18650 spot welding -how to- ULTIMATE REPOSITORY

999zip999 said:

With all those spot welds is it getting hot ?

Click to expand...

Nickel changes colors to what you see in about 400 and more deg. Celsius. Which means it must be getting hot (locally).

Obvious tip is to do only a single spot weld on one cell and then move to another cell, and come back to the first cell (for next weld) after finishing "the row".

What I noticed today is that (for recycled laptop cells) welding to the original tab-remainings is:
1. Much more random regarding the strength.
2. Much more colorful (welds look burned more)
3. Because of above: not recommended.

Best welds I get are on the original cell terminal area; the four spot welding points done by laptop manufacturer must be grinded (I use dremel). It is better not to grind the whole terminal but only these 4 points. These are my observations.

I also clean surfaces with extraction naphtha.

I imagine the shape of the tip and pluse or is it double pluse is it ?

999zip999 said:

With all those spot welds is it getting hot ?

Click to expand...

Yeah, the surface was hot, but this was just a laptop battery cell for practice only before the real deal arrives... Samsung 25r's. You think 3-4 welds would suffice?

Punx0r said:

Physical construction may be an issue since it looks like you're not using cell holders/spacers.

Click to expand...

I tried for a long time to come up with a design that could use cell holder but will not fit in bomber unless you do what what Sigmacom did.https://endless-sphere.com/forums/viewtopic.php?f=14&t=61604
so it will be the hot glue and stiff Glass Epoxy sheet method.

Offroader said:

Where are you going to get .2x8mm nickel? I'm yet to find a reputable source for 99% .2 mm nickel.

Click to expand...

http://www.aliexpress.com/snapshot/7361699930.html?orderId=73140737082115
We will see

Doctorbass said:

Excellent design!

Click to expand...

Thanks Doc
I think I will make a jig to do all the soldering first for the series links, then spot weld to the cells last. that should make it go faster too.

JohninCR recently commented that upon disassembling hot-glued packs the glue peeled easily from the cell heat shrink. This may be sufficient for your needs, but I have wondered whether it would be better to add a layer of conventional heatshink to the cells to improve glue adhesion and provide better protection against chaffing. The downside would be the effort, a little extra pack bulk and lower thermal conductivity for the cells.

Hello ES forum members. I'm new on this forum and after a month of reading a lot of thing about EV I decide to register. I already have an idea of beginners EV but of course I need some tools and equipment first.

I already have a 30W Ersa welder, I ordered some extra heat shrink tubes for 18650 cells. I didn't decide yet if I will use glue to make a pack or a holder http://g02.a.alicdn.com/kf/HTB1XQRh...olders-ebike-battery-holders-1000pcs-lot-.jpg for more industrial look. As I read here the heat will better transfer with glue than empty air spaces with holders. I think it's more personal opinion of what to choose, but you can save some space with the glued cells. I will also buy some positive insulation paper gaskets.

I read the whole topic about welders. Because I'm from Europe I need 220V-240V version. There is a lot of different version on the internet a lot of them from China.
At first I was looking this one: http://www.ebay.com/itm/Battery-Poi...662500?hash=item33a5015ea4:g:qI0AAOSwstxVUM1C
Then also recomended Suunko 709A and of course JP welder from user of that forum. But for now JP welder is not availiable any more.

I want to weld nickel stripes from 0.15 to 0.25 so I need a decent welder from which I won't get hurt. Any new (updated) suggestions?

Someone was asking for a link with CNC welder, here it is: https://www.youtube.com/watch?v=VtqGbpYbxaw

Hi Mailer,

> But for now JP welder is not availiable any more.

I am surprised the JP welder should not be available any more?
UPDATE: I just saw riba's post - I am sorry to hear that! Hope he gets well soon again!
Did you contact riba, who builds them with a PM?
He is a very nice guy - definitely worth supporting his project!
I have just bought a JP welder from him and it works beautifully!
I also like the idea of not having to store yet another bulky spot-welder box like the one you have linked.

If you cannot get a new JP welder, I could lend you mine once I am done welding my pack in the next ~2 weeks.
Depends on where you live - could send it fairly easy to France or Germany. The JP welder fits in a small package.
I can also offer this to others who read this later, if they don't want to buy a bulky device for a one-time project.

> nickel stripes from 0.15 to 0.25 so I need a decent welder from which I won't get hurt. Any new (updated) suggestions?

Here is a tip on Nickel: I would suggest using the recommended Aliexpress source from the first post: Best price/value.
But when I tried buying Nickel from them recently, they were temporarily out of stock (prob. Chinese New-Year)
So I tried this one from eBay, just to see if it is the real deal:
http://www.ebay.com/itm/271794492996
( If the link is outdated in the future, look for shop: superpower669566 )
They boast a certificate on their offer-page, and the Dremel/Spark and Salt-Water test were both good,
so it seems to be a valid alternative, in case someone doesn't want to buy via AliExpress or needs a small quantity.
The delivery took some time, but that was most likely also because of chinese newyear - just my bad timing.

Cheers, Ulli

Mr Lowbank said:

Hi All
Have been mining the Endless sphere for information to build a new pack for my Bomber #127. Thanks to forum members for all the great information.

I decided against 20s14p as it would be a tight fit, weigh more and I really don't think I need the amps. I'm hoping for Sanyo NCR 18650GA Cells but Tumich is out of the GA"s!!. Will be installing Max-E controller and have crystalyte 5405 and 5404 motors.
I have a JP spot welder on the way along with all the other bits I think I might need. After taking onboard all the info about building batteries I have come up with the following design. I'm interested in any comments.
Cheers Jon

Click to expand...

Edit: I have misread Lowbank's diagram. I thought he only had 0.7mm^2 for the series connections but it's 7mm^2!! That should definitely handle the 40A per line if the desired amperage is 120A from the 12P Sanyo 18650GA cells. According to Nobuo's chart you can get away with only 2mm^2 of copper for your design.

I think you're going to need more copper for your series connections if you don't want
to bottle neck your current. It looks like you have the total current running through three
main lines.[strike]The copper you're using can handle 20A so 20A x 3 lines = 60A. This means
your pack design can only handle 60A without creating heat.[/strike] If you're looking for 100A/3
lines that's 33A pre each of your three lines. You'll need at least 1mm^2 (17AWG) of
copper for your 100A. You'll probably want more copper to match your cell output.
Each of those Sanyo GA cells can give you 10A each and 10A x 12P = 120A. That means
each of those lines should be able to handle 40A.

I guess in that case if the strips were twice the width then that could solve that issue also and remain easy enough to handle, spot weld?

ps... like the layout!

Thanks for the comments Markisses. I did spend some time trying to get this right. but you have make me revisit this.
There are some rules of thumb that put copper at 2000A/in² which is 3.1Amps/mm² or 21.7Amps for my 7mm² copper strip. But this contradicts other information. Nobuo's table of tests with strips of nickel and copper wire has 1mm² copper wire at 19Amps. I will have 7mm² I understand that the current carrying capacity of a metal is effected by it's size and ability to shed heat and this depends on shape and thermal installation. A look on google tells me that 1/16 x 1/2 (20mm²) copper bar is good for 100Amp http://www.stormcopper.com/design/Ampacity-Quick-Chart.htm So my thinking was 7mm² should be good for the 33Amps. Also Kdog got 10Amps out of .1x8mm (0.8mm²) copper strip I will have 8 x that.

kdog said:

(The good news for myself and RTL is that 18amp through .1x8mm copper caused only very mild heating- prob comparable to about 8-10amps in the .2 nickel)

Click to expand...

I didn't want to go to much thicker or wider as I plan to solder the copper and thought the bigger I go the harder it would be to solder ( get the heat into it).
Have I got this right? has anyboddy else have thoughts on this.
Cheers Jon

How do you know which winding is the primary and which is the secondary when taking apart a MOT for a spot welding machine?

Mr Lowbank said:

Thanks for the comments Markisses. I did spend some time trying to get this right. but you have make me revisit this.
There are some rules of thumb that put copper at 2000A/in² which is 3.1Amps/mm² or 21.7Amps for my 7mm² copper strip. But this contradicts other information. Nobuo's table of tests with strips of nickel and copper wire has 1mm² copper wire at 19Amps. I will have 7mm² I understand that the current carrying capacity of a metal is effected by it's size and ability to shed heat and this depends on shape and thermal installation. A look on google tells me that 1/16 x 1/2 (20mm²) copper bar is good for 100Amp http://www.stormcopper.com/design/Ampacity-Quick-Chart.htm So my thinking was 7mm² should be good for the 33Amps. Also Kdog got 10Amps out of .1x8mm (0.8mm²) copper strip I will have 8 x that.

kdog said:

(The good news for myself and RTL is that 18amp through .1x8mm copper caused only very mild heating- prob comparable to about 8-10amps in the .2 nickel)

Click to expand...

I didn't want to go to much thicker or wider as I plan to solder the copper and thought the bigger I go the harder it would be to solder ( get the heat into it).
Have I got this right? has anyboddy else have thoughts on this.
Cheers Jon

Click to expand...

I think it's a matter of thermal dissipation.. in fact what limit the max amp is the temp rise of the conductive material versus the environnment it is used in. a 1mm2 cross sectional copper in a square shape should have a amp rating lower than a flat copper strip of the same cross sectional area. the flat strip have better surface to radiate all the heat witch allow more amp.

Doc

I just tried to weld .3 99% nickel and I have discovered that there is no way I'm going to use .3 99% for my build. It is too thick and too difficult to get consistent welds. The welds burn too easily also. Then when I tried to up the power of my welder I blew the fets.

So now I have to find an alternative way to weld my pack, either use thicker .15mm nickel or .2mm 99% nickel if I can find it.

But the question is what size nickel strips do I really need?

My Sanyo 3500 has a max of 10 amps continuous. But I will hardly ever use 10 amp continuous.

The other issue is I'm going to use mostly copper wire to connect my cells in series, and only nickel strip to connect cells in parallel.

Offroader said:

I just tried to weld .3 99% nickel and I have discovered that there is no way I'm going to use .3 99% for my build. It is too thick and too difficult to get consistent welds. The welds burn too easily also. Then when I tried to up the power of my welder I blew the fets.

So now I have to find an alternative way to weld my pack, either use thicker .15mm nickel or .2mm 99% nickel if I can find it.

But the question is what size nickel strips do I really need?

My Sanyo 3500 has a max of 10 amps continuous. But I will hardly ever use 10 amp continuous.

The other issue is I'm going to use mostly copper wire to connect my cells in series, and only nickel strip to connect cells in parallel.

Click to expand...

Paralleling connections use too low current sharing, taking into account you are using new cells for the same model, you don't need even 0.15mm pure nickel. 0.1mm is more than enough for paralleling connections.

You need only the right size in the main leads cathode and anode terminals, depending on how and where you solder the main (or mains) leads wires. 0.2mm x 7mm (or equivalent) of pure nickel would be the minimum right mass for each cell.

For serial connections, 0.2mm x 7mm of pure nickel for each cell is the minimum for a right conduction using Sanyo GA.

I made an updated chart on the first post that took a good time to me for calculate, use as a guide

Yes I did some calculations and figured it out, you are correct in that you don't really need all that thick of strip.

The main reason is that the parallel connections between cells are so short it is not worth the trouble to optimize.

For example, I calculated that with all the parallel cells tabs added together it works out to be 400mm of length for each 20s1p group.

in a 20s12p pack using .15 x7 mm of nickel you would have a .5 volt voltage drop.
in the same pack using .2 x 9mm of nickel you would have a .34 volt voltage drop.

So it doesn't make all that much sense to basically save .25% battery power by using really thick nickel.

So you guys pretty much have to make the calculations to weigh out the pros and cons of the nickel.

In my designed battery pack I use copper wire to the series connection so I avoid a lot of the difficulties with nickel plates,

Nobuo said:

Offroader said:

I just tried to weld .3 99% nickel and I have discovered that there is no way I'm going to use .3 99% for my build. It is too thick and too difficult to get consistent welds. The welds burn too easily also. Then when I tried to up the power of my welder I blew the fets.

So now I have to find an alternative way to weld my pack, either use thicker .15mm nickel or .2mm 99% nickel if I can find it.

But the question is what size nickel strips do I really need?

My Sanyo 3500 has a max of 10 amps continuous. But I will hardly ever use 10 amp continuous.

The other issue is I'm going to use mostly copper wire to connect my cells in series, and only nickel strip to connect cells in parallel.

Click to expand...

Paralleling connections use too low current sharing, taking into account you are using new cells for the same model, you don't need even 0.15mm pure nickel. 0.1mm is more than enough for paralleling connections.

You need only the right size in the main leads cathode and anode terminals, depending on how and where you solder the main (or mains) leads wires. 0.2mm x 7mm (or equivalent) of pure nickel would be the minimum right mass for each cell.

For serial connections, 0.2mm x 7mm of pure nickel for each cell is the minimum for a right conduction using Sanyo GA.

I made an updated chart on the first post that took a good time to me for calculate, use as a guide

Click to expand...

I looked at the chart, but don't really understand what it means. Exactly what is <4.7a referring to when it says optimal for .15mm nickel? I know it has something to do with the amount of amps crossing between cells. And in parallel you can have thinner nickel, but in a series you need thicker nickel. But how do you come up with the current that's passing between cells. I'm really not following...

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.

LikeToRideMyBike said:

Nobuo said:

Paralleling connections use too low current sharing, taking into account you are using new cells for the same model, you don't need even 0.15mm pure nickel. 0.1mm is more than enough for paralleling connections.

You need only the right size in the main leads cathode and anode terminals, depending on how and where you solder the main (or mains) leads wires. 0.2mm x 7mm (or equivalent) of pure nickel would be the minimum right mass for each cell.

For serial connections, 0.2mm x 7mm of pure nickel for each cell is the minimum for a right conduction using Sanyo GA.

I made an updated chart on the first post that took a good time to me for calculate, use as a guide

Click to expand...

I looked at the chart, but don't really understand what it means. Exactly what is <4.7a referring to when it says optimal for .15mm nickel? I know it has something to do with the amount of amps crossing between cells. And in parallel you can have thinner nickel, but in a series you need thicker nickel. But how do you come up with the current that's passing between cells. I'm really not following...

Click to expand...

In an ideal situation, if the cells would be true twins and have the same heat dissipation properties, and all the BUS conductors the same IR and length, the current passing through parallel conductors would be absolute zero.

In practical uses there are several variables as each cells batch (chemicals percentages, age, mechanic construction), BUS length and mass differences, cell position (cooling capabilities), etc. That makes the cells having small differences in its behaviour discharging with different curves, making the less efficient cells being paralleled charged with higher SOC voltage ones.

Depending on your construction concerns and discharge rate this current will be low, or very low, but never relatively appreciable if you are using the same model of cells, and the battery have a "standard" build.

For example I can measure between 1mA to 10mA current sharing with unmounted testing li-ion same batch cells discharging at 2A, inside an ideal situation.

You could expect around a maximum of 200mA current sharing between one cell to its next mate, on a very bad scenario where cells come from different batch, one of them is more inside the pack with a greater instant temperature while a 3C discharge rate, so during the discharge it performs more efficiently than its mate on the edge of the pack.

With different model of cells, extreme cold temperatures, not good nickel conductor patterns, you could have a more intense parallel current share.


The chart indicates how a single nickel strip of the size indicated behave under that current values, under real situations, working inside a battery

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

I have a beginner question. So, I'm studying the chart showing the size nickel you need for a given battery pack. How do I determine the current levels/values, so that I can measure the heat generated by the cells? Am I looking at max continuous rated discharge current per cell? Could someone explain this to me in layman's terms so that I can understand thoroughly. I would appreciate it greatly!

thank you kindly

What I still wonder about those amp ratings is they are based on continuous current. Nobody here is ever going to use continuous current on an ebike unless they took the bike on a highway and had OVS on their controller to actually push the cells to the max.

I guess that is why they have an acceptable rating. Kind of like an average of amp ratings you will use.

To figure out the current levels it will depend on how many battery amps your controller is set for, not the max rated discharge of the cells. Because your controller limits the current flow from your battery pack, and therefore each individual cell.

Then you will have to also determine how many amps flow between different areas on the pack, series connections may have more amps flowing through those areas than parallel connections. So you may very well have some nickel strips taking in more amps than others.

Offroader said:

What I still wonder about those amp ratings is they are based on continuous current. Nobody here is ever going to use continuous current on an ebike unless they took the bike on a highway and had OVS on their controller to actually push the cells to the max.

I guess that is why they have an acceptable rating. Kind of like an average of amp ratings you will use.

To figure out the current levels it will depend on how many battery amps your controller is set for, not the max rated discharge of the cells. Because your controller limits the current flow from your battery pack, and therefore each individual cell.

Then you will have to also determine how many amps flow between different areas on the pack, series connections may have more amps flowing through those areas than parallel connections. So you may very well have some nickel strips taking in more amps than others.

Click to expand...

Ok. That helps a lot, but could you give me an example?

thanks offroader

Can you give me your battery specs, like 18s10p . Also how many battery amps, or peak wattage you will be using with your controller.


An example would be if you have a 18s10p pack and you have your controller drawing 80 battery amps maximum. Assume 18s is 72 volts for your battery, you would have a peak wattage of 80 amps* 72 volts = 5760 watts.

Since you are drawing a maximum of 80 battery amps and have a 10p pack, each cell in your pack would draw a maximum of 8 amps each 80/10p = 8 amps.

Offroader said:

Can you give me your battery specs, like 18s10p . Also how many battery amps, or peak wattage you will be using with your controller.


An example would be if you have a 18s10p pack and you have your controller drawing 80 battery amps maximum. Assume 18s is 72 volts for your battery, you would have a peak wattage of 80 amps* 72 volts = 5760 watts.

Since you are drawing a maximum of 80 battery amps and have a 10p pack, each cell in your pack would draw a maximum of 8 amps each 80/10p = 8 amps.

Click to expand...

Ok, So my battery is a 16s8p pack with Samsung 25r cells. I'm not sure if the cell type matters. It's 60 volts and the controller with draw around 60 amps so my wattage will be 3600w. I'm not sure how wattage factors into it. So with an 8p pack at 60 amps I'll be drawing 7.5 amps per cell. Is that right?

Yes that is right. But since you vary the throttle you won't always be pulling the max continuous amount of amps.