Building 180Ah 12V LiFePO4 power pack

ShamusTX

Observer
I mentioned this in another thread - but I am not convinced there is an issue with BMS high voltage disconnecting. I have certainly not seen anything in running a LiFePO4 battery with a MOSFET based BMS for several years, and I have not heard of anything from the thousands of folks running Battleborn batteries with MOSFET BMS built in. Again if you suspect this might be an issue a clamping diode (TVS) across the output ports will protect against this.

I hate to say it, but I think your new setup will cause you many more problems than the Deligreen you had previously.

Charing based on SOC is a risky and unnecessary proposition. Coulomb counting SOC estimates are subject to integration errors - that is to say the small inherent errors in measuring current add up over hours, days and weeks to very large errors in SOC. The way Victron and other coulomb counting meters deal with this is to resync the SOC accounting on every full charge - ie when the voltage gets to some level that is indicative of 100% SOC (say 13.8V on a LiFePO4) then it resets the SOC to 100%. The problem is if you only charge to 90% SOC, you never have an opportunity to resync and the SOC measurement will get way out of sync from your batteries actual SOC.

It would be much more reliable just to use voltage to terminate charge - which is how everyone else does it. I am guessing you are trying to maximize battery cycle life by not charging to 100%, but this is more easily done by setting your charge voltage to 3.4 - 3.5 V/cell and is a secondary concern at best.

I don't know what your usage scenario is - but 8Ah a day to run the BMS and relays is a huge parasitic load and your battery will be dead flat if it sits for even a month. Typical MOSFET BMS have parasitic loads in the micro-amps. This may or may not be an issue for your particular usage scenario, but would be an issue for most camper batteries.

The Chargery BMS looks nice and gives you great feedback, but I would disable the SOC limits and switch to using MOSFETs (or Solid state relays) set on voltage limits. You could safely use the low SOC trigger to shut off the loads, but a low voltage limit should work fine for this too.


Good to know the High Voltage Disconnects may not be that big of a problem. I have not connected any DC loads, besides the inverter, to the pack. The fridge was the only dc load I was concerned about with HVD.

My plan was not to charge based on SoC. I guess that's what I made it sound like, but not my intention. Basically my solar controller will be set to cut out at a certain pack voltage. For example 14.0V ( charge limit or 14.0 and float at 13.5), still researching the solar controller settings. Then have the BMS set to open the charge relay at 3.55V per cell or 14.2V. The AC charger wants to charge to 3.6V per cell or 14.4V. If I want to allow the pack to charge to 100%, 3.6V per cell, then I can easily change that setting on the BMS display and allow the AC charger to fully charge the pack. An active balancer is always attached during charging or discharging.

I would also set the BMS to open the Load relay when the pack gets to low, say 3.10V per cell or 12.4V.

One of the reasons I switched to this BMS is that it did Coulomb counting and also provided a per cell monitoring. Two things I did not have with my original setup.

Yes, 8Ah a day is a large load when storing the batteries. I would not store the pack with the Relays closed. I would discharge the pack to 60% and open the relays by disconnecting the BMS and even disconnect the balancer. I normally don't store the battery in the trailer either and keep it in the garage.
 

hour

Observer
Dude... Holy shieeeeeet. I've never seen an ammo can like this. Damn shame that even a 50cal fat can isn't a hair wider.

BUT this means I can fit a whole bunch of stuff inside. I might even put my charge controller inside and cut out a port on the can that mates up to where CC backing is. In a pinch, use thermal compound adhesive and a low profile heatsink directly to the back of the charge controller and have that protruding a half inch off the can.

This surplus of space and knowing I'll be camping at 10,000 ft whenever the roads are navigable has got me to thinking about introducing heat in to damn near sealed box. Have you given this any consideration? Not sure if it's an issue in TX where I'll assume you live given your username...

I put my spare Victron 75/15 in a pelican case knockoff as a portable charger and hooked up a <$10 temperature controller to the load terminal. If load terminals are on, the temperature controller is on. If temperature controller is on and its reading exceeds 90*F (confirmed accurate), it turns on a fan that pulls air out of the box.

Thinking I could replicate that exactly but switch the temp controller mode to heat. There are some very affordable adhesive heat strips to be had in a whole range of watts, motorcycle heated grip kits, etc. In a nearly sealed box and operated when needed, and with respect to battery voltage, it sounds like it could be incredibly effective and allow for charging year round. I will absolutely be out on trips where the daytime temp never gets above freezing.

holyshiet1.jpg

holyshiet2.jpg
 

ShamusTX

Observer
It is a real nice box. I may get another one juts to store crap in. If you do need to color match paint, this stuff is a perfect match. When I drilled the holes I needed to paint the exposed metal to keep it from rusting.


Heating the box down in central TX is not really needed. But I have thought about this as well, since I do camp and hunt Elk in Colorado. That would mean heating the box during the night to keep the batteries above freezing. Then when the sun comes up the solar would begin to charge them, if they are above 40F.

My box is stored inside the trailer. https://www.expeditionportal.com/forum/threads/m1102-conversion-to-expedition-trailer.154336/
which is not heated and is exposed to outside ambient temperatures.

I was thinking about making a cover for the box using a water heater insulation blanket to add insulation. This would be used when it might get to freezing temps. Plus, like you mentioned, Adding a small heating device. The blanket would also allow the heating device to be more efficient.
 

ShamusTX

Observer
So I finished up the power box. At least for now. In the picture below I am staring to route the electrical wires. I am using 8AWG for the charging port coming for solar or AC charger. I am using 4AWG for the load port. The Load runs to the distribution panel which drives the inverter and lights. Then from that panel it runs to the other side of the trailer using 8AWG cable to another panel to drive the Fridge, DC charging station and at some point more LED lights.
IMG_6790.jpg

The bottom contactor in the pic above was moved to be directly under the top contactor to make more room in the front of the box for the power connectors.
Also I thought I had ordered a 100A BMS. But I got a 300A BMS. The only difference between the two is the size of the Shunt. The Shunt above is a 300A shunt and is larger then the 100A. No big deal, still made it fit.

IN the pic below I am finishing up the wiring and adding the balance wires and BMS monitoring wires / BMS power
IMG_6798.jpg

Cleaning everything up and making all connections
IMG_6800.jpg
 

ShamusTX

Observer
I then turn to the front of the power box. Originally I was using a Power Pole panel on the front. This was supposed to be a 45A power pole panel. But when I was charging with the AC charger at 15-20A it was getting very warm, not Hot,but very warm. So either I did not get a 45A or I had an Ohmic contact somewhere. So I decided to change and go with an Powerwerx SB50 for the charging power and an SB120 for the Load port.
IMG_6799.jpg

IMG_6804.jpg

I still need to do some touch up and repaint the box where I cut the new holes.
 

ShamusTX

Observer
I briefly attached the BMS display to make sure everything is working. I used the BMS to measure the cell voltages and compared it to a Multi-meter. The cell voltages measured by the BMS were 5-7mv below the multi-meter. One thing I'm going to do is use an external 12V DC supply to see if the BMS measures closure to the Multi-meter when the BMS is not being powered by the monitoring cables. Being off by 5mV is not that big of a deal, I just want to see if the difference is the BMS or due to the current draw through the 20AWG cables to power the BMS. I did replace the 26AWG cables that came with the BMS with 20AWG.

IMG_6801.jpg

IMG_6802.jpg

The display above shows another view from the monitor. I have and email out to the manufacture to see about the power consumption of the BMS. The BMS has to different boxes, the BMS and the Display. With the screen running I get a 3A draw with or withour the contactors closed. I have sent an email to chargery to see if they can provide power information when the screen is off and when the screen is disconnected. The first time the unit is powered the screen is attached to make sure everything is setup and program the BMS, but then the screen can be removed and the BMS will still function as long as power is applied.

I am also going to try and see if I can measure the current this weekend with the screen off and disconnected.

Overall I am happy with the build. I need to finish the connections to the trailer and reprogram the solar controller. I also need to set the parameters in the BMS to the profiles I want to run. I plan to measure each batteries resistance to see what they are compared to CALB specs. Maybe do a capacity testing soon too.
 

hour

Observer
Box looks awesome, you've given a lot more thought to where things (wires in particular) need to go than I did. Hard to commit to cable management during testing and with everything subject to change... may just go velcro cable wraps.

This bitty is full bluetooth (cell voltages, pack voltage, ah in/out, pv voltage and amps, balancing status) with all the data pushed to the cloud. Data gets logged, can view latest event to get current readings. Would have had this done by now if it weren't for nerding out. But in short, I can view and control it from camp over bluetooth, from my couch while sitting in my driveway over wifi, and anywhere in the world if I want to use data. Heating strips arrive today, already have relays to operate and logic written to microcontroller, tested by operating its onboard LED as if that were the heater. Seems legit. Just gotta commit, shorten and bundle wires, and close it up.

That's a bummer about your powerpoles. I was about to order some more 30 amp contacts last night for one-day delivery and all I could find that fast were knockoffs with reviews saying "Contacts are not like the originals / smaller". Thankfully I have a few real deals left and have thus far used all original Powerwerx stuff. I'm just down to the 45 amp contacts and need to watch a youpube video on crimping those properly as I've heard they're trickier.

I stacked two of the panel mount Powerwerx dual-15to45a things and the bottom most plug is flipped backwards and upside down for my charge inlet. Bottom two are 30 amp contacts (PV in, fridge) and I guess top two will have to be 45a. With MPPT in box (and vented) I don't think I'll experience any warming with ~40v input on a 15a controller. I may try dumb charging it at through a 45a load output @ 13.8whatever volts to see if that raises any alarms.

Check these out for vents if you ever decide you need em. 1" hole saw is perfect, use a rubber mallet to whack them in and they're very secure.

416gS67tB3L.jpg



Despite looking loose on sides, it's not. Bottom half of batteries are held tight on the sides via aluminum bar, threaded rod, and then some foam that bunched up because it was a half inch too long

battbox.jpg

Good tip on the paint earlier, had a cheap hole saw go apeshit.
 

ShamusTX

Observer
So I did not measuring this weekend to see what the current draw of the BMS8T is. I discovered that the 3.0A reading is a default reading when the Current shut is not connected. When I connected the current shunt to the BMS it would read 0A. I still measured the power of the BMS Setup by connecting it to an 80W switching power supply. I then set the power supply to 13.2V and powered up the BMS with the balancer cable plugged in and removed to see if there was any change in power. There was none, so the pictures below are with everything connected and powered.

The first pic is with everything connected.
IMG_6806.jpg

The next pic is with the screen module disconnected
IMG_6807.jpg

The screen consumes 100mA of current. with everything connected the total current is ~300mA. If the screen module is connected and the screen is off then the current is 250mA. These readings are when the BMS is set to external power supply. I contacted the company and they sent a response that when the unit is powered from the battery pack, with the screen running the current draw is 60mA and the BMS is only 20mA. I will have to measure that and see what I come up with. For now I will use 7Ah usage for a estimate over a 24 Hr period.
 

ShamusTX

Observer
I also did a internal resistance measurement for the CALB Cells. I hooked up a 250W Ship light to a 400W inverter. The inverter was plugged into my trailer. The solar controller and internal lights were also running. With the inverter on(shop light off), lights on and solar controller on, I had a base load of 1.2A. Then with the shop light added I had a load of 21.2A. I also had all the balancers disconnected or turned off. The cells have been sitting with the active balancer plugged in for a couple of weeks now.

The pics below are from the chargery display
IMG_6809.jpg

IMG_6811.jpg


Before taking the measurements under load I waited 30s after turning on the shop light to allow the currents to stabilize. It took about 15s to measure each cell and record each cell with the multi-meter. So the chargery data is about 45sec after applying the higher current load.
The battery pack was at approx 130 to 135Ah or 70 to 75% SOC.

Using a multi-meter,
Cell 1 -> 3.328 - 3.302 / 20 => 1.30 mOhm
Cell 2 -> 3.325 - 3.299 / 20 => 1.30 mOhm
Cell 3 -> 3.326 - 3.297 / 20 => 1.45 mOhm
Cell 4 -> 3.325 - 3.296 / 20 => 1.45 mOhm

Picture is from Chargery after measuring and recording all the cells with a Multi-meter
Cell 1 -> 3.315 - 3.281 / 20 => 1.70 mOhm
Cell 2 -> 3.319 - 3.288 / 20 => 1.55 mOhm
Cell 3 -> 3.318 - 3.285 / 20 => 1.65 mOhm
Cell 4 -> 3.316 - 3.281 / 20 => 1.75 mOhm


The image below is from the chargery when I first started to take the measurements with the multi-meter.
Under load the chargery and multi-meter were almost measuring the same.
IMG_6810.jpg


The last image is when I hooked the AC charger back up. Balancer is still disconnected at this point. The chargery displays the number of watts consumed from the last charge. It will count backwards to 0Wh when the AC charger is applied. I was going to put a request in to the company on the next firmware to have the option to show Ah in stead of Wh. The SOC meter will be more accurate after a couple of full charges with partial discharges. Right now the pack is 30% discharged.

IMG_6812.jpg
 

luthj

Engineer In Residence
I would be curious what the cell resistance spread is with something like 20A and 40A current references. Your numbers show roughly 10-20% spread between the cells. I think much of that variance could uncertainty due to the ~30mv total Delta-V. Your significant figures is limited to 4 due to the volt meters. So 0.00175ohm, which truncates to about 1.7 at best.

20% is acceptable for your usage. Higher C rate packs would be looking for 5% or less ideally. Typically matched cells for this purpose (CALB factory supplied for example) would be around 5-10%.
 

Forum statistics

Threads
185,541
Messages
2,875,674
Members
224,922
Latest member
Randy Towles
Top