diy solar

diy solar

Uh oh. lost the magic smoke

Real question here; if you find a cell, even if its 10 years old, but you do a capacity cycle test on it, and it tests within 90% of its original published spec, and you build your pack to within those acceptable tolerances... how is this bad? Yes, there might be calendar aging issues later on, but these cells can still provide years of service. At least that what I've always thought...
Agree 100% on 18650. Mainstream 2nd hand cells have specifications and there are pretty good libraries (and old sales listings) where you can find these specs - which will include capacity and discharge amps - regular and max.

For example - here's a 'hi amp' 2nd hand cell https://www.batteryclearinghouse.com/products/30-cells-high-grade-lg-mh1-10a-3ah-cells-in-10s3p-pack Notice the "Continous Discharge Rating 10a".
1652144213637.png
Here's a laptop cell - https://www.batteryclearinghouse.com/products/30-lgabb41865-2-6ah-18650-cells-in-10-grey-modem-packs
1652144706717.png

Next you have to design your battery. eBike packs typically use 'hi drain' cells (10a or 20a discharge specs) because the battery builder want's a small battery. Laptop cells (the most common type of second hand cells) typically have 0.5a standard discharge and 1 to 4a max discharge - so you can use them but you need 10-20x the number in parallel than a 10a discharge cells for the same current level. However, a 10-20x battery will last a lot longer so it's perfectly OK if you have room for a 10-20x larger battery.

Let's say you need 100a @ 48v battery. For 18650 that would be 14s and each set of parallel cells in the 14s needs to deliver 100a. If you have a 14s20p then 100a / 20p = 5a/cell. If you do 14s50p then 100a / 50p = 2a/cell. And so on. You just have to figure out the max load the voltage and max number of cells you have room for - e.g. the XsYp and then acquire the appropriate cells.

Test procedures will tell you if the cell is OK - regardless of 1yr or 5yr or 10yr or 12yr old history. Testing is straight-forward - charge, discharge at the battery design amp level you want (weeding out heaters). If near original capacity (85% or 90% or 95%) and IR is OK and no self-discharging after a few weeks - then you have good cells.

Note: If you overload a battery of 18650 cells - e.g. pull more current than the cell is designed for, you'll get extreme voltage drops. The more you overload the more you'll see. For example - pulling 5a/cell from 0.5a 'standard discharge' cell will lead you to see a full charge of 4.15v/cell drop to 3.5v instantly 'kind of thing' and then down to 3.3, 3.0 / BMS cut-off within 2 minutes And the cells won't last long in terms of cycles.

This post is just a long winded way of agreeing with @madmax Even new cells become 'old' in a year or 2. :)
P.S. I use this date code tool - it doesn't always work but more often than not for mainstream cells - https://batterybro.com/pages/18650-date-code-lookup-tool
 

Attachments

  • 1652144292489.png
    1652144292489.png
    43.6 KB · Views: 3
  • 1652144645540.png
    1652144645540.png
    25.8 KB · Views: 3
Last edited:
Somehow I knew that this would devolve into links to battery resellers. I should have known.

I'm out - continue on!
 
So finally got the electric cycle to a position where I could test it out.

Low "gear" worked great, slow wind up to about 30 mph. Flipped it to high and gunned it. Power cut out.

I was thinking I was pulling too many amps through my transfer switch between my 2 batteries, so I removed it.

Hooked up the 18650 pack and took it for a spin. Cut out again. So I looked down and saw flames shooting from my 18650 pack.

So these are not going back into this build. I'm just going to double up the SPIM08HP cells since they can pull more amps and have a much better connectivity method.
I just wanted to say thank for posting, it's hard to do and I appreciate it. Learn a lesson and move on, don't let those calling you names get you down.

What was the config (xSxP) of this pack?
 
It was 16s21p. When I was doing stand level testing it only pulled 20 amps at full throttle.

I completely forgot about the fact that friction and load would majorly adjust that amp load. That was my bad and no surprise that the pack flamed out.

The spims however are rated for 200amp steady load and 800amp spike, so yeah, that is where I am headed from now on.
 
make sure you compress the SPIM's.
And no, I don't believe that is the reason your pack flamed out directly. I capacity test each of my 18650's, and it takes like 10 hrs for a cycle. Then I have to let it sit for a week or two after testing to ensure that it can still hold voltage. Then have to type all of the capacity values in and then build parallel sets that have the closest series values. Even if you did all that, you can still have ones that die on you, days, weeks later.
It is so damn time consuming that I've given up on using 18650's pretty much once I've exhausted this set. I'll still use them for flashlights and stuff, but as far as using them for building packs, I'd order new (or tested by a reliable source), or just go to larger cell sizes.

Even at 5A a cell draw, 21p should have given you over 100A. It depends on how close a safety margin you'd like, but I would have went with a max 120A BMS. In addition, cell level fusing would provide protection for those few cells that are out of control. Air flow makes such a big difference when you get to the upper limits of 18650's for cooling, and you have one of those "stacked" staggered battery holders that save space, which hold in heat.

I'd give it another go, but go with cell level fusing, not the nickle 2P one you are using now, and use a 100A bms. It won't be as "fun" as a super power pack, but for just zipping around, its still a viable pack.
 
It was 16s21p. When I was doing stand level testing it only pulled 20 amps at full throttle.

I completely forgot about the fact that friction and load would majorly adjust that amp load. That was my bad and no surprise that the pack flamed out.

So not on a dyno?
Need some kind of drag. Spinning a propeller in a container of water would be good.

A current sense and regulating speed controller to limit max current (and max cell temperature) could help prevent failure.

Maybe supercapacitor would be a way to get more surge for launching.

1652190684772.png
 
I'm curious -- Where the pack blew out and flamed up -- was that where you were drawing your positive from the pack -- I find tapping into several areas with multiple wires distributes the heat much better --- on a 6P pack I'm building right now -- I'm using 6/14ga wires of the same length tied into a 10ga wire at the pos and bms connections -- when I did a test discharge of 35 amps for a complete cycle -- the heat difference was substantial
 
I'm curious -- Where the pack blew out and flamed up -- was that where you were drawing your positive from the pack -- I find tapping into several areas with multiple wires distributes the heat much better --- on a 6P pack I'm building right now -- I'm using 6/14ga wires of the same length tied into a 10ga wire at the pos and bms connections -- when I did a test discharge of 35 amps for a complete cycle -- the heat difference was substantial
My guess is that his distribution wires from the BMS to the bike controller were probably ok- if there was resistance there, the wires would get hotter and hotter and cause more resistance, which would cause the batteries to actually have less draw. I forget the math of things, but the general rule was something like Target gauge=2xgauge -4... something like that. Some actual electrician will chime in and correct my error. So like if you wanted a 6 gauge, you'd do like (10ga-4)=6 then 2 of them... You increase load carrying capacity and most likely reduce resistance, but by how much varies. Most of these things use multi cable because thicker cables are harder to bend.

He didn't seem to run it very long before flames. So you think its possible that the heat generated at the nickle strip to wire junction caused enough heat to cause the cells to rupture and catch on fire?
 
He didn't seem to run it very long before flames. So you think its possible that the heat generated at the nickle strip to wire junction caused enough heat to cause the cells to rupture and catch on fire?

Doubt it, poor thermal conductivity.

With series/parallel, resistance might cause one string or cell to carry more than its share.
That's where the suggestion of multiple wires from strings to BMS or busbar comes in. Matched lengths like headers.
 
Again, professionals make the same mistakes the first time around.

Back in the 1980's I was designing microprocessors at HP. I did some ESD protection, and heard of past experience:
ESD protection is multiple interdigitated fingers of conductor to clamping diodes. The part is tested with 5 zaps, if no change in input electrical leakage, it passed. But one more zap and it failed.

Trace resistance wasn't matched across the fingers, so current crowded into finger of least resistance, blew out one at a time instead of each dissipating 20% of the energy.

1652199206149.png
 
Doubt it, poor thermal conductivity.

With series/parallel, resistance might cause one string or cell to carry more than its share.
That's where the suggestion of multiple wires from strings to BMS or busbar comes in. Matched lengths like headers.
Ahh. I do recall for all the nickel flat ribbon that is used that it actually has a pretty high resistance value as compared to copper.

So.. equal length headers like for an engine so that the flow rates are all matched across all ports. I get the concept, but does it make that much of a difference if this pack had a copper bus rather than a nickel bus?
 
I did a T with the copper wire and soldered it down on the nickel strip. When I yanked the wire out it pulled half off. Once I fully disassemble the pack I will hopefully get more info. From what it looks like it was half the line, so I am thinking there was some sparking that melted a cell, caused it to vent start and spread the fire out.
 
I had a bunch of 18650's that I just kinda.. well lazy to shuck them from their cases. Finally did it.
These were (most) of the cells:
2VbuWtp.jpg

Brand:LG
Model:LGAAS31865
Capacity:2200mAh Rated
Voltage:3.60V Nominal
Charging:4.20V Maximum
1075mA Standard
2150mA Maximum
Discharging:3.00V Cutoff
430mA Standard
3225mA Maximum
Description:Blue Cell Wrapper
White Insulator Ring
18650 Form Factor
The mfg date on the pack were from 2011.

So far, the first batch that went into testing are all testing above 2200, which is awesome for what they are. Low buck, low discharge , medium capacity. I just dont want to waste these since I have em.

They will get built into 14P scooter packs (I have a bunch of swagtron aluminum fire resistant enclosures) for a total of 12 packs.
Each pack is capable of about 5a continuous draw... with 4 packs I can do 20A draw continuous, and well over 100Amax if needed.

sorry for the thread hijack. I didn't realize until I was done, but I am just trying to say like everyone else, use used cells wisely, and they will reward you with cheap power. Abuse them and they blow up on you. These are already 10 years old, and their charge capacity tests like new. Will there be some crappy ones? Absolutely!
 
He didn't seem to run it very long before flames. So you think its possible that the heat generated at the nickle strip to wire junction caused enough heat to cause the cells to rupture and catch on fire?
I did notice he put the Nickel strip right on the positive side of the cells, no extra insulating ring. Most of that heat shrink is very thing and not very heat resistant. If that nickel strip got real hot right there and melted the heat shrink, it would have caused a direct short
I wonder if he had the pack to controller Fused
 
Please - stop trying to justify the use of band-aids for trash.

Presumably this forum is to "learn" from, and not just a be a dumping ground for pics of stuff that "blew up real good".

I have stated before that this is not personal, and have given the reasons why this blew up on the op, despite his meticulous care of measuring capacity and internal resistance. Because with 2nd-hand cells that have unknown SEI-layer characteristics due to age/use/abuse, things can go bad in a hurry.

The op may have gotten away with it had this been a sub-c little general purpose unit to power small things.

But no, this is a MOTIVE POWER application, where overlooking things like this go pop.

I provided an answer. Something to learn from. And the basic fact is the use of sub-standard, randomized cells. Motive-power applications don't like that.

Again - would @Will Prowse give this the big thumbs up if he opened his Tesla and saw stuff like this inside his motive power application?

Just because you don't like the answer doesn't make it less relevant or true. However, as noted before, you are a targeted marketing demographic to dump these things on. If not by direct sales, then surely none of you go out and buy 10 brand-new Ryobi or other tool manufacturers to rip apart which is vastly more costly and time consuming. No, you either collect them from other's garage, or get them from guys which divert the waste-stream.

I've been down this road back when tools used to contain A123 LiFePO4. Then of course the counterfeits followed. Tip: are you sure your battery-reseller is actually an authorized dealer, or are they pawning off A123 counterfeits?

Unfortunately, many who don't like this answer are hoodwinked by proposed band-aids to cover up the fact that a random collection of 2nd-hand cells, despite consumer-level testing, are simply starting up on the wrong foot, especially in a motive-power application.
 
i greatly appreciate OP for sharing and for the insights provided by @Substrate and others.

wonder if there is ANY nondestructive way to evaluate SEI "condition" mainly to assess for dendrites since that's a failure mode on my radar.


most people probably don't have an xray available or CT scanner, and the use of one would pump up the cost, making the suggestion moot.

nonetheless, i do wonder what the limits of electrical assessment is (charge, discharge, coulomb counting, internal resistance calculation, etc) even up to a 5-10 cycle test that evaluates the small shifting of variables.

for reference, i try to source new or barely used cells, and wish to gather lots of knowledge about each component so the tradeoffs are more clear

thanks again to everyone contributing! ?️
 
One of the main problems with the OPs packs --- The cells he used ,No matter if they were Brand New out of the factory to his door step --They were not the right type of cells --- Laptop and Modem cells are not designed for High Amp output --- But on the other hand I see no problem with using recycled High Drain cells that are designed for this application ---To me they have to be the same manufacturer and model with test results close to a new cell -- I know I'm gonna hear my testing is inferior --- I test capacity , resistance with a 4 wire tester and I do simulated High Amp discharges --- It's the best I can do and I have had no problems ( other than burning down my house and workshop, I live in a van down by the river --- LOL just kidding )
 
Last edited:
Back
Top