diy solar

diy solar

LFP 126 amp-hour house battery design and build for a class A RV.

My next post goes into the final building phase. During my testing, I was able to get a pretty steady 111 amps for about 1 hour. The limit on the current draw was due to my inverter and load and not the battery. During the test, I was looking for any signs of conductor/connection heating.

High current is one component of I^2 * R heating. Any loose connection, high resistance pathway, or other construction issue will show up as a hot spot. Inside the battery, I found no hot spots. The only elevated temperature that I found was in the 8 awg wires. I used 3 wires in both positive and negative circuits. The wires felt barely warmer than ambient. I could not find any warm spot in the interior cell connections. Outside of the battery, I used 8 awg wires for my connections not because they were correctly sized, but because that is what I had. All of those wires got very hot to the touch.

The limiting device in my test appears to be the inverter. Initially, the heater that I had plugged into the inverter started on HIGH. The current draw went to somewhere around 160 amps. Both the inverter and the BMS safety circuits were in overload conditions. I was trying to get the load down before the protection circuits operated and I did not spend time documenting the high current excursion.

As I think about the test, I think that the inverter began to complain due to lower input voltage. I think that the voltage at the inverter was well below the output voltage at the battery due to voltage drops in my undersized connection wires. When I do my next test, I will measure voltages at several points.

Each cell is rated for both a 1C and a 3C continuous discharge rate. The 1C rate is 6 amps. Since the continuous draw during my test was approximately 88% of 1C, each cell would have been contributing approximately 5.28 amps. I believe that all cells in each section were contributing because the cell imbalance readings stayed very small. If a cell were not working, the other cells could easily makeup the current but the fewer cells working would cause a greater imbalance in the inter-section voltages.

Hope this helps. I will cover again when I post my next couple of sections.

Jay
 

The First 21P4S Battery​

Making the Heavy Current Connections

I want my internal battery connections to be able to carry high currents without heating. Copper is the best readily available conductor. But I do not have the machinery to weld copper.
My solution is to use copper 12 awg wires to carry the current and to use 0.3mm 8mm nickel flags crimped around the wire and soldered to give me a weldable connection to the individual cells. I cut the 12 awg wire into 8.5" pieces, chucked each wire into a drill, rotated at slow speed and sanded the wire to remove any coating and to facilitate soldering.
CamScanner 12-03-2020 16.29_27.jpgCamScanner 12-03-2020 16.29_29.jpgI made a wooden board with screws as a form to space the nickel flags prior to soldering.
CamScanner 01-15-2021 21.23_2.jpgI used a small table vice to hold the copper buss bars while soldering.
CamScanner 01-15-2021 21.23_4.jpgFor the series cell connections, the buss bars have 12 nickel flags, two per cell. For the positive and negative terminals, six flags are used for the cells. The Extra copper bar is used to make the output 8 awg wires.

I connected four red 8 awg wires to the positive battery cell terminal and three blue 8 awg wire from the BMS and one black 8 awg wire. Three of the red 8 awg wires to the battery output terminal and from the BMS, three black wires to the negative output terminal. The additional red 8 awg wire and black 8 awg wire will be explained later.

The temperature sensor for the BMS was taped to the side of a cell where the plastic coating was removed, and the balancing leads of the BMS were connected to the appropriate sections
CamScanner 12-26-2020 11.33_5.jpgI downloaded the BMS app and made a connection. I then did a final charge of the battery through the BMS to verify that the BMS was working. There are extra connections visible in this picture that will be explained later. Time for a test.
 

Attachments

  • Section 4 Making the Heavy Current Connections.pdf
    724.8 KB · Views: 1
Cool build, I just finished a 12v 48 cell build using the 5,000mah offerings from the same place. First discharge test netted 68ah with low cell voltage cutoff hit @ 2.8v. Figure that means my cells are clocking in at somewhere north of 5700mah.

I'm using an iCharger X6 so good to know that I may achieve better balance by reducing my charge current overall. As you've noted, it automatically throttles down to keep your highest cell from skyrocketing, but I figured the "Slow Balance" mode would be the best I could do. Maiden charge saw one cell group hit 3.6v way before the rest and it took a long time to balance. A few minutes after it dinged indicating the whole process was complete I was already off by 35mV and it remained that imbalanced sitting there overnight. They were pretty uniform in discharge though, 10mV or so.

Second charge resulted in a resting difference of 15mv. I'm doing another complete discharge now (showing about the same 10mV diff at 0.2c) but then I'll recharge at a lower rate. Or blast em as you did until they're mostly full and restart charge program with significantly less current. Hope it works for me as it did for you. Would have really been nice to come up with a way to parallel a ton of loose cells and charge as a parallel pack to 3.6v, but I couldn't think of a way to do that without welding/unwelding.

Any regrets on not buying prismatic cells? I wanted a fun project to learn spot welding on but these certainly weren't the best bang for the buck and take a lot longer to assemble + additional accessories. Not nearly as sturdy as a pack made of prismatic cells either - I'll likely forever worry that a spot welds going to break free if the pack shifts at all in its enclosure.
 
When I started to charge my new complete battery, my section variations were off by somewhere between 30mv and 120mv, as I remember. I did not worry about the exact amounts. I had planned to bring the battery up to a full balanced charge over several charge cycles. On the last charge cycle, the charger brought the entire battery up to full charge with the section variation of 0.0 mv to o.1 mv variation, and then the charger switched to 'KEEP', not 'DONE'. The charger held the battery steady at that point and the sections were showing 0.0 mv variations.

I DO NOT KNOW BUT GUESS that this process has allowed all of the parallel cells to finally balance out with the lower charged cells gradually reaching full charge.

I thought of a parallel charging fixture but have not built it because the cells that I have received I have found to be very good. The fixture I thought that I might try would be built as follows. I would get two wooden planks each approximately 6" by 24" by 1/2". Using a forstner drill bit slightly larger than the cell diameter, I would drill a pattern of holes approximately 1/4" deep for the negative and positive ends of the cells making sure that the holes would line up between the two boards. The forstner bit drills a fairly flat bottomed hole with a hole point in the center of the hole. In the center of each hole I would drill a through hole and fill it with a #8 round head machine screw. On the back of each board I would connect all the screws in a parallel connection.

If my idea works, and remember that this is an untested design, you should be able to set the cells on the bottom board, put the top board over the cells, strap the boards together and then connect to a charger to the parallel connections and charge the cells. If I found that the cells would not stand up on their negative ends in a shallow hole, I would double the thickness of the lower bard and drill a deeper hole to support the cells better.

I am still considering prismatic cells BUT part of my design consideration revolved around maximizing the volume of the space that I have. The energy density and simplicity of build of the prismatic battery is VERY attractive. What I looked at was the size and shape of the space that I intended to use. If I put a complete battery into the space and find that I have an unusable are left over, then I have lost battery space due to shape and size. Building my batteries has allowed me to more closely match each battery to the space available.

I am worried about the spot welds also. My approach to preventing them from breaking is to construct the whole battery as a solid structure where all parts are restrained from independent motion. The first step is to use cell holders that are solvent welded together with a tape backup. I have tried to size the current carrying conductors such that they will not heat up significantly when in operation.

Finally, when I put the battery pack into the case, I have secured it so that it should not move within the case.

Time will tell.

Edit,

As I continued to think about this, I realized that I missed a simpler solution. I am using individual cell holders. Lay out the cell holders on a board and solvent weld them together. Then hot melt glue the perimeter. Drill the center of each holder through the wood and add a flat head #6, #8, or #10 screw. If the head is not high enough to make contact with the cell, use a flat washer to raise the screw height.

Fill lower holders with cells, put upper board over top and clamp. Depending on whether there is good contact to each cell, a modification to upper board to reverse the screws, add nuts, and adjust each screw to the individual cell. Just a thought.

Jay
 
Last edited:
Back
Top