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

[tanglewood]

Originally Posted by CatNewBee

No, it is a deep discharge thing, when you leave your cells at low soc and they self discharge further, the bms disconnect cant save them because it is comming not from other external loads. It can also happen on overcharging and overheating.

So when doing things that will trash your batteries, you have to work a little bit harder to wreck them if you strap them up tight? :-)



But more seriously, part of what I'm wondering is whether the case design from other manufacturers provides the required support. Again, CALB doesn't say anything about adding support for the batteries, which leads me to believe they have designed a case that provides all the required support. But information is thin from all these vendors, so who knows for sure.

[CatNewBee]

Originally Posted by tanglewood

So when doing things that will trash your batteries, you have to work a little bit harder to wreck them if you strap them up tight? :-)



But more seriously, part of what I'm wondering is whether the case design from other manufacturers provides the required support. Again, CALB doesn't say anything about adding support for the batteries, which leads me to believe they have designed a case that provides all the required support. But information is thin from all these vendors, so who knows for sure.

Are the calb prismatic, or boxes with cylindric cells inside?
Cylindric cells do not need compression, because the cylinder itself cannot expand, so batteties built from cylindric cells do not need compression. Alternatively the box can be made sturdier and heavier to provide some rigidity or they just dont care about bulging.

[Checkswrecks]

Originally Posted by GoneDiving

Just asking: What is the role in trying to stop the swelling? In most chemical reactions pressure and containment promote and excite not reduce it.

Many explosives barely burn in the open air but detonate when placed in a bore hole or barrel.

Thanks

Whether or not to permit swelling has nothing to do with burning or detonation, it has to do with how the cell ages and will eventually fail.

Most cells have a nice smooth distribution of anodes and cathodes, so like I wrote previously the internal pressure results in pressure on the hermetic seals at the terminals; not a lot else.

That's not always the case though, especially in prismatic cells (rectangular cans).

There are basically two ways prismatic cells are made. The more common is highly automated and is simply layers of anode-plastic separator-cathode-separator-anode - . . . and on and on. (The separator is an extremely thin fragile porous plastic membrane.) The current collectors of the anodes are are one end of the cell and the current collectors for the cathodes are at the other end. When the cell swells all the layers simply move slightly apart.

Note - The above image is of a pouch cell but it nicely conveys the layering idea. CALB cells are made like this, which is one reason they cost less to produce and generally have less energy density but also good reliability due to the machine reproducibility of making every layer like the last one.

The other way prismatic cells can be made is to wrap the anode, separator, and cathode layers on what is essentially a toilet paper tube, then flatten the roll of toilet paper. A lot of the aluminum can cells are this way. When this type of winding is charged and discharged the stresses are not even and over a number of cycles wrinkles can develop in the layers. This doesn't happen much if the hole in the tube is vertical so the current collectors stay fairly even in distribution along the top and the layers kind of hang like curtains. Even so, in the front view you can see where wrinkles formed.


But if the hole in the tube is horizontal the layers tend to sag. Something else about this style is that the current collectors are not evenly distributed, so the wrinkling is pronounced at the ends of the collector fingers. You can see the wrinkles in the bottom third of the cell shown below, as that's where the stiff collector fingers end. Because those fingers are restrained at the top the bottoms can move a lot! These cells are intended to be restrained to minimize that movement. The problem is that sooner or later either dendrites develop to rupture the separator, or the movement that still happens will tear or otherwise damage the separator so the anode and cathode short out. Plus, the pressure definitely caused the hermetic seals at the terminals to leak in this design, as I mentioned originally.

(The cell above has three rolls in the one can and you are looking at the ends. The parallel electrical nature of the three separate rolls also resulted in an uneven charge/discharge distribution.)

You don't know which style you have unless you have access to x-ray or computerized tomography, so if the manufacturer says pack the tight, accept that the mfg knows the seals will hold. If they don't say, then I think allowing the layers to be at their most relaxed by not constraining them makes more sense. If somebody here wants to do it another way, it's their money and their choice.

[john61ct]

Originally Posted by tanglewood

CALB doesn't say anything about adding support for the batteries, which leads me to believe they have designed a case that provides all the required support.

That's a heck of a stretch!

The customers these makers care about, have their own engineers on staff.

My assumption is, **all** hard-case prismatics need compression plates as part of the standard installation. Within the big picture it's just not a big deal.

I am not thinking the aluminum shells are different either until I see hard evidence or explicitly told so by a user I trust.

[Checkswrecks]

Originally Posted by john61ct

That's a heck of a stretch!

The customers these makers care about, have their own engineers on staff.

My assumption is, **all** hard-case prismatics need compression plates as part of the standard installation. Within the big picture it's just not a big deal.

I am not thinking the aluminum shells are different either until I see hard evidence or explicitly told so by a user I trust.

LOL -
We posted at nearly the same time.

[nebster]

Originally Posted by john61ct

I am not thinking the aluminum shells are different either until I see hard evidence or explicitly told so by a user I trust.

Your thinking is incorrect.

[rgleason]

Evm1024 wrote:
Many battery monitors also monitor the start battery voltage. The shunt induced voltage inaccuracy of the start battery is opposite to the house bank. If the start battery is measured between the positive and the negative bus side of the shunt then the start battery voltage will be more accurate - just the opposite of the house bank.


This escapes me. Why?

[evm1024]

Originally Posted by rgleason

Evm1024 wrote:
Many battery monitors also monitor the start battery voltage. The shunt induced voltage inaccuracy of the start battery is opposite to the house bank. If the start battery is measured between the positive and the negative bus side of the shunt then the start battery voltage will be more accurate - just the opposite of the house bank.


This escapes me. Why?

I think that I will walk back the interaction of the start battery with out thinking more deeply on it. The battery monitor "should" give a start battery reading within its precision.

The house bank currents run through the shunt so there is a voltage drop across the shunt. The shunt is a resistor and the house loads can be viewed as a resistor. Thus the shunt resistance and the house load resistance form a voltage divider.

As the house loads increase the house "resistance" decreases and thus the ratio of the voltage divider changes.

The house voltage measured is across the house loads (and not across the battery) and thus it actually measures the voltage drop across the house loads. (not the battery voltage)

Wow, that was a great question. It got me thinking. If I were going to make a battery monitor like the Victron line (some type of uP in there) I would build in a compensation for the shunts voltage drop to get a more accurate reading.

I'm going to guess that they do and that the monitor reads accurately within the precision of the readout for both house and start batteries. I'll stand corrected.

It looks like it would only mess with any balancing of the cells.

[newhaul]

Originally Posted by evm1024

I think that I will walk back the interaction of the start battery with out thinking more deeply on it. The battery monitor "should" give a start battery reading within its precision.

The house bank currents run through the shunt so there is a voltage drop across the shunt. The shunt is a resistor and the house loads can be viewed as a resistor. Thus the shunt resistance and the house load resistance form a voltage divider.

As the house loads increase the house "resistance" decreases and thus the ratio of the voltage divider changes.

The house voltage measured is across the house loads (and not across the battery) and thus it actually measures the voltage drop across the house loads. (not the battery voltage)

Wow, that was a great question. It got me thinking. If I were going to make a battery monitor like the Victron line (some type of uP in there) I would build in a compensation for the shunts voltage drop to get a more accurate reading.

I'm going to guess that they do and that the monitor reads accurately within the precision of the readout for both house and start batteries. I'll stand corrected.

It looks like it would only mess with any balancing of the cells.

Ok question then if you have the install like thus bank to bms, to shunt, to load buss. The shunt should not cause any issues with causing a misread of cell voltages and affect balancing.

My bank is currently on the bench wired that way and charging for eventual balancing ( my only available charger at home is a 10 amp unit ) yes i know real small for a 240ah bank but i dont have to consider any issuea with high C rate charge right now)

[evm1024]

Originally Posted by newhaul

Ok question then if you have the install like thus bank to bms, to shunt, to load buss. The shunt should not cause any issues with causing a misread of cell voltages and affect balancing.

My bank is currently on the bench wired that way and charging for eventual balancing ( my only available charger at home is a 10 amp unit ) yes i know real small for a 240ah bank but i dont have to consider any issuea with high C rate charge right now)

Find you shunts resistance. Victrons typical is 50 mV for 500 amps. SO the resistance is 0.05 / 500 = 0.0001 ohms.

At 100 amps the voltage drop through that shunt is 100 * 0.0001 = 0.01 volts (10 mV).

At 10 amps 10 * 0.0001 = 1 mV

At 100 amps the #1 cell voltage could be off by 10 mV (low by 10 mV) and only 1 mV at 10 amps.

How the balancing (shunting) resistors play into this I could not say. But they will play into it.

One thing that could be done is to get it to the point it shunts and see how it plays out. It may be that the cells converge as the current decreases (full at shunting voltage so the major current is going through the shunting resistors is is "small).

I personally would not worry about a 1 mV induced error. I'm not even sure that I would worry about a 10 mV induced error especially if it factors out as the current decreases.

[newhaul]

Originally Posted by evm1024

Find you shunts resistance. Victrons typical is 50 mV for 500 amps. SO the resistance is 0.05 / 500 = 0.0001 ohms.

At 100 amps the voltage drop through that shunt is 100 * 0.0001 = 0.01 volts (10 mV).

At 10 amps 10 * 0.0001 = 1 mV

At 100 amps the #1 cell voltage could be off by 10 mV (low by 10 mV) and only 1 mV at 10 amps.

How the balancing (shunting) resistors play into this I could not say. But they will play into it.

One thing that could be done is to get it to the point it shunts and see how it plays out. It may be that the cells converge as the current decreases (full at shunting voltage so the major current is going through the shunting resistors is is "small).

I personally would not worry about a 1 mV induced error. I'm not even sure that I would worry about a 10 mV induced error especially if it factors out as the current decreases.

That would account for my .001v difference between cell 1 and the rest of them. Should go away while at rest

[CatNewBee]

Originally Posted by rgleason

Evm1024 wrote:
Many battery monitors also monitor the start battery voltage. The shunt induced voltage inaccuracy of the start battery is opposite to the house bank. If the start battery is measured between the positive and the negative bus side of the shunt then the start battery voltage will be more accurate - just the opposite of the house bank.


This escapes me. Why?

Misconception.

Voltage drop happens only if there is a current over a voltage divider (series of resistors).
The battery monitor measures the battery voltage on the battery side connector of the shunt (-) vs. the opposite pole (+) , the shunt is not in the current path of the volt meter for the board battery.

The start battery voltmeter goes to the (+) of the start battery, the negative is assumed common, it should be connected after the shunt, otherwise you won't be able to measure cross currents between the 2 batteries. By measuring the start battery the measurement current goes from Start battery (+) to the gauge to (-) from the shunt to common ground to the start battery (-). Because there is no significant current flowing from the start battery (other current path, separate circuit) trough the shunt (only few mA for the measurement) , there is no noticeable voltage drop even if the house battery and the start battery are under full load. Only if there is a cross-current between start and board battery there will be a noticeable voltage drops on the shunt when measuring the start battery voltage.

Without a cross current, the shunt is just a piece of wire on the voltage probe of the start battery and neglectable when measuring the start voltage, it is not part of other currents from the start battery.

[evm1024]

Originally Posted by CatNewBee

Misconception.

Voltage drop happens only if there is a current over a voltage divider (series of resistors).
The battery monitor measures the battery voltage on the battery side connector of the shunt (-) vs. the opposite pole (+) , the shunt is not in the current path of the volt meter for the board battery.

The start battery voltmeter goes to the (+) of the start battery, the negative is assumed common, it should be connected after the shunt, otherwise you won't be able to measure cross currents between the 2 batteries. By measuring the start battery the measurement current goes from Start battery (+) to the gauge to (-) from the shunt to common ground to the start battery (-). Because there is no significant current flowing from the start battery (other current path, separate circuit) trough the shunt (only few mA for the measurement) , there is no noticeable voltage drop even if the house battery and the start battery are under full load. Only if there is a cross-current between start and board battery there will be a noticeable voltage drops on the shunt when measuring the start battery voltage.

Without a cross current, the shunt is just a piece of wire on the voltage probe of the start battery and neglectable when measuring the start voltage, it is not part of other currents from the start battery.

Agreed, I walked that back - I over thought it and later realized that the starter battery did not use the shunt and that any current induced voltage drop was purely in the house circuit. Just a few posts back from this one. Also, it was clearly stated that you would need current ...

It is good to get fact check feedback.

[Oceanride007]

I've got a height limit of .24m (9.5") and I would like to install some where around 540Ah, one provider suggested laying his CALB 180 on edge. Not sure if this would work cos I would think access would be required for monitoring or whatever. Question to this group is access to battery tops required.

[CatNewBee]

No, you dont need to service then frequently, the bms gives you all you need to know usually. But you have to make sure the battery plates are not drying out, there are some orientations, that are not good for the cells. Check the datasheet or contact the manufacturer what is acceptable.