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

[diugo]

Originally Posted by ebaugh

This is not true. If you top balance even unmatched cells equally to 3.6V, the capacity of the bank is restricted to the lowest capacity cell. So long as you never draw that lowest cell below its full capacity, lets say something like 2.8-2.9V, the subsequent charge cycle will return all cells to full charge equally.

This is not true. The behavior you describe is true only for cells strictly in parallel. Significantly mismatched cells in series will not maintain any semblance of balance for very long. This why C- and D-sized flashlight cells are never wired in series, for example.

[diugo]

Originally Posted by ebaugh

To get an accurate reading, you probably need to be in the flat part of the discharge curve? Say 3.2 to 3.3V per cell?

I did look for more exact numbers for you. 12.92 to 13.15V at the pack, switching from a 31A load to 150A charge on my bank, better than I remembered. But I have no idea how to translate that to internal resistance with 48 individual cells.

Yes, you need a steady voltage for an accurate test, so you want to be in the flat part of the curve.

Your bank has 1/7 the internal resistance of mine, at 230/181 = 1.27 millohms. To figure out your average cell resistance, we can assume that each cell carries 1/12th of the total current, and comprises 1/4 of the voltage change. So 57.5/15.08 = 3.8 millohms---almost twice mine! But my cells are 2.6X bigger, so that is to be expected.

Note that even though your cells have higher resistance, your bank benefits greatly from having most of that resistance in parallel.

[ebaugh]

Originally Posted by diugo

This is not true. The behavior you describe is true only for cells strictly in parallel. Significantly mismatched cells in series will not maintain any semblance of balance for very long. This why C- and D-sized flashlight cells are never wired in series, for example.

Im quoting mostly what I've read about...I have literally hundreds of hours reading both research papers and practical experience from forum posts here and elsewhere. But maybe someone else can shed more light? The C and D flashlight analogy does not explain anything, they are typically not rechargeable, and if they are it's NIMH or NICAD not Li.

I've been led to believe that LiFePO4 theoretically moves electrons around perfectly, unlike other chemistries. With minimal degradation of anodes, cathodes and electrolyte over time creating a whole new ball game compared to others.

I see where what I quoted would also be true for parallel, but why not serial connections? Assuming there is only one power tap for the bank and current draw for discharge and charge are equal?

[ebaugh]

Actually thinking more about this...I'm not sure parallel would hold the same truth. The bigger cells would absorb more of the load as a bank discharged. In parallel the current draw on each cell would be different and at complete discharge of the smallest capacity cell, the bigger cells would charge the smaller ones if the load was reduced.

The main reason I make the point is based more on the practical experience of the EV community. They use 100-150V packs that are bottom balanced. That's a whole lot of serial cells compared to a marine house bank. Many are bottom balancing at install, and then not using any type of BMS or similar to maintain balance. The experience is showing minimal balance problems. With only 4 cells to keep track of, we should have fewer issues?

[diugo]

I thought my C vs D cell was quite apt. It does not "not explain anything". The C and D cells differ only in capacity---yet behave counterintuitively in series. Cell chemistry is irrelevant, but if you really want to model an ion cell as a perfect electron storage device, such a thing has existed for centuries.

It's called a capacitor---and capacitors in series simply do not behave like capacitors in parallel. If you want to know more, there are plenty of research papers on this subject as well

[deckofficer]

Originally Posted by diugo

On the other hand, if the cells are poorly matched to begin with, then any manual balance is just an exercise in futility, and you would probably want a balancer that does a top balance every cycle.

We shall see in about 48 hours. My 12 volt pack that I'm calling #1 took a discharge without any one cell dropping more than the average of the rest. It then charged at 50 amps until pack voltage reached 14.8 volts and the charger cycled off. No cell jumped the average of the others, so I deemed it fine.

Pack #2 is a whole different matter. Discharge had one cell drop to 2.5 volts while the others were 3.2 volts. On charging on the same 50 amp charger, I had to stop when the charger was down to 12 amps because a different cell had passed 4.0 volts. This is the pack I've bottom and top balanced and now has been put on the 30 watt load. I'll have results in about 50 hours.

[Kettlewell]

Another Boeing 787 makes an emergency landing due to smoke inside. Not sure what the problem was.

[dansken]

I have now ordered 8 pcs Sinopoly 200ah to become my new 24V 200A house bank and hereby consider myself a member of the club. Thanks to all the early adaptors that share their experiences and considerations here – you have convinced me to make the jump and get a LIFE.

Depending on temperament different people want to ad various automated controllers and switches, I will go simple: an initial manual balancing exercise and then add a CellLogg voltmeter in the final setup.

My two serpentine-belt driven alternators will be regulated by a Balmar MC624 with alternator temp sensor, the question is how should the regulator be programmed? (I see on the Balmar site that these now have a program for LIFEPO4 but I can’t find any info on what exactly that includes).

[goboatingnow]

Originally Posted by Kettlewell

Another Boeing 787 makes an emergency landing due to smoke inside. Not sure what the problem was.

Li Cobalt, thats what the problem is.

Dave

[mcarling]

Originally Posted by diugo

This why C- and D-sized flashlight cells are never wired in series, for example.

To the best of my recollection, every C- and D-sized flashlight I've ever had was "wired" in series. I just checked one of my Maglite flashlights and the D-cells are definitely arranged in series.

Originally Posted by dansken

I have now ordered 8 pcs Sinopoly 200ah to become my new 24V 200A house bank and hereby consider myself a member of the club. Thanks to all the early adaptors that share their experiences and considerations here – you have convinced me to make the jump and get a LIFE.

Welcome! Please let us know about your experience.

[goboatingnow]

Id like to pose some questions for users of large capacity prismatic LiFe, Based on som of my own observations and experience, primarily in small cylindric capacity Li Cells ( ie <10Ah), even though I have a large prismatic set to play with as well. ( one of the problems with the large prismatic Chinese is the spec sheets are very poor)

LiFePo4 shares all the normal specs of the Li Chemistry, but trades safety for energy density , ie about 14% less energy then LiCo.


(1)
I see little mention of pre qualification charge issues, in my case ( software based charger) I follow the recommendations that if charging below the LVC point , ( around 2.7-3V) all Li tech is spec'ed to only be charged at about 10 % of cell capacity at that point. ( less in some specs) This is a series issue for some people as on board they have no way to limit such current in a severely discharged cell. ( This was one of the reasons Sony originally had so much trouble with LiCo.)

(2) Many consumer cells can not be charged at below zero, I cant see if that preclusion applies to large prismatics? charging below zero can cause rupture


(3) While LIFeP04 technology seems to have a far higher shelf life then LiCo, as most people may know Li capacity deteriorates just doing nothing, far faster the LA in some cases. Anybody see any real feedback from LI Ferrous yet. ( Shelf life not charge cycle life)

(4) A significant point is that LI should not be float charged ie a constant top up charge aka used in LA chargers, all specs Ive seen suggest that on reaching the charge termination point, the charge should actually terminate. Then typically the recharge begins at the lower voltage threshold ( which can be very difficult to detect, with such flat voltages curves on high capacity prismatics, easier on cylindric). This is one areas I feel need to be highlighted in using conventional chargers originally designed for LA. The technical data suggests that a reduction in life occurs in Li is "floated", certain all the software charging algorithms published and the IC chargers do this. Any comments?

(5) most designs , including mine for small capacities, use a battery temperature sensor, but not in the same way as a LA charger does, I use it to ensure the case temperature stays with a nominal range , both on charge and discharge, ??

Im actually thinking of restarting a charge cycle on Amp counting rather then voltage. I wonder if allowing the battery to discharge significantly , since that counts as a cycle, is better then many small cycles . Looking at the technical papers I beleive it is, but theres little real world data.


This suggest that , despite what might have been said ( and by me) , that simply using a LA charger and just adjusting the voltages isn't sufficient, I really do think we need a ground up designed LI charger.


Dave

[ebaugh]

[QUOTE="goboatingnow;1131247"]Id like to pose some questions for users of large capacity prismatic LiFe, Based on som of my own observations and experience, primarily in small cylindric capacity Li Cells ( ie <10Ah), even though I have a large prismatic set to play with as well. ( one of the problems with the large prismatic Chinese is the spec sheets are very poor)

LiFePo4 shares all the normal specs of the Li Chemistry, but trades safety for energy density , ie about 14% less energy then LiCo.

(1)
I see little mention of pre qualification charge issues, in my case ( software based charger) I follow the recommendations that if charging below the LVC point , ( around 2.7-3V) all Li tech is spec'ed to only be charged at about 10 % of cell capacity at that point. ( less in some specs) This is a series issue for some people as on board they have no way to limit such current in a severely discharged cell. ( This was one of the reasons Sony originally had so much trouble with LiCo.)

(2) Many consumer cells can not be charged at below zero, I cant see if that preclusion applies to large prismatics? charging below zero can cause rupture

(3) While LIFeP04 technology seems to have a far higher shelf life then LiCo, as most people may know Li capacity deteriorates just doing nothing, far faster the LA in some cases. Anybody see any real feedback from LI Ferrous yet. ( Shelf life not charge cycle life)

(4) A significant point is that LI should not be float charged ie a constant top up charge aka used in LA chargers, all specs Ive seen suggest that on reaching the charge termination point, the charge should actually terminate. Then typically the recharge begins at the lower voltage threshold ( which can be very difficult to detect, with such flat voltages curves on high capacity prismatics, easier on cylindric). This is one areas I feel need to be highlighted in using conventional chargers originally designed for LA. The technical data suggests that a reduction in life occurs in Li is "floated", certain all the software charging algorithms published and the IC chargers do this. Any comments?

(5) most designs , including mine for small capacities, use a battery temperature sensor, but not in the same way as a LA charger does, I use it to ensure the case temperature stays with a nominal range , both on charge and discharge, ??

Im actually thinking of restarting a charge cycle on Amp

[ebaugh]

Originally Posted by goboatingnow

Id like to pose some questions for users of large capacity prismatic LiFe, Based on som of my own observations and experience, primarily in small cylindric capacity Li Cells ( ie <10Ah), even though I have a large prismatic set to play with as well. ( one of the problems with the large prismatic Chinese is the spec sheets are very poor)

LiFePo4 shares all the normal specs of the Li Chemistry, but trades safety for energy density , ie about 14% less energy then LiCo.

(1)
I see little mention of pre qualification charge issues, in my case ( software based charger) I follow the recommendations that if charging below the LVC point , ( around 2.7-3V) all Li tech is spec'ed to only be charged at about 10 % of cell capacity at that point. ( less in some specs) This is a series issue for some people as on board they have no way to limit such current in a severely discharged cell. ( This was one of the reasons Sony originally had so much trouble with LiCo.)

(2) Many consumer cells can not be charged at below zero, I cant see if that preclusion applies to large prismatics? charging below zero can cause rupture

(3) While LIFeP04 technology seems to have a far higher shelf life then LiCo, as most people may know Li capacity deteriorates just doing nothing, far faster the LA in some cases. Anybody see any real feedback from LI Ferrous yet. ( Shelf life not charge cycle life)

(4) A significant point is that LI should not be float charged ie a constant top up charge aka used in LA chargers, all specs Ive seen suggest that on reaching the charge termination point, the charge should actually terminate. Then typically the recharge begins at the lower voltage threshold ( which can be very difficult to detect, with such flat voltages curves on high capacity prismatics, easier on cylindric). This is one areas I feel need to be highlighted in using conventional chargers originally designed for LA. The technical data suggests that a reduction in life occurs in Li is "floated", certain all the software charging algorithms published and the IC chargers do this. Any comments?

(5) most designs , including mine for small capacities, use a battery temperature sensor, but not in the same way as a LA charger does, I use it to ensure the case temperature stays with a nominal range , both on charge and discharge, ??

Im actually thinking of restarting a charge cycle on Amp counting rather then voltage. I wonder if allowing the battery to discharge significantly , since that counts as a cycle, is better then many small cycles . Looking at the technical papers I beleive it is, but theres little real world data.

This suggest that , despite what might have been said ( and by me) , that simply using a LA charger and just adjusting the voltages isn't sufficient, I really do think we need a ground up designed LI charger.

Dave

I think there are more differences than density and safety between LiCo and LiFePO4...

1) I don't think the low charge rate at low SOC applies so long as LiFePO4 is kept within the specified normal voltage range. I have heard a severely discharged cell, below 2.5 V, should be charged very slowly initially.

2) I don't know. I'm in the tropics.....

3) I've heard they store well at 50% SOC

4) There is no good data about this issue. But...if you float charge, the charge has terminated. The charger simply holds a voltage and current into and out of the pack is essentially zero. I believe the question is what voltage is optimum, but don't know the answer for sure. My supplier said 3.35V maximum. I aim for that or less, and it's conveniently about the float voltage for LA. I do suspect there would be some cost in cycle life if floated too high, say at 3.6V.

So the implication and reality is you can't maintain 100% charge in "storage". More like about 85%, and 50% might be even better for months long time in marinas.

5) In high C applications (EV's) temperature is a bigger player. The BMS I use came set to disconnect at 150F. I see a 4-5F degree rise when charging at 150A, but that is only like .13C rate and may be more related to the fact the bank is adjacent to the genset. Nothing I've read indicates a need to vary charge profile based on temperature like LA.

I think for high rate fixed current chargers, assuming you can set charge voltages and defeat the acceptance phase and defeat the special float logic some chargers have that cycle to 12.8V periodically (Magnum for one) then there is little need for a customized Li charger since the bulk charge rate is fixed. But for low rate charge sources like wind and solar, I'm not as certain. If you look at what charge profiles are available, you can see that the termination voltage for 100% charge decreases with charge rate. You can take a stab at .1C, but .01C would only be a wild guess. An optimal solar controller would end the bulk phase at different voltages based on the present charge rate. But I have no idea what you would program it too. For now, I'd probably set it to some number near a float voltage.

There is data that suggests remaining on the flat part of the curve extends cycle life. Translated to remaining between say 20-85% SOC. But nothing I know that indicates if a deeper cycle is better than multiple shallow cycles or not. I hope we are splitting hairs here, but don't know.

[goboatingnow]

Originally Posted by ebaugh

I think there are more differences than density and safety between LiCo and LiFePO4.

Not really , its the primary reason Ferrous cells were developed. There are some side effects like longer shelf life. Do you know of others.

1) I don't think the low charge rate at low SOC applies so long as LiFePO4 is kept within the specified normal voltage range. I have heard a severely discharged cell, below 2.5 V, should be charged very slowly initially.

The data I have for cylindric cells all state that a pre-qualification charge is neccessary below a point. You also agree, but my concern is that typical LA chargers have no such ability. All the Integrated circuits for LifePo4 charging implement all the things Ive mentioned.

2) I don't know. I'm in the tropics.....

Sure but its an issue

3) I've heard they store well at 50% SOC

Yes they store well at various SOCs but for how long.Its Li achilles heel.

4) There is no good data about this issue. But...if you float charge, the charge has terminated. The charger simply holds a voltage and current into and out of the pack is essentially zero. I believe the question is what voltage is optimum, but don't know the answer for sure. My supplier said 3.35V maximum. I aim for that or less, and it's conveniently about the float voltage for LA. I do suspect there would be some cost in cycle life if floated too high, say at 3.6V.

The technical data I have, is very clear, Li should not be floated, Floating in effect feeds small currents into the cell on a continous basis, The data says the charge should terminate and remain terminated until some threshold is reached, Some data indicates quite significant loss of capacity if floated. Again all the ICs etc implement this feature.

So the implication and reality is you can't maintain 100% charge in "storage". More like about 85%, and 50% might be even better for months long time in marinas.

yes the great thing about Li is it doesnt care.

5) In high C applications (EV's) temperature is a bigger player. The BMS I use came set to disconnect at 150F. I see a 4-5F degree rise when charging at 150A, but that is only like .13C rate and may be more related to the fact the bank is adjacent to the genset. Nothing I've read indicates a need to vary charge profile based on temperature like LA.

correct , I was drawing attention to the fact that many chargers do not have temp sensing for Li as a safety feature, Ie you cant really use LA chargers for Li.

I think for high rate fixed current chargers, assuming you can set charge voltages and defeat the acceptance phase and defeat the special float logic some chargers have that cycle to 12.8V periodically (Magnum for one) then there is little need for a customized Li charger since the bulk charge rate is fixed. But for low rate charge sources like wind and solar, I'm not as certain. If you look at what charge profiles are available, you can see that the termination voltage for 100% charge decreases with charge rate. You can take a stab at .1C, but .01C would only be a wild guess. An optimal solar controller would end the bulk phase at different voltages based on the present charge rate. But I have no idea what you would program it too. For now, I'd probably set it to some number near a float voltage.

i think for sub 1C chargers I would pick a common cutoff voltage

[GordMay]

Greetings and welcome aboard the CF, dansken.