Spot welding fuses to individual 18650 cells?

ok I accept that. But I still don't see the point in welding fuse wire to the cells. It's actually very rare that a single cell would 'blow'... Almost all battery DIY major fails are related to mechanical failure, and by this I mean all types of physical damage - from wiring rubbing or some type of chafing, and connections overheating from poor current capability - possibly leading to the melting of battery or wiring insulation and even melting solder.

Build a solid pack, with good current capability through all connections... and pack it properly. Eliminate weak links, and keep it as simple as possible is my thoughts.
 
You both make excellent points. It sounds like I have a lot of testing and experimenting to do. The final battery pack for my go kart will likely be two modules of 20S10P connected in parallel using brand new Samsung INR18650-25R 2500 mAh cells. Maximum amps (short burst for a few seconds) can be as high as 150 amps so the 20 cells in parallel could get as much as 7.5 amps. The Samsung INR cells can put out 20 amps continuous but the fuse wire would have to be sufficient to handle this much current. If I use fuse wires (still undecided) I am thinking I would need 22 awg based on the table above but such thick wires would be more difficult to spot weld without upgraded equipment (which I am willing to do). The nickel plated copper bus bars look VERY interesting since I have experience in spot welding pure nickel strips. Either way, I sincerely appreciate all the great feedback and information. You guys are great. Any additional suggestions or ideas would be greatly appreciated.
 
RAD RAD RAD

rad rad rad, rad rad

Thread.

....

Now I remember why I spent so much time on Endless Sphere.... :)

...

It was not all that long ago that this technique was "SUPER SECRET SAUCE" :shock:

...

Little tiny fusable links FTW

Love the smashed copper pipe idea*
(My favorite is mapping needs to other isles in the store... who knew... plumbing for battery bus bars) !!

-methods
 
They smashed copper pipe idea has really worked well. It is also much cheaper than buying customized copper busbars off the Internet. It takes a while to cut the small grooves in the busbars for use with the fuse wire. I’ve used a dremel and can now make fairly straight/consistent grooves. I’ve gotten fairly quick at the process. Attached is a photo of an 18650 pack I created in the last few days using solid copper and bus bars, individually fused cells, and a 300 W inverter (not shown) that I’m using to power a small load (a couple of lightbulbs). The ”voltage drop” seems reasonable but I will have to see how it performs under a much heavier load. Thanks again to everyone for the feedback.
F79ED8CD-FEB1-4FCB-A0E5-8601FB65D8E9.jpeg
 

Attachments

  • 99E22ADD-0183-4A90-9C61-5D656CF0CE14.jpeg
    99E22ADD-0183-4A90-9C61-5D656CF0CE14.jpeg
    79.8 KB · Views: 2,182
  • 32E1410E-199D-4307-A22A-C101CD1F18C6.jpeg
    32E1410E-199D-4307-A22A-C101CD1F18C6.jpeg
    62.2 KB · Views: 2,182
  • 57A0638F-5675-4705-AC35-BCA2CE959AB2.jpeg
    57A0638F-5675-4705-AC35-BCA2CE959AB2.jpeg
    83.3 KB · Views: 2,183
If you measure the voltage across one of the fuse wires while the load is running and you know the load current, you can use that measurement and calculate what it would be at a higher load.
 
One thing that you might run into, where it comes to individual cell fusing, is the final bus pick-off point connection.

In a pack or sub-assembly, where fuses form the links, there will be one end of the pack requiring an independent supporting conductor, to mate with the final battery cell fuses and carry the buss terminal connection. This can add to the real estate needed and will present a specific mechanical and bonding issue, not necessarily benefiting from the mechanical forms and assembly techniques used for all preceding work.

Something to keep in mind while in the planning stages.

RL
 
fechter said:
If you measure the voltage across one of the fuse wires while the load is running and you know the load current, you can use that measurement and calculate what it would be at a higher load.

Thanks. Will run some test this weekend. I bought an inexpensive watt meter so hopefully it will do the trick.
 
legg said:
In a pack or sub-assembly, where fuses form the links, there will be one end of the pack requiring an independent supporting conductor, to mate with the final battery cell fuses and carry the buss terminal connection.

Yes. I added bus bars along side the positive (red arrow) and negative (yellow arrow) side of the pack.
I connected the main negative and positive wires (coming from the BMS) to these same bus bars. Probably a little late to be asking this question but “ is the the right way to do it?”
 

Attachments

  • BC2E7E45-6B98-4311-B080-0E8085FA5202.jpeg
    BC2E7E45-6B98-4311-B080-0E8085FA5202.jpeg
    168.3 KB · Views: 2,158
By the way you cannot see it from the photo, but there are two other busbars on the underside of the pack which are used to complete the “4S” connection.
 
Here is a photo showing the positive and negative wires coming from the BMS and attaching to the positive and negative BuSbar terminals on the battery pack. Any suggested design improvements would be greatly appreciated. Thanks again for all the information.
 
I have read a number of threads regarding the use of cell-by-cell fusing. I honestly can't see the evidence that this is required.
I work in the medical field and the Lithium Packs we use (Mostly 6S4P and 7S3P packs) have only two thermal fuses built into the packs, apart from the over current protection from the BMS. Scrapped versions of these packs are where I get my supplies of 18650s to build my own packs from.
These packs are professional and in use regularly in hospitals. Max discharge in these packs is around 15A, so not up there with most E-Bike peak discharge rates, the charge current is in the 2-4Amp range - unattended.
The charge/idle storage of the batteries is what everyone seems to be concerned over, and then their house burning down. Unless the constructor of the battery pack makes an uninformed and/or stupid mistake, I can't see it happening.
 
I have to admit that I have really struggled with the decision of whether or not to use individual fuses versus pure nickel strips. Plenty of smart people on both sides of the argument have made very compelling arguments. It seems to come down to a real or perceived issue of safety regarding potential thermal runaway and the very real issue of resistance that comes from using pure nickel strips or from using individual cell fuses. I’ve learned a lot by listening to this debate and I really appreciate everyone’s feedback.
 
For sure there are arguments for either style. Tesla opted for fuse wires, so they must have thought it was a good idea.
Your bus bars look great. The resistance of the heavy copper will be low enough to force good current sharing.

The only thing I would worry about is the fuse wires getting damaged from impact or vibration. If the cells are firmly held by the plastic holders and the housing prevents the wires from getting smashed, it should be good.
 
fechter said:
The only thing I would worry about is the fuse wires getting damaged from impact or vibration. If the cells are firmly held by the plastic holders and the housing prevents the wires from getting smashed, it should be good.
That is a very good point. That’s the next issue I’m going to address. You are correct in that the fuses could pop loose very easily with the significant vibrations that would occur on a go kart like this.
 
This is the first and only 18650 build that I have actually "liked" and wanted to emulate. :idea:

...

I was unaware of the strip vs fuse debate
To me - I would come to a very quick conclusion

...

I do not have a tab welder here
I wont have a tab welder on the side of the road
I do not wish to sink a bunch of heat into my cells soldering a big thermal mass...

...

It makes REALLY GOOD SENSE

REALLY GOOD...

To affix the lowest thermal mass possible
To allow for the least heating and impact on the cell
To allow the simplest tools to be used
To allow the easiest rework and up-cycling...

....

I mean... those tiny wires are probably good for 10A
The cell is hardly good for 10A... burst... but not continuous...
If you tear into a house you will find MANY wires that run HOT
Hot wires... for burst use... is NORMAL
(Folks dont know this)

... The actual power lost in the wires is below trivial ...

So thats my non-fused feeling about it... (ignoring the fusing aspect)
THEN
When we add in the fact that we get CELL By cELL fusing

Eh - for me - that is the obvious best bet.

...

As for Tesla -
You must consider that a major part of their decision was the PROCESS
How fast is it?
How well does it scale?
How well does it automate?

... They very well could have settled on the fusable link style simply because it improved thruput in production and rework*

...
Looking at this members build
Looking at how cheap the Amazon link is
Looking at the tools I have sitting on my desk...

I am inspired to duplicate his work!

Because... it can be done... and it can be un-done

The only thing I would change is:
1) Minimize glue
2) Avoid HOT GLUE at any cost and at ALL TIMES (it WILL reflow on you... do NOT trust it... even the high temp!)
3) Try to use mechanical methods which can be un-done - and redone - like 8" heat shrink

... Ideally we would mount these into an extrusion
Fill the extrusion with a type of expanding foam
The foam would hold...

The foam would have a solvant
One could un-pot the unit if and where needed
(this is much harder than it sounds - but could be done)

-methods
 
Here is a process suggestion which will save you a great deal of time.
This is not conjecture - we have done it - on the 420 controller run.

>...
Looking at the build I can see some time was taken to get a good blob of solder on the bus bars.

I would go about it in the following way:

1) Procure a Solder Pot
.
.
solderPot.jpg
.
.

2) Procure solder bars and some very high quality flux (not all flux is the same...)

3) Flux completely and then drag (using gloves or tongs of course) the buss bar across the solder pot (pre-heat only if required)

... Starting there... now you have your buss bar not only tinned... but protected... as raw copper will turn black and green. The solder covering it can be sacrificial... but more importantly... once you tin your fuse wires... they will adhere near instantly ... especially with a pre-heated bar...

4) Tin the fuses...

Since doing them one at a time would be silly -
Spec a fuse wire that is a gauge under
Do the same as the above
Get two guys... and "drag" the fuse wire thru the solder pot after fluxing it
Completely coat the fuse wire... so it is tinned... you can do linear feet per second this way

5) Snip up the fuse wire into the desired lengths

...

At this point
With simply a momentary touch of the soldering iron
You shoudl be able to mate the fuses to the buss bar FAST - like real fast - china fast - bap bap bap bap

...

Then you can push each of them down quickly
With your fingers (or better a jig - that sets a pinch or bend)

Then you can quickly spot solder to the cells.

...

I do not know how long your original process was - but I suspect - that tinning both the fuse wire - and the buss bar - will ease a lot of the pain you may or may not be having.

... This is how I would move that forward in production. I would drop the machining part, cut the fuse wires long, allow for slop...

-methods
 
P.S. I am super stoked on your build and documentation

Please dont take any suggestion as criticism
I think the way you are doing it is rad - and I will likely just copy your method
Just trying to suggest a way to scale it... as that is how my mind works now

-methods

Edit: And PS... I use the hot glue gun too :mrgreen: (it has just burned me - literally and figuratively - more than once)
 
Thanks for the great ideas. :D I really appreciate it. I will give it a shot. One question..... Do you think coating the entire bus bar in solder will increase the resistance of the pack? I like the fact that it would speed up the process and (more importantly) improve the consistency of how well each individually fused cell wire is connected. Thanks again for taking the time to make those suggestions. I will give it a shot and report back. I love the solder pot idea. Never even thought of that one.
 
Addy said:
garolittle said:
One question..... Do you think coating the entire bus bar in solder will increase the resistance of the pack?

No, it would actually slightly decrease the resistance of the bus bars because you're adding more conductive material to them.

Thanks for the response. I am very curious about this.
The solder I use has a 60/40 Tin/Lead ratio (see link below). Since both tin and lead have higher resistance than copper, how can the addition of this solder actually decrease the resistance of the bus bars? I certainly hope your comment is true because I am always looking for ways to reduce resistance.

https://www.amazon.com/dp/B073LT2PKB/ref=cm_sw_r_cp_api_0o8rBbG43BWH1

Thanks again for responding. :D
 
garolittle said:
Thanks for the response. I am very curious about this.
The solder I use has a 60/40 Tin/Lead ratio (see link below). Since both tin and lead have higher resistance than copper, how can the addition of this solder actually decrease the resistance of the bus bars? I certainly hope your comment is true because I am always looking for ways to reduce resistance.

https://www.amazon.com/dp/B073LT2PKB/ref=cm_sw_r_cp_api_0o8rBbG43BWH1

Thanks again for responding. :D

No problem! When you're adding solder over the bus bar what you're doing is like adding some solder in parallel with the copper. When you have resistors in parallel, the combined resistance is lowered.

https://www.electronics-tutorials.ws/resistor/res_4.html
 
I have so much to learn. That is so cool. So copper bus bars like this .....

25DC76BA-4C2A-4B24-B6C0-1DE60E66A472.jpeg

completely coated in solder is actually more conductive than just the copper alone. Cool. 8)

I am about to order a solder pot. Any brand recommendations for one that will make the coating process easier? Thanks again!
 
The oxide layers must be broken through or else it can affect the integrity of the spot weld , this is done by the force of the spot weld electrode pushing down on the strip/wire and battery end cap (more force is better). (and perhaps counter-intuitively, the smoother the two surfaces being spot welded, the better)

If the technique used is one electrode on the wire and one electrode on the end cap then it is essential that the electrode lengths are adjusted to ensure adequate force on the wire allowing for the inevitable deformation of the wire when the electrode pushes down on it, this is a bit trial and error.

Its also critical that the force applied is kept constant before during and after the weld, the single electrode on the wire technique tends to achieve this better than both electrodes on the fuse-wire (where the fuse-wire tends to 'roll' a bit)

The reason I mention this is that the weld may look good but could contain an oxide layer which will increase joint resistance.

I haven't seen it mentioned , it is important that the end-cap and wire is clean (no grease) , I tend to use iso-propyl alcohol (on lint free) because its quick and cheap !
 
asterysk said:
The oxide layers must be broken through or else it can affect the integrity of the spot weld , this is done by the force of the spot weld electrode pushing down on the strip/wire and battery end cap (more force is better). (and perhaps counter-intuitively, the smoother the two surfaces being spot welded, the better)

Agreed. I will use the techniques shown here..... https://youtu.be/H_b3OJuQpEE

If the technique used is one electrode on the wire and one electrode on the end cap then it is essential that the electrode lengths are adjusted to ensure adequate force on the wire allowing for the inevitable deformation of the wire when the electrode pushes down on it, this is a bit trial and error.

I had not thought about this but I really appreciate the information. It make sense that the electrode length needs to be adjusted so I will practice in order to get it right.

Its also critical that the force applied is kept constant before during and after the weld, the single electrode on the wire technique tends to achieve this better than both electrodes on the fuse-wire (where the fuse-wire tends to 'roll' a bit)

This appears to be the same approach in the YouTube video listed above and it make sense.
 
Addy said:
garolittle said:
Thanks for the response. I am very curious about this.
The solder I use has a 60/40 Tin/Lead ratio (see link below). Since both tin and lead have higher resistance than copper, how can the addition of this solder actually decrease the resistance of the bus bars? I certainly hope your comment is true because I am always looking for ways to reduce resistance.

https://www.amazon.com/dp/B073LT2PKB/ref=cm_sw_r_cp_api_0o8rBbG43BWH1

Thanks again for responding. :D

No problem! When you're adding solder over the bus bar what you're doing is like adding some solder in parallel with the copper. When you have resistors in parallel, the combined resistance is lowered.

https://www.electronics-tutorials.ws/resistor/res_4.html

That isn't exactly how it works. Adding resistors in parallel isn't actually lowering the resistance, it's adding another path for the current.

If you have a 5V source running through a 1Ohm resistor, the current through the resistor is 5A. If you add a 2nd 1Ohm resistor in parallel, you now have 2 resistors, each one has 5A through it, so 10A. The effect is like using a single 500mOhm resistor.

If you have a 1Ohm resistor, and a 10Ohm resistor, the total current would be 5.5A, the same as using a single 909mOhm resistor.

But plating a bus bar in solder won't work that way. You will be increasing its resistance, since it will act more like resistors in series. At the same time, you will also be protecting the bus bar from corrosion. So it's a trade off. The added resistance isn't that large, since the plating is usually very thin, like thousandths of an inch.

IMO, you would be better off buying a tin plating kit (and learning how to electroplate at home, it isn't difficult), than melting down solder, or silver plating if you are willing to spend the money. Silver will decrease resistance versus a tin plated copper bus bar, since silver has a lower resistance than copper. Both metals increase solderability.

Or you could just buy tin plated busbars in the first place.
 
Back
Top