LiFePO4 Puzzle...

[openplanet]

I will be switching from 17.5 kWh of FLA to 19.5 kWh of LiFePO4 (4 100ah 48v rack mount units with integrated BMS) in about 2 months. I've been studying LiFePO4 care and feeding, scouring the forums, etc. I'm determined to keep my bank between 80% and 20% SOC. In addition to the integrated BMS, I intend to install LVD and HVD devices. Here's what puzzles me: Let's say the bank is "fully" charged at 11:00. I now want to use the available array power to run whatever loads there are. But if I bring up a load greater than what the array can supply, I want the battery bank to help out. I don't see a way to accomplish this, even with intelligent devices. Any thoughts would be appreciated. Thank you.

[Mike90250]

Yes, the array will fluctuate all day long, it can go + - 15% with imperceptible to the human eye, changes in lighting conditions. So there will be loads that will exceed the harvest.

So, what are your loads ? 48VDC devices ? A off grid/BiModal inverter feeding a hydroponics tank ? bitcoin miners ??

The trick will be to set the Charge Controller float voltage, to a voltage that won't dump any more current into the batteries, and when a heavy load comes along, the batteries will buffer it out.
Defining that voltage will be key, maybe 75% of full ? Set Bulk & Absorb to a voltage that corresponds to 80% full, and then float is, say 75%. You will have to study your battery specs very closely and have a good source to calibrate or reference your meter to, so that you don't exceed the limits you are looking for.

The Li guys will be along with more of the details. Good Luck

[Ampster]

I guess the simple answer to what puzzles you is that the inverter and charge controller will behave the same way regardless of the type of battery. As Mike90250 alluded to, you will set the Absorb voltage to whatever you calculated to be 80% of capacity. I would not be surprised if the manufacturer did not already use some conservative settings. That would especially be true if they guaranty the battery for a number of years. Do you have a chart of the charge/discharge curves?
I would also see if the internal BMS has LVD and HVD built into it. Those settings will give you a clue and might eliminate the need for redundant devices. The Absrb voltage will be slightly below your HVD and your inverter presumably has a low voltage cutoff. I am not sure you need a third safety net. Do you use them with Lead Acid?

[openplanet]

Silly me! Yes, the BMS will take care of the LVD and HVD, so adding devices would just add unnecessary complexity.
And silly me again...of course, setting Float properly will take care of the issue.
You can be sure I'm going to watch these puppies very carefully at first. And at second. And at third!
Thank you.

[Ampster]

Lithium batteries dont like to be on Float for very long. That is why it was recommended to set Float below Absorb. The important thing in Absorb is the setting for time in Absorb or the Amp setting where Absorb cuts out. Absorb is the Constant Voltage phase and during that phase the Amps will taper off as the battery gets full. You can see that on a charge curve. Did you get anything from the manufacture?

[openplanet]

Not yet, But I will be.

[inetdog]

Just keep in mind that various Charge Controllers will differ in the exact current values that will trigger an exit from Float mode back to Bulk or Absorb. But there will be a current value (not a voltage value, since Float is constant voltage) that will terminate Float. That is fine, as long as the Absorb voltage is also not too high for a 90-% charged battery bank. (Assuming that your "full charge" limit is set to 90% or lower for maximum battery life.)

[openplanet]

I have 2 Midnite Classics and a Whizbang Jr., so fortunately there'll be good control.
The challenge is that the charge/discharge curve is so darn flat for LiFePO4s that the Classics might not be able to differentiate such tiny voltage differences for purposes triggering various charge mode changes...
Oh, meant to add that I've ordered a used Fluke 87 III to be sure I'm precise on the voltages. Having spent $$$ on the battery bank, makes sense to spend just a tad more for that.

[Mike90250]

............Oh, meant to add that I've ordered a used Fluke 87 III to be sure I'm precise on the voltages. Having spent $$$ on the battery bank, makes sense to spend just a tad more for that.

Depending on your skill level and budget, I suggest you verify the calibration of a meter before you trust it, unless the used meter has a metrology cert with it.

DIY: https://www.instructables.com/id/Pre...ion-reference/

ready made: http://shop.voltagestandard.com/prod...1&categoryId=1 ( Chose a voltage close to your cell voltage, for the scale you will use on your meter )

[inetdog]

I remember when I built a Heathkit VTVM (Vacuum Tube Voltmeter, with a 10 Megohm input resistance, for you youngsters). The manual said to use a brand new unused carbon zinc battery (C cell) for calibration. I think it was 1.54 volts, but it might have been 1.56. Since then, changes in the chemistry of the "standard" carbon zinc battery have changed the open circuit voltage somewhat.

[tom rickard]

Don't get too stressed about it. There are thousands of people out there with setups like yours that have been happily working for many years.

On the Midnite classic, set the absorb to 54.5v, time out after 30mins. Set Float for 53.5v

These settings are conservative, approx .5v under upper knee voltage. They will still give you over 80% cell capacity though.

No need for a super fancy multimeter, the Midnite will be fine.

Your inverter doesn't care if it gets it's supply from the panels (charge controller), or the batteries. The charge controller will maintain the voltage (SOC) at your preset level.

[Ampster]

........

On the Midnite classic, set the absorb to 54.5v, time out after 30mins. Set Float for 53.5v

These settings are conservative, approx .5v under upper knee voltage. They will still give you over 80% cell capacity though.

No need for a super fancy multimeter, the Midnite will be fine.
......

The OP never said how many cells were in his pack so I don't know how you arrived at that voltage. That is too high for 12 cells. Also I believe they are packaged LFP with integrated BMS so more info is needed before one would know the correct voltage.

[tom rickard]

Good point, although i haven't seen any rack mount LiFePO4 48v units that aren't 16cell.

For clarity, absorb at 3.4v/cell, float at 3.35v/cell.

Obviously it would help if the OP gave more info - but all the rack mount LiFe's like BYD etc will work well with these settings

[Ampster]

Good point, although i haven't seen any rack mount LiFePO4 48v units that aren't 16cell.

For clarity, absorb at 3.4v/cell, float at 3.35v/cell.

Obviously it would help if the OP gave more info - but all the rack mount LiFe's like BYD etc will work well with these settings

Do you think there is any risk of degradation floating LFP at 3.35? A few hours might no hurt but days on end without any loads is not often recommended. Usually a LFP charger will go into Contant Voltage until a minimum Amps or say a 30 minute timer.

[tom rickard]

Mine have been doing that for over 8 years, i know of others that have had that float setting for over 10 years.

In practice in an off-grid system there is always some load. If i was to disconnect my inverter, i would also disconnect my charge controllers.