Balancers installed

[dochubert]

Thought I would post a picture of my balancer setup now that it is installed.  So far, it is only connected to half of my battery bank.  My bank is Lifepo4 BYD 24v modules.  These are used modules, so they are something short of the 220ah rating they had when new.  16 of them connected for a 48v bank.  If new that would make my bank 1760ah at 48v.  The original 8 were installed several months ago, and I recently added 8 more to double the size of the bank.

This is a massive bettery bank, and being new to Lifepo4, I knew almost nothing of balancing when I started.  Tried Daly Bms, 1.2a and 5a active balancers with various degrees of disappointment.  They were just too 'small' to have the desired impact.  The result was keeping the charging set to 54v instead of the 56v or higher that is possible with these batteries.  They stay fairly well balanced at 54v and below, but I should be able to do better.

The first thing I should have done, and have done recently, is parallel all the cells.  This one action has done more to improve balance than anything else so far.  That brings us to today and Sid's balancers.  They are mounted on a massive heatsink, with a 5 inch fan on the backside, controlled by the temp controller you see on the right side of the pic.  16 balancers for 16 lifepo4 cells in a 48v bank.  If you zoom in you can read the cell volts on each balancer.  They are ready to go.

Honestly, I don't yet know how well these are going to work, because I only have half of the bank connected, and because charging is still set to 54v.  Ran out of wire.  In a few days when the additional ordered wire gets here, I'll connect up the other half.  Then I can start bumping up the charge voltage gradually, eventually settling at 56v.  That puts it at 3.5v per cell.  Any cell reaching 3.55v will start 'balancing', meaning it should bleed off voltage until it's back in range.  I'm expecting to be impressed with these balancers, and hoping the wire gets here soon.

Will update when the other half is connected

[kuhrd]

Nice neat setup for them so far, I will be interested in seeing how well they work for you on such a large bank.  I'll be happy if I can get my current bank up to 30Kwh within the next year.

[Sid Genetry Solar]

Whoa, that's a large battery bank...90kwh...!

Being LiFePo4, it should still be very feasible for the balancers to keep them balanced...though hearing that 5A-rated 'active' units couldn't keep up...

...my concern would be the wires between the balancers and the cells.  At 5A, you can get considerable voltage drop (well, relatively speaking!) across said wires if they're very long (or very thin).  It is very easy to lose 0.5v @ 5A over balance wires--and that's the difference between 3.55v and 4.05v on the cells.

According to a wire resistance chart, 18AWG wire comes in at 0.021 ohms/meter.  If the wires are approximately 3' on either side, that's 0.042 ohms total resistance @ 5A = 0.21v drop across the wires.  If the balancer is holding 3.55v at its terminals, the only way to reach 5A balance current is for the cell to be at (3.55 + 0.21v) = 3.76v.

Yeah, wire losses are brutal at low voltages 😉.

[Sid Genetry Solar]

It's also worth noting that with LiFePo4, it's very difficult to determine the SOC (state of charge) between approximately 20% and 80% without a cumulative watt-hour meter.  Having 1,760A of theoretical cell capacity does beg the possibility of cells being perhaps 10% SOC different...that's 176Ah to deal with @ 5A = could take nearly 2 days.

 

EDIT: With charge stopping at 54v, that's roughly 3.375v/cell (VPC I call it, volts per cell).  LiFePo4 cells have a very flat charge/discharge curve, quite unlike Li-Ion or even lead-acid.

• The bulk of their charge will occur below 3.4vpc; the bulk of the discharge above 3.2vpc. 
• When the cells reach near-full charge (I believe ~80% SOC), their voltage will start to very rapidly rise across 3.4vpc up to 3.6v (and beyond if not capped)
• When discharging, they will maintain above 3.1-3.2vpc until the cell is nearly empty (I estimate 20% SOC), at which point it will rapidly fall off under 3.2vpc.

Due to the very flat charge/discharge curve of LiFePo4, it's practically impossible to identify the SOC of each cell unless it hits one or the other end.  (I say "practically" because while there is a very small "cell voltage - SOC" connection, the cell voltage change by load/charge amperage is much more significant.)

In other words: holding the battery pack "float" at 54v really isn't charging the cells.  Trickle charging, perhaps.

I should qualify these above statements by saying that I have had a 12kwh LiFePo4 bank on my solar system for 4 years now (built with NOS--new old stock--cells).  Never any regrets or problems with it.

[dickson]

In other words: holding the battery pack "float" at 54v really isn't charging the cells.  Trickle charging, perhaps

 

Off Grid Garage  youtube Aug17  show  56v  is needed for charging  LiFePo4  at 16s .    

[dochubert]

Ok, everything is connected, all cells are paralleled on both halves of the bank.

Been watching for a couple of days and gradually increasing charge settings.  Float is now 56v and bulk is at 56.8v. Balance isn't bad.  Charging has been hampered some by sporadic sun and smoke from west coast fires.  After a windy night blew some of the smoke out, getting a good charge today at 95a.  Had to fix a couple small things, but now the balancer seems to be doing it's job.  The fan/temp control is doing it's job keeping the heatsink cool.  Still have 3 cells not quite coming up to the rest at this setting, but improving daily.  Will give it a couple more days at this setting and see what I get. 

[dochubert]

On 8/10/2021 at 2:34 PM, Sid Genetry Solar said:

Yeah, wire losses are brutal at low voltages 😉.

Hopefully, won't be too much of an issue with my setup.  Used all 14ga stranded high temp wire.  The longest wire is about 4 feet long..  Most are under 3 feet long.  Wish I could have made them all the same length, but just wasn't feasible.  Tried to arrange the layout so wire runs were short as possible.  Doesn't have to be perfect, just within workable range.

Seems good so far.  Might have to bump the settings up a bit more to get those last three cells to fill up, but will wait a couple more days and re-evaluate.

[Sid Genetry Solar]

On 8/19/2021 at 4:26 PM, dochubert said:

Ok, everything is connected, all cells are paralleled on both halves of the bank.

Been watching for a couple of days and gradually increasing charge settings.  Float is now 56v and bulk is at 56.8v. Balance isn't bad.  Charging has been hampered some by sporadic sun and smoke from west coast fires.  After a windy night blew some of the smoke out, getting a good charge today at 95a.  Had to fix a couple small things, but now the balancer seems to be doing it's job.  The fan/temp control is doing it's job keeping the heatsink cool.  Still have 3 cells not quite coming up to the rest at this setting, but improving daily.  Will give it a couple more days at this setting and see what I get. 

Glad to hear this; if your cells are rising past 3.5v (registered on the balancers or otherwise!), they're reaching a reasonably full charge.  Might want to double-check the actual cell voltages on the other end of the wires just to make sure there isn't a significant voltage drop across the wiring.

Kinda funny to hear about "a couple of days", but then I have to remember that you have a HUGE battery bank ;-).

[dochubert]

On 8/20/2021 at 5:09 PM, Sid Genetry Solar said:

Might want to double-check the actual cell voltages on the other end of the wires just to make sure there isn't a significant voltage drop across the wiring.

Thanks for the tip!

Checked for voltage differences between balancer end and battery end.  Didn't find any.  Wire must be big enough to overcome some unequal lengths, and not long enough to cause detectable voltage drop.  The hoped for result!

The 3 cells that are a bit low are improving so will watch without any change in settings for a couple more days.

[Sid Genetry Solar]

15 hours ago, dochubert said:

Thanks for the tip!

Checked for voltage differences between balancer end and battery end.  Didn't find any.  Wire must be big enough to overcome some unequal lengths, and not long enough to cause detectable voltage drop.  The hoped for result!

The 3 cells that are a bit low are improving so will watch without any change in settings for a couple more days.

Glad to hear.  If the balancers are running full blast (i.e. getting really hot), I'd expect to see some voltage drop; as long as the actual cells on the other end of the wires are staying below the desired maximum voltage, it should be alright. 

If the balancers are just "tugging gently" (i.e. < 1A), I don't expect any noticeable voltage drop on the wires you have.

[Sid Genetry Solar]

I personally run my LiFePo4 bank at 56.5 "bulk / absorption" and 56.0 "float"...staying down from floating the batteries at 100% SOC all the time.  Rumor has it that Lithium batteries don't particularly like being "floated" at high SOC...I dunno, I just know I'm really happy with the LFPs, worlds better than SLA.

[TheButcher]

My understanding of it is that at higher voltages the growth rate of dendrites becomes an issue.  Continual 3.65V will still see the cell do the rated number of charge/discharge cycles but if maximising cell life is the plan, the 80/20 SOC window is the way to go.

Charging up to 95+% SOC, holding for a while, then releasing back and holding a lower voltage is still recommended for most systems, unless you have a high current balancing system such as Sid's, as this allows a typical BMS with < 200mA of balancing current capability to bring all the cells to the same voltage.  The higher up the SOC curve you go the lower the balancing current.  I charge my own batteries up to 3.6V / cell, let them balance for a while up there, then drop back to 3.375V / cell.

[TheButcher]

That should be 3.55 not 3.6.

[dochubert]

Well, I'm calling it a successful operation!  I'll be leaving the charge settings where they are, 56.8 bulk and 56v float.  The cells are well balanced for the most part.  The 3 cells mentioned before are still a touch lower than the rest when at full charge, but still improving slowly.  Even if they get no closer, they are close enough.  Not new batteries, remember.  And I am able to utilize a 56v charge versus 54v before.  A significant addition to my bank's capacity just with that change.

So now I have to get another 8 balancers and put them on my 24v bank.  ( 6 more of the same type BYD modules all in parallel. - The fun never ends!)

[Sid Genetry Solar]

Curious what you notice with 56.8 bulk, 56.0 float.  Back when I was tweaking in my new LFP bank, I noticed that if the voltages were too far apart, the registered MPPT current would actually go slightly negative (Morningstar Tristar, seems to be normal for them??), and the inverter/loads would run off the batteries for several minutes while the voltage slowly fell to the new "float" setting--before the MPPT finally "re-took" the loads.  (In your case with a much bigger battery, it likely will take a lot longer.)

[TheButcher]

That's probably going to come down to where the charger is sensing battery voltage.

With lead acid this isn't really a problem since lead acid is held artificially high, typically 13.8v per 12v, so once the charger is no longer providing enough current the battery quickly drops about 1V.  Even with completely inappropriate wire gauge vs current there is always going to be enough voltage drop for the charger to go all-in and try to pull it back up by supplying all the current it can.

This all changes with lifepo4 and it's very flat main discharge curve.  Unless you are way up the top of SOC even 30mV of drop equates to quite a lot of energy in the battery.  If the charger is monitoring voltage at its terminals rather than the battery's all it will do is supply enough current to raise its terminals voltage by 30mV so until the battery drains enough to lower the voltage on its terminals to the point where any drop in the cabling / breakers / etc is exceeded the charger will never be in a position to supply all the current the PV permits.

[dochubert]

9 hours ago, Sid Genetry Solar said:

MPPT current would actually go slightly negative (Morningstar Tristar, seems to be normal for them??), and the inverter/loads would run off the batteries for several minutes while the voltage slowly fell to the new "float" setting

I'll have to look closer, but I haven't noticed a negative current.  The system does seem to 'coast' for awhile as it drops slowly to float.

 

3 hours ago, TheButcher said:

If the charger is monitoring voltage at its terminals rather than the battery's all it will do is supply enough current to raise its terminals voltage by 30mV

I see what you are saying, but;                                                                                                                                                           Seems to me if the charger is seeing a significantly higher voltage at it's terminals than at the battery, enough that it is not correctly charging the battery, then the charger to battery cable is too long or too small or both.

[Sid Genetry Solar]

3 minutes ago, dochubert said:

I'll have to look closer, but I haven't noticed a negative current.  The system does seem to 'coast' for awhile as it drops slowly to float.

So my point...just my opinion anyway...is that when the charger goes from "bulk" to "float", I didn't want to be actually discharging the batteries (microcycling) down to the lower voltage.  LFP cells will drop voltage slightly when the charge is disconnected, and I guess I just wanted to see the charge voltage drop slightly, yet without actually discharging the batteries to get there.  Of course, this is more of a personal matter; to each their own.  A lower "float" voltage might increase cell life, I dunno 😉

Negative charge current seems to be a phenomenon of the Morningstar Tristar; I don't understand the reasoning why, and didn't get a solid answer out of Morningstar--but it isn't hurting anything on a great MPPT, so I haven't pressed the question.

[TheButcher]

47 minutes ago, dochubert said:

I'll have to look closer, but I haven't noticed a negative current.  The system does seem to 'coast' for awhile as it drops slowly to float.

 

I see what you are saying, but;                                                                                                                                                           Seems to me if the charger is seeing a significantly higher voltage at it's terminals than at the battery, enough that it is not correctly charging the battery, then the charger to battery cable is too long or too small or both.

Not really.  Remember, we are only talking millivolts of drop.  If you have breakers odds on you are dropping at least 20mV across them at moderate amps.  Every joint, ring terminal, blah blah, it all adds up.  So long as the charger can pull the voltage at its terminals up to its regulation point, ie charger has reached the configured CV limiit, it will not provide any more current.  Grab a few meters, clamp, and observe your system in just this situation, assuming you aren't doing remote voltage sensing at the battery terminals.

You won't see it Sid.  Unless you make float = absorption you will always see some draw out of the battery until it reaches the lower voltage the charger is now targetting.  Even if you target the area with lifepo4 is in the fairly flat part of the graph 0.8V difference is 50mV.  50mV of movement at this point in the graph represents quite a lot of battery capacity.

Of course, once you load the system further unless the charger can see the battery terminal's voltage the battery will continue to provide a good proportion of the load even if solar is more than capable of fully carrying it alone.

Edited August 26, 2021 by TheButcher

[dochubert]

On 8/26/2021 at 4:44 PM, Sid Genetry Solar said:

So my point...just my opinion anyway...is that when the charger goes from "bulk" to "float", I didn't want to be actually discharging the batteries (microcycling) down to the lower voltage.

Seems the best way to manage that is to keep the settings for bulk and float fairly close together, which you are already doing.  Can't stop the microcycling completely, so minimize it with tight settings.  Making bulk and float equal defeats the purpose of having two setpoints, so back to personal choice.  What settings do we believe provide the best performance balanced with the best preservation of cell life.  At this point, until I see some persuasive evidence to adopt different setpoints, I like the results i'm getting with my current settings.  Time will tell....

[dochubert]

I ordered 8 more balancers on Aug 25 for my 24v bank.  The huge heatsink on the 48v bank works very well but is probably overkill.  For the 24v bank I don't have room to mount one that size so will take a different approach.  Planning to use individual heatsinks, so 8 of the type used for solid state relays should do very well in much less space.  Have 4 on hand so must run down to the electronics recycler store on friday when I'm in Boise for the other 4.  (They have a bin full of them - cheap, like me!)

Mounted in a straight line with one or two 4 inch fans for the lot.  Will put the temp sensor on the balancer that usually starts balancing first.  Tight temp settings on the big heatsink keeps it very cool.  28-26C (82-78F)  Should work about as well for the smaller setup on the 24v.

Still have to parallel all the cells and disconnect the individual active balancers on the 24v bank.  Probably take awhile to get the new balancers anyway.

[Sid Genetry Solar]

1 hour ago, dochubert said:

The huge heatsink on the 48v bank works very well but is probably overkill. 

Ha, yes...it seems to me that LFP are quite easy to balance.  My balancers got fairly hot the first few days--but that was it.  Now they never get even the slightest bit warm.  Probably don't even need a heatsink.

Li-Ion batteries on the other hand...they never seem to fully balance and are always running the balancers really hot.  Or else I've never had good Li-Ion batteries...