The Unofficial Spot Weld Resistance testing and discussion thread.

[SternWake]

I’d like to apologize for not having done this sooner as I promised.

In the Spot Welding copper/ Copper sandwich thread we discussed this, and @Battery_Mooch had graciously offered to test some samples with precision equipment and the skills to use it properly, so that we could stop guessing.

Nelvick @DIY500AMP.COM has donated 21700 cell samples and I am here, tools in hand, some time, and am ready to weld some strips to some cells Would like to visit the post office By Saturday.

I am contemplating the best methods to minimize the variables, and allow for comparable consistency in the future, as I am not going to be able to cover every base.

I have an AwithZ P20B 14.6 KW spot welder, that has proven capable of welding 0.3mm copper under 0.2mm nickel plated steel, using Nelvick’s welding flux at near its full available power. I have solidly welded as much as 0.2mm+0.15mm copper under 0.1mm stainless using flux, but my machine was incapable of 0.2+0.2 under 0.1mm stainless using flux at max power. it was close, but not quite enough. 0.07mm stainless atop 2 layers of 0.2mm copper made no discernable difference.

In general I Use 0.2mm copper under 0.1mm stainless steel using flux below the copper, using gear 385 to achieve very solid welds.
Without the flux I need to use about gear 550

Without Flux or Stainless on top, to weld 0.2mm copper I need about gear 840.

Weld consistency fluctuates with no sandwich or Flux.

The amount of pressure, tip shape and cleanliness, and temperature of welder and Leads playing a much larger role when there is no sandwich and no flux, at least on my P20B welder.

I cannot ‘raw dog copper’ thicker than 0.2mm(0.25mm maybe?) and personally do not mind having to use stainless or the Flux.

I know @Pecos has a bunch of cell samples, An AwithZ p60f(29.4kw) , or is it P90C (56.8kw) spot welder?, and is willing to Weld thicker samples that I cannot.

I have 0.1, 0.15, 0.2mm thick, 10mm wide copper strips

I have 0.15mm thick 10mm wide pure nickel

I have 0.07mm thick 304 Stainless

I have 0.10mm thick 304 stainless

I have 0.5mm thick 16mm wide Stainless, some 400 magnetic series.

I have 0.2mm Nickel plated copper, which behaved exactly like Non plated copper in my experiments

I have plenty of Nelvick’s welding Flux and extra welding tips, and about 15 21700 cells. Some tabless and some not. Mostly virgin cells but a few that have little nubs of welded strips I will need to remove.

I know some have no interest in using either flux or Sandwich, and have machines powerful enough to be able too without, whereas my P20B can only raw dog 0.2mm copper at high power levels.

I think any thinner than 0.2mm copper by self, is just too weak on the tear off tests, that it greatly benefits strength wise from 0.1mm stainless on top.

I have welded sandwich using pure nickel, and nickel plated steel, but both of these require the use of lot more power than using stainless steel.

If going for copper sandwich I personally see no point in using Pure nickel or Nickel plated steel on top. They require significantly more power. I use neither as conductors, only welding aids.

I would like to find out the difference between 2 pairs of welds, and 3 pairs, and 4 as well.

I would like to find out the difference between welds with flux, and without flux.

Copper sandwich welds, and non sandwich raw dog copper only welds, both with and without flux.(0.2mm limited)

I’d also not like to overload Mooch with a giant box of samples to test, if consistency is poor and results conflict.

I am really pretty dialed in with 0.2mm copper on my machine and I know that’s where I will be able to deliver the most consistently prepared samples.

[fearlesstvn]

Really cool of you to take the time to improve things for the battery community.

I’ve already picked up a bunch of good tips from you just from reading your posts. Appreciate you putting in the work on this, bro!

[Battery_Mooch]

My votes are for only “worst” and “best” cases in regards to interconnect material, layer, and thicknesses. Then for each of those, minimum and maximum number of welds.

All the in-between setups don’t matter until we know if the range of power losses is wide enough (between best and worst) to make exploring those other configurations worthwhile.

[SternWake]

I was thinking in the interests of consistency and repeatability, using 25mm long 10mm wide strips.

CELL A is a virgin Ampace JP40 rewrap.

0.15 mm pure nickel.

2 pairs of welds on Anode -

3 pairs of weld on Cathode +

i used gear 90

0.05ms preheating

02ms ‘intermittent’ this is delay between preheating pulse and first weld pulse

Double pulse with a 0.2 second delay between weld pulses.

Gear 90(of 999).

Nothing written in stone yet.

Am more than willing to modify this baseline.

[Battery_Mooch]

@SternWake

CELL A is a virgin Ampace JP40 rewrap.

0.15 mm pure nickel.

2 pairs of welds on Anode -

3 pairs of weld on Cathode +

For “worst” I’d love see 0.15mm nickel with…

Cell A = 1pr welds on neg/pos

Cell B = 4pr welds on neg/pos.

Let”s see if weld count makes a difference with “low performancel interconnects.

Then two more cells with the thickest and highest layer count metal, one cell with one pair welds and another cell with four pair welds.

Then we can directly compare thin Ni/SS metal and thick Cu/Ni metal setups. We can also directly compare the effect of weld count on different metal/thickness/layer setups. And we only need four cells.

To be sure about reading and effect of weld consistency on the results I recommend 2-3 of each of the four cells listed above. IMO, consistency can have a huge effect but I have no idea if it’s true.

Other testing can include different weld “quality” on identical cell setups. That is, how much does an easily peeled off strip affect power loss? Must we always go for insane levels of “grip” and consistency? Or does any weld pretty much perform the same?

[SternWake]

I’ll do my best to get you exactly what you want.

So, cancel cell A. maybe I’ll include it and label it cell A-Z:

Cells A1, A2, and A3 will have 1 pair of welds on both anode and cathode, using 0.15mm nickel

Cell B1, B2, and B3 will have 4 pairs of welds both anode and cathode, using 0.15mm nickel.

That takes care of 6 cells.

0.15mm pure nickel being the ‘ worst’ available to me. Other than nickel plated steel or Stainless steel as a conductor.

The ‘Best’ conductor I have on hand is 0.2mm copper.

I can weld 0.2 copper both with AND without stainless steel on top.

I can weld this with or without Nelvick’s welding flux.

I think ‘Best’, with 0.2mm copper, would be 0.2mm copper under 0.1mm stainless using flux and 4 pairs of welds.

Spot welding Copper / Copper Nickel / Copper Stainless Steel Sandwich

I carefully filed the electrode that had can steel stuck to its tip, and made sure it was same size and shape as the other. Flipped a cell over in holder and tried welding the + tab with the same settings. no Flooux No preheating, triple pulse, 2ms interval Two very even, weak welds lifted off very easily. Tiny holes in copper, didnt bother with photo Tried again with gear 750, 1ms interval, triple pulse. [IMG_7588] Notice the steel colored ejecta on the weld on left. That is not a wei…

Scanning the other copper welding thread, in post 44, it seems I used gear 750 triple pulse to weld 0.2mm copper with no flux and no Sandwich on an 18650 cathode. The copper tore but I was unimpressed with the tear off strength. I don’t often use triple pulse. I have the most experience using 0.2mm copper under 0.1mm stainless using flux gear 385, double pulse.

How about the strip length, Should I stick to 25mm, or should I make it some set distance from the nearest weld/ or the perimeter of the cylinder?

I’ll hold off welding anything more, for now, as to best utilize the 10 virgin 21700’s I have left.

Will await your precise instructions :)

I can clean nubs from previously welded cells, but that adds an unknown variable.

[Battery_Mooch]

@SternWake

How about the strip length, Should I stick to 25mm, or should I make it some set distance from the nearest weld/ or the perimeter of the cylinder?

I’ll hold off welding anything more, for now, as to best utilize the 10 virgin 21700’s I have left.

Will await your precise instructions :)

I can clean nubs from previously welded cells, but that adds an unknown variable.

25mm from center of cell works fine IMO.

I’m just not able to dive into all the details of this. If the folks here just don’t care enough about the results to chime in with what results they’d like to see then I guess it just doesn’t matter what we do.

IMO, we can’t use previously welded cells for any of these tests. Well, unless we’re testing what results we get with previously welded cells. But then there can be a whole new set of conditions to test.

[SternWake]

Your recent test of the Vapcell / Ampace JP30p1 rewrap, with the button top adding so much resistance and wasting so much energy as heat, is at the core of my interest in this.

I don’t care that the extra temperature gain might not push battery or pack or conductors into the danger zone, I don’t want current heating conductors period. Screw resistance. Within reason of course.

I’ve been of the mindset that I want to use as thick of a conductor as my welder can weld at ~80% or less of its max power, and as many welds as I can fit on a 21700 cathode, which is 4 pairs.

I am of the opinion that only the welds themselves can be counted on to pass current, that the strip adjacent to the welds is rarely in direct contact. This should be a goal of testing in my opinion.

I think that the welding Flux actually tins the cell and copper immediately adjacent to the welds, increasing surface area in direct contact between strip and cell, further reducing resistance.

Maybe 3 pairs of flux’d welds equal same resistance as 4 pair of unflux’d.

This is a great opportunity to answer some questions.

My 0.5mm stainless experiment seemed to weld the fluxed copper below so fat so flat and so wide with just 1 pair of welds, and the tear off strength was so extremely high, I’d love to see if this is potentially lower resistance than a more sane sandwich.

I’m ready to weld and mail samples and get this started, I just want to make the best use of the Virgin cells I have on hand.

[SternWake]

I got a FNIRSI HRM-10 internal resistance tester.

This model came with the clamps, not the cell holder or the 4 spring loaded probes.

Holding them open and pressing all 4 contacts onto a cell and getting a solid reading is a bit of a process, but there is no doubt when one achieves solid contact.

I did a little experiment with that cell shown above with the 0.15mm pure nickel which had 2 pair of welds on the anode. This experiment showed a third and fourth pair of welds did help reduce resistance, but I was not precise enough on the clamp locations and did not write down the results and decided to repeat the test as precisely and hopefully as accurately as I am capable of.

I cut 4 10mm wide strips of 0.2mm copper exactly 45mm long, and cleaned them with 91% IPA, along with the two test cells.

Each 45mm long 10mm wide 0.2mm thick Copper strip measured 0.43 mOhm with clamps on the very edges.

I dont have 0.2mm pure nickel for comparison, but a 10mm wide 45mm long piece of pure Nickel at 0.15mm thick = 2.43 mOhms

Cell A is a virgin Ampace JP40. It measured 2.70 mOhm ACIR at 3.4733vDC

Not using Flux. I welded the 45mm strips to the cells + and cell - under a 10x10mm square of 0.1mm thick Stainless steel. Gear 549 on my AwithZ P20B welder.

After welding I took extra care to position the clamps at the very edge of the 45mm long copper strips.

1 pair of welds on each 45mm long strip = 3.51 mOhm

2 pair of welds on each 45mm long strip =3.44 mOhms

3 pair of welds on each 45mm long strip = 3.30 mOhms

4 pair of welds on each 45mm long strip = 3.27 mOhms

So 0.43 mohm from 45mm strip + 2.70 mOhm from the JP40 = 3.13 mOhms

1 pair of welds added 0.38 mOhm

2 pair of welds added 0.31 mOhm

3 pair of welds added 0.17 mOhm

4 pair of welds added 0.14Mohm

I then took the back of a screwdriver handle and firmly pressed the strips against the cell ends.

4 pair of welds with sandwich pressed against cell ends added 0.23 mOhms

Cell B is another JP40. This one measures 2.77 mOhms ACIR at 2.6484v

Same 45mm, 10mm wide, 0.2mm thick strips under same 0.1mm SS squares

WITH welding FLUX. Gear 379 with P20b welder

1 pair of welds 3.59 mOhms

2 pair of welds 3.44 mOhms

3 pair of welds 3.33mOhms

4 pair of welds 3.30 mOhms

So 2.77 mOhm cell resistance + 0.43 mOhm from 45mm long copper strips = 3.2 Mohms

1 pair of welds with flux added 0.39 mOhms vs 0.38 without

2 pair of welds with Flux added 0.24 mOhms vs 0.31 without

3 pair of welds with flux added 0.13mohms vs 0.17 without

4 pair of welds with flux added 0.10 mOhms vs 0.14 without

I then did the screwdriver handle thing pushing the welded strips firmly against cell and measured 3.34Mohms

4 welds when pressed firmly against cell with flux added 0.14m ohms, vs 0.23 without

The decrease in resistance from going to 3 pair to 4 pair of welds was only 0.03 mOhm on both cells

Using the screwdriver handle to push the welded sandwich against the cell increased resistance on both cells by 0.09 mOhm on the flux free sandwich, and by 0.04mOhms on the Fluxed sandwich, compared to not pressing the sandwich.

So will 4 pair of welds be stronger, it seems so, but there is little reason to do so resistance wise.

Pressing the Strips against cell ends firmly, increasing the resistance, was unexpected.

I thought the ‘tinning’ that happened when using Flux, in a lot of my spot weld strength testing would reduce resistance some, and it did, a little bit, with 2,3, and 4 pairs of welds.

I really tried to be as equal as possible in this test, clamping the meters probes on the very end of the copper strips and welding them as equally as possible, although using flux under 0.2mm copper under 0.1mm stainless steel requires gear 380 with flux vs gear 549 without flux.

I can likely fit 5 pairs of welds on a 21700 Anode, but certainly not on the cathode. I wish that one preproduction Reliance RH60 fat cathode design were adopted by all cell manufacturers.

There was so little decrease in resistance going from 3 to 4 welds, with or without flux, one can argue that 4 pair is an unnecessary waste of time.

Pressing those sandwiches firmly into the cell, in both cases increased resistance to somewhere between 2 pair of welds and 3 pair of welds, compared to not pressing the 4 weld sandwich.

Perhaps there is some sort of shearing of the copper after it is welded, when it is manually pushed against the cell which increases resistance.

I have another JP40 cell with 25mm strips of 0.15mm pure Nickel welded 3 pair cathode, 4 pair Anode at 4.52 mOhm

I used same technique and tool to press This pure Nickel against the cell ends and 4.40mOhm.

So pressing nickel strip against the cell reduced resistance but the same process with copper SS sandwich increases resistance, in my testing.

Sorry that I never made a bunch of Samples to send to Mooch and get more professional results.

The lack of interest did not inspire me, so other projects moved up my list and this project was tabled.

I am glad I did it though.

I will likely keep using 4 pairs of welds and keep using Copper SS sandwich, and Flux, and I will not be trying to press.mold the welded copper sandwiches against the cells with the back of a screwdriver as this increased the resistance.

This does not answer the question as to how many amps can each weld can safely carry, or maybe it does if someone were good at math.

[SternWake]

I Did not perform these calculations, and A eye gets so much wroong, but Perhaps this copy and paste is not too out of bounds for 0.2mm copper:

Configuration (Per Cell)	Isolated Weld Resistance	Safe Continuous Ampacity (0.1W Heat)	Peak Esk8 Burst Ampacity (0.25W Limit)
1 Pair of Welds	0.195 mΩ	22.6 Amps	35.8 Amps
2 Pairs of Welds	0.097 mΩ	32.1 Amps	50.7 Amps
3 Pairs of Welds	0.065 mΩ	39.2 Amps	62.0 Amps
4 Pairs of Welds	0.048 mΩ	45.6 Amps	72.1 Amps

Thoughts?

[glyphiks]

I love this. Keen to read through properly after work

[Battery_Mooch]

@SternWake

I Did not perform these calculations, and A eye gets so much wroong, but Perhaps this copy and paste is not too out of bounds for 0.2mm copper:

Thank you for all that testing!

What were your criteria for setting the ampacities?

Single cells in open air or a pack under fishpaper+heatshrink?

[Battery_Mooch]

@SternWake

This does not answer the question as to how many amps can each weld can safely carry, or maybe it does if someone were good at math.

We would need the max weld temp for a couple of test scenarios that duplicated pack use (how the cells are arranged, how the pack is wrapped, etc.).

And a definition of “safe”. That temp (or current level) might be much higher than what would be accepted for long cell life and/or what heat shrink wrap can handle.

[glyphiks]

@SternWake

1 Pair of Welds 0.195 mΩ 22.6 Amps

So back in the dark ages when we used nickel, it was commonly accepted that one pair of welds would safely carry 5a

So with your resistance testing you had at a rough estimate of 4:1 nickel vs copper, I would think this seems fairly ‘accurate’ to our old ways of thinking.

5 amps per set of nickel welds, 20a per set of copper welds.

To be fair, i would have thought it would be less than 20

[SternWake]

I woke up thinking this testing was kind of irrelevant to the question of spot weld ampacity because the FNISRI hrm-10 measures ACIR, not DCIR.

It does not look like the AI bot was considering the battery being wrapped in kapton, Fishpaper, and shrink wrap when it ran the numbers I’d collected.

It does know this is for Esk8, and the Shoulder bent parallel tabs connected with silicone series wires that is seemingly Unique to Esk8 battery builds, but it does not seem that it took those factors into account.

I think it just calculated the 10Mm wide 0.2mm thick copper strip and the cell itself as the heatsink.

Anyway, here is another C&P from my convo with it regarding how it came up with the numbers it did with the numbers I gave it.

QUOTE:

You asked if someone could do the math to figure out how many amps each weld can safely carry based on your data. Here is the exact calculation.

To calculate real-world ampacity, we must isolate the resistance of a single weld pair junction. Taking your fluxed Cell B data:

• 1 Pair of Welds: Added 0.39 mΩ total resistance to the loop.

• Since the current passes through both the positive and negative strips, each single-weld pair junction accounts for half of that value: 0.195 mΩ per weld pair.

An industry-standard safety envelope dictates that a battery weld junction should never continuously generate more than 0.1 Watts (100mW) of waste heat locally to protect the internal chemical layers. Using Joule’s Law of heating (\(P = I^2 \times R\)), we can solve for Current (\(I = \sqrt{\frac{P}{R}}\)):

\(I=\sqrt{\frac{0.1\text{\ Watts}}{0.000195\text{\ }\Omega }}=\mathbf{22.6}\text{\ Amps\ Continuous}\)

Endquote

With the Big wide shoulder folded parallel copper tabs acting as a heatsink, seems the Ampacity of each weld could be increased?

One thing that really surprised me was that pushing the 4 pairs of welded strips against the cell increased the resistance on the 0.2mm copper SS sandwich, but decreased it on the 0.15mm Nickel.

The post weld flattening resistance increase was more without the flux than with it.

I wonder if copper spot welds are therefore more susceptible to fatigue and shearing.

Perhaps only after the back of the screwdriver handle , ‘flattens the Donuts’

[glyphiks]

@SternWake

wonder if copper spot welds are therefore more susceptible to fatigue and shearing

I would have thought so, given how soft copper is compared to nickel

[SternWake]

I will need to bust out the 0.15mm copper and directly compare to my 0.15mm nickel, but I have been thinking Nickel is more malleable than the copper in terms of pressing it down around the cathode’s button

I suspect the Sheets of copper are somewhat work hardened through the extrusion process. but annealed copper would likely be softer.

I don’t recall if I raw dogged 0.15mm copper, but I know I can.

Pretty sure I was welding 0.2mm copper no sandwich but with Flux, and was able to get excellent welds, but the strip tore too easily from the cell at the edge of the welds, IIRC.

Removing the CUSS sandwich easily distorts the cell can, especially with my 0.25mm copper.

Perhaps 0.25mm Cu no SS sandwich with flux is within range of the p20b

I will have to check the other thread to see how much ‘gear’ I was using for 0.2mm copper no sandwich with flux. I’ll likely need 33% more ‘gear’ than that.

Edit: looks like I stacked 0.1 and 0.15mm copper no Sandwich with flux at gear 850.

Spot welding Copper / Copper Nickel / Copper Stainless Steel Sandwich

Here is 0.15mm copper atop 0.1mm copper, with flux, gear 850, preheating 0.15ms, 04ms interval, double pulse. The peel off strength was good, but obviously 0.25mm stacked copper felt a bit flimsier than solid 0.25 copper would. I think this is very Viable nickel Sandwich free option with this welder. [IMG_75681copy] [IMG_7569] [IMG_7572] The other copper bits on the cell were half hearted attempts with 0.1mm copper atop 0.2 and gear under 750. I wasn’t feeling it, then decided to go d…

[SternWake]

Subjective malleability test. 0.15mm copper vs 0.15mm pure nickel.

My 0.15mm copper is only 8mm wide, the nickel is 10mm wide.

I placed both materials individually over the same side of this washer, as shown because the other side is a bit sharper edges and would yield different results.

I pressed the strips individually with my right thumb squeezing the table edge and then came in with my left thumb on top and pressed with both hands as hard as I could.

In my calipers the deformed 0.15mm copper is 0.24mm thick

In my calipers the deformed 0.15mm Nickel is 0.40 thick.

So the copper strip does seem less malleable than the nickel, in this test.

Does this mean the spot welded Nickel will fatigue quicker than copper, all other factors being equal?

I have no idea.

I have cut some 0.15mm nickel to 8mm wide, as best as i could with tin snips

I will compare an equal 45mm long 8mm wide 0.15mm strip of Nickel and Copper.

After running ACIR tests I will remove the strips from the cells and see which comes off easier.

Also subjective, and I am definitely biased toward copper but am trying to be as fair as I can, cognizant of the inherent bias.

[SternWake]

Second test.

2 copper strips, 45mm long, 8mm wide, 0.15mm thick.

Each copper strip 0.70 mOhm

With Flux, No sandwich. Gear 399 of P20B welder.

CELL A: Virgin Ampace JP40 2.77 mOhms ACIR

JP40 at 2.77 Mohm + 45x8x0.15 mm COPPER strip at 0.7 mohm = 3.47mohm

1 Pair of welds each side 4.22 mOhm - 3.47 = 0.75 mOhm

2 pair of welds each side 4.07 mOhm - 3.47 = 0.60 mOhm

3 pair of welds each side 3.90 mOhm - 3.47 = 0.43 mOhm

4 pair of welds each side 3.83 mOhm - 3.47 = 0.36 mOhm

5 pair of welds Anode only 3.79 mOhm - 3.47 = 0.32 mohm

6 Pair of welds Anode only 3.77 mOhm -3.47 + 0.30 mOhm

Welded strips pressed hard against cell…3.89 mOhm - 3.47 = 0.42 mOhm

CELL B: Virgin Ampace JP40 2.70 mOhm

45mm x 8 mm x 0.15mm pure NICKEL. 2.91 mOhm

2.70+ 2.91 = 5.61 mOhm

1 pair of welds each side 8.44 mOhm - 5.61 = 2.83 mOhm

2 pair of welds each side 7.94 mOhm - 5.61 = 2.33 mOhm

3 pair of welds each side 7.59 mOhm - 5.61 = 1.98mOhm

4 pair of welds each side 7.26 mOhm - 5.61 = 1.65mohm

5 pair of welds anode only 7.25 mOhm - 5.61 = 1.64mOhm

6 pair of welds anode only 7.21 mOhm - 5.61 = 1.60 mOhm

both strips Pressed tight to cell 7.30 mOhm - 5.61 =1.69mOhm

Notes:

Anode weld pair #5 was on the cell farthest away from where kelvin clamps attach

Anode weld pair #6 was on the cell closer to the kelvin clamp opposite the JP40 weld spiral which I avoided.

The pure nickel strips were 10m wide. I tried to cut it perfectly 8mm wide to match the copper strips and got it close, but not perfect.

Observations. 4 pair of welds reduced resistance over three welds more on the 0.15mm copper and 0.15mm nickel more than it did on 0.2mm copper

Pressing both the copper strips AND Nickel strips firmly against cell ends increased resistance more than leaving it unpressed.

It rose by 0.12 mOhm on the copper and rose 0.09 mohm on the pure nickel

Ripping off the strips required equal strength to get the strips to start tearing, but once the copper stared tearing it tore about 30% easier than the nickel until the strip completely tore off.

I Would not use 0.15mm copper by itself in an Esk8 or other mobility battery, unless it was capped by steel, preferably Stainless sandwich.

I plugged this data into A eye and it spat this out:

Quote:

0.15mm Pure Copper vs. 0.15mm Pure Nickel (No Sandwich Layer)

The following tables evaluate the raw connection interfaces by pulling out the material’s bulk resistance, focusing purely on the electrical efficiency of the weld points.

0.15mm Pure Copper (8mm Wide, No Sandwich)

• Isolated Strip Baseline: 0.70 mΩ

Configuration	Single-Junction Resistance (\(R\))	Safe Continuous Ampacity (0.1W Heat)	Peak Esk8 Burst Ampacity (0.25W Limit)
1 Pair of Welds	0.375 mΩ	16.3 Amps	25.8 Amps
2 Pairs of Welds	0.300 mΩ	18.3 Amps	28.9 Amps
3 Pairs of Welds	0.215 mΩ	21.6 Amps	34.1 Amps
4 Pairs of Welds	0.180 mΩ	23.6 Amps	37.3 Amps

0.15mm Pure Nickel (~8mm Wide) [1]

• Isolated Strip Baseline: 2.91 mΩ

Configuration	Single-Junction Resistance (\(R\))	Safe Continuous Ampacity (0.1W Heat)	Peak Esk8 Burst Ampacity (0.25W Limit)
1 Pair of Welds	1.415 mΩ	8.4 Amps	13.3 Amps
2 Pairs of Welds	1.165 mΩ	9.3 Amps	14.6 Amps
3 Pairs of Welds	0.990 mΩ	10.1 Amps	15.9 Amps
4 Pairs of Welds	0.825 mΩ	11.0 Amps	17.4 Amps

---

Core Takeaways:

1. The 4.5x Connection Advantage: At 4 pairs of welds, the pure copper joint interface resistance (\(0.180\text{ m}\Omega\)) is over 4.5 times lower than the pure nickel interface (\(0.825\text{ m}\Omega\)).

2. Where the Heat Generates: Because the pure nickel interface resistance is so high, pulling a standard Esk8 burst of 35 Amps through a single 4-pair nickel terminal would generate a massive 1.01 Watts of localized heat right on the cell cap (\(I^{2}R\)). The same 35A burst through the 4-pair copper terminal generates just 0.22 Watts—staying entirely within the safe burst window.

3. The Thickness Dynamic: Comparing this to yesterday’s 0.2mm copper sandwich tests, the 4th pair of welds offered a much more significant drop in resistance on the thinner 0.15mm strips. This empirically confirms that thinner busbar material creates a tighter bottleneck, requiring more weld paths to fully saturate the strip cross-section.

4. The Structural Verdict: While un-sandwiched copper completely dominates the ampacity charts, the destructive tear test showed it unzipped 30% easier once a tear began due to localized annealing in the heat-affected zone (HAZ). For high-vibration applications like Esk8, the ultra-low resistance of copper must be reinforced by a 0.1mm stainless steel overlay to act as a physical splint over those softened zones.

---

End quote:

[SternWake]

It seems that the copper sheets I cut into the Busbars is to some degree, work hardened from the process of cold rolling it into 0.2 or 0.25mm sheets, and is then annealed to some degree to some distance away from the welds in the Heat affected zone.

AI seemed to take some liberties in its conclusions in the above post, I was hoping for some replies as to what and where it missed the mark, rather than crickets.

I wonder how laser cutting the copper busbars that some people can do, affects the work hardening and the Annealing of the copper next to the welds and heat affected zone directly adjacent.

I am still a bit surprised that pressing the welded busbars, whether Copper or Nickel, or copper SS sandwich, down to the cell with the back of a screwdriver handle increased resistance.

I will definitely form the busbars to the cathode button before welding on all my future battery builds and leave the welds alone afterwards.