diy solar

diy solar

Testing first 24v build -need guidance

jjhdtv

New Member
Joined
Jul 22, 2021
Messages
35
Hello. I'm starting a new thread to seek assistance with troubleshooting and build advice on my new 24v 8s 2p build. I'll first explain the components and then get into some of the issues I'm having during testing. It will become evident very quickly that this is my first time building a Lifeo4 and I very much appreciate the wisdom found on this site.

Batteries: 16 total 180ah Lifeo4. Specifications below
Brand: LiitoKala
Product name: Lithium phosphate iron
Grade: Grade A
Quality: brand new
Material: Aluminum
Rated capacity: 180Ah
Minimum capacity: 160Ah-180Ah
Internal impedance: 0.1~0.5mΩ
Rated voltage: 3.2V
Size: 63*82*277mm (+/-10mm)
Weight: Close to 3.1kg±0.1kg
Recommended current constant: 180A (1C)
Discharge termination voltage: 2.0V
Recommended constant current: 180A(1C)
Charging voltage: 3.65V
Life cycle (90% DOD): 6000
Charging temperature: -5~60°C
Discharge temperature: -30~60°C

Batteries arranged in a 8S 2P configuration. Built a frame for each bank with tension rods. Banks placed stacked (with a 1/2" gap) in a battery box ("box.jpg" and "batt.jpg")
2 8s Overkill BMSs used (one BMS per 8s)
Fuses, wire /buss bar size, and disconnect as shown in photo. (batt.jpg)
All batteries were connected in parallel and top balanced to 3.65v

I did some testing previously with a new hybrid inverter- and noticed one of the cells (cell 6) in the bottom bank of 8, went low and the BMS went into protection. Assuming a loose connection, I took all apart and checked. Nothing appeared loose, and the BMS with no load, indicated equal voltage on all cells. I thought the issue was solved.

Today, I wanted to do a full test to verify If have something close to the advertised 180 Ah the China vendor sold these as.

Ran a 2Kw load for an hour through the inverter (shop vac and space heater - 110v)

I was periodically checking the BMS app - going back between bluetooth connection to what I call the Top Bank and the Bottom Bank

At some point I notice that the same cell No. 6 was drastically lower than all the others. I also noticed that the amp draw was different on each bank.
See "BottomUnderLoad.jpg", "TopUnderLoad.jpg" for the BMS values when I was pulling a 110v, 2Kw load through the inverter.

I decided to abort the test, and return here to ask questions. See "BottomUnderLoad.jpg" to see the BMS values after the load was turned off. The Cell No. 6 came back to voltage matching all the others.

Questions:

In the 8s 2p setup I have, when under load, should each bank be delivering equal current?
What diagnostic or fix is recommended to determine the problem with Cell no. 6?
After 1 hour of testing, all insulated wires and batteries were only warm, however, my diy copper buss bar (the bar making the parallel connection) was too hot to touch. Is this to be expected given the test load, or a bad design and needs to be fixed? (ref "batt.jpg)
Any other advice before I place the battery in my remote cabin (replacing failing lead acids).

Thank you, John
 

Attachments

  • box.jpg
    box.jpg
    111.8 KB · Views: 28
  • batt.jpg
    batt.jpg
    97.4 KB · Views: 32
  • BotomUnderLoad.jpg
    BotomUnderLoad.jpg
    67 KB · Views: 30
  • BottomNoLoad.jpg
    BottomNoLoad.jpg
    71.5 KB · Views: 23
  • TopUnderLoad.jpg
    TopUnderLoad.jpg
    67.8 KB · Views: 22
Was it the copper bus bar that was getting hot or those ANL fuses? Typically those are avoided.

How did you top balance the cells?
 
Verify the cell voltages under load to see if your meter confirms the bms readings.
Since the delta is extreme under load and relatively small at rest I would check the pack joinery.
Clean every connection with rubbing alcohol make make sure the fasteners are evenly torqued.
Are the packs sharing the load equitably?
Just be dead reckoning that bit of copper busbar looks big enough to handle 100 amps without getting too hot.
There are tables for busbar sizing on the internet that can be consulted.

Do you have an IR thermometer or similar?
Could you put one of the bms sensors on the busbar?
 
The cells were balanced using a DC variable power supply (amazon https://www.amazon.com/gp/product/B082HXVRNF/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1

I balanced 8 in parallel at a time to 3.65 volts. Waited until amps were close to 0.
After each 8 were balanced, I then connected all 16 in parallel and let them sit for a week. Cell voltage were all even when checked. I did get a .05 volt discrepancy with my meter vs the reading on the power supply when it was all said and done. Over a course of a week or so they all settled to about 3.5v.

It is the copper buss bar (made from 1/2" copper pipe) that is getting hot (not the anl fuses. I have a thermal gun - I will use to take a reading next time. I was already planning to use the bms temp sensor also to get a reading. I just don't know how hot is a problem. Typically if I cant hold it, I say it is too hot. I suppose I could use 3/4" pipe or buy some copper bar stock. But wanted to make sure I don't have some other issue that needs addressing prior to chasing bigger buss bars. As I hear the ANL fuses are discouraged, what is recommended?

Are the packs sharing the load equitably? No they are not. This was a question I had. I would assume they would but during my test, the top pack was taking about double the current of the bottom pack. (Shown in the two photographs previously attached). Does this point to a problem that is related to the low voltage on cell 6 during load? or something else?
 
Typically if I cant hold it, I say it is too hot.
agree
As I hear the ANL fuses are discouraged, what is recommended?
Class t fuse
Are the packs sharing the load equitably? No they are not. This was a question I had. I would assume they would but during my test, the top pack was taking about double the current of the bottom pack. (Shown in the two photographs previously attached). Does this point to a problem that is related to the low voltage on cell 6 during load? or something else?
The fact that the bar was so hot suggests to me that one or both connections were sub optimal.
Try and switch the 4 awg wires at the copper busbar.
If the current imbalance flips then it suggests the problem is the busbar.
The way you have it configured there will be slightly more resistance to the furthest connection.
It should not be huge though.
Terminal and lug mating surfaces should be meticulously cleaned with alcohol.
 
Reporting back progress.
I took everything apart and re-assembled, making sure all connections were clean and torque to the same value. In the un-assembly, nothing was obviously wrong. However, when testing after the re-assembly - a 1kW then a 2kW load through the inverter, all cells behaved as expected (no more low voltage under load). Really goes to show, that extreme care needs to be applied when assembling the packs.

As for the heating up bus bar, well that problem still exist. I put one of the BMS temp sensors on the buss bar. Under a 1kW load, all was good. Under the 2kW load, the BMS went into protection mode at 70C (after 10 minutes of 2kW load). I shut things down, and will be replacing my whimpy flatten copper pipe busbar with something more substantial.

I did notice that each pack under load was closer to even but not even. Top pack pulled 40 amps, bottom pack pulled 50 amps during testing of the 2kW load.

I would like to know if each pack needs to take equal load. If they do, then I would like to know how to change my design to achieve this. At least some suggestions. I'm assuming it is a problem with how I'm making the parallel connection. The beefier busbar should help the heat, but I'm not sure it will help with the load being equally distributed.

Again, the help is much appreciated, thanks.
 
I did notice that each pack under load was closer to even but not even. Top pack pulled 40 amps, bottom pack pulled 50 amps during testing of the 2kW load.
I postulate that the bottom pack is the the pack furthest away on the busbar, confirm?
I would like to know if each pack needs to take equal load. If they do, then I would like to know how to change my design to achieve this.
Your busbar has 2 battery connections and 1 system connection.
You want the resistance to be the same between the system connection and each battery connection.
Putting the system connection in the middle and the battery connections equi-distant on either side should make the resistance and therefore current flow more equitable.
At least some suggestions. I'm assuming it is a problem with how I'm making the parallel connection. The beefier busbar should help the heat, but I'm not sure it will help with the load being equally distributed.

Again, the help is much appreciated, thanks.
A beefier busbar would keep the heat reasonable.
 
Both packs have a 12" long 4 gage wire running to the individual 100 amp ANL fuse blocks. Both packs are identical 8S configuration. Negative at cell no. 1, positive at cell no. 8. So, I would say both packs are equal distance from the busbar. The only difference, is a slightly longer run on the combined busbar ( at most 1.5" of busbar length). Seems like a small distance to make a big difference. However, if I understand your suggestion correctly, I should move the 200 amp anl fuse block to the center of the two 100 amps (making a T). Correct?
 
Both packs have a 12" long 4 gage wire running to the individual 100 amp ANL fuse blocks. Both packs are identical 8S configuration. Negative at cell no. 1, positive at cell no. 8. So, I would say both packs are equal distance from the busbar. The only difference, is a slightly longer run on the combined busbar ( at most 1.5" of busbar length). Seems like a small distance to make a big difference. However, if I understand your suggestion correctly, I should move the 200 amp anl fuse block to the center of the two 100 amps (making a T). Correct?
One thing at a time please.
Did you confirm that the battery with the higher draw is closer to the system feed?
 
Ok, I took a few more readings and I did note that the battery closest the system feed does have higher amp draw. However, under 1kW load the differential is very small (less than .5 amps). I also found as I increase the load, the differential increases. Pulling a 2.2kW load, the differential is about 3.3 amps (with the closest battery showing higher). I also suspect things change as the crappy busbar is heating up.

Under a small charge (12 amps) the load into both batteries is very close to the same. Definitely will need to change the busbar. Anything else I should retrofit while I'm at it? thanks
 
So at 1k you notice even balance between batteries and a coolish bus bar.

But at 2k they drift and the bus bar starts to get scorching hot.

I think it’s time to pony up some coin for a real bus bar, vs DIY 1/2” copper pipe, or at least try 3/4” pipe ?
 
Ok, I took a few more readings and I did note that the battery closest the system feed does have higher amp draw. However, under 1kW load the differential is very small (less than .5 amps). I also found as I increase the load, the differential increases. Pulling a 2.2kW load, the differential is about 3.3 amps (with the closest battery showing higher). I also suspect things change as the crappy busbar is heating up.

Under a small charge (12 amps) the load into both batteries is very close to the same. Definitely will need to change the busbar. Anything else I should retrofit while I'm at it? thanks

That disconnect switch looks like an amazon special.
For 8s LFP batteries you should be running at least one Class T fuse.
You don't need 3 fuses on the battery sub assembly.
Have 1 class t per battery.
Since you said the wire is 4 awg and I will assume 105C insulation you can put a 150 amp fuse.
Confirm the 105C insulation rating though.

You don't need over current protection between the busbar and the disconnect switch.
You might be able to use something like this
to connect the batteries to the disconnect switch.

This is also a good place to get gear https://www.waytekwire.com
 
I took the suggestions, eliminated the 200 amp anl fuse between the disco, used a much heavier copper buss bar (1/4" x 3/4"). Keeping the other two 100 amp ANL fuses at this time. May upgrade to T fuses down the road.
Ran a 115 amp test load for 15 minutes - the 2 awg wires got warm as did the busbar, but never beyond a hand hold temp. BMS temp sensors on the batteries stayed below 25C. The 4 awg wires to each pack, remained close to room temp.

As for the differential in current draw between the two parallel packs, I still have about a 2 amp difference (regardless of how I connect them). It seems one pack draws slightly more current. Not sure if this is an issue to keep chasing or within what can be expected. thanks.
 
As for the differential in current draw between the two parallel packs, I still have about a 2 amp difference (regardless of how I connect them). It seems one pack draws slightly more current. Not sure if this is an issue to keep chasing or within what can be expected. thanks.
Is it always the same pack?
I still can't tell if you did the differential analysis I've been suggesting.
We want to know if the difference is the pack itself or its connection to the aggregation point.
 
It is always the same pack. I swapped the connections and nothing changed. I also did my best to move the "aggregation" point to be equal from each pack.
 
It has been a while since I last posted on this thread. I have my new battery in service but running into another challenge that I would assume others have solved.

The system consist of a 24 volt Magnum Energy inverter with the remote panel, and two individual Midnite mini charge controllers (each connected to a small independent solar panel array)

Eventually, I will upgrade to more panels, but currently my 24/7 loads and parasitic loses uses more power than I produce at my remote cabin. That works when I'm there because I can start the generator. However, when not there I count on the security cameras and computer equipment to run 24/7.

Recently, after about 2 weeks of cloudy skies, my new life04 battery drained below the bms setting and shut the batteries off. (the bms did its job). However, as the battery is connected directly to the Magnum inverter, the inverter senses that there is not a battery and will not turn back on to allow solar charging. Therefore, even when the sun comes back out, the batteries don't charge.

I need some ideas or suggestion as to how to make this remote system to power back up (un-assisted) once there is usable solar power being generated. With the lead acid setup, the inverter appeared to shut down inverting when the battery power dropped below safe levels, but allowed charging when the solar power came available. How is that accomplished with a battery controlled by a bms? Perhaps I need to adjust settings of the inverter that would keep the bms from shutting off. thanks again for commenting on my novice questions.
 
Is there a setting on the magnum to shut off all loads at a certain voltage?

I fear even then the magnum would still draw some load so it’d need to be a bit higher than the BMS low voltage disconnect. Right now I think you need to realized your system isn’t up for running a camera and computer with weeks on end of no substantial solar generation.

I think you hit the nail on the head, more PV is needed, specially in the winter time.
 
Back
Top