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

[diugo]

Well, after another full day of solar charging, some good news and some bad news. The good news is that the cell voltages are holding together very well. Three are all at 3.311V and the fourth is at 3.313.

The bad news is the way the cell voltages interact with the Morningstar TriStar charge controller. There appears to be significant voltage "lift" when charging, even at currents well below 0.1C. To illustrate, the battery was at 13.0V after sitting disconnected all night. But the moment the charge controller kicked on, the battery voltage immediately jumped to the setpoint voltage of 13.6.

This was extremely surprising. I had fully expected it to stay at or near 13.0 and slowly approach 13.6 as the state of charge increased, pretty much independent of charge current. With this new behavior, the controller is "deceived" and cannot make an intelligent decision as to when to properly terminate charge (or go to a lower float voltage). Note that it is NOT okay to just leave the battery at 13.6V indefinitely.

So stay tuned while I play with the TriStar's many programmable float settings over the next few weeks. But if I can't get it to play nice with the LiFePO4's, I may have to use a relay to open the panel circuit based on other criteria.

[ebaugh]

Diugo,

With a multimeter, check the voltage at the battery, and then at the charge controller. I suspect they may be different, Even with 4/0 cables from my chargers to the Li Pack, 14V at the battery equals 14.4V at the inverter charger. At 100A charge....the lower the charge rate, the closer the voltages match due to the resistance in the cables, connections and switches.

[ebaugh]

Hello and Dave,

Thanks for the alternator information...

Marine alternators are not routinely state of the art....so is it safe to pull the field and the battery at the same time? There is a version of the Tyco EV200 solenoid with aux connections. If the field ran through the aux contacts and the battery went away with the field via the aux, would that save the alternator? The thought here is the simple Celllog 8 can control the coil in a single EV200 relay to disconnect both the pack and alternator field, more or less at the same time. Is this a universal solution for any alternator that its not a one wire connection?

I don't see heat as a problem, the output drops to zero.

[diugo]

Ebaugh, my voltage observations were all done with a voltmeter at the battery only. Controller voltage wasn't relevant.

My point was, do you not see an instantaneous voltage rise across a quiescent LiFePO4 cell the moment a charge current is applied?

[ebaugh]

Diugo,

Not that much on my LiFePO4 bank. Sounds more like lead acid. I see about .3V rise going from 30A discharge to 150A charge when I start my generator and switch from invert to charge. For example, if the bank is 13V in discharge, it goes to 13.3 with 150A charge. All measured at the battery terminals. The inverter panel is a very different story,

[diugo]

Thanks ebaugh! Your 0.3V pack rise at 180A suggests an internal resistance of 0.3/180/4 = ~0.4 millohm per cell on average, which jives well with industry specs and is very encouraging. It may be that my older "clearance" cells have a lot more resistance---or it could well be due to loose jumpers, given that I don't have any bolts yet!

So I may have jumped the gun. On Tuesday I'll repeat the test with bolted straps and report back.

[ebaugh]

I just tested 20A charge to 150A charge. The voltage changed from 13.12 to 13.25. Can get more specific numbers if you need them starting at a discharge voltage, the original we're from memory, but should be close.

I have LiFeMnPO4 cells by GBS, but I don't think that should make much difference. I do have 12P 100AH cells, for 1200AH bank with 48 total cells, that might?

[Maine Sail]

Some interesting observations... I put my four 400Ah cells in parallel on Friday evening at about 5:00 PM. As of Sunday at 4:00 PM the movement of current between cells was finally out to the thousandths. It took a full 48 hours in parallel for them to essentially stop moving current and balance at the pack voltage. It got beyond the point where I could measure it with a clamp meter and I had to insert my Fluke 179 between cells to see what was going on. Initial cell voltages were;

Cell #1 = 3.275V
Cell #2 = 3.297V
Cell #3 = 3.285V
Cell #4 = 3.281V


So last night about 4:00 PM, with a resting pack voltage of 3.286V, I fired up my 30A bench top power supply and set it to 3.45V. I then dialed back the current to 25A. At 6:00 AM the pack voltage was still only 3.333V and at 6:30 AM my power supply shut down, on what I suspect is over heat.

I restarted the power supply and dialed it back to 18A of current to keep it a little cooler.

It will take a long while to get to 3.45V at 18A but I have plenty of time and would rather do this than series/400Ah for the initial push. Once at 3.45V I will disconnect and let them sit for a few hours in parallel, then un-parallel them and let them sit again for 24 hours and check each cells individual voltage. I will then bring them to 3.6V and again let them rest and then check cell voltages. Finally I will push for my top balance voltage of 3.8V per Balquon.

Top balancing a 3.2V 1600Ah pack takes a while......


I have decided on the CleanPowerAuto House BMS and cell modules, unless you guys have a good reason not to go this way.. I also looked at the Elithion and others but they are about 10X the money.

My main goal is to see if I can keep this bank to comparable pricing to Odyssey or Lifeline AGM's and still have it be a "safe" install with BMS protection.. It will be more than AGM, but not by a lot.

I think I am going to use the House BMS with the Tyco EV200 contactors or the Blue Sea version of the same relay. I probably get a better price on the Blue Sea version.

I would prefer to use the Blue Sea "magnetic latching" relays if Dimitri says the in-rush current is not to high for the House BMS board.

The Blue Sea magnetic latching relays are nice as they draw zero current when "made". They only draw current when changing from open to closed. Once closed they remain closed without need for current draw. One model even has a manual over ride.

The Tyco EV200/Blue Sea L-Series 9012 draws 13mA but I will have at least two, house loads and charging bus. I will have an LVE house disconnect, charge bus HVE disconnect. I may also have main pack disconnect which is only a final and last ditch step in cell protection with the House BMS module, but have not yet totally decided on the bank cut off contactor..

I have not yet decided how I am going to break the alt, field wire or power to the reg?

Solar may likely get its own relay as our boat is on a mooring and the rest of the system would be shut down except for the BMS.. Still playing around with diagrams cause there are sooooooooo many ways to do this........

[JayH]

Originally Posted by Maine Sail

I have not yet decided how I am going to break the alt, field wire or power to the reg?

Would it be safe to have the alternator field sense (or regulator power) disconnected by the HVE, but not the alternator output? Alternator would stop producing power when the field breaks, which accomplishes the goal of the HVE, but battery is still available to the alternator if it takes a couple seconds to stop.

[Maine Sail]

Originally Posted by JayH

Would it be safe to have the alternator field sense (or regulator power) disconnected by the HVE, but not the alternator output? Alternator would stop producing power when the field breaks, which accomplishes the goal of the HVE, but battery is still available to the alternator if it takes a couple seconds to stop.

Disconnecting the alt output, without first cutting field or power to the field, can easily cook the diodes. I don't know of any marine alts currently using avalanche diodes. Even if they were I would never cut alt output under load without cutting off the field first...

Cutting the field wire cuts the power to the rotor. This is safe and once no field exists no output power exists. With most all external regulators if you cut the power wire to the reg you cut the field too...

[JayH]

Originally Posted by Maine Sail

Disconnecting the alt output, without first cutting field or power to the field, can easily cook the diodes. I don't know of any marine alts currently using avalanche diodes. Even if they were I would never cut alt output under load without cutting off the field first...

Cutting the field wire cuts the power to the rotor. This is safe and once no field exists no output power exists. With most all external regulators if you cut the power wire to the reg you cut the field too...

Right, and I was suggesting you use the HVE to disconnect the field only and not the output. Sounds like a workable solution as long as the alternator output is on a dedicated wire.

[senormechanico]

Another "benefit" of cutting the alternator field during HV disconnect (at least on my boat) is the tachometer would go to zero, indicating the event.
So far, that's never happened, but if the single stage adjustable regulator ever went nuts, I've got it as protection for the batteries.

Steve

[NahanniV]

If your engine is running and your alternator, regulator or a cell in your LiFePo4 pack is malfunctioning surely you are there to hear an alarm and take some action (manually switch to other battery or shut down engine).

I do agree with automatically switching off other charge sources that might be left unattended (solar, wind, AC charger); Not because I don't trust their regulators, but because a single cell in the battery bank could go HVE before the bank voltage is at the regulators set point.

Would you disconnect your bilge pump on LVE ?

I've ordered a PowerLog monitor (Like CellLog plus current and temp) so that I can adjust the parameters for alarm conditions. If it works OK, I will interface it to charging and load disconnects. It was just $40, so no big deal if I don't use it in the end.

I plan to install the batteries in an area with poor access so I want to be able to monitor the cells remotely. A 5 wire (all fused at cells) cable to the PowerLog will let me do that. I can also switch out the PowerLog for something else using the same 5 wire cable.

Another BMS system I considered is from Elite Power Soloutions. It has cell monitoring (voltage and temp/cell) and state of charge by a shunt. It has a VGA output to display the information or a CAN bus output (don't know what can display that). The only cell boards I see for it are specific to GBS cells, but perhaps they have others. The price for a 4 cell system is <$200.

[ebaugh]

There is an attraction to non latching relays because of the reduced power. But consider all the failure situations you want to protect since it will take power to break the connection. I thought it better to make the default condition open.

[FlyingCloud1937]

I said it here http://www.cruisersforum.com/forums/...tml#post812652 date 07-11-2011, 11:02

Mainsail...you might be the one to answer some/most ?'s but how will your time translate into installed costs.

Just a note I support all of you here qualified to suss out the real solution....but I'm still not ready for prime-time...to a community of people looking/thinking they'll save a buck...

Especially the non-professionals that take a little idea and interpolate it into something really dangerous.

So Main.... and others the world waits.

don't try this at home

Lloyd

Originally Posted by FlyingCloud1937

I still don't believe this technology is ready for prime time. It certainly is in BETA testing by the DIY community, with some success stories and failure stories.

Just remember system cost, also include capable charging system, and safety systems. The real MTBF is only based on a computer algorithm. Remember when Gel and Agm's first hit the market? Many problems arose and I know that most if not all of the first to market batteries failed the manufactures claims.

How many of you have purchased CFL, CCFL, and LED lighting with manufacture claims of tens of thousands of hours lifetime, only to have to replace them in less then a thousand hrs? Those claims were based on a similar algorithm to guess the MTBF rate.

If you think lead acid batteries are unforgiving of mistake and neglect, we haven't seen anything yet. I know Prius owners that are having to by battery packs that cost over 1/3 the price they paid for the car, in less then 2 years. I know dozens of owners that can't properly manage a well set-up lead acid bank, I sure wouldn't recommend these bats to those owners.

I haven't heard anyone speak to the Over Current Fault Protection required of these high discharge capable battery banks. A large 600 amp Gel bank requires a OCP of 10-20 times the amp hr size of the bank. I don't think the ABYC has even developed a standards for these bats installed to a yacht. How does this effect your Yacht Ins. policy? I'm only guessing but I wouldn't be surprised if OCP will require 30-40 times the amp hr size.

It will certainly require a voltage protection interface between NAV gear, and any other voltage sensitive electronics. This includes LED lighting, and Halogen lighting.

Don't even think about using welding cable for any of the high current cables!

To get max benefit of charge current acceptance rate of these batteries, is going to require a very expensive solution.

I would not install these bat banks to a yacht unless they were in their own fire proof enclosure. With a dedicated automatic fire suppression system rated to the potential and chemistry.

I have seen two boat fires this year caused by lead acid bat banks. One was right across the dock from my boat. The bat bank was installed less than a month, by a well known battery manufacture's own in-house installers. In my lifetime I have seen dozens of melted/exploded lead acid battery banks.

I look forward to the day these batteries are vetted and ready for prime-time...but that day isn't today, I doubt it will be this time next year.

It's fun to look at the amorous projections of how efficient/cheap these bats are at a per KW value...but none of the projections I read include the total system costs.

Just talking out loud...somebody show me the way.

Lloyd