"Copper/nickel sandwich" buses for series connections

Finished my new split packs the other day which ended up being 46% heavier than my old 18650 packs for 60% more capacity (nominal). Packs are 2x 10S8P Molicel P42A.

Connections: DIY nickel-electroplated .20mm thick copper H-strip + .15mm nickel-plated steel H-strip
Welder: kWeld @ 70J

Results:
P42A Pack 01.jpg

P42A Pack 03.JPG

Overbuilt for the battery current I expect to be running, but I didn't want to make the same mistake I made with my 25R packs running out of throughput due to the connections when I eventually wanted to run more power.
 
.20mm thick Copper H-strips.

Need to be careful about the spacing though. These H-strips weren't an exact fit for the one piece 8X10 ABS celI-holders/brackets I used, but the ridge spacing was large enough on the brackets that the full 8P H-strip still just fit without riding on the raised ridges if that makes sense. The same ABS brackets in 9X10 or larger would have resulted in these copper strips riding up on the 1st or last bracket ridges.

The reason I didn't get 23mm spaced brackets is because I was trying to pack the cells in as tight as possible in this format and found one that got them closer than this.
 
Looks very clean - nice job. I don't know what kind of load you'll be pulling but using the H-strips reduces current carrying capacity compared to full-length strips. It's good to know the Kweld can do that kind of work at 70 joules. If I can ask, what do you use for a current source for the welder?
 
john61ct said:
Thanks.

Got a HowTo on your DIY nickel electroplating?

NP - I mainly just watched a few youtubes and went to town.

Have to say it wasn't exactly a fun process, but I can offer a few tips now at least.

-You'll want to use either a magnetic stirrer and/or a surfectant to ensure an even electroplating finish. I went with the magnetic stirrer only. In the trial phase, I tried plating one strip without the stirrer running and the finish looked like shit (that said, I don't think I did a very good job controlling the purity of the nickel solution).
-I saw another video where they bagged the anode(s) in cloth to contain the undissolved nickel impurities when making the solution, probably a smart thing to do, but I ended up just using coffee filters to filter the solution afterward.
-I needed a fairly deep plastic container to allow the H-strips to suspend vertically in the solution and consequently had to run the magnetic stirrer at the full 2000RPM to get the solution whirl-pooling as it should.
-In the video linked above he only has the sacrificial anode on one side of the jar, my container was large enough that I found it useful to put them on opposite sides and parallel the anode connection. This way I didn't need to worry about turning the H-strips for even plating.
-The time to plate will depend upon the concentration of your nickel solution, you'll have to experiment to get it right so be prepared to waste some copper strips. Mine needed about 4-5 minutes. Let the plating go on for too long, and you'll get an uneven finish with nickel dendrites.
-With my setup/container, I was able to plate 3 strips at a time without them getting stuck together.
-I used probably way more salt than I should have which apparently makes for chlorine gas. Lacking access to a good workspace for this type of thing, I ended up running the nickel solution-making phase outdoors.
-It doesn't hurt to wear a proper respirator.
 
Frank said:
Looks very clean - nice job. I don't know what kind of load you'll be pulling but using the H-strips reduces current carrying capacity compared to full-length strips. It's good to know the Kweld can do that kind of work at 70 joules. If I can ask, what do you use for a current source for the welder?

Thanks Frank. For sure, I'm only counting the 8.8mm x .2mm series connection part which puts them a little larger than 14AWG neglecting any contribution of the steel strips.

I took both of the Trojan 6V 225aH house batteries out of my camper van, wired them in series, threw a 6A charger on it, and spot welded in intervals until the electrodes got fairly hot to hold, then allowed time for cooling and recharging.
 
Nickel-plating is typically done to reduce oxidation, and make the copper parts corrosion-resistant. Raw copper is especially sensitive.

In a dry climate, bright shiny copper may just take on a dull brown appearance. However, if it is near the salty air of the ocean, it will corrode badly with the copper oxides taking on a green rough appearance.
 
I figured that was probably the case but I'm also curious if there might be some slight advantage when welding. I like the idea...
 
spinningmagnets said it better than I could've. While they probably would've been fine without, my packs/connections are only partially covered and not fully sealed so I just wanted to minimize my chances for corrosion problems developing over time. One less thing to never worry about..

For what it's worth, I saw no difference in welding when I tested the unplated copper combination which makes sense as the plated layer is very very thin, contributing practically nothing to the spot weld resistance path.
 
Just did another massive pack out of Samsung 50G 21700. This time I placed the top nickel to aditionally support the series connections. Last time I kinda missed that, so only copper was in between the series connections.

Cells: Samsung 50G 21700 5Ah (real 4,8Ah)
Pack: 20S7P (140 Cells)
BMS: LLT smartBMS (xiaoxiang) JBD-SP20S002 150A (only 42A used)
Capacity: 35Ah
Cellholders: generic 13-holders, custom cut
Case: sheet metal, partially powder coated

Welder: kWeld @ 110J
Battery: generic car batteries
Connections: 0.15mm Nickel + 0.15mm Copper

It's meant to be folded together with 3mm PVC foam sheet for isolation:



That's where it's meant to live in: [base: MINI DIRT BIKE XTDE PRO]
 
BEAU*T*FULL ... 20, 19-18, 17-16, 15-14, 13-12, 11 <> 10, 9-8, 7-6, 5-4, 3-2, 1 ... :bigthumb: ... (for both series and parallel connections)

Explain reasoning for using both "0.15mm Nickel + 0.15mm Copper" connections. Think i know, but not sure :?:
Does it have something to do with difference in voltage demand between parallel and series connections :?:
 
Copper conducts electricity with such low resistance, that it doesn't get hot enough to melt. The nickel has
"enough" resistance that it spot-welds easily, but it is a poor conductor.

The parallel strips don't need high conductivity, because all they do is balance the cell voltages. The parallel connections will flow less than 1A under all conditions. They can be copper, but copper is not "necessary".

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 hot enough to melt a spot of copper to the cell tip.
 
My limited understanding was if the controller is asking 40A from a battery pack made of 18650-25R (2.5Ah, 20A-max) with no BMS (or a BMS that is 45A or 50A, meaning the BMS factor is irrelevant in this case study) then a pack made of 10S2P is 36V-nom (42V-max) 5Ah 40A-max you are pulling 20A from each P group with 10 18650's in each series, it is really 2A from each cell, 10A from each Parallel string made up of 10 cells.

I think what you are saying is that you could build a pack that pulls 20A from the cell, so in the case above it would be a controller asking 400A from the battery pack with no bms or 500A BMS, again BMS being meaningless here, the battery needs 200A from each Parallel string with 10 18650's in each which means 20A from each cell.

The second situation the battery will not last long pulling 20A from each cell, the first situation the battery will last longer because you are pulling 2A from each cell. The lower the better, which is why building your own battery you must keep that in mind, along with the other factors, like draining the battery half way all the time then charging it compared to draining the battery all the way to minimum levels all the time. The biggest killer of batteries is storing battery at top voltage or keeping the battery at top voltage for to long before riding.

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.

Reading SM statement again I understand more clearly. The series string carrying amps would be divided upon how many parallel strings if everything is equal and it all depends on how your battery is built. In the situation above, the first would be a current of 20A on each series link, second situation would be 200A on each series link. Some batteries can push out very high amps like A123 LiFePO4 26650 2.5Ah, 20C(50A-max), Lipo 4Ah 60C which is taken from Hobbyking which is exaggerated 60C is 240A-max, 120A-cont.

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:

spinningmagnets said:
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.
 
This may sound odd, but...the battery pack would likely work just fine of it had no parallel connections, just the series connections. Over time, the cells would slowly get out of balance.

The Samsung 50G cell shown is rated by several internet sites as a 10A cell, so it is optimized for range. Temporary peak amps are higher than 10A. This pack is 7P, so the pack should be able to provide 70A fairly continuously.

The series connections for each cell should be able to handle 10A, and the series plate that connects a group of 7 cells to the next 7 cells should be sized in area and thickness to handle 70A.
 
Apparently the objective being ("0.15mm Nickel + 0.15mm Copper") speed bursts of at least 55mph ... :wink:

What myself and others would've liked to see is the not as Beau*T*Full other side showing the negative and positive terminal connections and guage of wire planned on going to the Controller) ...

s5YQSDOl.png


With its pack purpose being speed bursts (and endurance) what neg / pos terminal bonding method is best OR is high temp solder still OK for neg and pos terminal connections to copper wires on this high energetic pack ?

What if the cells were rated at 30A bursts and the Controller could handle that many pack amps. Would you still use high temp solder? What bonding method (neg / pos terminals) is better than soldering?
 
I wonder if...

very fine stranded tinned copper wire, e. g. UL1426 105c tinned "boat cable" type 3 rather than type 2

can be say 00 AWG or even fatter

Strip of a longer length of insulation than usual so you can "splay out" the fine strands over the copper plate to only one layer thick, or 0.010"

Cover with the nickel strip, maybe some conductive paste / adhesivr/ potting compound to keep the strands in place and fill in the gaps

and spot-weld the wire to the copper plate through the nickel

rather than solder.

In either case the insulated part close to the fastening point needs to be fixed, well supported against any movement

because the transition from the join point would be very vulnerable to strand breaking from the cable flexing.
 
The positive and negative terminals are at the top left and top right of the original picture. You can see the positive ear fairly easy. I have over-drawn where the bus-plates on the other side are.

BatteryCNS2.jpg
 
That overdrawing method is a superb way of clarifying how the current flows!

Should we assume that all the cells' orientations wrt pos/neg & up/down are the same?
 
I'm also curious about connection techniques to move amps from the pack to the load. Boat cable is good as it's very flexible and (I believe) often pre-tinned and I've often used it in other packs built with pouch cells or prismatics. I'm planning a 13P pack and wonder if a sub-optimal connection might load some cells more than others.
 
The following explanation is for a scenario with 200A Controller (20s7p -- 30A cell burst rating) for speed bursts of 5-10 seconds (at least 55mph).
john61ct said:
very fine stranded tinned copper wire, e. g. UL1426 105c tinned "boat cable" type 3 rather than type 2 ... can be say 00 AWG or even fatter.
Strip of a longer length of insulation than usual so you can "splay out" the fine strands over the copper plate to only one layer thick, or 0.010"

Cover with the nickel strip, maybe some conductive paste / adhesivr/ potting compound to keep the strands in place and fill in the gaps

and spot-weld the wire to the copper plate through the nickel

rather than solder.
Has anyone used the above technique and by chance have a closeup photo they could post with reason why soldering wasn't the way to go ... or ... a link to another website or ES thread w/photo and further explanation using the bonding technique explained above instead of soldering.


Appreciate the photos and comments that Thunder camel posts showing progressive build technique. Here he soldered the copper wire to the neg terminal (14S6P LG MJ1 -- 35mph top speed bursts).

"Pulling my negative wire straight down earned me a boo boo. I layered another nickel strip over that."
(could someone further explain boo boo and why he felt it necessary to layer another nickel strip)
yorkCeORh_H3Tt02-jc2HaNjzvrD4D2S67OupkWNPt9nzDu3NWRwjjtAVMbEpHhOLvyvRmjU61_VEquyDy8UeCzsgU07T1M4773FsQD8yrTlZm8dtavHn37X0VWlut2cc2ihk-qRYXI8CbAE-6jUbCPzrtCryeFkxHAv5lCG1P13OSooawL9VlShnKW2mNor-3DUDWIftwdY2BLXK8EpVKdVtf6bWtX2F65il6QwAftOx3JpNAt0r2br4gWjXk5pln9MqTDujAaAEelifHM23_mDsHoM5LyEqXwz6LcW2YT0uMQ5S5zwyDm-cb_3iEmg1RaZG8TXmCiptWWJYItEX5fmcafIY26Wn8a-7--Pj2heMypSohVrRfjLKGXXoY4k381ums9Ie_ZUScx9GwJwhIH_tmNmdG3pUKA0y_YZNNjKfuWcRbMn-9t23T6j7HS4Rn-ExJN03yV-8pvFDBldwTjVBJ0C2fID_xNlUS8HN__adsRZc4KCqxY9oDcmFSI8tlcRce7NJstFXMv2QaiYdopObM_0igcFDNlnqZvC7eDFtn2I2UMljIQNXe1tBTaAU9mO_6Uh7VxtBpFc_iNUFYu05iO4u20kaZqjRBeQYWI30TC9yzm-b9jyWDI-W0lSgWRiyCjR5IcaYrnPM-xrqxw5lfq_bNhlJZjf14kBUmbDYi4uV-mdz4SjZLAnlAVRhpM-rg8QAhHNMq0bUrdHz3YB=w4160-h2340-no


Frank said:
I'm also curious about connection techniques to move amps from the pack to the load. Boat cable is good as it's very flexible and (I believe) often pre-tinned and I've often used it in other packs built with pouch cells or prismatics. I'm planning a 13P pack and wonder if a sub-optimal connection might load some cells more than others.
Does anyone have a closeup photo they could post showing the bonding technique described above by john61ct ... and perhaps further explain ... "wonder if a sub-optimal connection might load some cells more than others".
 
The only thing worse than speaker wire on a battery build is copper clad aluminum speaker wire on a battery build. Thats beyond cringy, thats a fire hazard.
 
TrotterBob said:
The only thing worse than speaker wire on a battery build is copper clad aluminum speaker wire on a battery build. Thats beyond cringy, thats a fire hazard.
You're not aware that speaker copper wire comes in 12 guage ... https://www.skaraudio.com/products/12-gauge-oxygen-free-copper-audio-speaker-wire?gclid=Cj0KCQiAieWOBhCYARIsANcOw0yDOMUKw7GXU_J_kF-DmBqmM4pqJwASomPSXRPtE0rT3229rfzQcHwaAl1CEALw_wcB

Do you have a link or photo to share/post about the above spot-welding of copper wire bonding to terminal as described above by john61ct. It would be helpful if you could provide further explanation to Frank's previous question (see underlined portion) ... Thanks :)
Frank said:
I'm also curious about connection techniques to move amps from the pack to the load. Boat cable is good as it's very flexible and (I believe) often pre-tinned and I've often used it in other packs built with pouch cells or prismatics. I'm planning a 13P pack and wonder if a sub-optimal connection might load some cells more than others.
In other words 12 guage copper speaker wire could/would be sub-optimal for his 13P pack ... so what bonding technique would you recommend (post photo or link) if his pack were 20S with a 300 amp Controller.
 
There are two things to notice in the pic below. 1) The negative ends of the cells look smooth and show what appears to be bare metal. The positive ends of the cells have the non-conductive fiber washers, which are also heat-resistant.

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.

Doubling up on nickel strip in that one area will prevent the bus-strips from melting, but by using nickel on both of them, you are converting significant battery-watts into waste-heat.

BatteryCNS4.jpg

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. The drawing below is 5S / 2P

BatteryCNS5.jpg
 
Back
Top