Simple LiFePO4 system, soliciting comments

[marujo.sortudo]

So, I'm about to recycle my AGM's and go LiFePO4. I've decided to go with an electronically "simple" system with no BMS and a mere 100Ah 4S1P bank. Initially, I'll be trying bottom balancing and monitoring it to see if I'm comfortable with how the cell voltages spread. I'll be monitoring cell balance manually on a regular basis and choose very conservative charging voltages (bulk 13.6V to start.) A manual disconnect switch for the regulator "switch" wire will allow an additional means to protect the batteries from overcharging if necessary. The cell monitor will be me with multimeter in hand. A Sterling ProLatchR will be used to provide LVC/HVC protection with very conservative settings so that a cell failure might be caught just by looking at aggregate voltage. The goal here is to have some protection but very low energy usage.

I'm considering using our existing 1-2-Both switch in a very unconventional way which requires some explanation. "1" will run both the charge and load buses through the ProLatchR to protect the battery bank. "2" connects only the charge bus but bypasses the ProLatchR, presumably a feature to be used to restore the bank after a LVC event. "Both" connects the charge and load buses while bypassing the ProLatchR, presumably only to be used if the ProLatchR fails or an emergency necessitates risking the demise of the batteries.

A diagram is attached to clarify the above. I'm interested to hear comments from other LiFePO4 adopters and those with more electrical experience than me. I'm not interested in using a BMS or regular relays. I realize these might be controversial decisions, but I'd like to stick to them. As disclaimer, please don't use this wiring on your boat unless you fully understand all the issues and limitations involved.

[FlyingCloud1937]

Going against conventional wisdom is fine, but human error happens. So lacking cell level warning is going to put you at a higher risk, especially since your design will require human level attention on a regular basis.

You better at least find an automated way to cut the field to the alternator prior to the ProLatch R disconnecting the B+ on the alternator output. else you will loose the diodes in the alt, on the 1st cutout.

If you look at the website for Prolatch, their schematics all show an additional start battery connected direct to the Alt B+, and using the ProLatch as a charge relay between the house bat and the start bat.

Lloyd

Originally Posted by marujo.sortudo

So, I'm about to recycle my AGM's and go LiFePO4. I've decided to go with an electronically "simple" system with no BMS and a mere 100Ah 4S1P bank. Initially, I'll be trying bottom balancing and monitoring it to see if I'm comfortable with how the cell voltages spread. I'll be monitoring cell balance manually on a regular basis and choose very conservative charging voltages (bulk 13.6V to start.) A manual disconnect switch for the regulator "switch" wire will allow an additional means to protect the batteries from overcharging if necessary. The cell monitor will be me with multimeter in hand. A Sterling ProLatchR will be used to provide LVC/HVC protection with very conservative settings so that a cell failure might be caught just by looking at aggregate voltage. The goal here is to have some protection but very low energy usage.

I'm considering using our existing 1-2-Both switch in a very unconventional way which requires some explanation. "1" will run both the charge and load buses through the ProLatchR to protect the battery bank. "2" connects only the charge bus but bypasses the ProLatchR, presumably a feature to be used to restore the bank after a LVC event. "Both" connects the charge and load buses while bypassing the ProLatchR, presumably only to be used if the ProLatchR fails or an emergency necessitates risking the demise of the batteries.

A diagram is attached to clarify the above. I'm interested to hear comments from other LiFePO4 adopters and those with more electrical experience than me. I'm not interested in using a BMS or regular relays. I realize these might be controversial decisions, but I'd like to stick to them. As disclaimer, please don't use this wiring on your boat unless you fully understand all the issues and limitations involved.

[FlyingCloud1937]

Also you should move your alt sense wire to the battery. As wou will only be watching pack voltage, it will be confused by hi amp/charge discharge.

Lloyd

[Maine Sail]

Your v-sense and your regulator neg both need to go directly to the battery terminals or you will not sense voltage properly..

The simple system is one that works and works safely. I would highly urge, at a bare minimum, the House Power BMS and four cell modules. This is $126.00. A 70A auto headlight relay can break the regulator B+ and a piezo alarm buzzer can alert you to the rest. Personally I would rather see a LVC relay too but they get pricey for a good one... If you have solar a second $8.00 70A relay would be needed...

Don't really know why you want to bottom balance for a fractional "C" system....? Your risk of cell drift with a fractional "C" system is usually far more at risk at the top end rather than the bottom...

I would recommend 13.9V Bulk for 12 minutes then drop to 13.8V for however long you need, then to 13.2V or OFF for the Balmar...

I would also recommend a manual regulator cut switch for the regulator B+..

The Pro-Latch makes little sense as it is only looking at bank voltage not cell level....

Also if the Pro-Latch opens, while you are charging, say bye-bye to the alternator.... Alt cut really needs to be done properly.

Opening the Alt B+ is really not how it should be done. With the incorrect v-sensing you have drawn up that could very easily happen. To protect the alternator the B+ to the REGULATOR is cut NOT the B+/Output from the ALT.....

[marujo.sortudo]

First of all, thank you both for the excellent feedback.

I had hoped that the ProLatchR would be a simple LVC/HVC, but the diode destruction that you mention seems to rule out that attempt at a "simple" solution.

On this subject, I had included a manual disconnect to the switch input on the Balmar regulator, but both of you have suggested putting the manual disconnect to the B+ (power) terminal of the regulator. Is there a reason the B+ (power) terminal is prefered for manual disconnects over the switch terminal?

I didn't indicate it clearly in my drawing, but I did intend to run v-sense and regulator neg directly to the battery terminals.

Maine, I almost totally agree about top vs bottom balancing. The only advantage that I think *might* exist with bottom balancing is that a fully drained pack might survive. This seems to be outweighed by the fact that a top-balanced pack should much more easily avoid the charge curve's knees and therefor cell degradation, so while I did originally bottom balance this pack, I've been leaning towards top balancing more recently. Your voltage levels sound great for a top balanced pack.

The difficulty I have with relay based solutions has to do with power draw. Each relay adds a significant 24h current load for power misers like us (3.5+ Ah/day each.) This was the attraction to the ProLatchR (12 mAh/day.) Beyond that, I am involved in the DIY EV world and have a friend with the Mini BMS system and associated cell modules. While it works decently, cell modules and head end boards have failed for him from time-to-time, so I have some reservations about the system's reliability. My lead-acid, BMS free EV hasn't had these issues. To be fair, none of these BMS failures have resulted in a cell failure. I would have interested in a system that was able to do low power switching, but I haven't come across one yet. Any suggestions for low-power systems?

[Maine Sail]

Originally Posted by marujo.sortudo

The difficulty I have with relay based solutions has to do with power draw. Each relay adds a significant 24h current load for power misers like us (3.5+ Ah/day each.)

The HVC relays I suggested run normally closed so ZERO DRAW unless the pack is OVERCHARGED then they use power to open and break away the charge source. When over charged you havel plenty of capacity to open the relay.

They Tyco load bus relay or main pack disconnect, (Blue Sea also sells it) draws 0.13A when on but you wanted simple so I suggested audible alarms for an LVE and automatic break away for HVC... On our boat when I am not there, there is no solar charging and the BMS is turned off, thus no draw.. My bilge pump is wired to the AGM reserve bank when I am not there. These banks never need to get to 100% SOC so I don't believe in charging when I am not there, simply no need.

Originally Posted by marujo.sortudo

Beyond that, I am involved in the DIY EV world and have a friend with the Mini BMS system and associated cell modules. While it works decently, cell modules and head end boards have failed for him from time-to-time, so I have some reservations about the system's reliability.

If you properly wire and use the HPBMS you should NEVER be using the cell boards for balancing, CAN'T PHYSICALLY HAPPEN...... You will really only be using it for cell level voltage monitoring for HVC/LVC.

With the HPBMS HVC (14.2V HVC) occurs well BEFORE shunt balancing, This is by design.. Fractional "C" use is different than EV.. We have multiple charge sources thus HVC is usually the most critical.... An Ah counter can be used to keep your pack to 80% DOD max...

If you want to shunt balance with the HPBMS you will need to physically disable HVC and push the cells into the shunting range. I suggest a power supply to do this so you are able to adjust it to the LEAST amount of current to balance that you can. People damage cell boards by charging regularly at far to high a voltage and then feeding these cell boards with copious amounts of current. They can only dissipate so much heat...... If you did a proper top balance, and don't charge beyond 14.0V then you will likely not need to balance again for 400 or more cycles... My pack has not needed a rebalance for over two years and I am approaching 400 cycles, most to 80% DOD. However, I NEVER regularly charge it beyond 13.9V and stay well out of both knee ranges.....

Originally Posted by marujo.sortudo

I would have interested in a system that was able to do low power switching, but I haven't come across one yet. Any suggestions for low-power systems?

HVC cut only results in no draw but YOU are the BMS for LVC..... You could still use the Sterling for LVC but I would set it low enough so that it can't trip when the alt is running. If top balanced this low point could be potentially cause issues at the bottom if low enough.. I would probably set it for 11.6V or 2.9VPC (if they are still in balance at that point). I know mine are even with a top balance but every cell & pack will be slightly different.

A small solar panel could power the LVC / load bus relay (0.13A) when you are using the boat, and a simple ON/OFF switch on the BMS will stop any draw when you are not there....

[marujo.sortudo]

All great points as usual, Maine. I wasn't aware that NC relays could be used for HVC events, that alone makes the system more palatable to me.

I'm still unclear on the best way to disconnect the regulator either during an HVC event or for a manual disconnect switch. Various sources I've read have suggested cutting alternately: the field, the switch terminal, or the regulator B+. I presume that when folks suggest the B+ terminal they're referring to the B+ power terminal, not the sense terminal. They seem like they'd all have the desired effect to me. I guess that interrupting the field or B+ terminal might be more effective on a failed regulator? Clarification would be welcome on this point.

The EV200 0.13A drain is a bummer for us, alas. We have very limited and poor solar real estate, so, even this small current would have a real impact on us. I do like the idea of putting a ProLatchR on the load bus, though. No risk to diodes and combined with audible alarms for low cell voltage from a HousePower BMS could really be a pretty simple to run system (if more work/$ to wire.)

[roetter]

If you leave the starter battery connected to the alternator when the HVC cuts the LIFePo off from it, then there is no issue. The diodes do not get fried as the starter battery absorbs the spike. Have done this in four boats now.

[marujo.sortudo]

Good point, roetter. However, we have an independent starter circuit. A small AGM battery powers the starter and is charged by it's own alternator.

[Maine Sail]

Originally Posted by marujo.sortudo

I'm still unclear on the best way to disconnect the regulator either during an HVC event or for a manual disconnect switch. Various sources I've read have suggested cutting alternately: the field, the switch terminal, or the regulator B+. I presume that when folks suggest the B+ terminal they're referring to the B+ power terminal, not the sense terminal. They seem like they'd all have the desired effect to me. I guess that interrupting the field or B+ terminal might be more effective on a failed regulator? Clarification would be welcome on this point.

Per Balmar cut the regulator B+. NOT the brown/ignition, not the blue/field & not the red/v-sense..

Originally Posted by marujo.sortudo

The EV200 0.13A drain is a bummer for us, alas. We have very limited and poor solar real estate, so, even this small current would have a real impact on us. I do like the idea of putting a ProLatchR on the load bus, though. No risk to diodes and combined with audible alarms for low cell voltage from a HousePower BMS could really be a pretty simple to run system (if more work/$ to wire.)

A well laid out system would have both a load and charge bus... You can cut the "load" for an LVC but not interrupt the alt..

[marujo.sortudo]

Well, not as simple wiring wise, but the updated diagram should take care of protecting the cells fairly well to my eyes. I've added solar to the circuit, and used the "Field" terminals on the 1-2-Both switch as the manual disable switch for the regulator. Any added thoughts are welcome.

Right now, I'm leaving my Pro Mariner Ferro-Resonant Charger disconnected. I'm guessing it can't be made LiFePO4 friendly as it finishes at a nominal 13.8V without end, and contains the following warning about Lead-Calcium batteries in the manual: "Higher finishing voltages are seen because of the much reduced internal resistance of lead-calcium." Presumably the low internal resistance of LiFePO4 would also cause the finishing voltage to go higher than the nominal 13.8V. It's sort of a large beast anyway, and I wouldn't mind replacing it with something smaller someday if we ever spend some time dockside.

[marujo.sortudo]

As you probably know, there's still an active debate about whether cell level monitoring is needed at all. Personally, I'm not convinced that the clearly non-marinized electronics of the BMS create a better reliability situation than a simpler system with bank HVC & LVC and marinized electronics. Overall, the cost of the BMS including wiring, relays, terminals, etc. probably comes to $200 which might well make sense on a big ($$$) battery bank, but seems a harder sell on a smaller one like mine that cost $600. Beyond that, it *could* result in a less reliable system (I do realize there's lots of reasonable room for debate here, though.)

I thought some more and realized my original layout could use the Sterling for HVC & LVC but without the need for a BMS. The alternator diodes are protected by leaving the alternator wired directly to the battery bank. The sterling cuts power to the regulator and loads on a HVC or LVC event. I'm still trying to decide what the best voltages to use here would be. Being a bit more conservative than a BMS based system would seem wise. 1-2-both switch is normally used in position 1 so the batteries are protected for LVC/HVC, but either position 2 or both bypasses the Sterling for emergency charging, etc. Hooking an alarm up to the BMV-602s and programming it appropriately provides audible alarms as a further backup. Beyond what I've drawn here, having a disconnect switch on the solar would probably make sense, as unlike lead acid there is little benefit to having a fully charged bank. Your thoughts and feedback are welcome.

[marujo.sortudo]

I seem to be having a conversation with myself at this point, but perhaps my ideas will be useful to someone later on, so I'm continuing to post here. Feedback is still very welcome, though! Upon further reading of the ProLatch-R manual, I've discovered that while the LVC voltage is highly configurable, the HVC voltage is not. As such, powering the alternator regulator through the ProLatch-R will not provide HVC protection as I had hoped.

On the subject of LVC protection, I've have decided to use a very conservative voltage level of 12.3. The advantage of this is that it avoids the discharge knee in normal operation, gives some reserve capacity for emergency use via ProLatch override, and essentially acts as a cell-level LVC. That last point bears some explanation. Basically, assume a fairly terrible scenario of 3 cells 20% discharged and one failing really badly. At 0.3C discharge, a 80% SOC cell (CALB 100Ah) should be at about 3.26V. Three cells at 3.26V plus one at 2.5V would trip the ProLatchR. I'm using the ProLatchR this way to avoid using a EV200 contactor powered by the BMS and thereby saving a precious 0.13A which is important to power misers like us. Obviously, if you're using a contactor, this isn't needed.

I'm wiring up my existing 1-2-both switch unconventionally to serve as a LVC bypass for emergency use only. Position 1 will be for normal use, 2 or both for emergency discharging. Clear labeling and possibly a switch shield that must be removed to access 2/both positions should keep the emergency mode from being accidentally engaged.

Each charging source will be disconnected via relay on HVC events and each charging source will also be programmed to never generate an HVC event during normal operation. Manual disconnect switches for each charging source will serve as further insurance. The BMS will drive an alarm and I may even hook one up to the battery monitor if I get more paranoid.

[Hkalan]

Hello,

Your written words are not falling on blind eyes here ! I have been reading your thread today, and am quite interested in what your doing.

I too am using the House Power board and relays.

I am trying to incorporate wind that requires the wind turbine to be under load constantly, as well as solar charging, and seeing your diagrams. I am impressed.

It is a challenge to have a charge source that can be regulated to not activate the HVC on the BMS, and continue to charge and consume at a safe level. If we hit HVC. The relays stop all charge to the cells, then chances are high that LVC will occur when we are not watching.

Not an easy task when living off the cells 24/7/365 !!!

Alan

[durundal]

Originally Posted by marujo.sortudo

Beyond that, I am involved in the DIY EV world and have a friend with the Mini BMS system and associated cell modules. While it works decently, cell modules and head end boards have failed for him from time-to-time, so I have some reservations about the system's reliability. My lead-acid, BMS free EV hasn't had these issues. To be fair, none of these BMS failures have resulted in a cell failure. I would have interested in a system that was able to do low power switching, but I haven't come across one yet. Any suggestions for low-power systems?
...
Beyond that, it *could* result in a less reliable system (I do realize there's lots of reasonable room for debate here, though.)

This is my concern with the BMS units I've seen so far, they all appear to be un-ruggedized PCBs (maybe not even conformal coated?) that are diving board attached to one of the cell lugs. Even if they work as designed that's a fatigue failure waiting to happen, is there any sort of plug and play ruggedized unit with cell level monitoring similar to the house power setup? Any equipment capable of performing shunting functions is just another possible failure mode for small systems like this, just a bank shut off if something goes wonky is sufficient (in my case in a little coastal cruiser take it home and debug there). I also am setting up a 12V 100Ah system and am ok with the risk of the cells longevity being compromised but not with something falling off and possibly shorting across terminals causing a fire. It would be great to monitor cell voltages to catch a potential problem but not at the expense of reducing total system reliability.

I am only charging from a Genasun lithium MPPT controller which I think is reasonably likely to fail safe, but am considering one of those latching relays (thanks for the link, I had been hunting for something like that for a while!) to turn off the panel if it tries to overcharge somehow (two required failures to produce an overvoltage condition, genasun has to not limit voltage and the Prolatch has to not work). If the bank dies due to low voltage that's at least an energy-poor condition so less likely to set my boat on fire though it could kill cells, maybe worth a second Prolatch to avoid that case anyway.