LiFePo4 "upgrade" in progess.

[rickst29]

One LFP 150Ah battery (or two 100Ah batteries)

Instead of 2* "90Ah" Group 24 SLA on the tongue, I will have 1* "150Ah" LFP in the coach. (In my 2619, the only adequate area is the storage compartment under the couch - 2619 does not have a rear "battery compartment".)

I originally decided that I won't need more than 150Ah, and chose a "150Ah" battery from China. EDIT 9/7/2019: My Battery does not store a full 150AH, and I do not recommend that vendor (and I deleted the link). It is more like 100AH - I'll keep it, because it was cheap and refunds would be a hassle. I also paid less than the cost of any USA 100Ah battery, but it was mis-advertised.

With LiFePO4, you can use a bit more than 85% of the maximum capacity without harming the lifespan. But you should try to never run the output voltage all the way down to BMS limits (typically 10.0V), because that is less than 5% remaining, and DOES hurt lifespan.

EDIT 3/26/2020: I have decided to add a second battery.

With LiFePO4, I think that it is not "absolutely critical" for batteries to be perfectly matched: They have very low internal resistance. On discharge, they will have nearly the same voltage (12.8V) across a wide range of SOC (10% SOC thru 90% SOC). And for charging, BMS will protect from overcharging the one which gets full "first" (it will fall into overcharge voltage-disconect). When the current into the less charged battery falls below minimum for the "CV" stage, both have been charged. Because my batteries will charge differently, I am using two separate "Coulomb Counter" meters. An installation with identical batteries could get by with one. My second battery cost less than $400 (about $362 on sale), and they know I will be testing capacity to be more than 92 Ah when it arrives - and they guarantee that it will pass with flying colors. This one: https://www.aliexpress.com/item/4000595047815.html.

EDIT June 12, 2020: With Covid-19 affecting shipping prices (they wanted about $100 more, plus taxes and fees), I switched to a different battery vendor (on DhGate, with an identical conversation about my required minimum capacity). My 2nd battery arrived today, and it's now in discharge testing through a pair of high-wattage resistors (in parallel with each other). One reviewer says that his batteries only measured 51Ah. And, unfortunately, my load test agrees pretty well. I got 53.3 Ah of power from the battery, while dropping no-load voltage to 12.18v. That is about 12% capacity remaining, so the total capacity of my battery (brand new, and not driven below 10% remaining) was measured and calculated at about 59.9 Ah. Mine does not look like the picture, it seems smaller and doesn't weigh enough. From here: https://www.dhgate.com/store/product...rod|3232919042.

But - immediately following my test and complaint, this vendor agreed to refund $150 of my purchase price, to make up for nearly all "missing" capacity. The big online store (DHGate) has handled the distribution of funds and refund to me, resolving the issue. My slightly longer test report is here: https://www.trailmanorowners.com/for...3&postcount=39

A PD 4655L "MBA WildKat" 55A converter section replacement, with 'Wizard' pendant.

http://www.bestconverter.com/PD-4655...l#.XMtM_6RlDb0. Edit 9/7/2019: Surprisingly, my "55A Wildkat" actually puts out 65A when charging the battery.

The Wildkat has a DIP-switch selectable "Lithium" profile. In "Lithium" mode, "Bulk Mode" charges at maximum current while Battery Voltage increases, and then tries to perform a short "Bulk Mode Constant Voltage" stage at 14.7V to reach 100% State-Of-Charge (ending when current becomes extremely small). High Voltage allows the BMS to balance the cells. (Most 100Ah 12.8V LiFePO4 batteries are built with 4 cells "4sip", although larger batteries might use eight cells "4s2p".) I have not measured whether any "Float" stage occurs at all, and I have not measured to find the "Boost Return" Voltage setting. One of my batteries (the new one) will perform an overcharge voltage-disconect at about 14.65V, and won't become fully charged - but that's actually a good thing for LiFePO4. Unlike SLA batteries, the highest lifespan with LiFePO4 will be obtained by NOT pushing them to 100% or storing them at 100%.

Edit 9/7/2019: Additional info: Charging with a non-Lithium Converter:[/B] In PD's standard "Lead Acid" profile, "Bulk Mode" is the same, but switches into "Bulk Mode Constant Voltage" at 14.4V. charges at 14.4V (and completes a bit more slowly, even though the current will still be very large at the beginning of this "stage"). The problem will be: PD switches into "13.6V" mode at 90% SOC of a "Lead Acid" battery. This will happen too early. If you use the "Lead Acid" profile, you will need to use the "Charge wizard" Pendant to override PD programming and force it back to boost mode for a sufficient time to charge decently. So you could use a regular, non-Lithium PD Converter - if you're willing to whack the pendant button pretty frequently.

A "Coulomb Counter" battery meter.

You can't tell the State-of-Charge in a LiFePo4 battery from voltage, because the Voltage holds nearly constant (at 12.8V) through most of the discharge process. And so, I added this meter/analyzer. It measures power in and power out ("net" Ah), rather than just the Voltage. Amazon has something comparable, but it's more than twice the price - so I bought this from China. About $30 with an "on-sale" discount, and it includes the necessary 350A shunt:
https://www.aliexpress.com/item/8-80...885265096.html. Or, jere is an EBay Seller (perhaps the same "person"), who has shipped my second one by air freight - arriving barely than one week: https://www.ebay.com/itm/New-80V-50A...r/264451968056. If you will install a big inverter, as I have, then you need to pay a little bit more and buy the biggest shunt, 350A. The cord from the shunt to the meter, provided with the meter, is only 18" long, and might need to be replaced for easier viewing. An extension cord can be bouight separately, and is often included in sales.

[Craigrrr]

I was looking at these, but still too much $$$$ for me right now.. https://www.ebay.com/itm/LiFePO4-12V...fCf3trDL_n_nkA

[rickst29]

Originally Posted by Craigrrr

I was looking at these, but still too much $$$$ for me right now.. https://www.ebay.com/itm/LiFePO4-12V...h/273437658053

(3/26/2020) The 150ah model currently at $1099, the 200Ah costs $1359. They're made in China, but coming from France. Chinese prices have been falling fast, but that Ebay Vendor has not changed his prices at all. For only about $800, you could buy 2 * 100Ah batteries from China, and you could buy a single 4s2p 200Ah battery for about $780. (Tariffs and shipping fully paid by the sellers).

AliExpress has much better deals.

[rickst29]

The primary purpose is to disconnect the battery up on a Low-Voltage scenario, well before the BMS circuit performs its own disconnection.
https://shop.pkys.com/Victron-BP65-B...5A_p_6422.html

Edit 3/6/2020: Although I bought that battery protect, I also bought a big inverter - and need much higher current capacity for that new Inverter. The higher current Victron "battery-protect" devices cost much more, and I've got voltage monitors all over the place. I'll just watch what happens on the monitors, leaving the BMS limits (10.0V) in place as a "last resort" disconnect method.

[rickst29]

Since these things can accept a significantly higher charge current than my "old" Solar Controller can dish out - I have ugpraded to a new one - EpEver "Tracer 3215BN". Like my old one, it's programmable for LiFePO4. But the Display is much more useful, and it's capable of 50% more power.. But this was an expensive Upgrade, $185 from Ebay.

And also, since the new "Tracer" could try to pull can try to pull up to 450W from the Toyota when I use the dashboard "magical Zombie Switch" to turn the 4Runner into a big a "Solar Panel", through the Bargeman, I upgraded the Toyota Power Converter for about $40. (The "old one" is 480W max, which won't be quite enough headroom to "fill" the new Tracer without overheating. It was 24V*20A. The new one is 540W maximum, offering 36V with up to 15A.)

And finally, I bought myself a cheapo Inverter, this one: https://www.ebay.com/itm/EDECOA-Powe...V/232635461270. I suspect that the quality of "pure sine wave" is low, but it won't matter in my applications.

[rickst29]

At the upper left, I have the new Solar Controller and monitor/control panel (EpEver "Tracer" 3215BN with "MT-50" control panel). The connector cable and a battery temp monitor are shown in a package to the controller's right (upper middle), but the battery temp monitor is useless for LFP batteries. (Send me a message if you own a Tracer "BN" controller, and need a battery temp monitor.)

At the lower left, I have the new PD "Wildkat" Power Converter. It's plastic package contains the charge wizard and connecting cable.

In the lower center of the picture, I have the new inverter (powered up) with it's remote controller attached. The inverter provides AC and DC Voltage measurements on the blue screen, but the small Controller provides only dumb "battery status" LEDs, with a Converter status LED and a power switch.

This is wired into the battery, which is large but extremely light. On top of the battery, and not yet connected, is the new battery monitor. The big "shunt" must go on a negative connector into the battery, and I purchased a 1:3 "power distribution block" to split that connector for various "ground" input wires (the inverter, the TM frame, and the MPPT controller). This one: https://www.amazon.com/DS18-DB1034-1...dp/B00IORDM2O/. After the "Ground" distribution block has arrived, I will wire in the fancy monitor, recharge up to ~95% (13.2V), and then run the battery down to about 10% SOC (to about 12.0V) and use the coulomb counter to verify battery capacity.

Then, if it will only stop raining....

BTW, floor tiles are 18".

[rickst29]

Since I no longer have any low-efficiency light bulbs, I have attached an "air cleaner" to the Inverter. It has now been running for several hours, at about 70 watts, and battery voltage is beginning to drop. The Inverter has a display for "DC input Voltage", but it is currently showing only 12.8V (and that's not accurate). My fancy "Coulomb Counter" battery monitor shows 13.02V, and my "pretty good" Klein DVM agrees (within .02V).

I'll continue running load until the good monitors show 12.7V (that's an average of 3.175V per cell, corresponding to ~ 15% State-Of-Charge). Then, I'll "zero out" the Coulomb Counter by setting it at 15% SOC, 150Ah battery capacity - and plug in the charger.

I'll run the charger until Battery Voltage reaches 14.4V (3.6V per cell). This is not quite 100%, but it's pretty close (it's perhaps 95% SOC). If the Total power taken in from the charger was about 120ah, then the "150AH" battery is performing as advertised. (80% "useful" between 95% SOC and 15% SOC = 120Ah.) I know that hardly anyone actually tests batteries - but I've got the equipment, and the battery cost > $700. So it seems like a good idea.

[rickst29]

From 'moderately deep' discharge to full charge, the Coulomb Counter measured about 110ah - less than the 120ah I was hoping for, but not bad enough to demand a refund. After a couple of discharge-recharge cycles, the maximum charging capacity fell a bit lower (down to 100Ah) from roughly 10% SOC to 100% SOC.

While charging, however, the battery was willing to accept the entire maximum power output of the "55A" Converter. (And not just at low Battery State-Of-Charge: This 'fast' charging rate continued until more than 90% SOC.) The battery was fully recharged in barely 2 hours. And so, the notion of using #10 wire (30A) or even #8 wire (40A @ 60 degrees C) goes out the window. #6 will be required to do the job, for both "12V" positive and the grounding "-" return. Surprisingly, the Charge current was higher than advertised - about 65A.

I'll be running those wires underneath the TM floor, within 'liquid-tight' conduit. That is a highly insulated environment, which also contributes to designating #6 as the minimum requirement (even though the wire will be THWN/THN2 (capable of 90 degrees, 75A in a cerro-wire ampacity chart.). New Circuit Breaker size is 80A.

[rickst29]

Rewritten on 6/14/2020, using two larger "distribution blocks' with TWO LiFePO4 batteries.

In order to avoid a multiplicity of connecting lugs on the battery terminals, I purchased two of these 'Power Distribution Blocks' (XScorpion Model GDB20248G): https://www.ebay.com/itm/GOLD-POWER-...72.m2749.l2649

Each is made of solid brass, and provides for:
•Made of solid brass
•Two 0 gauge 'inputs' (back side)
•One 0 gauge, one 2 gauge, two 4 gauge, and two 8 gauge 'outputs' (on the semi-circle)
•Cover screws securely into place
•Smoked acrylic cover protects from surrounding environments damages (and nearly all accidental short circuits).
•Set screw terminals hold wire securely in place
•Fast and easy installation
•High power connection
•Multiple input/output design

They provide a total of FOUR connectors for AWG-2 and larger (which is hard to find in 'Distribution blocks of this type), plus 4 smaller connectors. Each is solid brass with an insulated plastic base and cap. The biggest set screws accept both Phillips #3 (which I hate, because the 45-degree angles on the screwdriver and screw cutout segments tend to push the screw driver out from the screw when you torque them) *and* square drive #3, which I greatly prefer. (The 4-AWG connectors accept square drive #2, the smallest connectors are phillips-only.)

The "Grounding" terminal block is slightly more complicated, basically sending all current into the Coulomb Counter meter "Shunt" for measuring power and into and out from the (new) pair of LFP batteries.

Both LFP batteries perform cell balancing at the same high charging voltage (up to 14.6V), and I don't care which becomes charged or discharged "slightly before the other": They are both 1P4S batteries, with nearly identical charge and discharge curves. As long as both reach zero current at 14.6V, they have been fully charged AND balanced under control of the two BMS systems.
- - - -
One battery has automotive-type tapered vertical post connectors , the other accepts large 'ring' terminals. I bought these 4-port Auto Battery terminals, and was slightly surprised to see that they are more capable than advertised: https://www.amazon.com/gp/product/B00NQ7U982/

The largest port of these battery connectors appears capable of #1 wire, maybe even #0. The second port does support #2. The last two ports (which I will not use) are identical and seem to support something larger than #8, although I did not verify. The tightening "arms" are definitely not lead, and some Amazon reviewers broken them while trying to stretch them out. I stretched and then tightened very gently, worried about the reviews - but they attached OK.

These auto-type battery connectors are each attached to 2 wires of AWG-2. On "+12V" one wire goes to the other battery (ring terminal) while the second wire goes into one of the big slots on the 8-port, +12V 'power distribution' connector. On "Ground" of the auto-style battery, one big port goes to "ground" of the other battery. But the second port goes into the "Battery - " terminal of the Coulomb Counter meter Shunt (which is rated for up to 350A, vastly more than I will ever use). The "Power - " terminal of the meter shunt goes to the Grounding/Return Current "12V Grounding Distribution Block", and receives all return current before forwarding it to the battery '-' terminals.

Per above, the "+12V" distribution block has an un-fused 2-AWG wire to the battery terminal (uninterrupted, length about 4"). 5 outputs from the "+12V" block are as follows:

1. 2-AWG to the Inverter, with 200A Circuit Breaker;
2. 6-AWG to the WFCO Panel and PD Power Converter, with 80A Circuit Breaker;
3. 8-AWG to the Solar Controller (now fused with a 40A Circuit Breaker);
4. 10-AWG to the new external 12V connector (now fused with a 30A Circuit Breaker), now on order);
5. 14-AWG to drive the Coulomb-Counter Monitor (now fused with a 20A Circuit Breaker, it uses less than 1/2 Amp).

On the "negative ground" return Distribution Block, I have matching wire sizes (2 * #2, plus one #6, plus one #8, plus one #10). No connection to the battery monitor is needed here. The #6 goes to the TM Frame connector over a fairly short distance, providing "Grounding" for the WFCO panel "12V grounding" bus bars and PD Power Converter.

The MPPT Solar Controller cannot put out more than 30A, and it's inputs (Solar Array and Bargeman Cable) are fused before reaching the solar Controller. I added a circuit breaker anyway, in case a "loose wire" resulted in a runaway short.

[FlyboyTR]

Rick,
I have enjoyed following your progress. Thanks for all the details!