18650 spot welding -how to- ULTIMATE REPOSITORY

@LTRMB
In my very rudimentary test, I was surprised at how much heat started to be generated once the optimal rate had been exceeded. OR is right we don't use cont current but every time you exceed the limit heat start being generated quite quickly. Given your battery is sealed (usually) heat builds up. Cheap battery packs get away with it cause the batteries can soak up the heat but they're not designed to be heat sinks, and, the hot spots on the battery ends are where the trouble starts.

The series connections take the full load but the parallel connections you'll have to figure out. In perfect arrangement ( ie. one series connection for each parallel group and a +/- distributing loom at each end) the P connections see virtually nothing- just balancing currents.
In my 20s5p build rated to 60A cont discharge I used .1mm copper for the series connections (5 of them- one for each P group) but I only needed a single strip of .15x8mm nickel for the P connections. I used a copper tab at each end and had a fat copper bus to join them so I had equal sharing in/out of the pack.
In a poor layout a P connection could see up to : total amps/ #P attached to it (ie full out put of those cells)
Size your connectors accordingly. Geometric shape ( space available) is a determining factor in how you can lay the pack out
The best way around the whole problem ( but only if you are not using holders ) is a really wide strip that does the P and S connection all at once- search Madin88 's builds.
K
 
Don't worry about volts or watts in terms of heat... Amps is only consideration
Spec your connectors for continuous peak controller output (batteries as well) using nobuo's guide.
Divide controller peak amps by # P groups for amps each cell delivers.
Estimate ( or test) P connection loads by using cell output as a guide.
S connector load = controller peak amps / no. of series connectors
Spec connectors as above
Pay attention to how you attach your power cables ( loom or bus is good) don't just solder it on to one end point and ruin all your good work!
Quick run down of battery I just finished
48v 23ah, 35a controller. 3C (10amp) rated cells
Block layout 13sx7p with holders. ( very easy, great current sharing layout)
Single .15x8mm Nickel P connectors
7x.2 x8mm nickel series connectors.
7 x.2 x8mm nickel tabs at each end soldered to a copper bus with 10AWG leads soldered to this.
Bomber!
K
 
kdog you mentioned in one of your builds using a 20s5p and 60a continuous discharge. I see you're using .15 x 8mm strip for the parallel groups. But thats not rated for the 12 amps the parallel cells will use.

I assume you just used the thinner nickel because the parallel connections are so short?
 
> But thats not rated for the 12 amps the parallel cells will use.

Maybe I am missing sth, but I would have thought the the nickel-strips only have to
carry half the Amps of the cells if you connect them at every point with the next strip.
Lets make an example: My GA cells are rated for max 10A, and I connect every Strip
with decent copper wire to the left and right of each cell:

cell-strip-amps.png
If someone connects the series with only 1, 2 or 3 strips instead of the copper wires,
then there will obviously be part of the strip that gets a lot of Amps flowing through...

Cheers, Ulli
 
True each cell delivers up to 12amps but... Each cell has a series connection so the current flow is from cell to cell via the series connector. The adjacent cell in the P group is at the same potential so no current needs to flow to it via the P connector. If I had only say two series connectors then the current flow through P connectors would be 2/5 cells output on its way to get to the series connection.
My cells are all the same batch and behave similarly so the balancing currents during a load are minimal.
 
Up0
Correct in my understanding. Even better is pop the S connector right on the battery terminal and largely eliminate P connector flow. I understand you don't want to solder there... I spot welded. Flow in P groups can be largely minimised if layout is perfect.
But battery housing space determines the shape of the battery and current flow is somewhat dependant on this.
Nickel is a pretty poor conductor but it's awesome to weld so that's why everyone uses it.
Copper is a way better conductor (4x) but difficult to weld and weaker. I spent ages learning to weld it, beneficial in high draw packs but .2 nickel is my preference in low power applications.
K
 
up0

I'm not sure if the way you have your cells setup is good. The first problem I see is that the current draw will not be equally distributed among the cells. The two outer cells have a lower resistance in the series connection because they have the left wire being only used for 1 cell. What will happen is when you draw power from the battery it will pull more power from those two outer cells.

I guess you could make that wire smaller to handle half the amps to even things out.

Maybe it doesn't matter at all since it is insignificant.

You could also just use wires between two cells, but this would increase the load on the nickel plate. As shown where the green wires are in the picture.

But you see that you can't just simply distribute the load on the nickel without causing problems elsewhere.

 
OR true but current will only flow where there is V difference and in this case the weighting towards the end cells is small (1/12 extra ottomh) so P flow in this strip of nickel would almost certainly be OK.
Balancing currents will restore any temp imbalance during load - P grp cells are always same V (at rest)
 
Cowardlyduck said:
Won't current sharing between cells depend on the placement of the series links?
i.e. if you only have 2 series links for 8 paralleled cells, then the cells not next to the series links will need to send their current through the nickel first before it reaches the series link.

Cheers

Current sharing between cells inside a parallel group depends exclusively of their differences in voltages. In your example you use the same conductor for parallel and for serial, so in that case the current is not shared but conducted from one cell to the next serial cell. That example makes more difficult the BUS pattern to be equally IR balanced to the current flowing level on each place. So it depends on how you configure the conductors paralleled pattern they will work only for current sharing, or main serial conduction

up0 said:
> But thats not rated for the 12 amps the parallel cells will use.

Maybe I am missing sth, but I would have thought the the nickel-strips only have to
carry half the Amps of the cells if you connect them at every point with the next strip.
Lets make an example: My GA cells are rated for max 10A, and I connect every Strip
with decent copper wire to the left and right of each cell:


If someone connects the series with only 1, 2 or 3 strips instead of the copper wires,
then there will obviously be part of the strip that gets a lot of Amps flowing through...

Cheers, Ulli

In this case there is no need to make the current pass through the paralleled nickel conductors, just only from one cell to the next matched one on the next series, you make a higher efficiency and best worth of the conductors if you place the wires directly over the cell anodes and cathodes
 
Nobuo, how is doubling up with nickel strip compare with a single thicker piece?

For example,

using two .15 x 7 mm strip
vs using a single .3 x 7 mm strip


Do you think that you would get close to 95% - 99% of the single .3 mm strip by using two .15 mm strips?

The reason I ask is because .3 nickel is just too difficult for me to weld and I am thinking about just using two .15 mm strips instead. I know there may be some loses in the welds.

What do you think? Any reason doing this may not be a good idea?
 
Its perfectly fine to double up, thats what I suggested.
A conductor is a conductor, but if I were you I'd use copper wire instead.
Copper is 4.15 times better a conductor then nickel plus copper wire is cheap and plentiful.
Copper is also easy to solder.

But if you do not have a soldering iron, then weld on that extra nickel strip.

https://en.wikipedia.org/wiki/Electrical_resistivity_and_conductivity
Conductivity Chart.jpg






markz » Tue Mar 01, 2016 12:00 pm
Use whatever nickel strips that weld good, ideally you want to go parallel first and series second. Series connection, you can solder on 8, 10 or 12 AWG wire, use whatever you have on hand, if its 14awg then double it up.

My suggestion would be to start with nickel strips, hook up to bike and ride it hard and see if the strips get hot.
If they do, start soldering wire. Thats how I would do it, skipping the riding part for testing and going straight to soldering wire, thats what I would do without researching it prior to this post.

I will be in your position within under a week. I bought some nickel from batterysupports which seem to be more expensive.
I have everything ready, just need to epoxy the top/bottom of the MOT (the Iron Core I believe its called), and build my electrode arms. Probably start that today, but it seems I am lazy these days.
 
Offroader said:
Nobuo, how is doubling up with nickel strip compare with a single thicker piece?

For example,

using two .15 x 7 mm strip
vs using a single .3 x 7 mm strip


Do you think that you would get close to 95% - 99% of the single .3 mm strip by using two .15 mm strips?

The reason I ask is because .3 nickel is just too difficult for me to weld and I am thinking about just using two .15 mm strips instead. I know there may be some loses in the welds.

What do you think? Any reason doing this may not be a good idea?

Basically almost the same conduction capability properties. For making a perfect IR BUS and making worth of the nickel material, I continuously weld layers in the needed places and never weld more than 0.15mm.
 
Thanks,

After trying to weld .3 nickel and having a very bad time with it. The welds were burnt, most of the time I was able to pull the nickel off the cell by hand. Couldn't get consistent welds no matter how many times I tried.

It just makes more sense to weld two layers of .15 nickel which is very easy.

The only issue with welding double layer is what happens if you weld the layers together over the edge of a cell? Will it weld through a cell protector like one of these? I need to weld there because i have to solder wire at that point, between two cells.

Free-shipping-font-b-18650-b-font-font-b-battery-b-font-font-b-gasket-b.jpg
 
Guys, about the 709A model, the head spot work very good, but the pen not really! I'm surprised nobody tell about it!
This model is recommended but the Pen work bad. I tried to spot 0.15 nickel and try some settings and no way. :?

Like I said, the head work great.

The problem is it only with the 110V version?

Is it possible to screw the Pen cable on the head? :?:
 
From what I observed during my initial tests, it is very important how you weld the second stripe. The current does not like to flow through a small spot weld and chooses to flow mainly through the first layer of the metal stripe.
I was testing that with steel stripes I couldn't reach 50% "efficiency of adding more layers". I can make some specific tests if you want.

To avoid this it is better to spot weld two stripes at once to the cell terminal. From what you said before it is rather clear you won't be able to do it in this way, but please - try and share the results.
 
Skrzypas said:
From what I observed during my initial tests, it is very important how you weld the second stripe. The current does not like to flow through a small spot weld and chooses to flow mainly through the first layer of the metal stripe.
I was testing that with steel stripes I couldn't reach 50% "efficiency of adding more layers". I can make some specific tests if you want.

To avoid this it is better to spot weld two stripes at once to the cell terminal. From what you said before it is rather clear you won't be able to do it in this way, but please - try and share the results.

Very good spotted, and forgot to mention, that the cells surface make the current prefer to choose that path because its higher conductivity. When you weld layer over layer, as the surface where you are welding (nickel) is the same IR as the material you are welding to (nickel), you need considerably a higher current, around the 50% more than welding directly over the cells
 
Skrzypas said:
From what I observed during my initial tests, it is very important how you weld the second stripe. The current does not like to flow through a small spot weld and chooses to flow mainly through the first layer of the metal stripe.
I was testing that with steel stripes I couldn't reach 50% "efficiency of adding more layers". I can make some specific tests if you want.

To avoid this it is better to spot weld two stripes at once to the cell terminal. From what you said before it is rather clear you won't be able to do it in this way, but please - try and share the results.


If I had to weld two .15 nickel stripes at once, then wouldn't that be like welding .3 nickel? Too difficult welding .3 nickel.

Seems like I may just have to use a really wide strip to get the proper cross sectional area.
 
In my experience ( not huge )I've not usually had a problem spot welding multiple strips. Occasionally I get a blowout on the 2nd/3rd strip (most annoying) and it usuall due to one of two things. Crud on my electrode, or, uneven contact between the layers resulting in a tiny gap. If I give a proud weld a quick scrape with a small blade or similar it usually fixes it.
If your tab and can are really flat/smooth, and your pressure and current are good, once welded the tab is so closely in contact with the battery there should be significant contact area (then not all the current (when riding )is forced to flow through the weld). Of course the top of the weld can be a bit rough leading to poorer contact on the second strip but with care it might be minimised.
K
 
i found placing the electrodes where they are to weld and rocking them on the point while under pressure helps considerably to make a good contact(with hand held electrodes that is)

i hate the feeling of being on edge from anticipating a blowout :shock:
 
The first blowout on my 10k amp capacitor discharge welder nearly gave me heart failure! The subsequent ones were worse because of the aforementioned anticipation.....there's always something!

I really slowed down after the first 2 or 3 and paid super close attention to the probe placement. It's just trying to get the probes on the plus side for 4 welds is tuff with the limited amount of real estate to work on. I would assume that is the only side they most probably would occur? I think on my first pack of 360 cells I ended up with 5 blow outs if memory serves.

Tom
 
Does cutting the nickel tabs like shown in the picture help with the welding?

What exactly is a blowout?

04ce59da53fbf1b72284acf4a04b5ee4.jpg
 
Actually in my case it was too much power with nothing absorbing it. When one of my tips was not directly above the battery case to be welded to. My nickel tabs were large and had to guess where the battery electrode was at below it.....I had a few bad guesses.

Tom
 

Attachments

  • image.jpeg
    image.jpeg
    56.9 KB · Views: 3,721
  • image.jpeg
    image.jpeg
    51.6 KB · Views: 3,721
Offroader said:
Does cutting the nickel tabs like shown in the picture help with the welding?

What exactly is a blowout?

04ce59da53fbf1b72284acf4a04b5ee4.jpg
When you apply spot welding electrodes to such stripes (that's actually my picture) the whole current must flow from one electrode through cell back to the second. It means 100% of power goes to make a weld. If you don't cut the stripes, something like half of your current goes through the stripe (burning it) and only the second half does the welding...
 
Skrzypas, did you find this a good solution?

How much has it helped you? It seems it may be the way to go. I have one of those heavy duty paper cutters that seems to cut right through this stuff fairly easily and straight.
 
All laptop manufacturers use this technology, see picture of a pair of cells from a brand name laptop. So I just learned from professionals.
84a6dcc2600ba4f04d30faf15816b19f.jpg


And this is the only way my spot welder welds .3 nickel.
 
Back
Top