Spot Welding Copper Strips to 18650 Battery Cells

mkp007

100 mW
Joined
Oct 28, 2016
Messages
48
Location
San Diego, CA
*
Nickel foil is the popular choice for joining cells together using spot welding.


The reason why it works so well is a combination of material properties such as electrical conductivity, thermal conductivity and melting temperature.

The thickness of the foil can be too thick in which the weld does not occur. For example, the 1.9kW Sunkko 709A spot welder has a max thickness of 0.3mm for the Fixed Copper Head and for pure nickel strip the max is 0.15 mm. My guess is the 0.3mm is for Ni plated steel which is likely easier to spot weld.

As for copper foil, they say it does not work. Why? Well, for one, the high electrical conductivity prevents the current from penetrating into the cell and creating the necessary heat to create a weld. Also, since the spot welder electrodes are made from copper, they may get too hot and weld to the foil prior to a weld at the cell-foil interface.Basically, the combination of lower thermal conductivity and higher electrical resistivity of the Nickel and Steel makes it easier to spot weld.

So is that the end of the story? Not exactly, there is a spot welder that claims 0.02" 110 copper foil. The 250i^2 EV by Orion (nice find Matador). It reminds me of a TIG welder, and indeed, upon reading the manual below, it employs tungsten electrodes and a shielding gas. Tungsten melting point is 3X more than copper. Makes me think about using tungsten electrodes on the cheaper Sunkko 709A spot welder. Anyways, check out some of these videos. Pretty darn cool.

http://orionwelders.com/wordpress/media/250i_usermanual_web.pdf

[youtube]tYvCZSNWNLk[/youtube]

I love the use of CNC for automated spot welding.

[youtube]weiBnbuKsYk[/youtube]

Here is 10mil copper foil.
[youtube]tqhJyy2hUdw[/youtube]


Ok, so spot welding copper strips to 18650 batteries is possible. Now what? A few things come to mind:
1) What is the cost of that welder?
2) Could retrofitting cheaper welders with tungsten electrodes make a cheap solution for DIYers?
3) Does corrosion of copper over time could weaken the weld?
4) Does plating the copper sheet with Nickel help solve issue #3 (ref Matador posts)?

Thoughts?

xxxxxxxxxxxxxxxxxxxxxxxxx

moderator edit: here is a link to copper sheet thicknesses (found on page 6 of this thread)
https://endless-sphere.com/forums/viewtopic.php?f=14&t=84680&p=1303321#p1303321
 
This is intriguing. I have seen some nickel strips that had a thin slot right over the end of the cell. I have been told that this creates a situation where the current passing through the nickel strip has a shorter distance to travel if most of the current passes through the metal end of the cell tip (instead of most of the current passing though the nickel strip from one welding electrode to the other, with very little current passing through the cell-tip)

Perhaps a thin slot in these copper strips would also help with the welding issues?

rBVaHFTHsRaASHqOAAXU8SF_L_w155.jpg
 
That is an awesome machine. I'm gessing at the very least 10 000 USD, it might be more...
If a reputable battery seller was building batteries with this specialized welder, guenine quality cells and thick beefy cupper-strips,
I'd definitely buy from them. Well worth paying a little more extra for that quality in my opinion.

Let's face it, motors and controllers are becoming more and more powerfull. Nickel strips aren't going to be enough anymore to handle these kind of power ranges.
Even my BBSHD would benefit from such a build.
Copper means less heat.
Less heat means longer cell life in the long run,
That can make a difference, especially if I'm pulling 30amps on just 4 parrallel cells.

Awesome thread ! Thank for creating it mkp007
 
Cheking on the net, Orion Pulse 250i EV is 12500 USD
The Orion Pulse 150i EV is 7500 USD.

Pretty sure somebody could reverse engeneer that thing... It probably does'nt need to be that fancy to be effective.
I could use it without the microscope oculars and the robot arm and the computer...
The really essential part is the electrodes and the power supply
 
Hillhater said:
spinningmagnets said:
This is intriguing. I have seen some nickel strips that had a thin slot right over the end of the cell.......

Yes many of the commercial packs i have seen are also have "split end " strips ....like this one from the previous thread..

IMG_20150915_152724_zpslsy3y2i4.jpg

Intresting indeed !
 
spinningmagnets said:
... some nickel strips that had a thin slot right over the end of the cell.....

That is interesting. I guess it forces the current to go down through the cell-foil interface. I like that idea. Thinking about creating a thin slit with a razor blade.

A note on the Orion welder, I guess this is the same company as Sun Stone Welders. The latter having a more up-to-date website. I have requested pricing on their 250i2 micro TIG system. They have some cool products but I bet they are $$$.

http://sunstonewelders.com/applications/batteries/ev-battery-welding/


Edit: There is a used 250i2 on Ebay for $10k. So, yes, expensive.
http://www.ebay.com/itm/Orion-250i-Pulse-Arc-Welder-/272408072771

Matador said:
...Pretty sure somebody could reverse engeneer that thing...
Matador, the technology is similar to a TIG welder. However, in TIG welding, you usually have the current go through the part. In this case, it is going from the electrode to the outer copper part. So a bit tricky.

I think the resistance welders (spot welding) is the workable technology. With the split, the Ni plating and tungsten electrodes, there might be a cheap and easy solution. :wink:






**
 
That orion welder can weld 0.020 inch (0.508 mm thickness copper strips) !
Imagine welding 10 mm width x 0.508 mm thickness copper strips. That would be 5.08 mm2 surface area --> Equivalent to around 10AWG copper wire (3.31 milliOhms per meter) !

When we weld the typical 10 mm x 0.15 mm Nickel strips, it's 1.5 mm2 in surface area --> Equivalent in surface area to around 16 AWG, but wait it's not copper... there's 4 times more resistance with Nickel... So instead of being like 16 AWG copper wire, if you think in term of resistance, it compare more to 22 AWG copper wire : That Nickel strip has a resistance of 46.6 milliOhms per meter.

If you check ampacity charts it's easy to understand. The 10 x 0.508 mm copper strip (equiv to 10 AWG copper wire ) can pass 55 amps for a given voltage drop (voltage drop means power in heat dissipation).
But the common 10 x 0.15 mm nickel strip (equiv to 22 AWG copper wire) can only pass 7 amps for that same voltage drop.

55 amps !!! Now imaging a pack with multiple copper stips layers combined with multiple seris connection for each cell in a parallel row. That could provide insane amperage. The limit now would actually be the limit of the 18650 cells themselves. We could build very high power packs for electrical motorcycles or electric cars.

I think copper is the way to go with high amp 18650 like the ones that provide 30A per cell.
 
mkp007 said:
With the split, the Ni plating and tungsten electrodes, there might be a cheap and easy solution. :wink:
**
That would be awesome !
I don't know if a 1500 Watt rewounded microwave oven transformer with one turn of 2/0 AWG copper wire as the secondary coil, a pair of tungsten electrodes and an Omron HC3R-A Solid-State timer (have one on hand) would do it ?

I'l have to find some 2/0 gauge wire, a pair of tungsten electrodes and some more parts...

Just like this : https://www.youtube.com/watch?v=DnOGyFideQc&index=6&list=PL7GtCwVvhYqjotqpgMZPslmgzRlxg3obV
But with tungsten electrodes instead of copper ones
And bigger wires, even bigger than this : https://www.youtube.com/watch?v=LrI0TT28SU4&list=PL7GtCwVvhYqjotqpgMZPslmgzRlxg3obV&index=2
 
Hillhater said:
Matador said:
That orion welder can weld 0.0020 inch (0.508 mm thickness copper strips) !
.
Errr ?..No, this time i think you mean 0.05mm thick
...unless you meant to say 0.020 inch ?

Yes ! I meant 0.020 inch (0.508 mm) not 0.0020 inch.
My bad :oops: . Let me correct my previous post by edditing :mrgreen:

I know 0.020" really sounds too good to be true, but that's really what they claim in their video.
Look at 30 seconds in the first video posted on this thread !
 
I just want to double check my understanding of proper current-flow in a pack. In a high-current pack, the nickel strips that connect the paralleled sub-packs can still be 0.020" thick, but it is the series connections between the paralleled sub-packs that must carry the full peak currents?

Meaning: only the series connections need to be much thicker? (also, I assume there should be several thick strips making the series connection?)
 
I understand the concern if copper corrodes when exposed to air, and I was intrigued to hear there have been some experiments in nickel-plating thick copper (for the series connections?). Let us imagine (for the sake of argument) that we have theoretical battery pack that is rectangular in overall shape, using 6P straight inline cell-groups that have been spot-welded using 0.20mm thick pure nickel strips. (edit: inch to mm corrected)

If we want high current from this pack, the series connections are thick copper bars, three between each paralleled 6P sub-pack. Located in-between every two cells, for intuitive current-sharing between each six-cell P-group.

Since a "bump" in size is cheap...we experiment by going for copper bars that are clearly thicker than required, so they run cooler than our "worst case scenario". What if these copper "series connection" bars are not nickel-plated? What if we sand the connecting part down to bare metal, and then immediately solder the copper-bar connections onto the nickel parallel plates? I'm sure the copper series-bars can be coated with "liquid tape", or some similar sealant (afterwards), but...what if a thin film of green oxidation formed on the outside of the thick copper bars? The soldered connection between the copper bar and the nickel plates that form the paralleled sub-packs would still be sealed, yes?
 
My DIY transformer based spotwelder does .1mm copper. Slots make it heaps easier. I laid down nickel parallel strips first then welder .1mm copper to this. I know others have done successful welding with copper. Pretty sure the JP welder will do .1 copper. Heaps of info in the nobuo's repository.
 
While I'm here I may as well devulge some of the info I've been gathering on nickel alternatives. Obviously nickels conductivity sucks ~25% IACS ( international annealed copper std for conductivity), and pure copper (98-105%iacs) is hard to weld bc it doesn't generate much heat and conducts it away.
Well according to the copper development assoc there are 871 recognised alloys of copper. You name it- it's there.any conductivity from 4-110%IACS, all tempers, weldability etc... So now I'm a Google expert in copper alloys.
If you take the selection of copper alloys with iacs of between 50-80% and screen for weldability you get about 15 or so that are right on the money. Good to excellent weldability and 2-3 times conductivity of nickel. ( FYI- pure copper rates as not recommended to spot weld ie below poor)
Sweet I'm thinking... Open a new Google search page type in copper alloy suppliers... Tick tock tick tock. Well 2days later I was yet to find a supplier who could supply less than aprox 20 tonnes.
The alloys I wanted do exist but if it's not one of say20 common commercially used alloys and you only want a small amount then you may s well smelt your own. ( I briefly considered this).
I still have a few vague leads that I didn't follow up on, cause I went cold on the idea. If one could find a low volume supplier of specialty copper alloys this would be really really interesting. I reckon nickel would be dumped pretty quick
Any metallurgists in the field want to share a hot tip on suppliers?
Nickel is easy, but as someone mentioned earlier, batteries are already superceeding the ability of nickel to conduct the cells available power.
Love to hear any thoughts on this matter
K
 
spinningmagnets said:
I just want to double check my understanding of proper current-flow in a pack. In a high-current pack, the nickel strips that connect the paralleled sub-packs can still be 0.020" thick, but it is the series connections between the paralleled sub-packs that must carry the full peak currents? Meaning: only the series connections need to be much thicker? (also, I assume there should be several thick strips making the series connection?)

In a well designed battery pack, every path is analyzed for voltage drop. The goal being to minimize it when possible to and for each cell to see the same resistance.


spinningmagnets said:
Let us imagine (for the sake of argument) that we have theoretical battery pack that is rectangular in overall shape, using 6P straight inline cell-groups that have been spot-welded using 0.20 inch thick pure nickel strips.

I'm pretty sure you meant 0.2mm thick (7.9mils).

spinningmagnets said:
If we want high current from this pack, the series connections are thick copper bars, three between each paralleled 6P sub-pack. Located in-between every two cells, for intuitive current-sharing between each six-cell P-group.

Sounds ok to me. I like that you pick up the current between the two P cells. This way, all the P cells see an equal and low voltage drop. A circuit diagram of a battery pack is a bunch of 3.6V loads connected by resistors. A battery pack designer should create one to be confident in their design. That's what this thread is for https://endless-sphere.com/forums/viewtopic.php?f=14&t=84412.

spinningmagnets said:
What if we sand the connecting part down to bare metal, and then immediately solder the copper-bar connections onto the nickel parallel plates?

Yes, I think this is fine based on eTrike's comment below.

eTrike said:
I live in a fairly wet climate and my main pack uses copper for series and parallel. The packs are half sealed but show no signs of greening in 4+ years in this config. A bigger concern is packs that use steel since that can grow in nasty ways. It is much worse on the coast. Aluminum can grow surprisingly in salt air. Of course solder protects against corrosion as you say. I like no-ox.

eTrike, I'm not sure I like the idea of an electrically conductive grease in my pack. Is this the stuff you use? How do you apply it so it doesn't get all over the place?
https://www.amazon.com/NO-OX-ID-Compound-Electrically-Conductive-Grease/dp/B00HSW341A

One technique that I have used for bond prepping aluminum is to scotch-bryte with a small mixture of acetone+epoxy. This way, as you abrade the surface of the aluminum, oxygen has no way of creating the oxide film as it has a film of acetone+epoxy. The acetone evaporates and the epoxy cures. The same technique could be applied for copper using the no-ox. Just put the no-ox on the scotch-bright and abrade. Wipe the excess but leave a thin film for protection. Two pass is recommended. The first pass is aggressive scotch-bryte only and is wiped clean with acetone. The second pass is a light abrade with scotch-bryte + the no-ox. Excess no-ox is wiped off leaving a thin clean coating.

kdog said:
....Well according to the copper development assoc there are 871 recognised alloys of copper. You name it- it's there.any conductivity from 4-110%IACS, all tempers, weldability etc... So now I'm a Google expert in copper alloys.
If you take the selection of copper alloys with iacs of between 50-80% and screen for weldability you get about 15 or so that are right on the money. Good to excellent weldability and 2-3 times conductivity of nickel. ( FYI- pure copper rates as not recommended to spot weld ie below poor)...

kdog, I think most if not all of the copper we see being welded to cells is the multi-purpose 110. I searched for the weldable 122 variety in foil and it is non-existent (at least with my 10min of google searching). At this point, I still think the most practical solution is Ni plated Copper 110 foil because Copper 110 is available in many thicknesses.

If using the parallel resistance spot welders, you need to make a slit in the foil to force current downwards to the foil-cell interface. The Ni plating may create higher resistance at the weld interface to help generate more heat.

The power of the spot welder is important. You need a high jolt of current. Low power will no work. Matador mentioned a "1500 Watt rewounded microwave oven transformer". The videos seem impressive. How does this compare with the 1.9kW Sunkko 709A for $200 on eBay http://www.ebay.com/itm/112190872309? 500A of current should easily spot weld 0.01" thick copper correct? Yet, they say it doesn't work. I'm going to probably buy this unit unless someone tells me different (hopefully soon :wink: ).
 
'
Exciting update. The following post by ridethelightning and kdog has confirmed the use of a tungsten probe to aid in spot welding copper.

ridethelightning said:
....the key to making it a smooth job i think is a tungsten tipped electrode. this will not melt like a copper one or stick to the tab. it will also create the resistence needed to create the heat for the weld, at the tip, similar to the effect kdog saw when only using light presseure with his copper electrodes....though...kdog knows all of this, hell, i banged on about it multiple times in a fever of excitement from seeing the copper tab actually welded properly with his MOT welder...
https://endless-sphere.com/forums/v...5&hilit=ultimate+repository&start=75#p1104439

:!: From what I have read on welding copper is copper fumes are very toxic. So you need to do this with a fume hood :!:

For TIG welding copper, the following settings are recommended: DC straight polarity, 2% thoriated tungsten and helium/argon shielding gas.

Here is a link to tungsten electrodes that are 3/32" diameter. They are easy to to cut using a metal cutting grinding blade and easy to sharpen to a point with a belt sander. Just keep the grind marks parallel to the electrode. The tip shape is very important. Based on the images below, I think a 90° angle with a 0.02" diameter blunt end would be best.

https://www.amazon.com/Welding-Tungsten-Electrodes-Thoriated-10-Pack/dp/B013KZNLGG

TIG-tung-shape.jpg


arczonetungstenangle.jpg


.
 
mkp007 said:
At this point, I still think the most practical solution is Ni plated Copper 110 foil because Copper 110 is available in many thicknesses.

If using the parallel resistance spot welders, you need to make a slit in the foil to force current downwards to the foil-cell interface. The Ni plating may create higher resistance at the weld interface to help generate more heat.

I analyzed a stripe from a pack from Sony laptop ca. 5-7 yrs old on a SEM/EDX microscope and:

1. They used tin-plated copper.
2. This material easily spot welds on my low power DIY-style MOT spot welder.
3. the stripes have poor fatigue strength (brake easily; one Sony pack I disassembled had perfect condition cells but the spot welds broke)

I asked in a medium company doing metal-plating and they do nickel as well as tin. Whatever I want. Question what would be cheaper in a large scale.
 
Skrzypas said:
I analyzed a stripe from a pack from Sony laptop ca. 5-7 yrs old on a SEM/EDX microscope and:

1. They used tin-plated copper.
2. This material easily spot welds on my low power DIY-style MOT spot welder.
3. the stripes have poor fatigue strength (brake easily; one Sony pack I disassembled had perfect condition cells but the spot welds broke)

I asked in a medium company doing metal-plating and they do nickel as well as tin. Whatever I want. Question what would be cheaper in a large scale.

Interesting. Tin, nickel and silver coatings on copper are quite common. Nickel seems to be better for various reasons. For one, "Tin tends rather easily to form hard, brittle intermetallic phases, which are often undesirable" (ref https://en.wikipedia.org/wiki/Tin).

However, there is this: "Tin vs. Nickel - Nickel is not solderable, so any items that may need to be soldered to the bus bar wouldn’t stick to electroplated nickel. Forming, nickel is not as ductile as tin would be and therefore may crack under stress if bent. Tin is better than nickel as far as conductivity is concerned. Nickel can and will grow a heavy oxide that may prevent conductivity over time as the oxide is more difficult to break through than tin." (ref http://ppc1904.com/resources/faq)

Difficult to make a judgement at this point.
 
Mp007- I've successfully welded up a 20s6p pack with .1mm copper, with a MOT welder ( it's now upgraded to a bigger transformer) so it's def possible. I don't have direct exp but I highly doubt that a cheapy suunko will do copper. Copper wil eat that 500a for breakfast, You'll need over 1200 pref >1800 I'd say. I think mine is putting out 1500ish but that's a guesstimate. 1.9kw is pretty dismal- check the threads on them I think they're good for .15 nickel max. Don't quote me on this though :?
 
I have a 800WS capacitor discharge welder I built which is supposed to be the equivalent of 10000 amps I was told. I reached out to Bryce from Sunstone who set me up with the Glid-cop electrodes I use for nickel, which were the perfect choice for my application. He said I needed tungsten tips,

Hello Tom,

In order to weld Copper, you will need to use Tungsten Electrodes. You can purchase them at the following link:
http://sunstonewelders.com/product/resistance-tungsten/

You also need to be aware that due to how much more conductive copper is, welding it is much more difficult that welding nickel. You will not be able to weld as thick of material when welding copper. Use thinner copper when you can. If you find that you are not making any welds, you can try grinding the electrode tips down to a smaller tip, this will focus the energy more and you will be able to weld through thicker materials this way.

I hope this helps.

Best wishes,

Bryce Bytheway
Welding Consultant

Sunstone Engineering
1693 American Way #5
Payson, UT 84651
Direct: 1-801-658-0015 x222
Office: 1-877-786-9353
Email:bryce.bytheway@sunstoneengineering.com
Web: http://www.sunstoneengineering.com

I'm going to order a set and I'll get back to you guys on how it goes. I also need to find some 110 but in the mean time I might just start with some copper from our local well stocked hobby shop.

Tom
 
kdog said:
Mp007- I've successfully welded up a 20s6p pack with .1mm copper, with a MOT welder ( it's now upgraded to a bigger transformer) so it's def possible. I don't have direct exp but I highly doubt that a cheapy suunko will do copper. Copper wil eat that 500a for breakfast, You'll need over 1200 pref >1800 I'd say. I think mine is putting out 1500ish but that's a guesstimate. 1.9kw is pretty dismal- check the threads on them I think they're good for .15 nickel max. Don't quote me on this though :?

According to a sales rep, the Sunkko 709A that is 220V has a little more power than the 110V version even though the specs are identical. This has also been confirmed in other posts: https://endless-sphere.com/forums/viewtopic.php?f=14&t=80644&p=1188754&hilit=sunkko#p1188754

Here is a neat article on replacing a blown triac in the 110V version: https://syonyk.blogspot.nl/2015/05/sunkko-788-welder-failure-and-repair.html
Here is a neat teardown of a 709A. Notice how both are 110V but the transformers are different. My guess is this one is a newer model. :?: http://www.eevblog.com/forum/reviews/sunkko-709a-spot-welder-teardown/

There are just to many variables to know how the 709A-220V will perform with copper strips. We know these things help:
- adding a slit in the strip between the electrodes helps with current penetration
- electrode tip shape plays a crucial role in penetration.
- Tungsten electrodes can help prevent sticking, create more heat and aid in stable arc formation
- Ni plating copper increases resistance at the weld interface to help create localized heating

I think 500A is sufficient. Just need to be smart with all of the available variables. I like how the 709A has some settings that you can play with. Like most Chinese electronics, be prepared to replace some parts or tweak the design. They always seem to get the hard stuff figured out but then fail with the easy stuff like mechanical relay switches vs solid state. If I pull the trigger on the 709A-220V welder, I plan on extending the arms another 2-3in and maybe switching out some cables.

Here is more from ridethelightning

ridethelightning said:
...the dynamics in the process of welding copper is quite different from that of metals with higher resistance like nickel.
because it conducts the weld current so well, and also conducts any heat away so fast, it requires much more current to get it to heat to the point where it melts and fuses to the cell. this makes it very difficult to weld, as it requires a huge amount of current in a very short pulse, and any factors such as bad contact or dirty surface could blow the arse of a cell to bits! my solution to this was is to use tungsten rod as the electrode tip,($3 from any good welding store (from tig electode) drilling a hole in the end of the copper dn10 electrode tips(as pictured earlier) and inserting some~15mm long piece of tungsten with a press fit into each electrode tip.
this has a very high melting point, quite high resistance and is very hard. the result is that the tungsten remains hard but glows cherry red for an instant as the weld pulse travels through it, producing the heat that creates the weld. because the copper is so good at conducting the heat away, the area where the weld is is barely warm to the touch afterwards. ( although the welding pens get warm after a while) im absolutely confident that it doesnt get hot enough to damage the cells, while creating a very high current connection....
ref: https://endless-sphere.com/forums/viewtopic.php?f=14&t=57291&start=775#p1165391


.
 
I'm a complete n00b but I find it odd that if I look at spotwelding battery packs I see that most of the time the two electrodes are placed side by side pressing the nickel or copper strip onto the cell before releasing the welding pulse. Doesn't that mean that most of the welding current just flows from one welding electrode to the other electrode through the nickel/copper strip?

If you want to spot weld two steel plates together you would have the plates on top of each other and together between the two electrodes. That forces the welding current from one electrode through the two plates to the other electrode. Making a spot weld between the plates.

When welding a battery pack by hand isn't it therefore better and wouldn't it make it possible to weld copper strip by putting one electrode directly on the cell top/bottom and the using the other electrode to press the nickel/copper strip onto the cell when spot welding?
Both on the same cell side of course, not letting welding current flow through the cell. Or is that still damaging to the cell?
 
SlowCo said:
...When welding a battery pack by hand isn't it therefore better and wouldn't it make it possible to weld copper strip by putting one electrode directly on the cell top/bottom and the using the other electrode to press the nickel/copper strip onto the cell when spot welding? Both on the same cell side of course, not letting welding current flow through the cell. Or is that still damaging to the cell?

You have a good point. So one electrode is off the strip touching the battery cell end and the other is on top of the strip. This forces the current to go through the strip/cell interface. This would produce one spot vs. two. But you could keep the one probe stationary and move the other one to different spots on the strip quickly.

One problem may be that if strip is contacting to much area of the cell it will cause the current to spread from multiple directions rather than just through the small location under the electrode. But if you keep the electrodes close together around the perimeter of the cell this would minimize current spreading.

Needs to be tested but I like the idea. Thoughts?
 
Back
Top