with all the information from my personal testing that of others on various other forums and others who have carried out cell destruction testing, some of unintentionally

T1 Terry
[/COLOR]

Great stuff T1...

[ebaugh]

Originally Posted by Maine Sail

Being that it is winter I am really getting a good chance to do some simulations of our on-water use. We only use about 17-30 Ah's per day (engine driven refrigeration) but I often supplement with the DC refrigeration when I don't want to run the motor in a quiet anchorage. This extends our silent time.

In those scenarios we can use 40-50Ah's per day (excluding the 140W panel which contributes about 20-25Ah's).

Out of curiosity I ran a 10.6A load for the last 5 hours and removed about 53.5 Ah's from the bank. At the end of 5 hours at 10.6A the bank was still at 13.25V and the cells perfectly balanced. Essentially at these tiny loads the battery voltage simply does not move.. This is a huge adjustment from LA.....

If top balanced like you did, the voltages in the middle operating range should line up exactly, probably for a very long time. An imbalance only shows at full charge or full discharge.

But they do move some. I beginning to think you might be able to approximate SOC by voltage. The full usable range is something like 12.0 to 13.6V. But the middle 50% is more like 12.7 to 13.2, which makes it hard, but perhaps not impossible. That's the test Deckofficer wanted to do.

If you can do a charge test at a multiple fixed rates of .1 .2 and .3 C, and measure the time between 3.4 to 3.45 to 3.5 to 3.55, it would give us some great data on the capacity sacrificed by cutting off the charge early. The final step could be to taper the current to like .05C at 3.55 and stop, but I don't think that will take long.

[ebaugh]

Originally Posted by Lagoon4us

Thought you'd like it, it gets worse up to 4.2v!

lithium & solar power LiFePO4, FAQ: Initial charging information The above is...

Terry gave the technical answer...but Ive only seen that charge suggestion on the GWL website. I think of it as a forming charge. This is done at the factory. The question is does it ever need to be repeated? I doubt Winston would suggest it today. Im not even sure Thundersky endorsed it before the breakup to Winston and Sinopoly. The Thundersky owners manual I have does not mention it. I think it may have been an early production recommendation subsequently dropped.

But based on that one source, I would not try it!

[daddle]

From what I have read the voltage does directly relate to SOC (State of Charge). The qualification being that it is the Open Circuit voltage after the cells rest. That makes the SOC problematic to sense while in use.

The only other way, as I recall, is to do some sort of chemical analysis akin to the LA specific gravity measurement ... Which is impossible with a sealed cell.

[Maine Sail]

Originally Posted by daddle

From what I have read the voltage does directly relate to SOC (State of Charge). The qualification being that it is the Open Circuit voltage after the cells rest. That makes the SOC problematic to sense while in use.

The only other way, as I recall, is to do some sort of chemical analysis akin to the LA specific gravity measurement ... Which is impossible with a sealed cell.

I don't think that is possible with much accuracy..

Last night I loaded my bank to 10.6A and let her rip for about 5 hours. I removed 53.5Ah's. The bank was still at 13.25V after 5 hours. Within 20 minutes of the load being applied it was at 13.26V - 13.27V.

Before I started the test my cells measured:

#1 = 3.339V
#2 = 3.339V
#3 = 3.338V
#4 = 3.337V

This morning (about two minutes ago) they read:

#1 = 3.337V
#2 = 3.338V
#3 = 3.337V
#4 = 3.336V


This is the change measured using the same meter after removing 53.5Ah's or about 13.5% of the capacity.. As we know measuring to the thousandths is not very accurate thus all the cells, even after 5 hours at 10.6A, on a 400Ah bank were still at 3.33V and unchanged from before the test unless we think we can measure 0.001V changes as capacity movement. I'm not sold on that idea....

[mrm]

Originally Posted by T1 Terry

I'm interested in how the latching part works, I know how to do it with mechanical DPST relay, switch the negative and loop the N/O terminal to the negative driver terminal, negative to common and load to N/C, when the relay switches so the common is linked to N/O terminal, the relay is self powered and will stay in the position till either the positive is cut or the loop between the N/O terminal and negative driver terminal is opened, the relay then resets.

And that is exactly what I had in mind. I like this approach, because it is simple and robust - well matched to harsh environment found in boats. Besides, I think simple discrete solutions have bigger chances of surviving a near lightning strike, which is always a plus in safety systems. If you like more integrated approach, a simple R\S\ latch built from NAND gates will work, or an equivalent transistor circuit. One could throw a small microcontroller at it (ATTiny from Atmel or equivalent) too and integrate more functions within along the way, but I think this would be an overkill.

Marius

[diugo]

Originally Posted by daddle

From what I have read the voltage does directly relate to SOC (State of Charge). The qualification being that it is the Open Circuit voltage after the cells rest. That makes the SOC problematic to sense while in use.

SoC is a function of the cell voltage and the current through it----and different whether it's charge or discharge. Only a microcontroller that can accurately interpolate a vendor-specific data set of voltage-current-SoC triples can accurately determine the SoC while the battery is in use. The greater the current, the greater the slope on the plateau, so accuracy is lower at low currents.

[mcarling]

Originally Posted by mrm

To avoid further thread drift I will only make the following remarks and stop with this subject.

You asserted, that DC-DC do not use transformers. This is plain wrong and a study of typical SMPS topologies I linked to earlier will tell you why. Btw, SMPS is a DC-DC converter.

Had you written 'DC-DC converters use power electronics AND transformers/inductors' things would have be fine.

This turned out to be only a semantic issue. I checked with a Polish colleague and it turns out the Polish words for transformer (przetwornica and transformator) define the thing based on what it is capable of doing (so two windings around a common core would be a transformer in Polish regardless of what it actually does i.e. how it is actually used) unlike in English where most or all nouns which end with the letters "er" define a thing based on how it is actually used i.e. what it is actually doing, independent of what it is capable of doing. For example, in English, being capable of publishing does not make one a publisher; only actually publishing makes one a publisher.

I hope this tangent has run its course.

[mcarling]

Originally Posted by diugo

SoC is a function of the cell voltage and the current through it----and different whether it's charge or discharge. Only a microcontroller that can accurately interpolate a vendor-specific data set of voltage-current-SoC triples can accurately determine the SoC while the battery is in use. The greater the current, the greater the slope on the plateau, so accuracy is lower at low currents.

I guess someone will be selling such SoC meters for LiFePO4 batteries within two years. With some clever programming, they could be self-calibrating. That seems to be what Chevy is doing with the Volt (automobile).

[NahanniV]

Originally Posted by diugo

SoC is a function of the cell voltage and the current through it----and different whether it's charge or discharge. Only a microcontroller that can accurately interpolate a vendor-specific data set of voltage-current-SoC triples can accurately determine the SoC while the battery is in use. The greater the current, the greater the slope on the plateau, so accuracy is lower at low currents.

State of charge can only be done well by counting current, it's the only way that gives useful results in a working system; This was also the case for Lead Acid.

I wish there was an integrated system that did Cell level battery monitoring of temp and voltage with alarms and a state of charge display (at a reasonable price).

The Elite Power Solutions BMS: http://elitepowersolutions.com/docs/EMS.pdf
looks good but only works easily for GBS cells.

I've been playing with the PowerLog 6s. For $40 it can do a lot: cell monitoring, 3 temp inputs, 1 alarm output, current measure 40A continuous and 140 peak, current counting, but no simple state of charge display and the current limit is too small for all loads (start, windless). I may Mod it to an external current shunt.

Anybody Know of another system ?

I need to "put a stake in the ground" and start planning an install soon. Only 70 days till launch !

[mrm]

Originally Posted by mcarling

This turned out to be only a semantic issue. I checked with a Polish colleague and it turns out the Polish words for transformer (przetwornica and transformator) define the thing based on what it is capable of doing (so two windings around a common core would be a transformer in Polish regardless of what it actually does i.e. how it is actually used) unlike in English where most or all nouns which end with the letters "er" define a thing based on how it is actually used i.e. what it is actually doing, independent of what it is capable of doing. For example, in English, being capable of publishing does not make one a publisher; only actually publishing makes one a publisher.

I hope this tangent has run its course.

I think it is more of electronics nomenclature than English/Polish differences, but let's put it to rest.

[mcarling]

Originally Posted by NahanniV

State of charge can only be done well by counting current, it's the only way that gives useful results in a working system...

The problem is that the current counter will drift away from an accurate reading over a number of charge/discharge cycles -- unless supplemented by another means of knowing the SoC (probably near the extremes) to self-calibrate when possible.