"Copper/nickel sandwich" buses for series connections

[john61ct]

Thanks!

Here is a quick and crude drawing that I hope will show how to figure out what the bus-plates on the other side of a battery will look like.

Assuming only one layer deep on the Z axis, that is 2P5S there?

 

[spinningmagnets]

Yes, 5S / 2P.

In the picture, you can see which cells are positive by looking for the fiber washers.

 

[eMark]

What's the largest guage of stranded copper wire that you've soldered to a nickel/copper bus bar before going with another bonding technique like that described by john61ct.

When using a copper bus plates have you ever had any success first soldering the copper terminal wires to the copper bus plate first ... then cleaning/preparing the bus plate and the cell surface for spot-welding a copper terminal bus plate to the cells.

 

[spinningmagnets]

You can bond thick copper stranded cable to the buses with high heat of you like. Let it cool, and then weld the bus to the cells. Part of the reason I like the "copper nickel sandwich" method is it doesn't transfer much heat to the cell ends.

I used thick welding cable for a 12V jump-starter suitcase, and the adaptation from round-cross-section stranded copper cable to flat tab was accomplished by inserting the cable into a tightly-sized copper water tube, crimping, and then the outer half was flattened and cut on one side to allow the flat tabs to slide in.

The drawing at the bottom left is a cross section of the crimp. I drove a dull screwdriver into the center of the copper tube, then put the tube in a vice to smash the top and bottom together. If you want to add solder, then one option is to use flux and solder paste on the copper strands before insertion, then heat after crimping.

I don't have a pic of that yet. Try whatever method appeals to you, and post the results.

https://endless-sphere.com/forums/viewtopic.php?t=101153

 

[john61ct]

Properly crimping the boat cable to Ring terminators and drilling/punching a hole in a string bit of the metal plate would allow for a high-torque join.

Be wary though of regular nut/bolts loosening from vibration, but plenty of proven solutions there.

 

[Frank]

Properly crimping the boat cable to Ring terminators and drilling/punching a hole in a string bit of the metal plate would allow for a high-torque join.

This is what I've done for smaller packs, but where would you attach the ring terminal in a "high P" pack? In the middle? At the "north" end of the pack for the +'ve and "south" end of the pack for the -'ve? In other words, how to ensure that cabling does not inadvertently load some cells higher than others?

 

[john61ct]

I do not think that is possible using these plates / cross-strips methodologies.

If the copper conductive plate gets thick enough that resistance is super low, would help.

It's a challenge even with cells only connected by wires. http://www.smartgauge.co.uk/batt_con.html

 

[eMark]

In this case, the washers did exactly what they were designed to do, they prevented a "shoulder short" when the nickel strip melted from high local current. Without the fiber washer, the hot nickel bus-strip would have melted the green insulative sleeve, and caused a short between the negative shoulder and the central positive nipple.

GOOD EYE :thumb:

Looks like there may have actually been a shoulder short on the lower left cell ... causing two dead cells ?? ... YES or NO ...

What i really appreciate is a builder the likes of thundercamel and others that always show both sides of their DIY pack build ...

The top photo shows the Postive and Negative terminal side of thundercamel's 14S6P LG MJ1 battery pack. Some DIYers only show what they consider the more Beau*T*Full side. Doing so doesn't tell us nearly as much about their build as do the progressive build photos by thundercamel and others that are revealing the other more interestng (and instructive) side that's just as beautiful in its own way :thumb:

Thumbs UP to thundercamel for always showing us the more interesting, revealing and instructive side of his battery builds ... :thumb:

My original Beau*T*Full (tongue-in-cheek) post was that the owner apparently just wanted to show-off what he considered it's best side. His 20s labeling was OFF when he began by referring to the first series of his 20S7P as "0" instead of "1".

20, 19-18, 17-16, 15-14, 13-12, - 10 -, 9-8, 7-6, 5-4, 3-2, 0
Was he even the builder with his wrong labeling with "0" instead of "1" and "10" instead of "10" & "11"? A true DIY builder would also show the more interesting and revealing side of his battery pack. Not just show-off what he considers it's more glamorous BEAU*T*FULL side as if that's the most important side.

The following instructive photo my Ron gives one an idea whether or not one is still a (wet-between-the-ears) newbie. I'm still a newbie because i wasn't able at first to figure out which was the negative and positive terminal depending on the side. I assumed that 1 & 20 are the negative and positive terminal, but wasn't sure until Ron's instructive photo ...

That aside Ron is assuming that the side showing in the above photo is the terminal side, but we don't know that for sure until we see the other side. If the negative and positive termnals are already located on the other side then the left side of the battery in the above photo wouild be the negative terminal connection (on the other side) and the right side the positive terminal connection (on the other side).

Because of the OPs wrong numbering with "0" instead of "1" and "10" for both "10" & "11" the terminal connections (for all we know) may already be intact on the other side of this battery pack. If that were indeed true then the numbering would be from left to right ... 1, 2-3, 4-5, 6-7, 8-9, 10 -- 11, 12-13, 14-15, 16-17, 18-19, 20

The other problem with the BEAU*TFULL pack is 4 of the 20s groups on each side have only 2 series connections between them and the others only 3 series connections between the remaining 20s groups. Thundercamel's battery design has improved current flow having 7 series connections between each of the 14 series groups of 6p ...

Thus the downside of some triangular pack layout designs or that of an eye-catching BEAU*T*FULL butterfly layout design ...

Thundercamel's battery layout design is superior to that of the BEAUT*T*FULL butterfly layout design. The most efficient battery pack is the one that has as many series connections as parallel connections. So thundercamel's layout design is the winner when it comes to being more efficient for energy flow than the BEAU*T*FULL butterfly battery pack ...

From left to right ... 14, 13-12, 11-10, 9-8, 7-6, 5-4, 3-2, 1 ... 14s6p with as many series connections as parallel connections :thumb:

AND Its Easy To See Where Each Of The Plates Would Be ... if one were to instead spot-weld copper bus plates. WHY even consider copper bus plates IF the outcome isn't a more efficient battery pack that gives the operator the most c/d life cycles possible ...

 

[eMark]

That aside Ron is assuming that the side showing in the above photo is the terminal side, but we don't know that for sure until we see the other side. If the negative and positive termnals are already located on the other side then the left side of the battery in the above photo wouild be the negative terminal connection (on the other side) and the right side the positive terminal connection (on the other side).

Late last night the light came on and i felt so stupid realizing that the side showing in this photo posted by Ron ... DEFINITELY ... has to be the terminal side. No need to see the other side. Now it seems fairly straight-forward to figure out which side of the battery build photos shown in this thread is the terminal side. Kind of embarrassing (once the light came on) that it's not all that difficult to figure out which is the termnal side of the copper-clad build photos in this informative thread .

My conclusion (having followed this thread) is that copper bus plate series" build technique is reasonable IF ...

The [only] reason for doubling up with both nickel bus bars and copper bus plates is for "Raw Performance" when pulling mucho amps. Otherwise why go to the extra work and expense.

This cooper-clad method requires larger size copper cable wires ... yet only one photo in this entire thread pointing this out ...

Choosing and attaching the right size cabling is VERY IMPORTANT and yet very little mention of it with photos. It's as if everyone is enamored with their copper bus plates more than the cabling that is needed to handle the higher amperage and whether or not to solder or use another bonding technique that hasn't been discussed until my recent posts and Frank's post with Ron's followup post.

Please correct me if wrong : The main or maybe ONLY reason for using both 0.15mm nickel bus bars with 0.15mm copper bus plates is for more efficient (at least 150 amp) "Raw Performance" ... otherwise doing so is not worth the extra expense and labor intense [meticulous] time consuming effort ... as evident in three of the following beautiful copper-clad battery packs ...

Terminal Side. Would have liked to see a photo of the neg and pos cable connected to the copper bus plate ...
(or maybe i should have said what looks more like copper foil than 0.15mm copper plate) ?

20S6P Non-Terminal side. Can you know for sure which will be the Pos & Neg terminal group from this side ?

The two groups at the top are the end-collector cells for the positive and negative cables to the controller. If you look close, you can see the top right bus has the edge bent over the top, and there is a horizontal copper plate that will somehow have the positive cable for the whole pack connected to it.

Sure would have liked to see a photo of that positive cable connection. If it is 24S wouldn't you need 24 balance leads from the BMS ?

My only defense for my recent posts is being 77 years young and still interested in learning ... WHY and HOW ... even when wrong.

So can we conclude that 0.15mm copper bus plates in combination with 0.15mm nickel bus bars are worth the extra effort and expense in the long run for both efficiency and endurance when pulling 30A bursts from your 20A rated 18650 cells ?

 

[john61ct]

I never heard of a pack having a "terminal side".

If you mean where the external power leads are attached, your idea of a "terminal group" is valid I guess, but the negative lead is just as likely to be attached on the completely other side of the pack from the positive.

Depends on whether the S-count is even or odd I suppose.

Remember, cells are not groups, even if they are the same voltage.

The fact that high amps requires lower resistance, thicker plate and fatter AWG should be obvious, and really deserves its own thread.

And of course, for low C-rates it would be a silly waste to use the same "fat" materials required by high C-rates. Especially with the price of copper these days...

 

[eMark]

The fact that high amps requires lower resistance, thicker plate and fatter AWG should be obvious, and really deserves its own thread.

MUCH needed being that there seems to be a consensus that soldering on these high performance copper-clad packs pulling mucho amps may be frowned upon by some builders. Does anyone spot-weld their Neg & Pos cable wire connections instead of soldering ? Do you have a closeup photo (spot-weld cable wire bond) on one of your copper-clad DIY build that you've previously posted on this thread ?

Isn't it just as important to discuss and post closeup photos of how the builder attached the Pos and Neg cable runs? Even moreso on their high performance copper-clad packs whether soldering or spot-welding when bonding the Pos & Neg cable wiring connections.

My take FWIW (having followed this thread) is that copper bus plate series" builds with 0.15mm nickel bus bars is reasonable only IF ...

The [only] reason for doubling up with both nickel bus bars and copper bus plates is for "Raw Performance" when pulling mucho amps. Otherwise why go to the extra work and expense.

This cooper-clad method usually requires larger size copper cable wiring ... yet only one photo in this entire thread pointing this out. Would be helpful discussing and including closeup photos showing cable wiring to the Pos & Neg connections using spot-welding if in fact it's a better bonding technique. For all the extra effort and expense of a copper-clad pack with nickel bus bars then the correct bonding of the Pos & Neg cable wiring should not be minimized as it seems to have been in this thread.

This is the only photo in this entire thread that even shows the cable wiring and even then only as a visual after-thought ...

Choosing and attaching the right size cabling is VERY IMPORTANT and yet very little mention of it in this thread and no closeup photos of Pos & Neg connectins to copper-clad bus plate.

It's as if everyone is enamored with their copper-clad bus plates more than the wiring needed to handle the higher amperage and whether or not to solder or use spot welding on the copper bus plate. John's suggestion is well-worth considering ... start another thread on spot-welding cable wire connections on the Pos & Neg copper bus plate when soldering isn't sufficient. Or does most everyone building high performance packs prefer soldering instead of spot-weld bonding of the Pos & Neg cable wire connection ??

So from this thread we can conclude that 0.15mm copper bus plates in combination with 0.15mm nickel bus bars are worth the extra effort and expense in the long run for both efficiency and endurance when pulling 30A bursts from your 20A rated 18650 cells ?

Perhaps it's worth reading once again portions of the previous thread post by markz...

When I was making home made battery packs I would use 10awg solid core house copper wire for the series and the 14awg solid core house copper wire in the same cable for the parallel. There is a building design question where the Makita's were 5S2P so you would put two in series to make 36V 10S2P, then you either parallel the positive and negative ends to attach another 10S2P to make 10S4P or you parallel each Parallel group. I had 10S8P in 10S2P groups and did it both ways. Parallel each parallel group was not successful for me as it could kill the entire 10S8P pack rather then on block of 10S2P. That was what happened to me and I was not keeping track of balance then.

Design of the battery build needs to look at all factors including the resistance of the parallel and series link, positive and negative leads of the battery pack itself, connectors have resistance, mechanical or solder connections of connectors, cells themselves have internal resistances which all add up. Look at it from an over all view bottom up, top down. Crappy cold solder joints, some generic crappy connector, net enough tab welds on the tab, undersized series/parallel gauge conductor will not be good.

One last comment is to make sure you have good insulation protection to protect the end of the cell from touching the tabs.

Be sure the series connections are the right gauge for the amps running through them, controller rating, if your battery has a bms take note of those specifications. Always keep that in mind when changing the controller because you dont want the controller to ask to much from the battery and you dont want the bms to always kick in. The less stress on the battery the better :thumb:

Controllers can be cheap, batteries are not cheap even the cheap batteries are not cheap and why would you roll the dice on a cheap, back alley, basement built, unkown, p.o.s. someone bought on ebay, alibaba, aliexpress? Yes there are some reputable builders who sell on there and nothing wrong with buying from a known, reputable entity but it will not be cheap like the eye catching rock bottom blow out sale batteries made with who knows what, counterfeit, faked components. That is why people build their own :thumb:

The series connections flow the full amps per-cell, which can be as high as 20A per cell for fairly common cells (30A temporary peak per cell is possible).

If you put nickel on top of the copper, the nickel gets hot enough to melt in a tiny spot, from the spot-welder. The nickel-heat is hort enough to melt a spot of copper to the cell tip.

In all the copper-clad photos it appears the copper is laid over the nickel bus bars. Are you saying it's better if a 0.15mm nickel bus bar is on top instead of under the copper plate or copper foil ?

 

[Frank]

For sure the copper will reduce any losses due to higher amps. I like to think in terms of wire size, for example, I'll be building a pack of 21700's that will have to occasionally generate 450A for short bursts (seconds). This application is currently (sorry) well-served by #2 cable, so I know that works for the higher draw even if it's oversized for regular use.

#2 has a cross-sectional area of ~33.6 sq. mm. If I use .15 copper I know that to obtain the same cross sectional area, I need to have a "series conduction length" of 33.6/.15 = 224 mm. If my cell holders space the cells 22.7 mm apart I will need 224/22.7 ~ 10P just to have adequate conduction area for those short bursts (assuming no "bunching" of P-groups as shown in the pics above). Then I have to make sure my cells can handle that load: 450A/10P or 45A per cell. I'm actually using 13P so I'm pretty sure my cells can handle the load and that there will be adequate conduction area.

I'm still working on the best way to make sure I don't inadvertently load certain cells more than others but I'm confident a good solution will present itself. :thumb:

 

[spinningmagnets]

I haven't discussed the transition from the terminal buses to the fat pos/neg cables, or the attachments for the BMS because that is another subject, and other threads can discuss that. Low-amp battery packs seem to work just fine with 0.20mm pure nickel buses, so if a builder is happy with that, then I won't spend any time trying to change their mind.

If anyone is building a battery pack that they hope to draw high amps from it, I recommend studying the copper/nickel sandwich method. It's almost impossible to weld only pure copper to the end of a cell (laser copper welders are expensive), especially if using the common affordable spot-welders.

If trying to use a "double layer" of nickel to increase the bus capability to run higher amps, it is my feeling that it would cause too much voltage sag and heat at the cell tips. The positive cell-tip suffers the most from this. This causes a lot of battery watts to be converted to waste-heat instead of applying those watts to spinning the wheel...

Nickel strips are expensive and have to be ordered to be delivered. Copper sheet is actually fairly affordable, and can sometimes be found locally in major cities. Only a tiny square of nickel needs to be placed over the cell-tip to create the local heat-spots to accomplish the welding.

Even if your supplier of copper-sheet charges more for copper than for nickel, copper is four times as conductive, and has very low resistance. This means that for performance, the pack would have very low voltage sag. The low resistance of copper means that there is very low heat at the buses when running, plus...the copper acts as a heat-sink to pull heat away from the cell which makes the copper bus act like an air-cooled radiator.

 

[Tomblarom]

...just wanted to show-off what he considered it's best side. His 20s labeling was OFF when he began by referring to the first series of his 20S7P as "0" instead of "1".

I was in a rush. This was a one-night build (one of many!) and I finished it in the early morning hours. Rarely captured any pictures, beside them. Stick to the ones I provided and appreciate (or not) my efforts wanting to share my experience & knowledge! Thanks

Was he even the builder with his wrong labeling with "0" instead of "1" and "10" instead of "10" & "11"? A true DIY builder would also show the more interesting and revealing side of his battery pack. Not just show-off what he considers it's more glamorous BEAU*T*FULL side as if that's the most important side.

I started at 0, because the BMS is actually labeled like this (and I do a lot programming 8)). The commonly used TI chips themselfs (e.g. BQ76940) are labeled like this. Have a look into the datasheet!

..but we don't know that for sure until we see the other side

This is false. Having an even cell-count means + and - are exiting on the same side. The opposite applies to an odd cell-count: + on one and - on the opposite side - Check for yourself, there is no other solution. --> Therefore it's possible to determine the opposite side by having only the provided side.

Additionally you can count the weld-connections. 20/+ and 0/- only have 7, because of 7P.

So thundercamel's layout design is the winner when it comes to being more efficient for energy flow

You totally missed, that there are 0.15mm nickel, additional to the 0.15mm copper connecting 4 parallel cells. Sure designwise his layout may be better, but I had to work with what my requirements were: fast solution, robust and range (not power).

And btw, this is how the butterfly looks folded:

 

[john61ct]

I started at 0, because the BMS is actually labeled like this (and I do a lot programming 8)).

Starting at zero is a very common convention in many fields, and especially with BMS where the total count of balance wires is +1 the cell/group count.

Not a mistake, and just ignore statements from anyone who says so.

 

[mbgjt1]

Hi all

I have the Sunkko 709AD.

Will I be able to spot weld 0.15mm pure nickel ontop of 0.1mm copper sheet?

I don't have a kweld unfortunately.

The sunkko is rated to be able to weld 0.15mm pure nickel but not sure how the copper will affect this

 

[ossivirt]

Hi all

I have the Sunkko 709AD.

Will I be able to spot weld 0.15mm pure nickel ontop of 0.1mm copper sheet?

I don't have a kweld unfortunately.

The sunkko is rated to be able to weld 0.15mm pure nickel but not sure how the copper will affect this

Sorry but I dont think you have any chance. If you have the materials then you can try and post results.. Anyway if you try 0.1mm nickel plated steel and 0.1mm copper there is a chance off somewhat decent welds but you have to try to figure that out.

 

[mbgjt1]

Yes I'll give it a go with nickel plated steel

Have to order and wait 4 weeks for the material to arrive !

But the sunkko can do 0.25mm nickel plated steel so maybe it can do 0.15mm with 0.1mm copper

 

[spinningmagnets]

The thinner 0.15 "nickel-plated steel" caps for welding would be better than the thicker material. Thin steel takes fewer amps from the welder to get hot.

I suspect the 0.10mm steel would be even better.

 

[mbgjt1]

The thinner 0.15 "nickel-plated steel" caps for welding would be better than the thicker material. Thin steel takes fewer amps from the welder to get hot.

Thank you,

The Sunnko would output 800A. and a 5ms pulse time, which means it can output a max of 40 Joules. Some people are using that power in the kweld to spot weld 0.15mm nickel plated steel to 0.1mm copper, so it may just work.

Fingers crossed! I've got the Sunkko 709AD 220V version which is said to have abit more 'umph' than the 110V version

 

[Voltspa]

Thanks so much for this thread...this forum is great.

I just got a K-weld and tried it out with 0.1mm copper and .15mm nickel plated steel at only 53 Joules with a 12volt 450 cranking amps battery on some old AA cells since that's all I have for now, and it worked great! Very strong welds, can't peel off with my fingers, ended up ripping the copper trying. I think I will try some 0.2mm copper for my build.

I'd really appreciate any constructive feedback on my cell layout. 20s12p using Samsung INR 25R 18650s. Since these are rated at 20A continuous I feel like I need a robust design and that copper is a good idea for keeping resistance down. I plan to use copper sheet to cover each grouping. It is in this shape because I am fitting it into a round pan that will be a fake spare tire! 8) I'll be powering a 5kw hub motor in a 1974 Vespa 125 Primavera. I'll start a thread for that soon.

What do you think about this spiral layout?

View attachment Circle 15%22 240-18650.pdf

 

[Amperry417]

Hey Guys,

Have been following this thread and many other battery building threads for a while, educating myself for my battery build. Have bought many different materials in preparation for testing and ultimately my final selection of materials. Last night did my first test welds with copper and wanted to share my findings and settings.

Welder being used is Malectrics V4 with 12V PB Marine battery. 875 CCA at 0°F charged to 2.4v per cell /14.4V.

The pictures show .1 copper and .1 nickel plated steel tabs at 27-29ms. Extremely happy with the results. 27ms makes a good weld that tears but very slightly inconsistent. 29ms will bend the battery can/positive button and the welds are insanely strong and out of 20-25 welds, every one was great.
I also successfully welded .15 copper using the same .1 nickel plated steel tabs at 30ms but after a handful of test welds, the results were somewhat mixed. Some very strong and tearing welds and some that didn't melt so great. The ones that didn't weld perfectly were still in the vicinity of acceptable. Not really necessary for me to use .15 and introduce excessive heat. However, the .1 copper gave the most consistent results.
I feel it's better (in my particular case) to use the .1 with consistent results than the .15 and guessing if/when the welds are spot on.
Another interesting note is regarding the tabs for welding aid. I first tried with .15 pure nickel slotted tabs. Welds were good, more than acceptable. Then I tried the .1 nickel plated steel (no slot) and the results were dramatically better on the same settings. So, the plated steel is a much better welding aid than pure nickel, no question. That has been confirmed previously and it was my experience as well. Adding a slot to the plated steel would increase it's "aid" function. Going to test with slotted plated steel tabs this evening.
The welding area was warm to the touch as fast as I could get my finger onto it, not "hot" in any way. This leads me to believe the 29ms setting is safe for the cells and the copper is helping to dissipate the heat quickly.
I'm using Samsung 48X cells, so will be in neighborhood of around 15-17a absolute max from each cell, more like 8-12a continuous. Using the copper (rather than nickel) for heat dissipation, low resistance and low voltage sag reasons. My application is a scooter, so battery is sitting inside the cast aluminum deck with zero flex and minimal vibrations. The mechanical weakness of .1 copper isn't as much of a concern as it would be in an off road e-dirt bike or esk8 with deck flex. My battery will be secured in place firmly.
The size of the parallel/series "plates" are large. Each plate is 165mm wide. I have 9p and 9 series connections by default of using the "plate" rather than strips. 165mm wide plate / 9 cells/connections = 18.3mm width for each cell to cell connection...correct? In any case, this should be more than ample for the battery I'm building and cells I'm using.
Will post some updates once I start getting the actual battery welded up.
I'm a noob but have been planning this project for a while and have a fair amount of problem solving skills and common sense. I try to apply what I've learned and take precautions at each step to make sure I don't look back and wish I would have done something very differently.
Thank you everyone who has contributed to this (and other) topic. It has helped me find my way and educate myself greatly. Any feedback/comments/criticisms/questions are more than welcome.
**Side Notes**
--The plates in the pictures have since been trimmed back. I realize they are extremely close to eachother in the picture I posted. I opened up the notches a bit with a round sandpaper drum on a Dremel to allow more breathing room between the plates themselves. They're also sitting loose on the battery, so they may have even shifted slightly. Before I weld anything permanently, I will be sure there is plenty of headroom between the plates!
--On the Malectrics V4, if you hold the foot pedal pressed after the weld, it will show you how many amps are being pulled/used and at what voltage. Even though my battery is 875CCA, the readings are coming back around 550a of actual welding power. Not sure if this is loss from battery to welding tips or if it's an approximation. In any case, the welds on .1 copper are great. Something seems a little off with those numbers however. At 550a * .029 that works out to be only 16 joules. Using the battery rating of 875a * .029 that works out to be around 26 joules. Maybe it's the nickel plated steel that's allowing for such low power and good welds? Or... my math or readouts are way off. Again, not much of concern because the results speak for themselves, I just found that interesting.

 

[spinningmagnets]

Well done!

 

[Voltspa]

Ok let’s try this again…

I got some 0.2mm copper and received my Samsung INR 25R 18650 cells.

I did several tests with 0.1mm nickel plated steel over 0.2mm copper and different power settings on the K-weld:

105 Joules was just OK, one side was good and one side only welded the tiniest spot.

110 Joules was great on the positive button, but again not great on the negative end. I also tried a split nickel/steel over the copper and that was OK but not great.

115 Joules is better, but then I tried a split copper under nickel plated steel and that was very nice!

The K-weld read that I was using 1180-1204 amps during these tests.

 

[Frank]

Split copper, great idea; I don't know if anyone's tried that before. ??

It would even be possible to maintain series-current-capacity by using a thin copper strip placed towards the outside edge of the cells.