PSA: dont underspec your strips

flippy

1 MW
Joined
Aug 12, 2015
Messages
2,351
....because of you do you will end up with something like this:

njjl5h2l.jpg


melty....

the customer came to me because often he would smell something "funky" after pulling full amps.
this is from a pack that was built by "a friend" for him.
he used the cheap 0.15 nickel plated steel and simply started stacking strips and using too low welding current to make connections and increase the current capacity.
i could simply pull them off in a single light tug.

i think we can all imagine what would happen if he kept going and ignored his nose on this one.


please listen to the advice given by the peeps here that built lots of batteries, only you can prevent battery fires.

C9IaplqU0AA31Dj.jpg
 
This is another reason the slit in the bus strip is important. It is helpful in reducing localized stress in the spot welds.

Constant heating and cooling cycles cause metal expansion and contraction. If the bus strips are thin, they will run hotter than the cell tip, leading to uneven heat-expansion between the two connected parts.

Add to this the constant vibration and bumps from riding an ebike, and cracks on the spot-weld can start and grow until it separates. Then, the remaining spot welds on each cell-tip have to handle more current than they did before, leading to the surviving spot -welds running hotter, in a downward death-spiral...
 
the slit helps if you have a cheap low power welder. if you have something more meaty then you can just compensate with more joules/amps.

i did some delicate questions and it seems that the guy has a sukko 7xx welder. too bad he does not know how to use it and/or has worn out electrodes. that welder should be able to weld 0.15 real nickel, the steel fake crap he used should not have been any problem. my guess he welded at way too low power settings or his electrodes are shit.

he used holders, so the welds did not have any mechanical stresses on them, if those holders were not there the pack would have probably shorted out and burnt out long ago. this is why i always recommend using holders on mobile applications were shocks and viberations are a real problem.
the tape was also melted at some places due to heat buildup.

about 2 dozen of the 180 cells had serious heat marks where most of the power went tru one or 2 shitty weld marks instead of 6 decent ones. those cells were completly shot, they had massive IR and held less then 1500mAh.
there was heat warping visible on the cell holders and foil from the cells heating up WAY beyond their rating. those cells are tossed in the bunker outside coverd in sand.

its pretty amazing the abuse those samsung 29E's can handle. most cells were still salvagable but i still recommended he dumps the lot and start over, wich he will.
 
Funny, i was talking to a guy who worked on the Boeing 787 the other day. Stated cause for the battery failures on the plane was just basic interconnect shit like this :shock:
 
Nope, still need strain prevention somehow, deeper welds just tear the nugget right out of the can, if they don't already cause pinhole failures just from welding.

Spot welds are how to fail at making a battery.


flippy said:
the slit helps if you have a cheap low power welder. if you have something more meaty then you can just compensate with more joules/amps.
 
[Following] :mrgreen:
 
liveforphysics said:
Nope, still need strain prevention somehow, deeper welds just tear the nugget right out of the can, if they don't already cause pinhole failures just from welding.

what stress is a nickel strip under when proper cell holders are used?
 
I have to agree with LFP that ultrasonic friction bondng individual fuse wires is best, and spot welding is second-best.

The reason I am frequently posting about spot-welding techniques is because I realize a lot of home garage pack builders are going to spot-weld, and I hope to help their results be "less bad".

I started a thread a while back to provide s public parking spot for any info we can find on possibly re-purposing some type of ultrasonic transducer into a DIY fuse-wire friction-welder. It faded fast, even though I still feel it has great potential.

I am not worried about home pack-builders hurting the production of ebike battery packs. 99.9% of ebike packs are purchased, rather than built by the user, and reducing that number by 1/10 of one percent will not affect anyone who sells company-built packs.
 
flippy said:
what stress is a nickel strip under when proper cell holders are used?

There will be expansion/contraction from thermal cycling. It may take hundreds or thousands of cycles to fatigue but they can eventually fail. If the strips are undersized or crappy material, they will heat more and fail sooner. Properly sized nickel strips have a pretty good track record. There are other things you can do to prevent fatigue failures, like having a flexible section somewhere that accommodates the expansion without creating high stresses.
 
Most folks will know that I'm a fan of the pouch cell and the convenience that the tabs offer for interconnects.

But 18650 cells are popular and they have good energy density, so if you're going to build a pack with them, you might as well do the best you can. As the OP has noted - the overall resistance of the interconnect path through the cells and out to the load is critical.

Nickel is a lousy conductor, and spotwelded nickel to stainless steel is worse. Thicker nickel is harder to weld, and you risk damaging the cells by trying to pump more joules into the weld.

Trouble is, ultrasonic wirebonding is out of reach for pretty much all of us. So what's the solution? Do the best you can with what you got I suppose. Use pure nickel (not nickel plated steel) and make your welds solid. Use cell holders that keep them from touching each other, and make sure the currents being pulled from the pack are no more than the cells specs (preferably following the half-and-half rule; whatever the quoted continuous and peak ratings are, halve them).

Or get pouches that deliver more amps for the same Wh/kg - you'll find the results are about the same.
 
fechter said:
There will be expansion/contraction from thermal cycling. It may take hundreds or thousands of cycles to fatigue but they can eventually fail. If the strips are undersized or crappy material, they will heat more and fail sooner. Properly sized nickel strips have a pretty good track record. There are other things you can do to prevent fatigue failures, like having a flexible section somewhere that accommodates the expansion without creating high stresses.

Wow, i never thought of that factor.
i imagine when you weld a battery up, you're creating tension.. a cell might expand a millimeter or two during discharge.. that's enough to make those tiny pin prick welds get stressed.

I think if you overshoot the C rating by a mile, you get less expansion, so the connection is less stressed by that. But then there's the vibration of the vehicle itself you have to contend with..

If you have too small of nickel strips, and not much headroom in the discharge, you've got the battery AND the nickel strips heating up.. and that's a recipe for failure over time.
 
spinningmagnets said:
The reason I am frequently posting about spot-welding techniques is because I realize a lot of home garage pack builders are going to spot-weld, and I hope to help their results be "less bad".

Harm reduction is a great thing. It's kind of a shame people have had so many problems with 18650 packs catching fire and such for silly reasons.

18650 cell construction is a hell of a lot better than our ol' dodgy hobbyking lipo pouches, but it seems like bad pack construction with 18650's makes the battery pack just as dangerous.
 
Matador said:

That looks great for conductivity but bad for brutalizing the cell with heat from soldering.
 
neptronix said:
spinningmagnets said:
The reason I am frequently posting about spot-welding techniques is because I realize a lot of home garage pack builders are going to spot-weld, and I hope to help their results be "less bad".

Harm reduction is a great thing. It's kind of a shame people have had so many problems with 18650 packs catching fire and such for silly reasons.

18650 cell construction is a hell of a lot better than our ol' dodgy hobbyking lipo pouches, but it seems like bad pack construction with 18650's makes the battery pack just as dangerous.

but if you think about it is nothing.

thermal expansion of nickel is 0.00000013 per degree per meter kelvin.
so a 2cm strip would extend 0.0013mm for a 20 degree rise in temperature.
and still, only lowering temperatures would give problems in that regard. something that is easely absorbed by the loose nature of manual spot welding. only if you were welding under tension it could be a factor.


so yes, in theory it is something. but unless you have nickel strips 100ft long in sub zero temperatures it wont be any realistic problem.
the lithum would lose power before the nickel welds even might become an issue.

so unless you are planning to use an ebike at the north pole i dont forsee any issues.

Chalo said:
That looks great for conductivity but bad for brutalizing the cell with heat from soldering.

my thoughts exactly.
 
Chalo said:
Matador said:

That looks great for conductivity but bad for brutalizing the cell with heat from soldering.

Totally agree. But this pack costs less than a spotwelder itself would cost me. 107 USD for 60 cells.
14S4P Battery (439.5 Wh 52V 8.6Ah LG MF1 cells from 3 new old stock hoverboard battery packs..

I pre-tinned the cells using generous amounts of Rosin Flux to minimize time applying the iron.

OF COURSE, I used Thick cardboard insulating rings on the positive terminals, to avoid very predictable short circuit from the shrinkwrap potentially melting under heat, thus exposing the negative edge of cell to copper connecting the positive terminals and creating a massive short ciruit and hard to control fire/explosion. Hence why I soldered near 0% state of charge in case of accidental short.

Then I drilled the copper with small clusters of holes and applied rosin flux to the drilled areas.

Then I soldered the copper with solder which permated through the drilled holes to join the solder already on the cells.

More detail in pictures of how I soldered them with my 120W soldering iron here if your curious : https://endless-sphere.com/forums/viewtopic.php?f=2&t=93576&start=75#p1414963

Still very brutal indeed. But for now the capacity does not seem to be affected. I bottom balanced and hooked a BMS

Assembled discharged groups then almost fully charged to 93% (before BMS has a chance to kick in for the first time) were:
01 - 3.23 --> 4.06
02 - 3.23 --> 4.06
03 - 3.24 --> 4.08
04 - 3.23 --> 4.06
05 - 3.23 --> 4.06
06 - 3.23 --> 4.06
07 - 3.23 --> 4.07
08 - 3.24 --> 4.06
09 - 3.23 --> 4.07
10 - 3.23 --> 4.06
11 - 3.24 --> 4.07
12 - 3.24 --> 4.07
13 - 3.24 --> 4.06
14 - 3.23 --> 4.05
Measured with a crappy Mastercraft DMM with an almost dead 9V battery...

So I've come to ask myseflt this question: Should I really invest in a spotwelder, to weld suboptimal nickel strips...
In practive, I wonder if poor/questionnable nickel spotwelding would not be worst that equal current share privided by soldered copper , even if soldering heat can harm...

I had a lot of pleasure soldering copper. I would spotweld copper if it could be done reliably. But for nickel... seems like crap...
I want the total amount of connectors in a pack I build to be equal or less than on hundredth of the internal resistant of one single cell... That's my rule of thumb

Matador
 
Matador said:
Chalo said:
Matador said:

That looks great for conductivity but bad for brutalizing the cell with heat from soldering.

I pre-tinned the cells using generous amounts of Rosin Flux to minimize time applying the iron.

OF COURSE, I used Thick cardboard insulating rings on the positive terminals, to avoid very predictable short circuit from the shrinkwrap potentially melting under heat, thus exposing the negative edge of cell to copper connecting the positive terminals and creating a massive short ciruit and hard to control fire/explosion. Hence why I soldered near 0% state of charge in case of accidental short.

Then I drilled the copper with small clusters of holes and applied rosin flux to the drilled areas.

Then I soldered the copper with solder which permated through the drilled holes to join the solder already on the cells.

More detail in pictures of how I soldered them with my 120W soldering iron here if your curious : https://endless-sphere.com/forums/viewtopic.php?f=2&t=93576&start=75#p1414963

Still very brutal indeed. But for now the capacity does not seem to be affected. I bottom balanced and hooked a BMS

I had a lot of pleasure soldering copper. I would spotweld copper if it could be done reliably. But for nickel... seems like crap...
I want the total amount of connectors in a pack I build to be equal or less than on hundredth of the internal resistant of one single cell... That's my rule of thumb

Matador

This is very much what I intend to do in assembling an unusually shaped custom job. I'll also be testing a bit on some scrap cells to determine best flux, solder, and tinning coats to minimize that "brutality". AND, I'll tin my copper before applying, to minimize heated contact time between strips and cells even with the strips permeated with "solder holes",... I want good solid and fast contacts! Still worry a bit of movement and cold joints though. Jus always in the back of my mind in such critical work.

My experience spans everything from various soldered metal "boxes" and sheet work with furnace heated copper irons, to delicate board work before the days of wave-soldered systems. Todays technology is SO much improved, but I'm not investing in a welder for jus a couple personal "one-off" projects when I'd see a better return and greater use of updated soldering equipment. Would be different if I were building for others or in greater capacity and numbers.
 
Low melting point solder, dudes. Sometimes I use pure indium, MP about 300F. Its electrical and thermal conductivities are much better than lead-tin solder. Sometimes I use indium-bismuth alloy, MP about 165F. Both are expensive, but a little goes a long way.
 
Chalo said:
Low melting point solder, dudes. Sometimes I use pure indium, MP about 300F. Its electrical and thermal conductivities are much better than lead-tin solder. Sometimes I use indium-bismuth alloy, MP about 165F. Both are expensive, but a little goes a long way.

So if i were to pull some serious amps on a hot day all the solder joints come apart.

Yup, solid advice.
 
I see your point about getting a battery pack hot on hard discharges, but that is also on my list of things to avoid ahead of time, by using design choices.

No cell, whether 18650 or flat pouch, should ever reach 140F. So, 300F solder might be a good choice or a bad choice, however if your pack reaches 300F you would have other major problems to contend with, and the type of solder is not one of them.

Edit: I would be satisfied if my pack got warm to the touch at its data-logged max, maybe 100F?
 
flippy said:
Chalo said:
Low melting point solder, dudes. Sometimes I use pure indium, MP about 300F. Its electrical and thermal conductivities are much better than lead-tin solder. Sometimes I use indium-bismuth alloy, MP about 165F. Both are expensive, but a little goes a long way.

So if i were to pull some serious amps on a hot day all the solder joints come apart.

Yup, solid advice.

A relatively large, soft, semi-liquid contact might actually carry a greater current than those tiny poorly bonded welded contacts depicted in the photos of the OP. As I noted previously, the technology available today in regards to soldering, is MUCH improved over my lifetime. While coppers have always been easy to work and identify, many other metals and alloys aren't quite the same and have seen some GREAT technological changes. Even identifying the qualities of the various steels in 18650 constructions may be a bit more elusive, and adhesion to these are most important. Flux, cleaners, AND solder choices for these materials should considered more closely than the previous choices of jus a glob of lead alloys and rosin dabs.

By chance would anyone know the specifics of those materials of cans and buttons in the 18650 housings??? Steel grade?? Surface Treatments?? Anything???
 
DRMousseau said:
A relatively large, soft, semi-liquid contact might actually carry a greater current than those tiny poorly bonded welded contacts depicted in the photos of the OP. As I noted previously, the technology available today in regards to soldering, is MUCH improved over my lifetime. While coppers have always been easy to work and identify, many other metals and alloys aren't quite the same and have seen some GREAT technological changes. Even identifying the qualities of the various steels in 18650 constructions may be a bit more elusive, and adhesion to these are most important. Flux, cleaners, AND solder choices for these materials should considered more closely than the previous choices of jus a glob of lead alloys and rosin dabs.
By chance would anyone know the specifics of those materials of cans and buttons in the 18650 housings??? Steel grade?? Surface Treatments?? Anything???

the pictures i posted were done by an amateur with shit materials who clearly had no idea what he was doing and are in no way a representation of a proper spot welding job. a proper welding job is superior to solder in almost any regard. first reason being that the welds will never let go during heating and will stay in place until the melting point of the metal is reached.

and a material with a low melting point loses its structural strength long before it melts. it becomes buttery well before the official melting temperature. this is a REALLY bad thing in the case of cells where the entire outer shell and the rolled top edge of the cell is the negative. as soon as that material becomes soft and is squeezed or shocked into the pocket between the negative edge and the positive top you have a massive dead short on your hands that only the internal cell's last line of defence (the PTC plate) can stop. but at that point your battery is probabyl already engulfed in smoke and burning plastic.

the 18650 cans are usually nickel plated steel. the quality and thickness of the nickel and steel is up to the manufacuter. usually panasonic and samsung have the best quality in this regard.
 
flippy said:
DRMousseau said:
A relatively large, soft, semi-liquid contact might actually carry a greater current than those tiny poorly bonded welded contacts depicted in the photos of the OP. As I noted previously, the technology available today in regards to soldering, is MUCH improved over my lifetime. While coppers have always been easy to work and identify, many other metals and alloys aren't quite the same and have seen some GREAT technological changes. Even identifying the qualities of the various steels in 18650 constructions may be a bit more elusive, and adhesion to these are most important. Flux, cleaners, AND solder choices for these materials should considered more closely than the previous choices of jus a glob of lead alloys and rosin dabs.
By chance would anyone know the specifics of those materials of cans and buttons in the 18650 housings??? Steel grade?? Surface Treatments?? Anything???

the pictures i posted were done by an amateur with shit materials who clearly had no idea what he was doing and are in no way a representation of a proper spot welding job. a proper welding job is superior to solder in almost any regard. first reason being that the welds will never let go during heating and will stay in place until the melting point of the metal is reached.

and a material with a low melting point loses its structural strength long before it melts[b]. it becomes buttery well before the official melting temperature. this is a REALLY bad thing in the case of cells where the entire outer shell and the rolled top edge of the cell is the negative. as soon as that material becomes soft and is squeezed or shocked into the pocket between the negative edge and the positive top you have a massive dead short on your hands that only the internal cell's last line of defence (the PTC plate) can stop. but at that point your battery is probabyl already engulfed in smoke and burning plastic.
[/b]
the 18650 cans are usually nickel plated steel. the quality and thickness of the nickel and steel is up to the manufacuter. usually panasonic and samsung have the best quality in this regard.

With all due respect. I think these consideration are not really relevant.
I used Sn60Pb40 solder, which has a melting point of 188°C (370°F).
There is absolutely no way I will ever bring these cells higher than 50°C during their cycle life, even if I pull very hard on them.
If you think about it one second, if these 18650 cells were volontarly brough to such high temperatures, they would probably vent (aka, go into thermal runway) waaaayyy before the solder would even start to melt.

So I think the idea that the solder could melt under high discharge rate is not a real life problem. It's irrelevant.
 
Matador said:
So I think the idea that the solder could melt under high discharge rate is not a real life problem. It's irrelevant.

Yep - I used soldered busbars on Voltron and we were pushing 28 C peaks, or 280 amps from the 2p pack. Nothing melts, but the cells get warm.
 
Matador said:
With all due respect. I think these consideration are not really relevant.
I used Sn60Pb40 solder, which has a melting point of 188°C (370°F).
There is absolutely no way I will ever bring these cells higher than 50°C during their cycle life, even if I pull very hard on them.
If you think about it one second, if these 18650 cells were volontarly brough to such high temperatures, they would probably vent (aka, go into thermal runway) waaaayyy before the solder would even start to melt.

So I think the idea that the solder could melt under high discharge rate is not a real life problem. It's irrelevant.

my comment is about the indium-bismuth alloy Chalo uses with a melting temperature of 70c
 
Back
Top