Author Topic: Xantrex SW2512MC inverter / charger with lithium batteries  (Read 1679 times)

Gene Obie

  • BAC Member
  • *
  • Posts: 140
  • Thanked: 33 times
Xantrex SW2512MC inverter / charger with lithium batteries
« on: February 06, 2022, 07:01:10 AM »
Here's an update on my efforts to get my Xantrex inverter to properly charge my new LiFePO4 coach set.

This is turned out hella long and  may only useful to someone trying to charge lithium's on older vintage Xantrex/Trace inverter chargers. If you have specific knowledge about Xantrex inverter bulk charge algorithms (other than what the manual "claims") I am very interested in learning more. I haven't been able to get the normal bulk charge algorithm to work for me like I want.

My inverter is Xantrex SW2512MC (Rev 4.01 option 1)
I'm charging 3-12V SOK 206AH LiFePO4 batteries

I've been putting my 3-206AH SOK LiFePO4 thru charge cycles trying to get my inverter settings where I want and having some issues getting batteries to 100% with the bulk charge algorithm the charger uses. The charger always drops into Float mode before the batteries are fully charged. Batteries never actually get to the bulk charge voltage or to the proper SOC before this happens (I have very nice data logger on a shunt right on the battery monitoring it all).

SOK recommends the following inverter settings:
Bulk 14.4-14.6V
Float 13.6-13.8V (I subsequently backed this off to 13.4 as didn't really need to keep bank at this high of SOC when sitting on shore for long time)
Absorption time: 10min

I've disconnect the temperature sensor (unplugged the BTS sensor at the inverter). I've set the Max Charge Amps AC to charge at rate no greater than 120A (SOK spec is 50A max/40 recommended for each battery in my bank of 3 parallel- so 150A max).

I've been discharging the batteries down to about 50% SOC and kicking off charge cycles. The charger is not acting exactly like explained in the manual. It basically kicks out of the charge routine and goes to float when my battery is about 80% (+/-)… it's independent of my charge current (50A/80A/120A), bulk charge voltage (tried 14.2-15.0), or absorption time. The battery voltage never reaches the bulk charge voltage. Once in float the batteries still accept a small current (10A or so) and overnight they do get to 97% or so. There are no errors on the remote when the charge cycle enters Float. I've checked the inverter fan and temperatures during charging and all seem fine.

I spent some time on the phone with Xantrex support and this charger is quite old so he didn't have specific detailed information on it's internals…only that the inverter settings were sometimes fixed/tuned for a particular coach/battery config.  And the algorithm is a little more involved than the simple explanation in the manual - many of their inverters (he wasn't sure about mine)  actually try to estimate battery resistance during the charge cycle to determine if batteries need more charge. Evidently resistance in the wire connections to the batteries can significantly affect this algorithm if it's more than the design expects as it throws off the inverter's "math". He suspected the much lower internal resistance of my Lithium bank was causing the issue - but couldn't explain exactly what was going on. He did specifically say there is no ability to set a "tail" or termination current with my particular model as the inverter does not sense a termination current. I've observed when my batteries are full the charge current drops of rapidly - within a few minutes at the max charge voltage spec for the battery so this would have been a nice feature.

For the normal charge cycle the inverter manual says that the inverter enters the absorption phase when bulk charge voltage is "near" the bulk charge settings and stays in bulk charge as long as the voltage is "near" (and I think this fuzzy "near" and some other undocumented algorithm is the issue). And I'm not able to find a way around it. However, for Equalization charge the manual specifically states the voltage must exceed the equalization voltage to enter equalization phase (so equalization charge algorithm appears very  different). So I got to thinking, why not program the equalization charge with normal bulk charge settings  and test that way. So I kicked off equalization charge with Equalization voltage of 14.4V with 10min equalization time and the charge cycle was almost perfect. Battery current was steady 80A (where I had set it with AC input limit) until bank was about 97% then the voltage increased pretty rapidly from 13.7V  to 14.4V, and currents dropped to 35A at 14.3V, then to 10A at 14.4V within a couple minutes or so. The inverter then terminated the 14.4V "equalization" as expected and voltage dropped to float of 13.4 and the bank showed 100% SOC on my monitor app. At 13.4V float the cells discharged at a low rate for a bit until SOC drops to 98-99% and it just stays there (bobbing around +/-1A or so). In reality I probably could set the equalization time to 0 - when the batteries hit that 14.4V they were pretty much done wanting any charge. Next test I will set equalization time to 0. So behavior of the bulk and equalization charge with the same voltage settings was very different.

So my thinking is to operate my setup as follows. I'm either going to be boondocking and cycling batteries pretty deep between charge cycles or on shore with no battery cycling.
   1) Keep identical settings for "normal" bulk charge and for equalization charge.
   2) When on shore power , I really don't want a full charge cycle anyhow as my batteries will likely be fully charged. And I know if a bulk charge is kicked off (each time shore AC powers up)  the charger, for some still unknown reason, will start a charge but just terminate early and inverter will go to the float voltage. If batteries happen to be at low SOC, the batteries will still charge at the float voltage up to the SOC corresponding to that voltage (high 90%'s based on my experiments). So only possible downside is I don't have batteries as close to 100% as I may like. There's also an issue of cell balancing that I've read happens once voltage exceeds about 14.2V, so LiFePO4 BMS won't be doing that either.
   3) When boondocking I expect batteries will be down to 10-70% at the end of each 24hr period (depending on my usage). So after starting the generator I will just manually kick off an "equalization" charge as I know that will charge and terminate properly at 100% SOC. Note I'm not equalizing my lithium's per se- just using that equalization charge algorithm to complete a normal charge.

BTW, I'm using a Thornwave Labs shunt/BT monitor with companion phone app where you can monitor real time and also track statistics. It does offline logging so when I come back to the coach and open the app it just updates the logs/graphs and statistics. And I can monitor the entire charge cycle start to finish and analyze it afterwords. Without it I would have been almost totally blind here. So highly recommend this if you're a DIYer trying to do a conversion like this. It's reasonably priced as well for what it does. It can also measure chassis battery at the same time and control a relay if I were to want it to control or over-ride the coach/chassis isolation relay for example. I  logged both voltages to test BigBoy modifications  and have another (probably long) post I could make on the isolation relay topic and on discussions I had  with Intellitec on modifications I made to my BIRD controller - but very happy with how that worked out. If all this goes well I will not need to change inverter or BIRD/BigBoy for my lithium conversion. But got to say these batteries rock. My older lead acid were shot and the lights would dim after an hour on battery.. The voltage on these things barely budge after 18hrs and they are cranking out periods of 200-300A when instahot and microwave happen to be on together- and not a flicker. Prices seem to be slowly dropping and when you consider cycle life, weight/size, and power capability they are starting to look more attractive all the time.'

Here's a lnk to the battery monitor I'm using. There's also a youtube video somewhere of an RVer installing it (where I got the lead).

https://www.thornwave.com/collections/smart-products/products/bluetooth-battery-monitor-dc-power-meter
 
Holy cow - sorry for the long post. I must be bored. Sometime soon I will post some pics of my battery tray restoration and reconfiguration to accommodate SOK.
-Gene

2005 Beaver Marquis, Ruby 40, C-13 Cat 525HP, Allison 4k
Towing 2018 F-150
Washington

Fred Brooks

  • BAC Member
  • *
  • Posts: 1617
  • Thanked: 749 times
  • RVIA Certified Luxury Technician 49 years, Retired
Re: Xantrex SW2512MC inverter / charger with lithium batteries
« Reply #1 on: February 06, 2022, 01:36:27 PM »
   Gene,
  Sounds like a dangerous experiment with the price of those SOK batteries! I can appreciate all the effort you are struggling with, but the reality of the situation is that you have mis-matched technologies. Trace inverters were pretty sophisticated in their day but can't match today's lithium technology. You may also be fighting the internal battery management that is inside the batteries. I would just bite the bullet and buy a new Magnum MS 2800 pure sine wave inverter with the lithium remote control algorithms and quit struggling with 3% SOC (state of charge) to full. PS, how do you sleep at night? Fred
Fred & Cindy Brooks
2000 Marquis, Jasper
C-12 Wild Cat (U of A)
2014 Honda CRV
Proverbs 3: 5 & 6
The following users thanked this post: Joel Ashley, Bob Stone