Tuesday, April 28, 2020

Re: [electricboats] AGM or LiFePo4

I'm not so sure that "the BMS should only be used in case the charger fails"---if the charger fails ON, even a BMS will probably be useless to stop it.  Same story for a charger that fails by allowing pack to discharge into the charger after shutting down.  Ignoring the charger failure scenarios, a BMS is useful during charging in 2 or 3 ways:

·       Identifying that a cell has reached a maximum voltage and informing the charger so it can shut charge current OFF

·       Cell balancing (typically by bypassing current around full cells)

·       Allowing charging to restart after cell imbalance corrected

 

The basic CV/CI charger is blind to what is happening at the individual cell level, while a BMS cell monitor/manager is cognizant at that level.  And while it is important that the BMS is commonly referred to as being used to "balance" all the cells, they often can and should provide critical information, data or control to the charging system to ensure against overcharging a single cell.  Yes, wildly out-of-balance cells should not be expected if a BMS has been managing cells regularly or at least occasionally.  But they can happen.  Fortunately, the outliers are usually "leaky" or perhaps most common, self-discharging or discharging thru leaky BMS mosfets.  And so a single or a few low cells won't cause much of an increase in charged cell voltage for the rest of the majority of cells at the final charge voltage.  If, however there were widespread self-discharging but a small number that did not, then those few cells are at risk if a blind charger does not respond to cell voltage reaching a maximum first.  And so, a BMS sending out a stop command to the charger can be useful, and not because the charger itself has failed.

 

In my case, I have some 240 cell pairs, all with BMS connections but only maybe once every 12-18months do I ever power these BMS circuits.  I also have not generally been using the BMS to control and limit charging.  Mostly I charge with a CV/CI power supply, start it charging and come back the next day to shut it off.  I do make sure ahead of time that there are no outlier cells that are significantly higher voltage than the rest or that there are too many low outliers which would add up to allow the high cells to get overcharged.  And I stop charging at 46-48v while the pack is rated for over 49v.

 

While this may sound crazy, I also only am charging the pack about once per 2-3months, so we're only talking 4-9 charge cycles.  I trust the cells themselves more than I trust the leakiness risk of the BMS mosfets (and standby power).  The downside is that I do see 6 out of 240 cell pairs that significantly drift (again, either self-discharging or discharging thru an unpowered mosfet/load resistor).  And 3 of these are significant---between 100 and 350mv decay in voltage over a year.  It's important (given I am not powering the BMS circuits always) to consider these decay magnitudes when letting my pack voltage drop to its lower levels.  Yes, I could just trust the BMS cards, and I do, but I do not want to stress them for no reason and risk my pack life due to failed BMS self-discharge.  I personally think it's more risky to power these BMS cards 365+ days when I really only need to power them before or after the few times I'm going to charge the pack.

 

I recently found myself speaking on the phone with a sales engineer for a noted lithium battery manufacturer.  I thought it was an engineer working for my company so I was frank in my opinion about the batteries in question.  I was quite vocal about my opinion that "no self-respecting engineer would design a built-in BMS that is always powered and will self-discharge a battery before the box was ever opened, sitting on the shelf".  The sales engineer for the battery manufacturer appreciated my frankness.  I said, it's fine to have a BMS that monitors cells when there are no loads on the battery, however it's not fine to design a system that continues such loading to the point that the low cell voltage is reached.  Worse, these common 12v lithium batteries will at that point, either blow a fuse or open a mosfet internally that denies ever being able to recharge the cells or use them again.  I.E. the BMS itself "bricks" the lithium battery.  I pointed out that even the cheapest of kids toys where "batteries are included" include a plastic tab that the user has to pull out before use.  This is to ensure the battery doesn't self-discharge thru the toy/lamp/whatever.  I also pointed out that this was a "design choice" and that a responsible design would instead open up a mosfet that disallowed any further discharge (even internally), but would allow charging.  Laptop computer batteries at least do this somewhat.

 

Many or most modern sealed 12v lithium batteries w/self-contained BMS have this same design flaw.  And they internally draw enough current that within 1-2years on the shelf, they become bricks.  So much for long life of lithium.

Our BMS circuits will similarly draw power and one needs to consider this.  Mine draw about 1watt each.  If I kept all BMS cards powered continuously, a full pack in my boat would discharge thru the BMS cards in 1500 hours (not to mention the DCDC losses to make the 12v for the BMS cards).  While this sounds like a lot of time, it's only 2 months!  So it's important to understand when your BMS is powered, what power it takes, how often you really need it powered and can you sensibly power it only when it's needed.

 

-MT

 

From: electricboats@groups.io [mailto:electricboats@groups.io] On Behalf Of Matt Foley
Sent: Tuesday, April 28, 2020 8:00 AM
To: electricboats@groups.io
Subject: Re: [electricboats] AGM or LiFePo4

 

To piggy back on what Tom said, 

 

Its best practice to have two means of protecting batteries. On the charge side, the charger have the correct charge profile and the BMS is secondary and should only be used as back up in case the charger fails. On the discharge side, the motor controller/inverter should be programmed with the correct voltage settings and the BMS is the failsafe. 

 

Unless you are charging at a high C rate (most of us in DIY electric boats are not) you do not need to worry about CVCC. You can just charge with CV and stop. Unlike lead acid, which prefer to be fully charged and even over charged occasionally, lithium prefers to be less than fully charged. They would be perfectly happy sitting at 50% SOC for years.  If using a lead acid charger, turn off the float charge or make it as low as possible. Be certain it does not have an equalization charge. Overcharging lithium just once can and most likely will ruin them, sometimes in a dramatic way. 

 

For lifepo4 I stop charging at 3.5 VPC. There is almost no capacity above that voltage.  If you look at a discharge chart you will see what I mean. Lifpo4 voltages should not be confused with lithium chemistries with a 4.2 VPC cutoff 

 

I typically use  90% as upper limit and 10% on the low end. 

 

 

 

Matt Foley 

Sunlight Conversions

Perpetual Energy, LLC

201-914-0466

 

 

 

On Tuesday, April 28, 2020, 09:40:12 AM EDT, THOMAS VANDERMEULEN <tvinypsi@gmail.com> wrote:

 

 

MARTIN: Since you have expertise, I hope you'll agree that the acronym 'BMS' stands for battery management system, but that what some people are referring to as a BMS is actually a battery protection system.
A BMS that's suitable for a 16 cell, LiFePO4 prismatic pack, such as might be found aboard a sailboat -- such as the Orion Jr. or the Dilithium BMSC from Thunderstruck -- would provide for measuring and reporting individual cell voltages; balancing the cell voltages within the pack; and also often the ability to measure cell or pack temperature.  The Orion Jr even implements charger control in a single unit, while the Dilithium units from Thunderstruck separate the two functions -- BMS & charge control -- into two different units, but enables using two different modes of communication between the two devices, including CAN.
in contrast, the typical "BMS" that's embedded in 48v battery packs used for electric bicycles, for instance, provides only for the overcharge protection to shut off charging when voltage reaches the safe operating limit, and low voltage protection to shut off discharging to load when voltage drops to the safe operating level for that battery pack.
In the case of battery protection circuits, it's possible for individual cells to get out of balance relative to one another, such that one cell will reach full charge voltage before the others in a pack.  The protection circuit cuts off charging due to the high cell voltage cut-off, and the remaining cells will be less than fully charged.
Many of the ElectricBoat members are familiar with everything I've just mentioned, but it may be useful to repeat the information as new members come aboard.
Fair Winds to all!
[-tv]
Tom VanderMeulen
"Grace O'Malley"
Cape Dory 27
Monroe, Michi.

No comments:

Post a Comment