Lithium battery upgrade questions

[rubicon327]

All 4/0 DC wiring from Xantrex inverter/charger to lithium battery is now complete. Takes quite a bit of time when working with such large wiring. Since it is welding cable it is very flexible so was pretty easily routed. Used the existing seat support and added a plywood backplate to mount the main disconnect switch. 6AWG from inverter chassis ground lug routed through the floor and attached directly to the frame. Left to finish is AC wiring connections, Victron DC-DC charger on tow vehicle charge line and programming of Xantrex inverter/charger. Getting close to the finish line.

[Brian B-P]

Looking good.

Just curious: why is the main disconnect switch buried in the cabinet, instead of recessed into the cabinet face? When things go wrong, moving a cushion and digging into a cabinet doesn't seem like an attractive idea. And there could be storage space there instead.

I assume not mentioned in the to-do list is painting and sealing the frame ground connection to reduce the inevitable corrosion. Is the frame tapped, or is there a nut installed in the frame wall?

[rubicon327]

Thanks Brian.

Originally Posted by Brian B-P

Just curious: why is the main disconnect switch buried in the cabinet, instead of recessed into the cabinet face? When things go wrong, moving a cushion and digging into a cabinet doesn't seem like an attractive idea. And there could be storage space there instead.

Thought about that. Maybe not totally ideal but I saw many installations by AM Solar where it is in a storage area with the other components. The wooden seat support is there anyhow and it seemed like a logical use of space. The 120V AC feed to the Xantrex is shut off at the dedicated breaker in the power center.

Originally Posted by Brian B-P

I assume not mentioned in the to-do list is painting and sealing the frame ground connection to reduce the inevitable corrosion. Is the frame tapped, or is there a nut installed in the frame wall?

It's actually on the rearward area of the frame support piece that was added as part of the recall on the 3" frames. I tried to tap it and that didn't go well. It goes through with a nut. Can you suggest the best way to paint/seal. I knew it needed something but wasn't 100% sure of the best route.

[Centex]

Yes, looking good, Dave.

Originally Posted by Brian B-P

.... Just curious: why is the main disconnect switch buried in the cabinet, instead of recessed into the cabinet face? When things go wrong, moving a cushion and digging into a cabinet doesn't seem like an attractive idea. ....

Brian, my plan calls for location of the analogous 'main disconnect switch' in a less-than-immediately-accessible location similar to Dave's. My rationale, FWIW, is:

• Minimizing length of the high current capacity battery cables
• The "when things go wrong" protection is provided by the various breakers and fuses in the system, e.g. the fast-acting Class-T fuse when things go wrong in a big way, other circuit-specific devices for individual circuit issues (I may or may not be present or have the presence of mind to timely flip the main disconnect when things go wrong)
• The primary purpose of my 'main disconnect switch' is battery isolation for long-term storage and when working on downstream circuits (both relatively infrequent and 'planned' events)

Your insights are always valued, your thoughts on my 'rationale' in this case would be appreciated too.

[Brian B-P]

Originally Posted by rubicon327

Maybe not totally ideal but I saw many installations by AM Solar where it is in a storage area with the other components. The wooden seat support is there anyhow and it seemed like a logical use of space.

It will work there - no problem.
For a possible future upgrade, an outside access hatch could provide access to a box (open only on the hatch side) which could also serve as the seat support.

Originally Posted by Centex

Brian, my plan calls for location of the analogous 'main disconnect switch' in a less-than-immediately-accessible location similar to Dave's. My rationale, FWIW, is:

• Minimizing length of the high current capacity battery cables
• The "when things go wrong" protection is provided by the various breakers and fuses in the system, e.g. the fast-acting Class-T fuse when things go wrong in a big way, other circuit-specific devices for individual circuit issues (I may or may not be present or have the presence of mind to timely flip the main disconnect when things go wrong)
• The primary purpose of my 'main disconnect switch' is battery isolation for long-term storage and when working on downstream circuits (both relatively infrequent and 'planned' events)

Your insights are always valued, your thoughts on my 'rationale' in this case would be appreciated too.

I think that logic all works, but it would also work with the switch in an accessible location. The cable path length via a switch mounted in the face of the cabinet forward of the inverter/charger would be the same as (or likely shorter than) the path length via the switch location on the seat support.

[Centex]

I note that Dave is using “Flex-A-Prene®” for his large-gauge cables, an unquestionably a high-grade product, presumably sourced from AMS. In the case of the 4/0 pictured it's comprised of 1,950 strands of 30 AWG wire and the "HD" has an insulation that's extra-thick compared to the non-HD 4/0 Flex-A-Prene®. All of the Flex-A-Prene® cable from #8 AWG and larger (even non-HD) have a rated (insulation) temperature range max of 105°C (221°F). Flex-A-Prene® spec sheet attached FYI.

It seems that a max (insulation) temperature rating of 105°C (221°F) relates to, among other things, ABYC Marine Guidance for conductors installed in engine-room and other similar environments where higher 'ambient' temperatures may be encountered.

For several reasons I am considering use of this Class M welding cable for my large-gauge cables. Class M being comprised of more strands of thinner 34 AWG wire, it is even more flexible than Class K for comparable cable gauge (4/0 AWG Class M being comprised of 5225 strands of 34 AWG wire). But, the max (insulation) temperature rating is 'only' 90°C (194°F).

Y'all's thoughts with regard to use of cable with an insulation rating of up-to 90°C (194°F), as opposed to 105°C (221°F), in our RV 'utility bay' applications, would be appreciated!

[Centex]

Originally Posted by Brian B-P

... I think that logic all works, but it would also work with the switch in an accessible location. The cable path length via a switch mounted in the face of the cabinet forward of the inverter/charger would be the same as (or likely shorter than) the path length via the switch location on the seat support.

OK, thanks for that input.

BTW, in the case of my particular component arrangement plan (5.0 installation in passenger-side rear dinette bench) moving the main disconnect to the face of the bench would result in the +leg between the battery and inverter being ~3x longer (my physical arrangement specifics differ from Dave's, in my case that +leg as planned is much shorter).

Lacking your mention of any particular flaw in my logic I'll stick with that part of the 'plan'.

[Brian B-P]

Originally Posted by Centex

BTW, in the case of my particular component arrangement plan (5.0 installation in passenger-side rear dinette bench) moving the main disconnect to the face of the bench would result in the +leg between the battery and inverter being ~3x longer (my physical arrangement specifics differ from Dave's, in my case that +leg as planned is much shorter).

Ah, right, I was thinking of Dave's cabling in the 19'.

[Brian B-P]

Originally Posted by rubicon327

It's actually on the rearward area of the frame support piece that was added as part of the recall on the 3" frames. I tried to tap it and that didn't go well. It goes through with a nut. Can you suggest the best way to paint/seal. I knew it needed something but wasn't 100% sure of the best route.

So it's a plate, not a box section - that's handy for the current installation because it allows access to the back side for the nut.

It's pretty thin to tap, even with a fine thread, and the frame tube wall would be as well.

I might try installing the bolt without the cable, but with a suitably sized washer on each side to act has a mask for the area to be left bare, for painting. After reinstalling the cable, there is spray-on battery terminal coating which goes right over the completed connection. It's commonly red or yellow, and is available in various brands of automotive chemicals (e.g. CRC, Permatex/Gumout).

A possibility to consider instead of the current hardware is a blind rivet nut (or "nutsert"), which is permanently installed in the hole (making a good connection to the bare steel around the hole), then can be painted over, later removing the paint from just the threads and the face of the nut - that would leave no bare frame steel to corrode. This is functionally similar to the welded-in nuts used for electrical connections to the vehicle structure in cars. These rivet nuts are easiest to install in the aluminum nut version, but of course for an electrical connection in steel it should be a steel nut to avoid corrosion due to dissimilar metals. I would still want the protective coating with ordinary hardware, although with a plated cable lug it wouldn't be required.

[Centex]

Originally Posted by Brian B-P

... A possibility to consider instead of the current hardware is a blind rivet nut (or "nutsert") .....

Gotta love all the 'incidental detail' discussions in this thread

I've been thinking about the ground issue but not having my trailer in-hand yet for close examination I've not pursued this critical detail; a 'nutsert / rivnut' hadn't come to mind in any case. That notion may have some merit for me, so thanks yet again for the 'Brian-storming'

(would need to ensure the lug-hole is larger than the slightly-elevated face of the rivnut to ensure tight / intimate contact with the cleaned metal surface in which it's installed, the better primary electrical path I would think)

[Brian B-P]

Originally Posted by Centex

(would need to ensure the lug-hole is larger than the slightly-elevated face of the rivnut to ensure tight / intimate contact with the cleaned metal surface in which it's installed, the better primary electrical path I would think)

I think that I would go the opposite way. An insert nut has a substantial flange on the side from which it is installed, and I would want the lug hole to be as closely fitted to the bolt diameter as practical, so the lug would seat against the nut flange and not the frame surface. That way there would be no unpainted steel frame surface involved.

In a car, where a welded nut is used to attach a cable lug to the structure for a "ground" connection, the electrical contact to the structure is through the threads, not the structural panel or even the face of the nut; the nut and panel are painted after the nut is welded in.

[rubicon327]

Originally Posted by Centex

Y'all's thoughts with regard to use of cable with an insulation rating of up-to 90°C (194°F), as opposed to 105°C (221°F), in our RV 'utility bay' applications, would be appreciated!

Alan: I think you might just be splitting hairs. Our applications don’t even come close to these limitations. 4/0 is already really big and therefore expensive. Unless there was some very compelling reason I would just go with the more economical choice.

[Centex]

Originally Posted by rubicon327

Alan: I think you might just be splitting hairs. Our applications don’t even come close to these limitations. 4/0 is already really big and therefore expensive. Unless there was some very compelling reason I would just go with the more economical choice.

That works for me, Dave, Thanks.

[rubicon327]

Finished the A/C wiring. Ceremoniously pulled the shore power plug. Let my son turn the inverter on and the Freedom XC display lit up! Tested an outlet, microwave and mini-split A/C in battery (inverter) mode. All working well. Mini-split drew close to 80A with the compressor ramped up when set point was lowered. Once set point was met it went down to about 5A. I tried the 1000W microwave simultaneously as a bit of a battery stress test. Pulled around 200A for about 20 seconds and the system worked well. With the microwave running only I turned shore power back on and the unit transferred seamlessly back to shore power without any hiccup with microwave operation. Just a click at the unit as the transfer switch initiated. Battery was charging at a high rate but XC is not programmed yet and I realized I left our PD charger in the WFCO power center live so both were trying to charge the battery. I’m studying up on programming the Xantrex properly for the lithium iron phosphate batttery but initial tests were all successful. I have to admit it is pretty neat to be able to unplug from 120V power and still be able to run a quiet A/C off battery!!

[Centex]

WooHoo that's just outstanding, Dave!

[dstreight]

Nice job!

[rubicon327]

Originally Posted by Semievolved

As people set up their lithium systems, I wanted to share the graph below because I have found it to be the single most helpful set of data I have run across in my LFP adventure. The key takeaway is the importance of charge rate in setting absorb voltages (and also setting low voltage cutoffs). If your charge rate is low, say 0.05C, as it almost always is in solar systems (except maybe Centex Alan's with his huge collection capacity!), setting an absorb voltage at 3.45V/cell (13.8V for 12V battery) results in something around 98% full battery and virtually eliminates the potential for cell overcharge/damage, which is not likely until cell V is over 3.6V. It also will greatly increase the expected life cycles for the battery compared to charging to the oft recommended 14.6V. Also important to note that at low charge rates, by the time the battery hits the absorb V, it is essentially charged and the absorb time can be minimal. I use 10 minutes. Anyway, I hope others find it useful. If picture doesn't come through, I'll try something else. Please let me know if you want to see it and can't. I can email it.

here's a link to it online:
Bing
https://www.bing.com/images/search?v...00&vt=2&sim=11

Dave: Hoping you can help as I set up my Xantrex Freedom XC to treat the Bestgo lithium properly. I have great flexibility in the parameters which is one of the reasons I went with this unit.

Low Battery Cut-off: adjustable from 10.0 - 12.8V (will likely set at 12.0V which is 3V per cell). Based on an earlier post I believe that was your suggestion to not get anywhere near the 10V discharge floor of the internal BMS. As a point of reference I did notice the Xantrex readout did get down to 12.1V when I ran the mini-split and microwave together yesterday which was a draw of about 200A. I will rarely if ever be pulling that many amps.

Battery Type and charge voltages: Not sure what defaults the unit has for LFP but if I choose custom instead of LFP I can set an absorption voltage from 12-18V and float voltage from 12-18V in 0.1V increments. Default for these settings is 14.6 and 13.5V respectively. From the above it sounds like you are recommending 13.8V for absorption. How about a float voltage? 13.3V?

Charger current: Adjustable from 5 - 80A in 5A increments. Currently I have it set on 30A (0.075C) for no other reason than you mentioned that a lower charging rate is better for the battery. I do want to balance optimal life of the lithium pack with the desire to recharge "quickly" from a generator if the mini-split is run overnight. When I studied what Bestgo recommends even my 80A charge limit is equal to their most stringent "suggested" charge rate (C/5) at higher and lower temperature range (see attached pic). So while I'm pretty sure this can be bumped up towards the 80A limit of the charger I'm just not positive where to land.

[rubicon327]

DC-DC charger is installed and tested. Mounted nicely within the DS bench seat after adding an add’l 1”x2” wood strip on inner bench face. Advertised as an 18A charger but only provided 10A to the battery in my limited testing. This is likely due to undersized tow vehicle wiring (14 AWG?) which was predicted. It will still provide some charge to battery from tow vehicle when needed with proper charge profile for LiFePo battery. Wiring is now all complete! Just need to do some clean-up and reinstall benchtop, table, etc.

Jon V. if you see this I PM’d you but figured it out in the meantime. The trailer light ground stayed on the tow vehicle side stud with the 7-pin white and I extended a new pigtail from the junction box to the Victron negative input. Then I just took the original chassis ground and routed it to the negative on the output side of the Victron. The positive coming from tow vehicle was intercepted and put on positive input of Victron. Then a new pigtail was made to connect positive output on Victron to 40A auto reset circuit breaker which then goes through disconnect switch and then on to the battery. It all seems to work fine.

[Semievolved]

Hi Dave, apologies for the length of this diatribe, there’s not a one-size-fits-all answer. Bestgo (and many LFP batteries) were designed/built for electric vehicles (EV). Their durability is good for us in RVs. But, vehicles have very high charge/discharge rates and the charge parameters Bestgo recommends are based on high charge rates. I can understand the hesitance to use 13.6-13.8 as the absorb/boost voltage when Bestgo recommends 14.6. Here’s a graph of charging the Bestgo 430Ah at 5A (0.0125C) if this image doesn't come thru here it should be attached below:
[IMG]file:///C:/Users/dharr/AppData/Local/Temp/msohtmlclip1/01/clip_image002.jpg[/IMG]
There’s very little capacity to be gained by charging above 14.0 (at low C). I may try a test 10A (max I can do with my testers) and see what the curve looks like. The Winston test curves I posted earlier should be representative of the Bestgo. Perhaps you could run a capacity test at 30A, plot V vs time up to your cutoff of say 14.4V, and pick your cutoff in that shoulder where the voltage vs time goes exponential.?
If you set a high absorb (say 14.4 or 14.6V), but then charge much slower when you are on solar only, the battery will be essentially full at lower V (under 14). That might sound fine but it requires that the balancing system within the battery behave perfectly and not allow any cell to runaway. I have watched this happen with my home solar – as the bank nears full, one cell gets to the magic 3.45V or so and starts to go beyond the charge shoulder while other cells are below 3.42 and are still taking on charge without appreciable V rise. A full cell takes such small increments of power to cause ever accelerating V increases that the cell voltage “runs away” while other cells are still trying to catch up. It happens fast! Two things limit the phenomenon: 1) a well made battery will be constructed with cells as closely matched (top balanced) to each other as possible so that they all should chart the same charge profile and reach that exponential shoulder at the same time, and 2) a balancer internal to the battery that either shunts power away from charged cells toward undercharged cells or applies resistance to keep cells from charging too much while their neighbors catch up.
Although Bestgo is a great battery, they did not design it for slower charge rates so I cannot speak to how much variance their balancing system allows among the cells and that is not a problem at high charge rates. I also do not have the Xantrex or other nice converter system and I rely solely on solar to charge my battery. At this time, my maximum charge rate is around 10A (0.025C).
Look at this link https://www.pbase.com/mainecruising/image/155043664
[IMG]file:///C:/Users/dharr/AppData/Local/Temp/msohtmlclip1/01/clip_image004.jpg[/IMG]

They consider this Winston cell to be fully charged when it hits 13.88V at 0.4C. You can see why everybody gets concerned about pushing a charge beyond the “knee” where the risk of overcharging a cell increases very quickly.
There is no cookbook way to set parameters for our Bestgos and other LFPs. Unfortunately, the best choice is dependent on YOUR system, primarily your charge rates.
Your lower charge rates should guide your absorb V to protect your battery. If your charge rate is under 0.1C, I conclude that a prudent absorb V would be 13.8 or below. Tony at Bestgo will never agree with this because his perspective is from the EV side and he will tell you that 13.8V will result in well below stated capacity. If he ran low charge rate tests, he would understand but he still couldn’t recommend 13.8V because that would grossly undercharge for the EV application. If you go to the forums (DIYSolar, Solarelectric, etc) and follow the discussions for off-grid residential where typical charge rates are under 0.1C you’ll find a consensus for absorbing only to 3.4-3.45V/cell. RVs are a difficult hybrid because they have slow(er) charge rates on solar but can be lots higher on shore power. You strive to balance between getting your battery close to capacity while ensuring long-term protection from overcharging; not simple for a system with a wide range of charge rates.
Low V cutoff; if you are likely to hit that with high discharge rates, 12.0V should be reasonably conservative and protective. That’s probably what I would go with. At low C, 2.8V/cell (11.2V battery) is definitely on the downhill side of the discharge knee, 2.6V is edge of the black hole, and 2.5V is oblivion! Here’s the Bestgo testing out at 430Ah with 2.8V cutoff at 10A (0.023C) attached below:
[IMG]file:///C:/Users/dharr/AppData/Local/Temp/msohtmlclip1/01/clip_image006.jpg[/IMG]
This curve would shift downward at higher discharge rates so your LVCO would be more protective. I dunno exactly where cell damage occurs and I don’t want to find out. If your battery ever does get below 3.0V/cell, the recharge should be at very low C until it gets back up above 3V/cell. Maybe you could try a LVCO of 12.2V and if you have repeated cutouts, lower it in decrements of 0.1V until it’s not an issue (assuming that the battery is at temporarily depressed V when under large loads and springs back up to good levels as soon as pull terminates).
I hope this is somewhat clear. Please ask for clarification where I muddied it. I am not sure the images will come through so I may try to attach them as well.

[Semievolved]

Sorry Dave, I forgot to address float V. This one is a bit easier, you just want that set to maintain the battery at the highest possible capacity (V) without risk of a cell running away while at float. In practice, that means setting float to a little ways below the charge shoulder; 3.375V/cell or 13.5V battery should be pretty safe