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

[ebaugh]

Originally Posted by Maine Sail

Unless an I/C is your primary charge source I usually consider it a load bus item. If it was to be your primary source of charging then it could go on the charge bus. The AC input can still be interrupted by an HVC relay to cut charging, despite it being on a load bus. Many ways to wire them in..

I believe all sources of charging, primary or otherwise need to be hooked into the LFP monitoring system.

If you use the AC transfer switch inside the I/C, and most installations do, when you pull the AC input, it will switch to invert mode. Not desirable, and further, it will bind the neutral to ground, a no-no with simultaneous shore power. Especially if wired with the inverter only powering a sub panel, not the entire boat.

[goboatingnow]

Originally Posted by CharlieJ

Lloyd-

I am sorry, but I do not understand your point. Several of the LFP cell manufacturers lists a float voltage on their spec sheet. They also provide nominal voltage, nominal charge current, some of them provide pulse charge current, standard charging method; e.g., "C/3 constant current to 3.65VPC then Constant 3.65VPC with current taper to C/20"; cutoff voltage, etc., etc. Taken as a whole, I am assuming that this is "charge cycle management". But it includes a value for float voltage so, my take is that the manufacturers of the cells intend on the users of the cells to maintain them at the prescribed float voltage.

Li tech should not be floated , the anode starts plating up

Dave

[s/v Jedi]

Originally Posted by goboatingnow

Li tech should not be floated , the cathode starts plating up

Dave

I think, if you float it, at least use some venture capital!

[ebaugh]

Originally Posted by goboatingnow

Li tech should not be floated , the anode starts plating up

Dave

Can you point to research that shows this for LFP? It is a serious issue for LFP adoption if true.

Where floating means maintaining a CV in the 3.35-3.55 Vpc range.

[goboatingnow]

Originally Posted by ebaugh

Can you point to research that shows this for LFP? It is a serious issue for LFP adoption if true.

Where floating means maintaining a CV in the 3.35-3.55 Vpc range.

All the charge circuits I've seen for small cylinder cells , finish charging and disconnect from the cell. The re-engage on the proportion of terminal voltage anywhere between 65to 90%.

This is also to prevent lots of mini cycles.

LFP is Li. So the same rules apply

Here's a non technical summary http://batteryuniversity.com/learn/a..._ion_batteries

Any why float Li. What does it achieve. LA is floated because self discharge is high.

Dave

[daddle]

On the subject of switching off an alternator:

There are two considerations when opening the battery connection of the alternator. The most dangerous one is a one-time positive spike of voltage at the positive terminal of the alternator. This is from the energy stored in the magnetic field of the alternator. The voltage can much higher than it is practical to design the alternator diodes to block. They will typically fail by their reverse blocking voltage rating being exceeded. This is the famous failure mode caused by battery switches. It is a one-time event at the instant of the switch being opened. The Zap-Stop devices, varistors I think, work well to protect against this.

The second problem is that a disconnected alternator at high RPM can have a rather high terminal voltage. Maybe 100 volts, I might guess. This might also destroy the rectifier diodes.

Any switch or relay in the battery leads of an alternator is problematic. Disabling the field or regulator is the prudent method. As a poster asked above: it would not be safe to switch the field an battery relays at the same instant. I would guess you would want a small time delay from field disable to battery switch opening. Hard to guess but 100 milliseconds would be my first estimate.

Sorry about the generalities, but I don't have any alternator specifications to calculate from.

Interestingly, for geeks, DC motors have the same switching spike problem. But it is trivially solved with a reverse diode across the motor terminals. This does not seem to be possible with an alternator because of the AC windings of the device. Maybe?

Perhaps it is best to only disable the field on an Over Voltage alarm.

[neelie]

What about using the HV alarm to trigger the early opening of the relays which need it. The rest get disconnected by the HV relay trigger.. Perhaps some 0.02V later?

Question for gurus... How best to wire 2 or 3 relays in parallel given the limited current available from a BMS or Cell Logger?

[diugo]

Originally Posted by Maine Sail

This is what CV charging looks like on my pack at 13.8V when the battery has reached 13.8V.. 0.0A

I've been in the shop now for two hours and not seen the charger need to kick in to supply any current.... These cells hold voltages tremendously well. It does happen when the voltage falls off of 13.8V but the current only kicks back in for a couple of minutes at best. Once back to 13.8V there is no need of any more current because the bank is "at voltage" so the current delivery stops until the voltage drops again....

This behavior is likely why Genasun says 14.2V CV is suitable for their packs...

Contrast that to the many LA batteries I have in the shop. They will keep taking current indefinitely and never shut the charger down...

This is a vitally important differentiation and perhaps the keystone of proper float management for LFPs.

Virtually any off-the-shelf LA wall charger can do "forced float", which is really nothing more than a DC power supply. This kind of charging is bad, bad, bad for lithium ion.

In contrast, what Genasun apparently is doing with its "CV" solar chargers---and perhaps more accurate testing should be done to verify this---is effectively disconnecting the panels, then closely monitoring the battery voltage, only reconnecting the panels when battery voltage dips slightly. In this way, the battery avoids the constant "polarizing electric field" conducive to undesirable plating.

Going a step further, it would be also be interesting to see whether this good charging behavior is an innovation exclusive to Genasun---or if other popular LA solar charge controllers already exhibit similar behavior when hooked up to LFPs. Easy to test---we are looking for truly zero current during float, as opposed to near-zero. PWM duty-cycle-wise, no more than a few percent at most, I would think.

BTW thanks Daddle for that master's thesis. Looks like I've got some good reading to do this weekend!

[Dsanduril]

Originally Posted by neelie

Question for gurus... How best to wire 2 or 3 relays in parallel given the limited current available from a BMS or Cell Logger?

The traditional answer to this is to wire a single "pilot" relay to the BMS/alarm contact that cannot handle much power. The alarm then only has to power the coil of one small relay. A second circuit is then run through the contacts of the pilot relay and feeds the coils of the main contactor(s).

[ebaugh]

Originally Posted by neelie

What about using the HV alarm to trigger the early opening of the relays which need it. The rest get disconnected by the HV relay trigger.. Perhaps some 0.02V later?

Question for gurus... How best to wire 2 or 3 relays in parallel given the limited current available from a BMS or Cell Logger?

What's needed with a battery management app for marine is a series of relay triggers that can be set to any voltage desired and any time delay desired. That way if you wanted to both stop charge and a disconnect, you could use the same voltage, but set the charger disconnect to 5 seconds and the battery disconnect to 10 seconds. The "alarm" would be set to a lower voltage so the scenario would be alarm first, 15 to 60 minutes later depending on charge rate and loads, pull the charge sources, if the problem corrects itself the disconnect is avoided, if not a few seconds later the pack is disconnected, There is no system you can buy to do this today that I know of.

[s/v Jedi]

Originally Posted by ebaugh

What's needed with a battery management app for marine is a series of relay triggers that can be set to any voltage desired and any time delay desired. That way if you wanted to both stop charge and a disconnect, you could use the same voltage, but set the charger disconnect to 5 seconds and the battery disconnect to 10 seconds. The "alarm" would be set to a lower voltage so the scenario would be alarm first, 15 to 60 minutes later depending on charge rate and loads, pull the charge sources, if the problem corrects itself the disconnect is avoided, if not a few seconds later the pack is disconnected, There is no system you can buy to do this today that I know of.

You could buy a BASIC Stamp and put anything you want together. Once you have it perfect find a geek like me to convert that to straight microcontrollers. Oh wait, this is what the BMS guys do...

[ebaugh]

Originally Posted by diugo

This is a vitally important differentiation and perhaps the keystone of proper float management for LFPs.

Virtually any off-the-shelf LA wall charger can do "forced float", which is really nothing more than a DC power supply. This kind of charging is bad, bad, bad for lithium ion.

In contrast, what Genasun apparently is doing with its "CV" solar chargers---and perhaps more accurate testing should be done to verify this---is effectively disconnecting the panels, then closely monitoring the battery voltage, only reconnecting the panels when battery voltage dips slightly. In this way, the battery avoids the constant "polarizing electric field" conducive to undesirable plating.

Going a step further, it would be also be interesting to see whether this good charging behavior is an innovation exclusive to Genasun---or if other popular LA solar charge controllers already exhibit similar behavior when hooked up to LFPs. Easy to test---we are looking for truly zero current during float, as opposed to near-zero. PWM duty-cycle-wise, no more than a few percent at most, I would think.

BTW thanks Daddle for that master's thesis. Looks like I've got some good reading to do this weekend!

Maine tried to make this point before, but Genasun may do what you describe with their solar controllers, but their LFP battery systems and related management systems don't do this. Their LFP battery systems have an overvoltage solenoid and an under voltage one. And unless it's an undocumented feature, the HVC relay only activates for HVC in their BMS, not to regulate charge. If its an undocumented feature, it could explain the 14.2V max charge recommendation.

Boats at a marina have DC loads, more if occupied, but even some if unattended. Nothing I know of is equipped to stop/start charge for LFP. If multiple small cycles is better than a CV float, it needs to be done. But Im still waiting to see documented evidence of this behavior if the cells are CV floated at some reasonable number below 100% SOC. Say 90% for discussion.

[ebaugh]

Originally Posted by s/v Jedi

You could buy a BASIC Stamp and put anything you want together. Once you have it perfect find a geek like me to convert that to straight microcontrollers. Oh wait, this is what the BMS guys do...

Im planning to build one around an industrial PLC. And perhaps a very basic one around a Flyport MCU if it works. The issue with the real basic micro controllers is differential voltage sensing needed and the noisy environment if not isolated from the boats other systems.

[diugo]

Not sure if a time delay is really necessary. Why not just a unique voltage for each charging source? Example, at 14.6V, disconnect alternator field. At 14.8, disconnect shore power charger. At 15.0, solar charger.

The right tool for this job is the microcontroller. It has multiple analog inputs to read voltages (scaled by a divider) and multiple digital outputs to drive disconnects (through pilot relays or FETs). Cost, under $30. Drawbacks? Requires programming in a high level language like C. Also, all the "brains" are in one unit, which could fail.

Alternative? Individual voltage-based disconnects for the marine market, such as maybe LVA Low Voltage Alarm and Relay | Technologies LLC , for each charging source. Drawback? At $85 each, quite pricey---especially if you need more than one.

[Singleprop]

There are many BMS's and cell loggers mentioned in this thread, but very few links to useful relays, solenoids, "make before break" switches. Could any of you post these links?