48v 8 seat golf cart LiFePo4 conversion: Mistakes were made

MateoMX

New Member
Background
I run a dive center in the Mexican Caribbean and we use an 8 seat golf cart to move customers wearing gear from our dive shop to our boats. A bunch of lead acid batteries were near end-of-life and I had spent waaaayyyy to long watching Will's YouTube channel not to consider LiFePo4.
New video idea for Will: Come to Mexico, help me out with this and I will teach you to dive and take you on some incredible dives, including underground caverns for free. You can also spend your time here in Mexico investigating how solar should be better used in poorer but sunny countries.
I'm not kidding - if you can make it here, I'll sort out the dives for you. We're one of the longest running and experienced dive shops in the area, you'll have a great time.

I opted for LiFePo4 because...
  • I don't trust myself with 'spicy' lithium chemistries that have higher fire/explosion risks. I also wanted to see a big improvement of charge cycle life over lead-acid.
  • I hadn't really seen much about Lithium Titanate and these cells seemed very expensive
  • I wanted to move away from Lead Acid since I charge the cart indoors and didn't want hydrogen off-gassing. I also wanted to avoid battery maintenance.
  • I wanted to reduce the weight of the cart to improve hill climb speed with a full load of divers
  • I liked the idea that with liFePo4 most of the capacity would be between 52.8-54.4 and with the DC drive of the golf cart that should help keep motor performance as high as possible.
The classic 48v golf cart battery pack is based around 8x 6v 225Ah Lead Acid batteries. I didn't need a particularly long run-time since I often make short trips with the ability to re-charge back at the shop whenever necessary, therefore given the rule of thumb 50% DoD for Lead Acid, I thought 100Ah should be fine.

Based on patchy data I was able to find online, the golf carts draw a max of around 100amps when accelerating or going up-hill. Spoiler alert: It turns out with a full load the current is probably closer to 200amps at times. It really can't be too much for too long because most of the main cables are only 4AWG. But with most LiFePo4 cells having a peak discharge of 3C I figured this should be OK for a 100Ah cell.

What did I buy and why?
Will is clear with his advice - Buy high-quality, matched cells, from reputable suppliers

Unfortunately, I live in Mexico, where...
  • there are no domestic suppliers (as far as I could find)
  • import duties are incredibly high and unpredictable
  • not all suppliers are willing to ship here
This meant my options were very limited. In the end I gambled on these 100Ah cylindrical cells from a pretty much unknown seller via Amazon. I knew there were risks, but I figured that by ordering through Amazon at least I would get some support if things went very wrong, or if I got stung for import duties (since this has to be covered by vendors using Amazon Mexico's import program). The 16 cells cost approximately $1450 USD including 10 day delivery and import duties. This was a fantastic price by comparison with other options here in Mexico. In the end, the import duty was over $200, so it was a good that I was able to claim this back through Amazon!

The build
December 2019
  • The rear section of the golf cart has plenty of space for the batteries under the seat. I built a rudimentary fiberglass box using bits of fiberglass left-over from a recent boat re-modelling. This box would help keep the batteries away from the metal from of the cart, which seemed a reasonable precaution to me.
  • All the cells arrived at about 3.3v, so I figured, great - they are all "the same". I proceeded to add them in a 16s configuration using the included busbars, which, didn't actually allow the cells to touch one-another which meant building a strong interlocking structure wasn't possible.
  • I had only been able to get my hands on a crappy looking 50amp BMS, which I installed and used to control a 300amp relay, so the real load went straight through the relay, rather than the BMS.
  • I added a 48v > 12v DC-DC buck converter to power the 12v systems of the cart (horn, lights, GPS tracker).
How did it go?
I charged the pack using a LiFePo4 charger through the BMS. The charger soon cut-out claiming a full charge, but the pack voltage was still low. So something about the separate-port BMS caused the LiFePo4 'smart' charger to think the pack was full before it really was.

To fix this, I connected the charger directly to the 16s pack and allowed it to charge, reaching a total voltage of around 58v before the charger stopped.

The BMS's mighty 0.25w resistors were bleeding off some of the cells by the end of the charge, but considering this was a 10amp charger, this 0.25w of discharge was negligible.

The cart worked really well! With significantly more acceleration and torque compared to the previous Lead Acid batteries (although granted they were near end-of-life).

After each short trip, I kept an eye on the pack voltage and everything looked pretty reasonable without much of a drop seen. Then, unexpectedly, during a trip the BMS hit its Low-Voltage Protection and cutout the relay and therefore removing power from the drive system of the cart. I knew this was a risk, so I attached a jumper that would allow me to force the relay to re-engage and I drove the short distance back to the shop. Once back at the shop I checked the pack voltage at it was still at around 52.8v (3.3v per cell), well above the low SOC voltage of 49.6, so I naively left the jumper in place and continued to use the cart for another couple of short trips.

I eventually saw the pack voltage dropped down to around 49.4, so I figured it was now time to re-charge. Before re-charging, I checked the voltage of the each cell in the pack and found that actually most of the cells were between 3.2-3.3v, however a few cells were 1.9-2.6v - so they certainly seem to be at very different levels of SOC.

Balancing
After going back to watch more of Will's videos I then started to understand the important of paralleling cells prior to use and particularly bottom balancing. I had incorrectly assumed that the cells being delivered at "about" 3.3v is actually not a good way of testing the true SOC of each cell. I followed Will's advice and left them in parallel for a day while I took a dremel to the busbars to allow me to reinstall the batteries in a more robust configuration.

I re-installed the pack and added better balance leads for a 2amp active cell balancer that I had ordered from AliExpress and I was awaiting delivery. I partly wanted the balancer for its balance functionality. Although, to be honest, I also wanted the ability to monitor the individual cell voltages via bluetooth so I could see voltage drops during use - kind of difficult to do manually with a cart full of people driving on the streets on Mexico! Believe me Will - you won't believe how jealous I became of your inverter + heat-gun loads. That would have been a much easier than trying to monitor batteries parameters while dodging pot-holes at 20mph!

Once re-installed, I charged the pack again and this time kept a closer eye on cell voltages between each use and topping up where required.

Problems continue...
It started to become clear that some cells would discharge faster than others and some cells would reach higher voltages during charging compared to others.

Once I received the balancer, this proved to be really handy. After running the cart low, I was able to do a bottom balance again, but this time it was much easier compared to removing the entire pack and putting all cells in parallel (albeit taking quite a long time). However, despite the efforts of occasionally bottom and top balancing the pack using the balancer, I realized there were serious differences in the characteristics of these cells.

Now, in July 2020 these issues to continue and I regularly see about several cells reach between 4-4.3v during charging, meanwhile others only reach 3.35-3.6.

Obviously, these 4v+ values are not good news at all and this is why I am reaching out here for advice.

What is the cause of the problem?
In your opinion what is the cause of the problem?
  • Bad cells from the start
  • Damage caused by initial use without bottom balancing
  • Wrong cells for the job - 70amp constant, 150-250amp peaks
  • Incorrect charging method
  • Damage cause by running the cart after the BMS had hit LVP
  • Something else
What can I do?
If you were me, what would you do?
  • Charge the pack differently (would different charge profiles allow the cells to charge more evenly without a few of them raising in voltage disproportionately to others)
  • Request a refund (not a realistic option now I have had the cells this long and made mistakes myself)
  • Give up and buy some of Will's recommended cells (likely to be over $2,500 for 48v@100Ah after delivery and taxes in Mexico)
  • Only some of the cells seem bad. Buy a few more and switch them out.
  • Buy an entire extra set and go for a 200Ah config (16s2p), this will lower the C rate particularly for discharge and will mean the cells with behave less erratically
  • Start a YouTube channel all about dive shops, golf carts and life in Mexico so I can afford to buy nice things rather than going for the cheapest options
  • Get Will down to Mexico to help me out while I teach him to dive and he makes some great solar and battery videos
 

zaps

New Member
Congratulations, that's a great story. (y)

I couldn't tell you what caused the problem, but my guess is that there's a reason those cells were so cheap, and they appear to have no markings on them.

Anyway, I think the "current" problem is due to your cells having mismatched capacities in amp hours. You'd need to check each cell individually to identify which ones are outliers, and reject them from your battery bank. Putting cells in parallel doesn't make them matched, it only makes them have the same voltage.

Of course, I'm new to all of this, have never done anything like it, and could be 100% wrong. Despite that, here are my thoughts on what you could do to solve the problem, in the same order as your bullet points:
  • Charging the pack differently wouldn't make the cells have the same capacity, so I don't think that would fix your problem.
  • As you already suspect, getting a refund is unlikely.
  • Buying good cells should work, but for some reason they cost more than the questionable ones. 🤔
  • I don't know if buying more questionable cells and switching some out would work. I'd suggest first measuring the real capacity of each cell you already have before deciding whether you want to buy more.
  • Buying an entire extra set doesn't automatically fix the problem of cell capacities being mismatched within a series.
  • Starting a YouTube channel seems like an interesting idea. You too could become a star!
  • Getting Will to help you out sounds great, but I don't think it's the best time for any of us to be crossing borders.😷
Best of luck!
 

Steve_S

Offgrid Cabineer, N.E. Ontario, Canada
Welcome, to the forum MateoMX.

Good story and excellent write up of the issues you're having. If it would be possible, posting a photo of your battery configuration would be handy, just in case there is an obvious something lurking there.

The CHARGER you are using ? Brand, Model & Specs (or link to it) would also be helpful. I hope you are using a charger which supports LFP, most standard Lead Acid Charger's are not really good for LFP, the charge profile is not the same and this can cause serious issues. From your description, my "gut guess" is the charging system is not a proper CC/CV which is sensing resistance relative to charge.

I checked your link to the cells you used. Sold by SHUNBIN ! Ohhh My !
I am sorry BUT there is NOTHING GOOD TO SAY ABOUT SHUNBIN !
See link in my signature - I am afraid you are likely a Victim of this Shlock-Dealer.
Sorry to be the bearer of bad news but I & several others have been phookered by this company and for very serious cash too.

Capture-shunbin.JPG
 

MateoMX

New Member
Thanks so much for you comments guys. I was very nervous making my first post just holding my hands up to a catalogue of errors. But I figured if it helped anyone avoid the issues I have faced then that is a good thing. I really wish I lived in the USA, where you guys have such easier access to quality cells!

The charger I am using is this one: https://www.amazon.com/gp/product/B07CXQBQ8D/

Some more questions...

Is there any point in capactity testing? It already seems as if we can be pretty sure we have a capactity mis-match. Will knowing the actual capacities help me? I guess if I do buy an entirely new pack then I could split these in to 4x 12v packs and match the capacities within these packs. I could use them from my UPSs??? But aside from know the capacities for re-use, is there anything that knowledge can do to help me fix this pack?

Should I buy some more unknown cells? I know it sounds crazy, but instead of buying an entirely new set of cells, what do you think about buying maybe 5 more cells from the same vendor. Capactiy testing all 21 cells and using the closest 16 of them? Or do I just need to face reality and buy all new cells?

Can I just keep using this as it is? Due to the short runs I do with the cart and the fact I can charge it daily, it is still useful as it is right now. Are there any significant safety issues in continuing to use and charge the cart as I am (with several cells reaching 4.0-4.3v)?

The photos
If you trace the series route you will notice it takes a kind of weird path, this is because of the angle of the bottom of the cart and the clearence with the seat. I needed the cells to physcially fit in a specific configuration and I ran the busbars accordingly.

20200110_183520.jpg
Parrelleling the batteries in the early days

20200706_073945.jpg
The battery pack in place under one of the seats.
The current monitor visible on the left hand side no longer works due to water damage.
You can see the balancer on the top of the pack

20200706_073936.jpg20200706_073925.jpg
The view from the left hand side of the golf cart. The humidity is definitely not doing some of the connectors any favors. But, only the plus side, low-temperature disconnect is not something I have ever had to consider! :)

20200706_073854.jpg20200706_073839.jpg
The view from the right hand side of the golf cart. All the balance leads are exactly the same length and I run them through the gas in the pack so they all emerge on this side of the pack. You can also see the 90amp breaker which allows for pretty big surges for prolonged periods (I'm guessing it is heat based?) and hasn't tripped the whole time I've used the cart, so the load cant be too crazy.
 

BiduleOhm

Electronics Sorcerer
Is there any point in capactity testing? It already seems as if we can be pretty sure we have a capactity mis-match. Will knowing the actual capacities help me? I guess if I do buy an entirely new pack then I could split these in to 4x 12v packs and match the capacities within these packs. I could use them from my UPSs??? But aside from know the capacities for re-use, is there anything that knowledge can do to help me fix this pack?

Should I buy some more unknown cells? I know it sounds crazy, but instead of buying an entirely new set of cells, what do you think about buying maybe 5 more cells from the same vendor. Capactiy testing all 21 cells and using the closest 16 of them? Or do I just need to face reality and buy all new cells?

If you buy another lot of 16 cells and capacity test all of them you can pair them up so each pair has the same capacity as the others but that's expensive of course. Buying less and hoping to have enough of them to make good enough pack will be less expensive but a lot more risky.

You can also find a high current (at least a few A) balancer (an active one would be less wasteful of the energy but there's plenty of crappy ones so be careful, see Will's video about active balancers for more info) which will be a lot cheaper.

A cost free solution in the mean time would be to sort the cells from the highest capacity one closest to the negative side of the pack, to the lowest capacity one closest to the positive side. Then you can run all your 12 V loads from the 4 cells closest to the negative side (and you don't need the DC/DC converter anymore...) and hope they'll use some of the excess capacity of those cells so they match better the other ones. It's obviouly not a perfect solution but it only require some time to rearrange the cells and should improve the things a bit. And you can also power some 24 V loads from the pack too (between negative and the middle of the pack) if you have some.

Can I just keep using this as it is? Due to the short runs I do with the cart and the fact I can charge it daily, it is still useful as it is right now. Are there any significant safety issues in continuing to use and charge the cart as I am (with several cells reaching 4.0-4.3v)?

If you don't use all the cell capacity between each charge I would recommend to not charge them all the way so the lowest capacity ones stay under 4.0 V (ideally under 3.65 V). Else they'll have a very short life...
 

Solar_Mann

Solar Enthusiast
I’d guess you overcharged the pack right from the start which didn’t help anything. You said the charger cut off while charging through the separate port BMS. That was probably when your first cell hit 3.65, who knows how many got well above 4v. Safest bet is to charge and discharge through the BMS. Even if they are unbalanced you won’t hurt anything running it that way. Might just be one of those things where you run it as is and call it good while ignoring the fact it isn’t perfect.
 

nosys70

Solar Addict
you said the cells are 100Ah but occasionally the load could go high up to 200Ah.
Possibily these cells are not designed to get a 2C discharge rate ?
on the other hand, discrepancies between the cells are common, so BMS is mandatory, but sometimes a BMS does the job, and cut the battery
because this discrepancy between cells is too high.
So you should return cell that are out of specs because obviously there is a problem.
a cell going under 2.5V is considered as damaged, so if yours have gone under 2V, there is a problem.
 
D

Deleted member 4428

Guest
From the photos I see: Very thin buss bars with a big hole in the middle. BMS leads should be on top of the washer. Corroded fasteners.
 

JoeHam

Photon Sorcerer
If it were my lazy butt I would run it until it dies.

I am waiting on a “drop in” 48v, 100 ah battery for $1250 as described here:


I hope to report good things when I have it in my hands but you never know.
 
Top