LFP FLA hybrid

[tanglewood]

Originally Posted by Dockhead

Well, I'd want to be sure to think through all these issues and be quite sure you don't cause damage even in an abnormal LVC event.



You are right that these are issues not unique to LFP power systems.



For example, my system has remote control contactors for isolating both battery banks, which was part of the original design of the electrical system. These contactors have large diodes on them to dissipate flyback currents. My bow thruster control box also has some kind of snubber in it.


I suppose everyone knows that you have to protect an alternator from sudden isolation from the load. There are plenty of ways to solve this. The simplest way is to just whack a big Zener diode on it -- which is all the old Zap Stop system was.

Clamping diodes are pretty common in large relays, but I'm pretty sure their purpose is to prevent welding of the relay contacts. Extinguishing the arc is a real challenge with DC switches, and is why you will frequently find switches with very different AC and DC current and voltage ratings. AC is inherently self-extinguishing because there is a zero voltage crossing every 100th, or 120th of a second depending on the AC frequency. Not so with DC, so a lot harder to extinguish the arc, especially as the voltage climbs.


Regardless, I totally agree that diligence is required.

[Dockhead]

Originally Posted by tanglewood

Clamping diodes are pretty common in large relays, but I'm pretty sure their purpose is to prevent welding of the relay contacts. Extinguishing the arc is a real challenge with DC switches, and is why you will frequently find switches with very different AC and DC current and voltage ratings. AC is inherently self-extinguishing because there is a zero voltage crossing every 100th, or 120th of a second depending on the AC frequency. Not so with DC, so a lot harder to extinguish the arc, especially as the voltage climbs.


Regardless, I totally agree that diligence is required.

Ah, yes, you're no doubt right about the diodes on my contactors. I had forgotten about that issue.

[evm1024]

A bit off topic - In past years I wrote a model of an entire power grid. 12,000 miles of lines and every bushing, insulator, load break etc. This model produced a "Bottleneck" report that was (is?) used to figure out what was going to blowup when we had a disconnect and large transients.

Arcing on the load break disconnects is a major problem and various ways are used to minimize it. The most fun is air blasts to blow "out" the arc.

Here is a link to a video of a not too big load break disconnect closing then opening. Look at the disconnect arc near the last of the video.

Same thing happens (to a smaller degree) in relays, switch and such on out boats.

https://www.youtube.com/watch?v=Av7d94LT4hE

[tanglewood]

Originally Posted by Dockhead

Yes, the power in the inductor is in proportion to the magnitude of the current, but the voltage of the spike is a function of the RATE OF CHANGE of current, not the magnitude of the current:



Attachment 177218




It's a corollary of Faraday's Law:


Attachment 177219


Where epsilon is electromotive force and ΦB is magnetic flux.


What this means practically is that high induction loads of very small current can produce very high voltage inductive spikes.

OK, so it's been a very, very long time since I did this stuff, so I could be remembering incorrectly.... so please tell me where you think I skip the rails....



With a steady state DC current, di/dt is zero, so no induced voltage. But there is established magnetic flux, and the magnitude of that flux is proportional to the current. All the stored energy is in that flux, so stored energy is proportional to current.



Now, take away the current. That creates the di/dt that creates the back EMF. The question is what determine di/dt?


Take a coil or inductor that will suddenly become open circuit. I drops from it's initial value to zero immediately, slowed only by capacitance of the remaining circuit. The bigger the initial current, the bigger the di/dt. Now consider a circuit that has some resistance in it. di/dt will be determined by the resistance, but still, the larger the initial current, the larger di/dt.


Now look at it another way. There is stored energy in the inductor when the steady DC current is flowing. When the current stops, that energy needs to be dissipated. If the circuit is high resistance then current flow will be limited, so voltage will be high. And if circuit resistance is low, like with a LA battery connected, current can be high and voltage will be lower.

[Dockhead]

Originally Posted by tanglewood

OK, so it's been a very, very long time since I did this stuff, so I could be remembering incorrectly.... so please tell me where you think I skip the rails....



With a steady state DC current, di/dt is zero, so no induced voltage. But there is established magnetic flux, and the magnitude of that flux is proportional to the current. All the stored energy is in that flux, so stored energy is proportional to current.



Now, take away the current. That creates the di/dt that creates the back EMF. The question is what determine di/dt?


Take a coil or inductor that will suddenly become open circuit. I drops from it's initial value to zero immediately, slowed only by capacitance of the remaining circuit. The bigger the initial current, the bigger the di/dt. Now consider a circuit that has some resistance in it. di/dt will be determined by the resistance, but still, the larger the initial current, the larger di/dt.


Now look at it another way. There is stored energy in the inductor when the steady DC current is flowing. When the current stops, that energy needs to be dissipated. If the circuit is high resistance then current flow will be limited, so voltage will be high. And if circuit resistance is low, like with a LA battery connected, current can be high and voltage will be lower.

"Slowed by REACTANCE", not capacitance, no?


I suppose the rate of change of current is related to the starting magnitude of the current. It's been a while since you've done this -- me too.

[ramblinrod]

Originally Posted by Dockhead

"Slowed by REACTANCE", not capacitance, no?


I suppose the rate of change of current is related to the starting magnitude of the current. It's been a while since you've done this -- me too.

So in far fewer words and in practical terms, if a person or BMS disconnects an LFP bank (without disconnecting load and charge sources gently first) there is a very good chance of:

A) A significant transient voltage spike with normal cruising sailboat loads (refrigerant compressors, autohelm, magnetron radar, etc.)

B) A huge transient voltage spike with high current loads (bow thruster, windlass, inverter).

Question: When is this disconnect most likely to occur?

Answer: When the bank is low, someone is not paying attention, and operates one of the high current loads.

That big bank of low impedance batteries is no longer there to absorb it because it was just disconnected!

Result: Zap! All electronics are gone. It was like they just took an indirect lightning strike.

[evm1024]

Originally Posted by ramblinrod

So in far fewer words and in practical terms, if a person or BMS disconnects an LFP bank (without disconnecting load and charge sources gently first) there is a very good chance of:

A) A significant transient voltage spike with normal cruising sailboat loads (refrigerant compressors, autohelm, magnetron radar, etc.)

B) A huge transient voltage spike with high current loads (bow thruster, windlass, inverter).

Question: When is this disconnect most likely to occur?

Answer: When the bank is low, someone is not paying attention, and operates one of the high current loads.

That big bank of low impedance batteries is no longer there to absorb it because it was just disconnected!

Result: Zap! All electronics are gone. It was like they just took an indirect lightning strike.

This is a non-starter.

In a properly designed LiFePO4 system the BMS will not disconnect the bank before it has warned the operator for some time.

In the case of over-voltage the BMS will shut off charging long before the BMS feels the bank needs to be disconnected (if you allow your BMS to disconnect at all. I do not).

Additionally, we all know not to switch the battery disconnect off when the engine is running. That unlikely event does not change the results regardless of the battery type. Switch out a LiFePO4 bank or FLA bank when the alternator is running and ... This is an operator error. Oh wait. The engine is running so we must have charging so we do not have a low voltage disconnect. (or our operator has been ignoring their SOC for a few days.)

Additionally LiFePO4 is capable of supplying full current right up to very low SOC. Again a non-starter.

In my own system I do not allow the BMS system to take the house bank off line. I do allow it to sound some loud alarms.

[Dockhead]

Originally Posted by ramblinrod

So in far fewer words and in practical terms, if a person or BMS disconnects an LFP bank (without disconnecting load and charge sources gently first) there is a very good chance of:

A) A significant transient voltage spike with normal cruising sailboat loads (refrigerant compressors, autohelm, magnetron radar, etc.)

B) A huge transient voltage spike with high current loads (bow thruster, windlass, inverter).

Question: When is this disconnect most likely to occur?

Answer: When the bank is low, someone is not paying attention, and operates one of the high current loads.

That big bank of low impedance batteries is no longer there to absorb it because it was just disconnected!

Result: Zap! All electronics are gone. It was like they just took an indirect lightning strike.

In even fewer words: not really.


I have remote control battery isolaters in my system, operated by switches on my main electrical panel.



The batteries have been disconnected a few times from the loads -- as suddenly as a LVC would have done -- by crew pressing the wrong button.


No "zap", no electronics gone, no lightning strike, no problem.


The Zener diodes on the contactors may or may not have dealt with with any inductance spike, or else there was simply not enough inductance in the loads to create much of a spike.


In any case, this is a straightforward electrical design issue. It's good the be aware of it and to be sure that sufficient account has been taken of it in the design process, but it's silly to be afraid of it.

[Dockhead]

Originally Posted by evm1024

This is a non-starter.

In a properly designed LiFePO4 system the BMS will not disconnect the bank before it has warned the operator for some time.

. . .



Additionally LiFePO4 is capable of supplying full current right up to very low SOC. Again a non-starter.
. . . .

Exactly. If the fear is that a LVC event may occur because of operating heavy loads when the bank is at a low SOC -- lithium has almost no voltage sag and won't trigger a LVC unless the battery is really empty.


As MaineSail said -- we have to forget what we know about lead batteries.

[ramblinrod]

Originally Posted by Dockhead

Exactly. If the fear is that a LVC event may occur because of operating heavy loads when the bank is at a low SOC -- lithium has almost no voltage sag and won't trigger a LVC unless the battery is really empty.


As MaineSail said -- we have to forget what we know about lead batteries.

Well there is certainly "disconnect" somewhere, because Mainsail won't even consult remote customers on LFP anymore because too many were zapping their electronics and trashing their LFP banks.

[evm1024]

Originally Posted by ramblinrod

Well there is certainly "disconnect" somewhere, because Mainsail won't even consult remote customers on LFP anymore because too many were zapping their electronics and trashing their LFP banks.

A quick look at Mainesails website shows:


Beyond all that we’ve had a LFP bank on our own boat since 2011. I am a huge fan of LFP banks, for many reasons, but I will not act as a “fan boy“, that would be unfair to you the reader… I will give you both sides of the story, not just the pluses.

And

I no longer do any long distance consulting or design work for LFP batteries or systems. If the boat is not readily accessible here in Maine, Casco Bay to be specific, I can no longer do LFP consults. Why? Far too many owners insisting on cutting corners and multiple situations where end user implementation was not done to the design or design criteria have forced my hand. If you happen to glean something from this article, that’s great, but please don’t email me with further consulting questions on LFP.

So the disconnect is in understanding why mainesail does not do long distance consulting.

The disconnect is when someone uses what they know about FLA to make choices with LiFePO4. Simple things like assuming that the LiFePO4 voltage will sag at low SOC (what is low SOC? 20%?).

[Delfin]

Originally Posted by Dockhead

Yes, I'm coming around to this view. The more I study it, the more straightforward the issues seem to me.



Moreover, it's not like managing a lead bank for off grid use requires no thought or care. Far from it! How much time have I spent in my cruising life worrying about finishing charges, looking for shore power because I haven't had a good charge in a few days, equalizing and worrying about overdoing it, bla bla bla. A lead acid battery bank is NOT a low maintenance system on a cruising boat -- far from it.


Lithium does require some knowledge to design and install, but probably requires less knowledge to actually manage the finished and working system, than lead does -- just, it's different knowledge, and what you know about lead doesn't help you, so you have to start over from zero.



I guess a well-executed lithium power system will greatly reduce the work load on the cruiser, flowing from the power system. Lithium is somewhat complex to implement, but once you have it, it's like a fuel tank -- don't have enough power, pour some in. Have some power? Take it out. Much simpler than lead. Equalization? Finishing charge? Sulfation? What's that?

The fuel tank analogy is how I describe the LFP mindset to others. Take out some energy, put some back in. Sometimes you put it all back, sometimes not, but the fuel tank doesn't care. The bigger the nozzle the faster you fill the tank. And you shouldn't just stick the nozzle in, turn it on and walk away. The auto shutoff might not work, so you pay attention to the refueling process. There is nothing on a world cruiser that doesn't require attention. Sails, charts, radar, water usage - all require attendance. The desire to install LFP and then another it seems pretty pointless to me.

[rgleason]

I am going to correct my earlier post

Hybrid System

FLA Flooded Lead Acid 225ah - Existing, Replace every 6-8 years.

LFP 90ah 1 x $800-1000
Sterling Pro Ultra B2B $350 + wiring.
Change settings of charger appropriately.

LFP LithiumFePo System
LFP 160 ah with BMS under/over voltage limits. 2 x $1000 = $2000
Would still get a shore charger probably $350.

I knew the argument for a full LFP system would be made and the argument for simplicty is rational. I am still not so sure there are are not good reasons to go with a hybrid system:The Lithium battery type seems solve the 100%SOC issue for FLA.

1. The high Lithium battery acceptance rate can be used to shorten charging time, with long times at high amps, but shorter periods, which supports greater sailing rather than motoring.
2. The FLA batteries as "House Bank" provide an inexpensive, known and common battery type that now has a much better chance of lasting for 6-8 years, providing the source for the house loads.
3. The Lithium Battery would become the "Charging & Reserve Bank" acting also acting as "Emergency" in the event the House Bank is discharged or fails.
4. The overall system might be very efficient and effective.
5. Solar PV would not have to be as large or maybe is not really necessary because the Alternator is used for the most part.
6. The thing that might put this system in the trash is:

* How often would the LFP batteries have to be brought up to 100%?

(Not often, not necessary)

* How long would that take for a small LFP?
(Dependent on the Alternator)

* What is the LFP typical:
- Max Discharge level C (Between 70%-50%SOC 2500-5000 cycles.)
- Bulk level C -max absorption amps & voltage level. (14.2-14.4v quite high 500a )
- Absorption level C -CV? amp ranges, trailing amps? (Not an issue with LiFePo)
- Float level C & voltage?



I need to read the rest of this string, next, but I thought I should correct this.

[rgleason]

Originally Posted by rgleason
Tanglewood
What is "Most cell specs allow for 2C discharge, " this? Unheard of.

John:

No that is normal.

Again, not drop-ins where a cheap BMS gets in the way.

Direct access to and from the prismatic cells themselves.

2C = 2 x C? I do not understand. Incomprehensible.

Is this value actually supposed to be 0.2C?

[john61ct]

No 1C is when current = AH capacity

So 100A charge rate into a 100AH bank.

Or discharge.

2C is 200A

.2C is 20A

.02C is 2A

EV world, most of the industry, is violent high C rates, pushing SoC for max range.

House bank is gentler, sub-C rates, better able to focus on longevity.

Better explanation than the conspiracy one maybe.