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

[transmitterdan]

You start out saying you are only speaking for yourself and then condemn anyone doing it not your way as "wrong headed". You can do it any way you like. But it is not "wrong headed" to protect the alternator, have LV and HV cutouts with continuous voltage monitoring of each cell group in series. To suggest that this minimalist set of features is "wrong headed" is going a little too far IMO. I believe that set of features is pretty minimalistic. KISS is in the eye of the beholder.

[Highland Fling]

Originally Posted by transmitterdan

You start out saying you are only speaking for yourself and then condemn anyone doing it not your way as "wrong headed". You can do it any way you like. But it is not "wrong headed" to protect the alternator, have LV and HV cutouts with continuous voltage monitoring of each cell group in series. To suggest that this minimalist set of features is "wrong headed" is going a little too far IMO. I believe that set of features is pretty minimalistic. KISS is in the eye of the beholder.

I said there was a lot of wrong heded thinking......re LIFePO4 battery installations. I deliberately did not criticise anything about that design/installation.

Where in my post did I condem anyone....you need to read the words that I wrote..... IF you read about my system it meets all these conditions / situations. Without getting seriously and unneccessarily complex.

i really don't know where you got all of that from, try and understand what is written here before replying please..

I WROTE and you obviously did not read or understand these simple words.

Just another simpler way of doing things.....KISS. Also I am NOT Looking for an argument here. as to which system is 'best'

THIS IS My System you can design yours to meet your wants desires needs.

[OceanSeaSpray]

Originally Posted by Highland Fling

Nice design and write up of a rather complex complicated LIFePO4 system.

I see KISS was thrown out of the window in fact it looks like it was thrown out of a window on the top floor of the Burj Khalifia.

... I love the design, write up and the diagrams BUT for sure most of this I see as unneccessary for MY BOAT.

.....................

Yes, KISS contains "stupid". A better understanding of lithium (and lead acid) batteries...

Mbartoch's design shows what needs to be done to prevent the charging bus from getting directly connected to the load bus with no battery present if the BMS drops the bank.
It is its main discerning feature and every other design I have seen with a HPBMS had that vulnerability and when this does happen, it tends to cause a great deal of damage and expense.

Things simplify considerably if one stays away from a BMS that is in fact unsuitable and inadequate for the application (same when it comes to its even poorer Australian copy, some look-alike packaged ones etc), because those were never intended to control a split bus system.

Mbartosh's design transforms the totally unsafe "power to disconnect" design into a fail-safe "power to connect" (by the way of two additional auxiliary relays) so if the bank drops, all power is lost and the LVC and HVC NO contactors drop open (instead of closed before) and everything gets disconnected. As it should.

I wouldn't call this a cute, very optional way of doing it.

[exMaggieDrum]

Originally Posted by missourisailor

Hi All,
Just to give another perspective on a LiFePo setup, here is a simplified diagram on one of my boats.
What I have that is different than other systems that have been discussed here, is the Time Delay circuit in the OV (over voltage) disconnect system.
This is put here solely to protect the alternator system.

What happens is when initial power is applied, the alt reg power is delayed for 5sec after the command to close the OV solenoid is initiated. This ensures that the batteries are connected before the field is energized.

On a OV event, the OV solenoid is held closed for 5secs after the open command is issued. This gives the alt.'s magnetic field time to collapse. Thus saving not only the diodes but also lessening the effects of a sudden disconnect on the reg..

The time delay circuit is all SS (solid state) and is about the size of your thumbnail. It is rated at 5amps.

Just another way to do it.

Okay, for some reason the image is not coming through.. so here is a link.
http://www.midwestsailing.com/CF/LiFePo-simplified.pdf
http://www.midwestsailing.com/CF/LiFePo-simplified.pdf

Thanks for sharing your diagram and explanation. Can you give the brand/model/spec for the TD relay?

[exMaggieDrum]

Originally Posted by OceanSeaSpray

Yes, KISS contains "stupid". A better understanding of lithium (and lead acid) batteries...

Mbartoch's design shows what needs to be done to prevent the charging bus from getting directly connected to the load bus with no battery present if the BMS drops the bank.
It is its main discerning feature and every other design I have seen with a HPBMS had that vulnerability and when this does happen, it tends to cause a great deal of damage and expense.

Things simplify considerably if one stays away from a BMS that is in fact unsuitable and inadequate for the application (same when it comes to its even poorer Australian copy, some look-alike packaged ones etc), because those were never intended to control a split bus system.

Mbartosh's design transforms the totally unsafe "power to disconnect" design into a fail-safe "power to connect" (by the way of two additional auxiliary relays) so if the bank drops, all power is lost and the LVC and HVC NO contactors drop open (instead of closed before) and everything gets disconnected. As it should.

I wouldn't call this a cute, very optional way of doing it.

I can appreciate the logic of your explanation. Since they are NO and not NC it seems it would be best practice to keep a spare BMS on hand in case of a failure, and perhaps a spare contactor as well. I think mbartoch had mentioned he has spares. If the BMS fails, no juice to the house. You could always jump around a contactor on a temporary basis as well although you would definitely have to beware of any LVC and shut all down if needed. Same for the charging side too.

[exMaggieDrum]

There are many ways to make these installations more complicated and more expensive of course, and KISS is always good. I was looking for a way to alarm before the BMS sends a LVC or HVC signal which shuts down either side. I think it was wishful thinking but I saw "alarm" on a BMS spec sheet and thought that might do it but it is not the same.

Mbartoch says he programmed his battery monitor system to generate an alarm just for this, so that it would alarm before LVC got to that point and same for HVC.

Do some BMS's come with that built in? It is the same logic and sensing only an additional piece of circuitry and output terminal (at additional cost I'm sure).

[OceanSeaSpray]

Originally Posted by exMaggieDrum

I can appreciate the logic of your explanation. Since they are NO and not NC it seems it would be best practice to keep a spare BMS on hand in case of a failure, and perhaps a spare contactor as well. I think mbartoch had mentioned he has spares. If the BMS fails, no juice to the house. You could always jump around a contactor on a temporary basis as well although you would definitely have to beware of any LVC and shut all down if needed. Same for the charging side too.

Hmmm... my explanation was aimed at highlighting the difference in the design. NO or NC doesn't make much difference in terms of spares or no spares really.

The BMS can fail in several different ways, the auxiliary relays can fail, the main contactors can fail... the more stuff you have, the more likely something could play up. If it fails safely, not that big of a deal.
Now before someone jumps on the "nothing-is-best" Have-It-Stupid bandwagon again, you still need to have enough to make the system safe and resilient, by opposition with riding luck until something nasty happens - which is by far the more likely outcome.
I would rather bypass something temporarily if ever needed as you describe, rather than being exposed all the time: that choice is pretty clear-cut.

Start instead with a decent BMS/BPS that has two outputs: "Ok to load" and "Ok to charge" and all the crap hanging off the HPBMS falls off, there is no single main contactor causing headaches any more and you can have two sensibly controlled buses: KIS.

On the average HPBMS installation with a main battery switch in series with the main contactor (not a great idea), all you need to do is turn that battery switch off and you have now linked your solar panels, wind generator and what-not together and then directly to the loads on the boat, with the voltage regulation usually gone out the door, because in most instances the battery is needed by design for the regulators to operate properly.
A simple PWM controller with a 36-cell panel will send spikes at 22V into your electrical system and some mechanical generators will do far worse.
If you turn that switch off while charging with inductive sources like generators or MPPTs, you might really hit the jackpot. The initial spike could be spectacular. I know of one that recently blew electronics rated for up to 50V.

If you don't want surprises down the track, you need to ask all the "what if" questions and solve the issues, not just make it work when everything is normal and erupt in self-congratulations.
This is why LFP tends to turn into a bit of an engineering project each time.

[darylat8750]

"OceanSeaSpray
Marine Service Provider
Start instead with a decent BMS/BPS that has two outputs: "Ok to load" and "Ok to charge" and all the crap hanging off the HPBMS falls off, there is no single main contactor causing headaches any more and you can have two sensibly controlled buses: KIS."

Can you recommend a BMS/BPS that has ok to load and ok to charge signals?

[mcarling]

Originally Posted by OceanSeaSpray

Now before someone jumps on the "nothing-is-best" Have-It-Stupid bandwagon again, you still need to have enough to make the system safe and resilient, by opposition with riding luck until something nasty happens - which is by far the more likely outcome.

...

This is why LFP tends to turn into a bit of an engineering project each time.

All of this logic applies at least as much to lead-acid batteries as to LiFePO4, but I don't see anyone calling for voltage monitoring of individual lead-acid cells, despite the fact that lead-acid batteries are more failure prone and tend to fail more dangerously than LiFePO4 batteries.

[OceanSeaSpray]

Originally Posted by darylat8750

"OceanSeaSpray
Marine Service Provider
Start instead with a decent BMS/BPS that has two outputs: "Ok to load" and "Ok to charge" and all the crap hanging off the HPBMS falls off, there is no single main contactor causing headaches any more and you can have two sensibly controlled buses: KIS."

Can you recommend a BMS/BPS that has ok to load and ok to charge signals?

I build my own when I need them and this is what I have been using. It is technically a BPS and I designed it specifically for this application.
I am designing a new model (BMS) at the moment based on everything I learned along the way, but it hasn't reached the stage of the first prototype yet. An alternator control project got in the way, but this is now near finished.

Not long ago I needed to identify a pathway for a 24V system and after reviewing a great many of them, I found that you could use an Orion Junior for that with benefit...
It would work for 4-cell/12V too and it has the capability of monitoring many pack temperatures etc. Unlike the other one, I see it as a decently engineered product.

Check it out and I hope this can help...

Eric

[OceanSeaSpray]

Originally Posted by mcarling

All of this logic applies at least as much to lead-acid batteries as to LiFePO4, but I don't see anyone calling for voltage monitoring of individual lead-acid cells, despite the fact that lead-acid batteries are more failure prone and tend to fail more dangerously than LiFePO4 batteries.

You need to read about a concept called "shuttle reaction" and learn how each type of battery deals with current once fully charged, why LAs are able to equalise and LFPs can't etc. They simply have nothing in common...

What you are saying above is completely wrong.

[mbartosch]

Originally Posted by OceanSeaSpray

Start instead with a decent BMS/BPS that has two outputs: "Ok to load" and "Ok to charge" and all the crap hanging off the HPBMS falls off, there is no single main contactor causing headaches any more and you can have two sensibly controlled buses: KIS.

Agree with the notion that the House Power BMS is not an optimal choice for marine applications, but you can work around a few of its limitations.

For example you can achieve this "OK to load" and "OK to charge" completely without a main contactor.
It's actually in my design but a bit hidden - you could leave out the main contactor in my schematics and only retain the LVC and HVC NO relays. I included it because I actually wanted a way to completely isolate the bank.

Now this is how you get OK to load and OK to charge with a HPBMS:

1. you need small auxiliary relays as described here (you can copy the schematics, you don't need the box of course)

2. you then wire the "common" contact of the two auxiliary relays to the E output of the HP BMS. This E output pulls to ground whenever the BMS is active and the potential Main Contactor should close (i. e. normal operation outside protection voltage)

3. then you use the NC contacts of the auxiliary relays to wire your actual LVC and HVC contactors. Connect the other side of these contactors to +12 V (fused)

What you achieve: the LVC and HVC contactors close when the BMS is on and no error has occurred. If LVC happens, LVC contactor drops. If HVC happens, HVC contactor drops. If the BMS detects an error (or goes offline) both contactors drop.

Now if the E output is no longer connected to ground this also causes the HVC and LVC contactors to open immediately.

This way the LVC and HVC contactors actually act like the "OK to charge" and "OK to discharge" outputs without the need of a main contactor - and protecting loads from spikes on disconnect.

(BTW, I actually sat down and created freaking logic tables when first designing that, and converted them to relay logic - only to find out it's that simple )

HTH

Martin

[missourisailor]

Originally Posted by exMaggieDrum

Thanks for sharing your diagram and explanation. Can you give the brand/model/spec for the TD relay?

There are actually two different timers. One delay on and the other delay off. Both combined will fit into a 1/4"x1/2"x1" enclosure, with 5 wires hanging out.

The SMC boards (Surface Mounted Component) come from Curious Tech-TMR1

CuriousTech Digital Solid State Timers

I had a batch special made for me. Cost about $10 each.

[exMaggieDrum]

Originally Posted by OceanSeaSpray

Hmmm... my explanation was aimed at highlighting the difference in the design. NO or NC doesn't make much difference in terms of spares or no spares really.

The reason I mentioned NO vs NC (for the contactors) is that the fail would be to NO so you would not have house power if there was a problem. With NC you would have a different problem. Sometimes your biggest worry is to power the radio or lights on a temporary basis. So I am just playing what if as you suggest. I am a bit of a spares fanatic, especially for something as critical as my house battery.

The BMS can fail in several different ways, the auxiliary relays can fail, the main contactors can fail... the more stuff you have, the more likely something could play up. If it fails safely, not that big of a deal.
Now before someone jumps on the "nothing-is-best" Have-It-Stupid bandwagon again, you still need to have enough to make the system safe and resilient, by opposition with riding luck until something nasty happens - which is by far the more likely outcome.
I would rather bypass something temporarily if ever needed as you describe, rather than being exposed all the time: that choice is pretty clear-cut.

Yes

Start instead with a decent BMS/BPS that has two outputs: "Ok to load" and "Ok to charge" and all the crap hanging off the HPBMS falls off, there is no single main contactor causing headaches any more and you can have two sensibly controlled buses: KIS.

Not sure I understand what a "decent" BMS is. Is the HPBMS not a decent BMS? It does have "OK to load" and "OK to charge" if I understand it right. But I may be misunderstanding something here. If installed with separate charge and load contactors, as mbartoch has done, I would have that right?

Thanks for sharing your BMS pic. Do you have a schematic you could share?

On the average HPBMS installation with a main battery switch in series with the main contactor (not a great idea), all you need to do is turn that battery switch off and you have now linked your solar panels, wind generator and what-not together and then directly to the loads on the boat, with the voltage regulation usually gone out the door, because in most instances the battery is needed by design for the regulators to operate properly.
A simple PWM controller with a 36-cell panel will send spikes at 22V into your electrical system and some mechanical generators will do far worse.
If you turn that switch off while charging with inductive sources like generators or MPPTs, you might really hit the jackpot. The initial spike could be spectacular. I know of one that recently blew electronics rated for up to 50V.

I personally like the idea of a manual main battery shutoff in order to isolate the batteries to do work on them (replace, repair, cell boards, etc.) without worrying so much about an accidental short circuit to ground. It would normally be left on in an installation while the BMS and relays and contactors manage the LV and HV. I don't think I would shut it off while charging, although the solar panels would be "on" during daylight. I just rechecked my Outback MPPT solar charge controller and it gives no warning about disconnecting the battery while charging, and I have done that with on my previous boat by pulling the fuse and never noticed a problem. It may be I missed the bullet?

I haven't given it detailed thought yet, and still reading what mbartoch says about his rationale for where he put his sensors, but I would want (in general) for them to be on the battery and not after a switch or contactor.

If you don't want surprises down the track, you need to ask all the "what if" questions and solve the issues, not just make it work when everything is normal and erupt in self-congratulations.
This is why LFP tends to turn into a bit of an engineering project each time.

It certainly has turned in to a engineering project! Pretty fun right now but I haven't started spending the money yet. I am very conservative when it comes to making my systems idiot proof as I am not convinced I am not an idiot, at least not all the time.

[Windship60]

We are changing out our AGM batteries for lithium ion. We are currently looking at the 12V 300 amp hour ReLion LB300 batteries. We are going to go with two banks of 12 volt in series for 600 amp hours. They are on the cheaper end of lithium batteries and we know that cheaper doesn't always mean worth buying. We are in southern Florida. Does anyone have any comments/reviews on this battery?