DIY Lithium Packs, Proposal and Discussion

[luthj]

Here is a quick circuit diagram to help work out if limiting current with cable gauge is feasible. This requires a datasheet for the battery in question, or some experimentation.



Since all components in this circuit must have the same current, and all voltage drops and sources must sum to zero.
Valt-Vbat=I*(Rwire+Rbatt)

Any battery can be approximated (for steady state DC) as a voltage source with a resistor. In this example I am using 1mohm, but this will be a fair bit higher with drop-ins. In the circuit above we can see that if Rwire is zero, than the charge current through Rbattery is 13.8-12=I*0.001 I=1,800A. Now, there are a few caveats. First, the alternator cannot supply this much current, so Valt would sag until equilibrium is reached. Second there is always wiring resistance, so Rwire is never zero.

In the case of a 200A alternator in this circuit, that would be Valt-12=200*0.001 Valt=12.2V. Now, the starter battery may contribute a bit of current at 12.2V, which could be 20-50A. So this is a significant simplification for calculations sake.

In order to limit current to say, 100A, then we solve backwards for Rwire.

13.8-12=100*(Rwire + 0.001) Rwire = 0.017 ohm.

Now, what if you don't know Rbattery? Rbattery is a dynamic value which will change with SOC. To determine Rbattery, discharge the pack to the typical minimum SOC. Measure the resting voltage (Vbattery1). Then apply a known charging load, say 20-50A (higher is typically better). Measure Vbattery2 at the terminals.

Rbattery= (Vbattery2-Vbattery1)/I So for example Vbattery1=12V and Vbattery2=12.05V, and current I=100A Rbattery=(12.05-12)/100 = 0.001ohm. Another option is to apply your desired charge load at low SOC, and measure the Vbattery. This loaded voltage can be used with Rbattery to determine Rwire.


Now for a real world example.

Battleborn 100AH batteries have an internal resistance of 6.2mohm.

Assuming Valt of 13.8V and Vbattery of 12V. Note that I chose Vbattery from an estimate. best to measure at the lowest SOC expected. To limit current to 100A is as follows.

Valt-Vbat=I*(Rwire+Rbatt)

13.8-12=100*(Rwire +0.0062) Rwire = 11.8mohm

So to work wiring path. As a rule of thumb every ring terminal connection adds 0.25 mohm, fuses add 1mohm, and relays will add another 0.5 mohm. Maybe 1 for ones under 100A. So in a typical charging circuit you will have say ~8 ring terminals. A relay, and two fuses. So 4-5mohm not counting the cable/wire. So 11.8-4= 7.8mohm. Looking at the table below we see that 6AWG is the minimum wire size based on ampacity. It has .39mohm per foot. So 7.8/.39=20ft. So a 20ft run of 6AWG plus the fuses/relays will result in a ~100A limit. Now its important that the ENTIRE charge circuit is taken into account. This means the engine ground strap, and any other wiring to the battery. If an existing wiring system is installed, the resistance of the components can be measured all together by applying a known load, and measuring the voltage drop. Then using ohms law to work out the resistance.

AWG wire size (solid)	Area CM*	Resistance per 1000 ft (ohms) @ 20 C	Diameter (inches)	Maximum current** (amperes)
0000	211600	0.049	0.46	380
000	167810	0.0618	0.40965	328
00	133080	0.078	0.3648	283
0	105530	0.0983	0.32485	245
1	83694	0.124	0.2893	211
2	66373	0.1563	0.25763	181
3	52634	0.197	0.22942	158
4	41742	0.2485	0.20431	135
5	33102	0.3133	0.18194	118
6	26250	0.3951	0.16202	101
7	20816	0.4982	0.14428	89

 

[luthj]

Here is the busbar layout I decided on.




I am making bubars from 2/0 flat braided and tinned ground strap. At each contact point I am soldering on a 1" square 1/16" copper tab (both sides).



After a bit of experimenting, I made a flat spiral out of rosin core solder, and put it under each tab. Then clamped them tight. I heated the tabs with a torch until the solder flowed. The bond looks good as far as I can tell, and the braid is compressed a bit for better contact.







This method is a bit labor intensive, so I am not sure if I would repeat or not.

 

[luthj]

Got a notification that the cells have been delivered to Fedex. Guess I had better get this battery box finished...

 

[luthj]

Got 6 of the cells today, should get the others tomorrow. Packaging was good, and cells all measure 3.298V.






Got the busbars drilled and surfaced flat.






I am about done with the battery box, just need to weld the hold down blocks to the bottom, and paint it.

 

[luthj]

Coming along. Wiring is quite tedious as usual. Got the BMS powered on, and it appears to be in working order. I need to work out a number of wiring runs between the main electrical cabinet and the battery box. I am still working out how to attach my negative cable to the terminal. There just isn't much space. I could make a 90 degree bus bar I guess, but that's not ideal.

IMG_20200818_145757426

IMG_20200818_145802222

IMG_20200818_151602669

IMG_20200819_155032927

IMG_20200819_163811132

IMG_20200820_093839882

IMG_20200820_130332662

 

[luthj]

Installed the heating pad to the pack. When I removed the backing paper, I discoverd the heating element stops 1-1.5" from the edges. So I could safely trim the foam down for a better fit. A was also able to remove some material from the side with the wire entry, and relocate the wire retaining clip. Hopefully I didn't damage it during the install.


You will notice the thermistors siliconed onto the battery case. I have one at the top center of the pack, one on left back, and the other on the right front, both closer to the bottom.









I checked the metal plates for isolation from the battery cases/terminals, and discovered it was showing connectivity to one of the cells on the end. So I pulled the plates off, and added a bit more rubber sheet. Ended up being a burr on the threaded rod which dug into the cell. So put some vinyl tubing over the rods.


I am making a few final adjustments to the boxes mounting area on the van. Should have mounted tomorrow. I need to run the wiring into the box and splice a couple dozen wires, and the positive/negative cables. Since I already have 4/0 cables run near this location, I am just using butt splices and my hydraulic crimper to get them into the box. Its freaking tight, but everything should fit Though unplugging the BMS connector is going to be a pain I think. It shouldn't be a regular occurrence. I will have canbus run inside the van for the bluetooth monitoring, and a serial connection as well, so I can make software changes without opening the box.

 

[luthj]

Got the box dropped in. All the heavy wiring is done. Still need to do comm and control wiring. Plus reprogram my inverter, monitor, and solar controller. You can see the blue sea RBS, and my ACR. I am using the ACR as a standard relay though, as its voltage logic is no good for a lithium pack.

 

[Alloy]

How many sides is the heating on?

 

[dreadlocks]

looks like just the bottom, should be plenty

 

[luthj]

looks like just the bottom, should be plenty

Yes, just the bottom.

 

[DiploStrat]

Dumb question. What do you measure across the battery terminals when the BMS has shut down?

Could you not simply drive a relay across the + and - terminals of your battery? BMS disconnects, relay opens? N.B. I am using the Overkill Solar BMS.

In my case, I am considering the Victron 150/45, which appears to have a dumb remote. >3v at the terminal, controller on, <2v at the terminal, controller off.

Victron on the subject: https://community.victronenergy.com...-damage-if-the-battery-is-disconnected-b.html

 

[luthj]

Some of the FET BMS have a small leakage current ( a few nano amps or so) when off. If there is no load, the terminals may read a volt or two. If necessary you can add a high value resistor across the output terminals, which will pull it down when the BMS is off.

The BMS I am using does not have any disconnect hardware, its performs balancing and logic. I am using the BMS control output to drive a blue sea mag latching (auto releasing) relay. It only consumes a few mA when active, and will automatically release when the BMS drops the control signal.

 

[DiploStrat]

I am reaching out to Overkill, but, at least at first glance, to be functionally an intelligent relay with balancing. (I can hear John61ct cringing!) There are a few other terminals, temp, and, I believe, a Bluetooth card.

There is also something labeled as a "Discharge Switch," normally jumpered. Its function is said to be: It shuts off discharge current, like a battery shutoff switch. So, if I get it, it would allow me to add a manual or remote battery kill switch. But this is a signal in, not out. That is, it tells the BMS to shut off, does not tell a remote device that the BMS has shut off.

So, ignoring leakage, if the BMS opens, the battery should be an open circuit? No?

How do you wire the Blue Sea latching relay? Across the main terminals, or in line with the BMS on the negative lead? I have never used a Blue Sea latching relay, only their "S" series. Do you have to give them a pulse, or can you simply wire like a normal relay?

 

[luthj]

So, ignoring leakage, if the BMS opens, the battery should be an open circuit? No?

For all intents no current will pass to or from the battery.

The latching relay can interrupt either the positive or negative feed. For ease of wiring mine switches the positive feed. The Blue SEA Remote Battery Switch has its own internal logic board. It is wired like a normal relay. Supply 12V to the control wire and it remains closed, remove signal and it opens.

 

[DiploStrat]

Looking at the Blue Sea site, it appears that you cannot control the ML relay with a simple voltage present/absent circuit, but rather you need to give it pulses.