My 4s6p A123 M1 car starter battery

[John in CR]

My lead POS car battery crapped out a few months ago. Since I had some A123 M1's not in use, I decided to make my own car battery. All of my M1's were balance charged about a year ago after capacity testing them, so I picked 24 that were the closest in capacity and current voltage. Matching based on capacity and self discharge seemed good enough for me.

Build procedures:
First, I used polyurethane glue to the cells in a 6p x 4s arrangement. They're the original paper wrapped kind.
All had tabs on both ends, so soldering six 4s strings was simple.
Then I stripped the insulation off of some 14ga zip cord to parallel all 6 at the cell level.
First I tinned a spot on the tabs for the parallel connections between the tab weld to the cell and the edge of the cell with the
positive end (the flat end on M1's) at that series connection. The button end have + from the can and - at the button in
close proximate, so I minimize work there.
Then I soldered a single piece of the bare 14ga to all 6.
At the positive and negative ends, I used 6 equal lengths of 10ga, 1 connected to each tab, to join together as one with a
piece of normal car battery cable that already had an automotive size ring connector. These are taped up to prevent shorts
all the way until the battery is ready to install in the car.
For the balance tap I attached a normal 4s plug and wires used for RC.
Then I painted it with a couple of coats of electric motor high temp spray paint, and gave it a good double wrap of duct tape
for good measure.
The 4x6 fits perfectly in a 3L soda bottle with smooth straight sides, so I cut the top off one and slid the pack in.
Using a stovetop burner it shrunk down nicely, saving the open end for last. After feeding the terminal leads through holes,
I trimmed and folded the end so it was mostly closed.
Then I used silicone to seal it up, and let that cure for 48hrs, so no acetic acid vapors would be sealed inside in the next step.
Then I cut the bottom out of another bottle, slid the terminal leads through the top and slid the whole battery in and shrunk it
and sealed the bottom end in the same manner as the first bottle. I also used silicone to form a plug around the wires exiting
the soda bottle top.
After the silicone was cured a nice wrap of electrical tape near and on the soda bottle top holds everything in place and limits
the flex of the terminal leads so they can't get worn through over time on the lip of the soda bottle.

Sorry I forgot to take pics while building it, and I waited until having it in service for a few months without issue to post about it. It has only a human BMS. I checked it for balance after each use at first, then weekly, and now that dry season is starting it will be when I think about it....monthly at the most frequent. For a long trip I'll just connect to the balance tap with a long run and have an HK cell level alarm in the car. It's been perfect and functions far better than lead, with the minimal sag really showing in a faster cranking starter. It ain't pretty, but it's protected enough that with a bit of attention to seal the terminal connections and balance tap, I'd have no problem using it in a boat or on an ATV.

The straps and battery terminal covers were required to pass vehicle inspection here, which it did with flying colors.

 

[liveforphysics]

Nice!! I bet that starter spins up a bit faster than before.

 

[John in CR]

Nice!! I bet that starter spins up a bit faster than before.

Yeah, the lack of sag helps a lot. I'll get my buddy to test it on his car battery tester when I bring a load of stuff to his auto shop. Just one trip with the car, and then ebike commutes to use all his good toys, I mean tools, that I don't have.

 

[wb9k]

Nice work John. I assume these are 26650 cells. Were these cells recovered from BAE bus packs or Hymotion packs? Salvaging the tabs and soldering to them rather than to the cell itself works quite well. With newer cans (with PET sleeves instead of paper) it's important to leave some physical space between cells that are members of different cell groups because the can is at cathode potential. If you chafe through the sleeve on two cans from different cell groups and they touch, you now have at least one cell group shorted dead out. It's not terribly protect against this, but I always point this out for people thinking about using A123 cylindrical cells.

The other thing I worry about when making a car battery like this with no BMS is, what is the maximum voltage that can come from the alternator? Many, if not most, automotive alternators put out between 15 and 16V--too much for LFP in 4S. At 16 Volts for example, if your battery were perfectly balanced, every cell group would be reaching 4.0 when reaching maximum charge in the car. This is high enough to risk cell rupture, and it's definitely detrimental to the longevity of the cells. If your pack is a little out of balance, now we're talking 4.1, 4.2 Volts for some cells, and this is time for emergency cutoff of charging energy. A max charge voltage of about 14.2 Volts is ideal for a no-BMS LFP starter battery in good health. This allows for a bit of imbalance before anyone risks landing above 3.6 Volts. I like to put sense harnesses on packs like this so I can use a cell-log to quickly check voltages. Better to keep it easy to make quick checks on packs like this. One could also use the alarm output feature of the cell log to control a contactor to prevent overcharge. Lots of options, but this is the biggest gotcha when making a Li ion starter battery, IMO.

 

[John in CR]

I have only the old M1's from Dewalt packs, all wrapped in paper. I wouldn't glue the plastic wrapped ones like this for the reasons you mentioned. I haven't seen the pack get past 14.5, so I think I'm ok in terms of the alternator output. I have some of those cheap 2s-8s cell voltage display units with alarm (LVC only though), and will run an extension into the van to keep an eye on it for longer trips. I always have ebike tools and stuff when I travel, and worst case I left the car battery terminal connectors in place, so getting stranded for long isn't possible.

My biggest concern is high temps in the engine compartment.

If I get just 250hrs of usage or 250 starts out of them, that's 4-5 years for me, I'll be tickled pink.

 

[dnmun]

you can move the battery out of the engine compartment too.

you can use 4 channels of an 8 channel BMS to protect it. i have not seen a 4 channel BMSs yet.

 

[wb9k]

I have only the old M1's from Dewalt packs, all wrapped in paper. I wouldn't glue the plastic wrapped ones like this for the reasons you mentioned. I haven't seen the pack get past 14.5, so I think I'm ok in terms of the alternator output. I have some of those cheap 2s-8s cell voltage display units with alarm (LVC only though), and will run an extension into the van to keep an eye on it for longer trips. I always have ebike tools and stuff when I travel, and worst case I left the car battery terminal connectors in place, so getting stranded for long isn't possible.

My biggest concern is high temps in the engine compartment.

If I get just 250hrs of usage or 250 starts out of them, that's 4-5 years for me, I'll be tickled pink.

OK...Temps under the hood are probably a bigger concern for your climate than here, though in the summer it can be a potential problem here as well. If you can't get them out of the engine compartment (and who wants to run big copper through 1/2 the vehicle?) heat shielding or a small fan might be viable. Monitor temps inside the pack if you can, and you can make an informed judgement. As long as you're not getting above 40 C or so for extended periods of time, there should be little impact. If you reach those kind of temps just sitting parked, you'll probably want to find a shadier spot to park. If you can avoid lots of high temp time and the cells are in decent shape, I would think 4-5 years should be easy for the duty you describe.

 

[dnmun]

he would not need to run large cable through the vehicle. he could mount the battery under the dash along with the BMS and then pull the battery cable inside the cabin through the firewall and it would be plenty long enuff to reach the starter. the charging cable would not have to be substantial and he could cut a bunch of the shunt wires off of the BMS to prevent the alternator from charging the pack too fast. then for the starter current he could bypass the BMS on the output and use the discharge mosfets for an led signal light providing some info about the status of the pack itself such as over discharged.

 

[John in CR]

Vehicle won't pass annual inspection here without the battery strapped down on the battery tray.

FWIW I live in the mountains where almost every day is like a spring day where you live. Lower latitudes doesn't mean it's hotter, especially when the Caribbean and Pacific are less than 100mi apart. The sun can be brutal, but that's what trees are for.

 

[wb9k]

Sounds like you;ve got it licked! Let us know how it holds up over time. No salty breeze where you are, is there?

I've never seen a torn-down DeWalt pack with A123 cells. Almost all of the Hymotion packs and the very earliest BAE hybrid bus packs used the white paper-sleeved 26650s. Those both used pure nickel weld straps that were resistance-welded in place. What's the material used in the DeWalt pack weld straps? I assume it's fairly easy to solder to.

 

[Doctorbass]

John,

In the case of my integra ( when i had it) I have moved the 4s4p A123 pack in the trunk of the car and connected it thru the 2 big 3/0 AWG car audio cables i had in the time i was making SPL cometition... it run fine!

All you have to do is finding some cheap big cables and put it inside the car...

I used my A123 pack for 5 years with only a RC Blinky Balancer that balance at 70mAh all the time. I was anxious about the job it wodl do but it worked perfect!

The same battery was also tested to a 5000W rms subwoofer amplifier with awsome results!

These A123 did an excellent job at ANY temp... I mean here in Canada.. we have -30C in winter and +30C in summer :lol:

Doc

 

[John in CR]

Thanks Doc, I have some 50mah balancers, so I might add one to the equation. These old cells drop any surface charge pretty quickly, so what I generally see right after a drive is all 4 groups around 3.45V. After a week of sitting they drop to about 3.4V, and then self discharge all but stops. After over a year these cells were all 3.30-3.35V. Amazing cells for sure, but too bad they can't figure out how to get them to 2X the capacity.

 

[okashira]

Do you live in a warm climate? Lithium Iron Phosphate cells will not last long in the constant heat.
NCA chemistry cells would likely last much longer, especially if you can limit voltage range to ~3.2 - 4.05V.
Such as NCR18650PD or INR18650-25R.
I guess you will want 4s, which will keep voltage low, but with INR18650-25R you will need like 6p to get enough amps, depending on the car.
4S would put cells at ~3.0V - ~3.45V, you might get 10+ years out of them.
Would be expensive, though :-/

 

[dnmun]

i suspect the huge charging currents do not help either. that was why i was thinking of how he could cut the shunts off the BMS to get the charging current down. so it would charge when the alternator slowed down enuff to drop the input voltage below that which would cause the shunts to cut off the mosfets for hi current.

 

[wb9k]

The temps John cited are not going to have a big impact on cell life--in fact, they're about ideal for striking a good balance between performance and long-term calendar life. The charging currents in the car are going to be a walk in the park for these cells, just like the big discharge currents are. These are the same cells used in the Orion hybrid buses. That's a 20 Ah pack that sees current from regen spike as high as 900 Amps. Those cells just keep on chugging, even under what--for most cells--would be considered abusive conditions.

 

[dnmun]

i always think that charging currents have to be kept low to prevent the precipitation of the lithium metal away from the anode.

max of 1C for lipo and .5C for lifepo4. i believe in that more than the hysteria about charging lipo to 4.20V.

 

[wb9k]

John,

In the case of my integra ( when i had it) I have moved the 4s4p A123 pack in the trunk of the car and connected it thru the 2 big 3/0 AWG car audio cables i had in the time i was making SPL cometition... it run fine!

All you have to do is finding some cheap big cables and put it inside the car...

I used my A123 pack for 5 years with only a RC Blinky Balancer that balance at 70mAh all the time. I was anxious about the job it wodl do but it worked perfect!

The same battery was also tested to a 5000W rms subwoofer amplifier with awsome results!

These A123 did an excellent job at ANY temp... I mean here in Canada.. we have -30C in winter and +30C in summer :lol:

Doc

It's funny you mention the audio application. These cells are a real natural for high-end audio amplifier use. My main audio system at home (I've been both a pro musician and audio electronics specialist over the years) has an amplification chain that is entirely powered by batteries (again, the ubiquitous M1 26650). Super low noise and a super-low impedance power source for great transient response and steady deep bass power from even relatively large power amps.

Sorry for getting a bit OT....

 

[wb9k]

i always think that charging currents have to be kept low to prevent the precipitation of the lithium metal away from the anode.

max of 1C for lipo and .5C for lifepo4. i believe in that more than the hysteria about charging lipo to 4.20V.

A123 cells are rated for a continuous charge rate of 4C and pulses of 20C. You can exceed even these numbers, but you begin to eat away at calendar life more appreciably. Thermal management becomes more critical as well. In fact, this is part of the calendar life problem, but not all of it.

 

[John in CR]

Do you live in a warm climate? Lithium Iron Phosphate cells will not last long in the constant heat.
NCA chemistry cells would likely last much longer, especially if you can limit voltage range to ~3.2 - 4.05V.
Such as NCR18650PD or INR18650-25R.
I guess you will want 4s, which will keep voltage low, but with INR18650-25R you will need like 6p to get enough amps, depending on the car.
4S would put cells at ~3.0V - ~3.45V, you might get 10+ years out of them.
Would be expensive, though :-/

okashira,

These are real A123 M1 cells harvested from warranty return DeWalt 36V cordless toolpacks that I bought several years ago. They're the 26650 cells that A123 rated at the time for 30C discharge and 10C quick charge, though the nickel tabbing used in those tool packs limited current to below the cell capabilities. The 6p4s pack cranks the motor over so much better than a standard car battery that I think 6p is overkill, and if I make another one I'm going the route Doc did and use just 4p4s. In fact, since I still have plenty of leftover cells and 4p worked for Doc in Canadian winters, I'm going to make a little 4"x4"x2.5" or maybe even just 3p4s jump starter pack to keep in the glove compartment. Whipping out one of those with short leads just long enough to reach car battery terminals and spring clips modded for high current would be too cool the next time someone asks for a jump. Plus it would work better than cheap jumper cables could ever dream of.

wb9k,
Why should I be concerned about salt air? The pack is seal in a double layer of PET bottles with silicone completing the seals, so it's submersible as long as I don't dip the terminal leads or the balance tap plug.

 

[Punx0r]

Interesting

Looking at the spec for the current A123 26650 cells:

Cell Dimensions (mm): Ø26 x 65
Cell Weight (g): 76
Cell Capacity (nominal/minimum, Ah): 2.5/2.4
Voltage (nominal, V): 3.3
Internal Impedance (1kHz AC typical, mΩ): 6
HPPC 10 Sec Discharge Pulse Power 50% SOC: 200 W
Recommended Standard Charge Method: 1C to 3.6V CCCV, 45 min
Recommended Fast Charge Method to 80% SOC: 4C to 3.6V CC, 12 min
Maximum Continuous Discharge (A): 70
Maximum Pulse Discharge (10 seconds, A): 120 A
Cycle Life at 10C Discharge, 100% DOD: >1,000 cycles
Operating Temperature: -30°C to 55°C
Storage Temperature: -40°C to 60°C

It all looks pretty rosy for a 4S4P pack. The only place it loses out to the typical lead-acid battery is on capacity for small or parasitic drain, but that's a small concession considering the A123 version would be a fraction of the size and weight.

Does anyone know more about the HPPC 10 sec specification? It seems pretty weak?

I see A123 are now also advertising a 14Ah prismatic pouch and a 4.4Ah 32113 cell

I think I need a li-ion car battery... Just need to find some cells... I have a plan involving a discharge balancer activated by ignition live, so it only balances when the car is in use, not when stood idle.

Regarding audio use, I assumed batteries would be a very clean source of DC, but later read that the chemical reaction in the cells creates considerable noise in the supply. Perhaps that's just characteristic of lead-acid?

 

[dmwahl]

Regarding audio use, I assumed batteries would be a very clean source of DC, but later read that the chemical reaction in the cells creates considerable noise in the supply. Perhaps that's just characteristic of lead-acid?

Careful what you read regarding audio use, companies marketing to audiophiles are notorious for misinformation and exaggeration. Noise due to the chemical reaction in the battery sounds like major BS to me.

 

[wb9k]

Interesting

Looking at the spec for the current A123 26650 cells:

Cell Dimensions (mm): Ø26 x 65
Cell Weight (g): 76
Cell Capacity (nominal/minimum, Ah): 2.5/2.4
Voltage (nominal, V): 3.3
Internal Impedance (1kHz AC typical, mΩ): 6
HPPC 10 Sec Discharge Pulse Power 50% SOC: 200 W
Recommended Standard Charge Method: 1C to 3.6V CCCV, 45 min
Recommended Fast Charge Method to 80% SOC: 4C to 3.6V CC, 12 min
Maximum Continuous Discharge (A): 70
Maximum Pulse Discharge (10 seconds, A): 120 A
Cycle Life at 10C Discharge, 100% DOD: >1,000 cycles
Operating Temperature: -30°C to 55°C
Storage Temperature: -40°C to 60°C

It all looks pretty rosy for a 4S4P pack. The only place it loses out to the typical lead-acid battery is on capacity for small or parasitic drain, but that's a small concession considering the A123 version would be a fraction of the size and weight.

Does anyone know more about the HPPC 10 sec specification? It seems pretty weak?

I see A123 are now also advertising a 14Ah prismatic pouch and a 4.4Ah 32113 cell

I think I need a li-ion car battery... Just need to find some cells... I have a plan involving a discharge balancer activated by ignition live, so it only balances when the car is in use, not when stood idle.

Regarding audio use, I assumed batteries would be a very clean source of DC, but later read that the chemical reaction in the cells creates considerable noise in the supply. Perhaps that's just characteristic of lead-acid?

Bear in mind, this is the next-gen M2 26650 that is spec'd here. We are talking about the older M1 cell that isn't quite as pretty on paper, but it's pretty close. I'm not as sharp on the HPPC test as I'd like to be, but if we calculate for the max power the cell--at 50% SOC--can deliver for ten seconds while falling to half of OCV under the load, then the numbers looks to work pretty well to me:

(3.3V/2) * 120 Amps = 198 Watts

I may well be wrong about how the number is derived, but the above makes sense to me. Surely someone will chime in to correct me if I'm wrong.

Regarding audio use, I've never heard of "noise" in audio devices induced by lead-acid batteries. For preamps and phono stages, lead acid is a popular choice and it works fine. A Li ion battery is much smaller and lighter, which, perhaps coupled with longer life, is to me the only real appeal in that segment vs. lead acid. One could get great performance from a pack of 18650's in 4S1P in a pack a fraction the size of the 5 and 7 Ah SLA's that are common in that market today. In power amps, sagging voltage under load is going to be a real concern with LA batteries in many cases. Sagging rails impose distortion on the musical signal in the form of dynamic compression. It's not going to be perceived as "noise", rather the music will be less visceral because fast transients and big, loud (power-hungry) passages are potentially going to be scaled down in amplitude, so the full impact of those events is lost to the listener. A good cell with fast rebound and low output impedance can have a real performance advantage in that context.

Note also that the typical THD rating, for example, won't tell you jack squat about the scenario I just described, nor any other of a myriad of real-life performance concerns in an amplifier. I once attended a talk sponsored by the Audio Engineering Society called "Lies, Damn Lies, and Audio Specifications". The one-number "ratings" assigned to audio amplifiers and, to an even greater extent, loudspeakers are virtually useless. Another poster cautioned above to beware of audio gear-makers' claims, and this is sound advice--no pun intended. The only thing I can really compare directly to that level of skullduggery is battery specs To be fair though, I think the truth of the matter is that the honest, full characterization of both audio systems and batteries is a complex issue. These things perform differently under different conditions; their behaviors are dynamic. Most people buying a stereo, or for that matter building an electric bike, aren't interested in that level of detail. They just want a number they can plug into the formula they are using (or think they're using) to tell them everything is OK.

 

[wb9k]

Why should I be concerned about salt air? The pack is seal in a double layer of PET bottles with silicone completing the seals, so it's submersible as long as I don't dip the terminal leads or the balance tap plug.

If the paper sleeve becomes impregnated with salty water, it can become conductive over time. Now you're back to needing physical separation of the cell groups. You're right though, I think given the level of splash protection you're affording here is plenty good enough that this should not be a concern for you.

 

[Punx0r]

Thanks for the clarification - the 200W figure makes sense when you figure in the voltage sag...

Regarding noise in batteries used as a power supply, I must confess I was mistaken. I happily ignore 99% of technical "information" in the audio world as BS, but I thought the source in this case was Rod Elliot of ESP, which I consider to be a reliable source. However, having just checked properly, I see he says no such thing and in fact states the opposite: batteries are a good power supply for preamps as "the DC is completely smooth, and batteries are essentially noise free".

 

[John in CR]

Why should I be concerned about salt air? The pack is seal in a double layer of PET bottles with silicone completing the seals, so it's submersible as long as I don't dip the terminal leads or the balance tap plug.

If the paper sleeve becomes impregnated with salty water, it can become conductive over time. Now you're back to needing physical separation of the cell groups. You're right though, I think given the level of splash protection you're affording here is plenty good enough that this should not be a concern for you.

Thanks for pointing that out, because humid salt air soaking into the paper never crossed my mind. I still have lots of cells with the paper wrappers. If I was building for marine use, which would you suggest, painting the paper, or duct tape (or similar) as extra insulation between each series level, or both?