Testing HWC 18650 form-factor LiFePO4 (3.2v/1.4ah) Cells

I mentioned earlier over in the "even newer 24 channel BMS" board that I would be getting in a sample set of HWC 18650 format cells for testing in an EV project. I have received my cells and have had a little time to tinker with them. The cells are available online at Volts Phreaks - a distributor apparently already popular with other forum members for their isolated chargers. I was attracted to these cells because that vendor has a good price break when purchasing over 50 units. My end goal is to possibly use these cells as the traction pack for a converted VW Cabriolet (more details on that can be found over on my project blog).

I would like some input from the other members of this forum on the best ways to test and document the performance of these cells for me and other users of this forum. How do people normally load-test cells? I would like to "verify" in some way that these cells do perform close to their 1.4 ah rating. How do I test that? Is it as simple as (1) charge to capacity, (2) place a known load on some set of cells, (3) watch the voltage and amps over time, (4) chart + graph the data at fixed time intervals, (5) repeat for many cells?

Here is a shot of the cells as delivered:



I only have 2 relatively cheap volt/ammeters (one digital, one analog) at my disposal for making measurements. I do have a 5.5 amp load I can put on the batteries, as shown in the second picture here.



Here is my first impression of these cells:

These cells the come with no warranty (as they are intended for engineers). To me that means there is a nonzero failure rate for them and they must all be tested upon arrival. In general, LiFePO4 batteries shouldn’t be discharged below 2.5 volts and cells below 2.0 volts are often considered dead. Of the 80 cells delivered, 75 arrived with charges of 2.5 volts or higher. 4 came with charges of 2.0 volts to 2.5 volts. These cells perked up after a good sit on my home-made charger (an LM350 circuit). The last cell arrived at 0.6 volts and while it took a charge to 3.6 volts and then handled a load it had more thermal dissipation (read: it got hotter) that the other cells in it’s pack during testing did not present. I marked this cell bad. One bad cell out of 80 - that’s not too bad for "cheapo" cells.

Cheers,
--Adam

ahambone said:

I would like some input from the other members of this forum on the best ways to test and document the performance of these cells for me and other users of this forum. How do people normally load-test cells? I would like to "verify" in some way that these cells do perform close to their 1.4 ah rating. How do I test that? Is it as simple as (1) charge to capacity, (2) place a known load on some set of cells, (3) watch the voltage and amps over time, (4) chart + graph the data at fixed time intervals, (5) repeat for many cells?

I only have 2 relatively cheap volt/ammeters (one digital, one analog) at my disposal for making measurements. I do have a 5.5 amp load I can put on the batteries, as shown in the second picture here.

Click to expand...

Adam,

Check out this thread to collect info about cell testing: The LiFePO4 Headway cell 38120P TEST REPORT inside

Although your manual method of testing could be OK if done right, I think you will need something more automated, maybe like the CBA II some of those guys use. You'll want to spot the weak ones if they deviate too far from the others. Having a device that will automaticaly measure DC internal resistance would be nice too, or you could find it manualy using two different loads.

Hope they turn out well for you, and please let us know! I was spoting these since they aren't too expensive, and also the higher power/lower capacity ones also available at voltphreaks.

Thanks for the advice, Zap.

I've spent a few evenings setting up tests for these cells on my workbench. You are right - for a formal test I would need some kind of measurement equipment that I don't have. For the time being I decided to make do with the two multi-meters that I have and use the old car headlamp I have as a load. I do have an ADC L91-4003 DC motor on my bench too, but it eats power for lunch! My string of (3p)*4s A123s will turn that huge motor happily, just not for very long.

My real goal here is to see if these cells perform anywhere close to their rating. I am hopeful that they do, because they are so cheap that stringing groups of them together in a 10p with copper or nickel strips is doable for my EV (car) application. These cells might just turn my dream of not using lead-acid into a reality.

I do not have a good way to test individual cells, and I think that testing them one at a time would take forever. A "12V" battery made of these cells was the best I could come up with. I have run 4 nights of tests against a set of 12 cells, each day setting all of the cells in parallel for 12 hours and charging them to 3.53 volts (that's the best charger I have on hand for now). Then I would connect the cells in a (3p)*4s configuration.

The cells "should" be able to push 4.2A without issue, but I found that when I asked the cells to push more than 2A each they started heating up. At ~4A the heat was significant enough to make me be worried about damaging the cells. I decided to keep the per-cell draw under 2A for the purposes of testing. Maybe I should order some of those higher output (10C) cells voltsphreaks and try them out.

The first couple of nights I screwed up the process by wiggling connections between the cells while touching the cells to see how warm they were getting. The last couple of nights I got my system down so I could "measure" temperature without interrupting the flow of current. I think I have my testing procedure's kinks worked out.

Here's the data from my most recent test (attached):


I might upload my test results from my earlier runs later; I haven't keyed them into a spreadsheet yet.

The graph doesn't show it but these cells exhibit the typical LiFePO4 "cliff" at about 2.5V/cell. At that point the power output drops dramatically. One of my earlier bench tests ran the batteries down to a 2.1V/cell LVC and the last 2 minutes showed a very quick voltage drop. The cells got warmer the closer they came to 2.1V, as well. For safety's sake I would suggest being conservative with LVC (2.5V) for anyone using these cells.

Here are my conclusions so far:

- These cells are close to their 1.4 ah nominal rating. It's actually more like 1.3ah or even 1.2ah of usable energy.
- Don't expect "heroic" output from these cells. They handle up to 2A just fine but over they they start to heat up. They are rated at 3C (4.2A) but at that output they get quite warm. I don't recommend pushing these cells to the limit of their rating.
- I still think the bang for the buck is good, especially for people who want to make packs out of them using spot-welders.

I am currently in the process of making a capacitor + SCR spot welder to try on these cells. When bought in quantity, 10 of these in welded in parallel will have something like 13ah of capacity with a 20A to 40A load capability - for about the price of a single A123 M1. Not stellar performance, but certainly worth investigating.

Cheers,--adam

Thanks for posting your test results, Adam. The results aren't looking too great so far, so maybe the high rate cells from voltfreaks might be better and closer to spec?

So far, it would look like something like the newer headway cells might be a better option, specially for a big pack like a large EV would need.

Thanks for the tests!

I'd really like to see how the high power cells perform. I'm considering building a 12s9p pack of those 10C rated cells. The price really is great.

I will be posting some updated test results soon. I have ordered a set of the VoltPhreaks chargers. After reading the posting below I suspect that my charger may be the reason we were seeing capacity about 15% lower than expected.

With LiFePO4 cells, the voltage rises at a slow steady rate, as it charges, until the cell gets to about 3.65-3.70V, which if you stopped there is about about 85% full. Above that point the voltage rises at a much higher rate. It will take only seconds to get to 4.0V, or higher. In order to get the last 15% into the "tank", you need to hold the voltage at somewhere between 3.65-3.70V, and let the current taper off. When it gets down under about 50mA, or so, the cell is about as full as it is going to get. With LiFePO4 cells, the crossover point voltage is not all that critical. I've tested cells using a VP charger that has a crossover set to 3.84V, and then tested the same cells using our BMS, which has a 3.68V crossover point. Once the surface charge is burned off, which with some cells happens on its own, but with others you need to put a very slight load on the cells for a few seconds, to get rid of the surface charge, the cell ends up with the same "full" voltage, and has the same capacity.

-- Gary

Click to expand...

.

I think my problem is I used an LM317 to charge to 3.60V and did not really handle the "topping charge" that Gary described. If his hunch that the last 15% or so comes from the tapering-off charge then these cells may hold up closer to their actual rating. That would be really good for me because VP lowered their cost significantly when bought in quantity.

I will post new test results once I get the chargers in and test out some cells with a proper charge.

Also, as a word of advice: NEVER SOLDER DIRECTLY TO THE ENDS OF THESE CELLS. I knew better but for some reason decided to try that out last night ... to disastrous results. I ended up throwing a bulging cell outside about 10 seconds before it exploded and blew the can a few meters away from the electrolyte. I have more photos of the carnage if anyone is interested and I am glad that nobody got hurt.

View attachment canRemnants-small.jpg

Cheers, --adam

ahambone said:

I ended up throwing a bulging cell outside about 10 seconds before it exploded and blew the can a few meters away from the electrolyte. I have more photos of the carnage if anyone is interested and I am glad that nobody got hurt.

Click to expand...

Wait, what? Thought stuff like that wasn't suppose to happen with LiFePO4?!?

And... yes... more photos please!

Yes, we want more pictures of the carnage But mostly, what is the copper colored ribbon made of, is the black coating lithium iron?

That could have been dangerous, how long did you leave the heat on the positive terminal?

Technically VP says on their site "will not vent with flame". I guess there is a reason they do not say "will not explode"

Here is the story:

Every discussion board I’ve read mentioned that spot-welding is the correct way to attach tabs and nickel strip to the ends of raw cells. Last night I spent several hours working with my CD spot welder, adding tungsten tips, tinkering with voltage, etc. all in an effort to get a good spot weld on these HWC cells. Things were not working so well with the spot welder but that is a topic for another thread.

After a while and lots of frustration I thought, “hey, why don’t I just tin the underside of the nickel strip and then use the iron to attached the tinned strip to the cap of the cell?”. Bad Idea Jeans. Really bad idea jeans. More like bad idea bleached acid-washed jean cutoffs.

After tinning the nickel strip I placed it on the cap of a cell. I started applying heat. At first nothing happened and I also got no adhesion from the solder. After applying some more heat to no avail I pulled the iron and the nickel strip off.

At that point I got my first sign of a problem - the blue heat-shrink wrapping of the cell shriveled back from the top of the cell. I knew something was wrong. The cell started hissing and the cap (+) end swelled up away from the can by about a millimeter. I threw the soldering iron on the tile floor and then managed to open the back door and throw the cell out on the patio. At this point I thought I was okay because, at least according to all of the videos I had seen of LiFePO4 batteries online, the cell is supposed to gas and vent. I was happy with the cell doing that outside.

It was about 10 seconds later when the surprise came. Bang. Like a someone set off a black cat firework. The cap blew off the can in a small explosion. I couldn’t find the can until this morning in the daylight - it had blasted a few meters away from the cell.

You can see in my earlier picture that there is some printing on the side of the cells. I might pull the blue heat-shrink material off of one of my weaker cells just to see what that writing shows.

When I threw the cell out on the patio it vented for a few seconds before popping. When it popped it left a nice little explosion mark on the patio, as seen here.

View attachment blastPoint.jpg

The electrolyte guts and what looks like copper sheet lining were spewed against the wall of the back of my garage. Here is a shot of the battery innards a few minutes after the pop. Most of the innards landed about 1.25 meters away from where the cell was originally resting.

View attachment shrapnel.jpg
View attachment 3

I could make out the connector that used to connect the positive cap to the innards of the cell.

View attachment electrolyteShrapnel.jpg

The remnants of the cell casing plastic/insulator was found shredded a small distance from where the cell popped.

View attachment theOldWrapper.jpg

I didn't find the can until the next morning. It was a few meters away in the bushes. It was surprisingly in tact and I could tell that the cell clearly blew the cap off and then send the guts flying one way and the casing the other. Yay conservation of momentum

View attachment lookingDownCan.jpg


It might be worth my time to purposefully short a fully charged cell, leave it on the patio, and watch what happens. I saw a video on youtube where a guy took a weak lifebatt and shorted it to show the venting-not-exploding failure case for shorting the cell. In my case I had a hot iron on the cap of the battery. The iron may have put more heat into the cell than the engineers thought it could produce when shorted. If that is the case then I should be able to short a cell and watch it vent but not pop. If these things pops when shorted then I have some concern - I don't want to put 1,500+ of these things in my Cabriolet if they pop when shorted!

Cheers, --adam

That could have been dangerous...

Click to expand...

I am fairly certain that it _was_ dangerous!

I left the iron on the positive terminal for about 3 to 5 seconds - through the nickel strip. I had already tinned the nickel strip on the bottom side. I was hoping the iron could get the nickel hot enough to melt the solder and get some to stick to the positive cap. Before this cell popped the strip and solder pulled away completely without any connection to the positive plate, so not enough heat was present in the iron-nickel-solder-cell sandwich to melt the solder.

According to VP's web site the positive cap is galvanized steel and the can is (apparently non-galvanized) steel. I've had problems getting my 1.45 F 12V spot welder to weld anything to the positive caps of these things. The negative end seems to work much better. Any suggestions on how to make my welder work better are welcome!

Cheers, --adam

I've soldered a lot of 18650 LiMn cells and now I'm starting to wonder if maybe that's not such a good idea either. :lol: Thanks for sharing your story!

Is there any chance the solder liquified and went under the cap ? many of the 18650 cells i've seen have a " cap " or 2ndlayer on the + side ... i have no clue but just curious...

Quick thinking !! great to hear it turned out ok ( all things considered )

Yesterday I soldered 48 cells to a 8s6p pack directly to the + and - poles with a 150W solderiron. I took a 1,5mm² copper wire twice and soldered very short and had no problems with the High power LiFePO4 18650 cells. The solder iron should have at least 100 - 150W ore more.

I left the iron on the positive terminal for about 3 to 5 seconds - through the nickel strip.

Click to expand...

Here are some photos from soldering my battery. There are no problems if the iron has 150W but don't solder longer as half a second, then there should be no problems with overheating!

I am certain the high-power cells are more tolerant of heat than the standard VP blue cells. Maybe I should consider upgrading to the higher power cells even though they have lower capacity. I was planning on having so many cells in parallel that it wouldn't matter that the cells were high-C output or not. At that point I was maximizing for capacity. That being said, if the high power cells are easier to work with then maybe it is worth the higher-cost, lower capacity trade-off.

Cheers, --adam

Hi Ahambone,

Glad no one was hurt! By the way these cells will not pop when shorted, they aren't capable of enough current. It was definitely the heat of the solder that caused this. Without solder, some ways the cells can pop the way yours did is due to overcharging (to 5V, etc). I've never seen or heard of it happening so violently as your case. Usually there's a "pop" and the top pops open, nothing ejected anywhere. Sometimes the cell still works (obviously you wouldn't want to keep using it), sometimes it's open (zero volts).

- Tony

P.S. Want to add, the ONLY times I've heard of these cells popping is overcharging. They just aren't capable of enough current to heat up hot enough during use. I've never heard of one of the higher power cells popping or getting damaged, even with overcharge (doesn't mean it never happens, just that no customer has reported it, and I've never experienced it myself). The venting can be considered an "explosion" in that it can be violent enough to eject material, but there is no flame. It's not the same chemical reaction as lithium cobalt. The same thing happens to A123 cells when heated, the top will pop open, and it can happen violently if the heat and change in temperature (solder heats it up very fast) happens quickly.

The good thing about lifepo4 is that there's no thousand degree flame that burns even without oxygen and keeps burning even with water poured on it. But a lifepo4's battery contents will expand when heated up, and like any airtight container can react violently!

Hi Tony,

Glad to see you are checking in! Thank you for the update on the experiences you have had. My "explosion" was definitely a pressure burst and not a hot-metal-oxidizing kind of explosion; nothing kept smoking or burning after the pop. Hearing that the cells can not really produce enough heat when in a healthy or shorted circuit is good to know.

And thanks for the warning on overcharging. I haven't charged past ~3.6V with these cells yet, and don't intend to :wink: My current plan is to get my DIY CD spot welder in better working order and use that with these cells. Much shorter burst of heat that way.

Cheers, --adam

Hi Adam, Can you update what has happened to these HWC 18650 size LiFePO4 cells.
I checked out your VW build and see you changed to using the larger Headway Cells.

Did you get any documentation with the original samples of the HWC 18650 size LiFePO4 cells.
HWC 18650? 3.2V << what is the character in the ?
1400mAh GB14 FL25

I was wondering if they are really LiFePO4.

Did they find useful home?