(semi)-solderless design: Attaching cells with solder tags using brass bolts

kilou

10 W
Joined
Aug 25, 2014
Messages
87
Hi,

I'm investigating a semi-solderless design where 18650 cells are bought with soldering U tags (8mm x 0.15mm thick pure nickel) already spotwelded but these tags are then attached together to form small 2P groups using M2 or M3 brass bolts. The idea is then first to combine these 2P groups into larger (multiple of 2) parallel groups using a thin copper bus bar (e.g. 0.5mm thick, 10mm wide) and second, make the series connections using thicker copper bus bars (e.g. 1mm thick, 10mm wide). The parallel connection would sit at the bottom while the series connection would sit at the top, with the solder tags from 2 parallel cells sandwiched in-between. Everything being compressed by those M2 or M3 bolts.

This method would involve drilling a 2-3mm diameter (elongated) hole in the nickel strips attached to each cell to accomodate the brass bolt. On a 8mm wide strip, this would leave 6-5mm of nickel thus slightly compromising the current carrying capacity of the strip. On the other hand, the brass bolt is conductive and should compensate for that as it "fills" that hole.

Some advantages:
1) no need to buy a spot welder or solder cells. Cells are bought with U tags already professionally spot welded (e.g by nkon.nl). It is complettely solderless for the user but semi solderless in reality since U tags are spotwelded on the cells.
2) pack can be assembled/rebuilt easily: only involves unscrewing a few bolts and nuts as with the Vruzend kit (less bolts actually, see 3)
3) fewer connection than Vruzend kit (2 cells are joined with only 2 screws while Vruzend would require 4 screws, one on each terminal of each cell)
4) 100% reliable cell contact since the U tags are professionally spotwelded
5) Relatively cheap: only involves a modest fee for buying cells with U tags (+0.5$ per cell). All the rest is easy to find (copper sheet, brass bolts/nuts, cell holder etc) and DIY

Some disadvantages:
1) you need to buy cells with U tags, which may limit vendor choice and slightly increases cell price
2) since one end of the U tag is attached to the cell and the other end is attached to the copper bus bar, the limiting factor for current somehow remains the size of the U tag (no matter how thick the copper bus bars is). On the other hand the length of the tag is pretty short. In my application, each cell should provide max 5A (10S6P with max 30A) so 8mm x 0.15mm U tags appear sufficient.
3) only works for even number of cells in parallel (2P, 4P, 6P etc) but something similar could potentially be done for odd number of cells by using Z tags instead of U tags I guess.
4) requires drilling holes in the nickel strips (U tags) to accomodate the bolts= takes some time but is easy with regular tools.

Some questions:

- Do you see any issue with this design?

- Is the hole compromising anything electrically (structural stability provided by a cell holder)?

- Is an M2 or M3 brass bolt sufficient to provide enough compression for good electrical contact between the copper parallel/series bus bars and the u-tags nickel strips that are attached to the cells? Max tightening torque for such small brass bolts would be pretty low I guess and some threadlock will probably be needed (best would be nylock brass nuts but I don't know if they exist in M2-M3)...

Would be interested to know whether this approach has already been attempted and would be sensible.Thanks!

EDIT: I will add a scheme later on to illustrate the design
 
kilou said:
Some disadvantages:
1) you need to buy cells with U tags, which may limit vendor choice and slightly increases cell price
To me that is the fatal flaw.

Hard enough to locate top notch quality cells and a trusted supply chain with reasonable pricing, delivered.

Unless just a once off, need a pack building scheme that works with the bare cells.

IMO

 
If you have. “top notch quality cells and a trusted supply chain”. .....why would you need or want to compromise pack integrity with hundreds of non permanent connections ?
Take a lesson from one of the accepted lead pack builders...Tesla..
However,.. if you are assembling salvaged, or otherwise imperfect/unreliable cells, ..then maybe look to incorporate these reliability reducing assembly methods of connection. :roll:
 
Nothing is permanent.

Tesla has more resources for R&D, and I do not see how their production welding methods relate to anything forum members here are doing.

The ability to easily test and replace individual cells as needed,

along with the likelihood that my welding or soldering will reduce cell lifespan,

together make it worth figuring out a no-weld / solderless solution.

But I agree with OP, **if** there are excellent cells trusted from trusted suppliers available with tabs already welded

that would be preferable to attaching my wiring directly myself using heat.
 
kilou said:
So you mean there is no alternative to a spot-welded pack?
No, there are alternatives, ultrasonic, lazer, even solder if done properly.
( has anybody seen a professional test report on the effects of correct solder connection on cells ?)
These days, spot (resistance) welding is probably the simplest and quickest assembly method for home use.
But my point is really that any non permanent “contact” connection....screw, spring pressure, clamp, bolt, etc etc ....introduces a potential increased failure point, and increased resistance potential.
 
I agree that spot welding is probably the way to go when it is done professionally. The problem is that DIY spot welding also brings its good share of risks IMO. I've been interested in the malectrics spot welder but have also seen video showing that its welds were not always consistent. Honestly I may have more confidence in tightening a brass nut on some copper bus bars that doing spot welding by myself, even with the proper tools. You can sure make a few trials on dead cells first but you can't be sure that all the welds you're doing are good with similar resistances. The Kweld seems like a good alternative in that respect since it quantifies the energy going into each weld rather than solely relying on a timer. But honestly, most DIY will built one pack every 5 or so years (mine currently is 7 years old). At the price of a Kweld for the very occasional builder, I guess you're better off paying for a ready-made pack then.

I also like the NIB magnet approach with copper sheet sandwiched inbetween but I thought that bolted connections between U tags and bus bars would be the way to go if a reputable cell dealer can be found (such as nkon.nl in EU where I'm located).
 
kilou said:
...
.... most DIY will built one pack every 5 or so years (mine currently is 7 years old). At the price of a Kweld for the very occasional builder, I guess you're better off paying for a ready-made pack then.
Exactly,..so the primary reason for non permanent connections is not justified.
If you have an economical source for “tabbed” cells you can safely solder the tabs to each other , or buss bars, and eliminate the added risks of compression joint resistance, and hundreds of extra components inside the pack
 
Hillhater said:
...the added risks of compression joint resistance

I do get your point but is it not a bit overrated? I mean there are people (e.g. Shawn McCarty with batteryblocs) who have been using magnets and compression for years without issues apparently. Also we're all using XT90-like connectors to connect/disconnect our batteries. Those connectors are capable of handling >90A yet they are pretty simple "springy" solderless mechanical connections in which joint resistance should also be questionned then. So why would it be reasonable to rely on such mechanical connection to handle the whole pack power and be shy about using mechanical (e.g. bolts/nuts, magnets, compression etc) to just handle the e.g 5A each cell is supposed to provide? To me the fact that things like XT90 are unquestionably able to handle such high currents means that a mechanical connection shouldn't necessarily be ruled out as a mean to make intercell connections. The downside force applied by an M3 bolts on cell tabs sandwiched inbetween two copper bus bars should be well enough to provide a strong connection, at least as good as what a XT90 connector is capable of doing, yet I just need to pass 5A, not 90A. What's your opinion on that?
 
kilou said:
Hillhater said:
...the added risks of compression joint resistance

I do get your point but is it not a bit overrated? I mean there are people (e.g. Shawn McCarty with batteryblocs) who have been using magnets and compression for years without issues apparently. Also we're all using XT90-like connectors to connect/disconnect our batteries. Those connectors are capable of handling >90A yet they are pretty simple "springy" solderless mechanical connections in which joint resistance should also be questionned then. So why would it be reasonable to rely on such mechanical connection to handle the whole pack power and be shy about using mechanical (e.g. bolts/nuts, magnets, compression etc) to just handle the e.g 5A each cell is supposed to provide? To me the fact that things like XT90 are unquestionably able to handle such high currents means that a mechanical connection shouldn't necessarily be ruled out as a mean to make intercell connections. The downside force applied by an M3 bolts on cell tabs sandwiched inbetween two copper bus bars should be well enough to provide a strong connection, at least as good as what a XT90 connector is capable of doing, yet I just need to pass 5A, not 90A. What's your opinion on that?

Very good point
 
kilou said:
....... The downside force applied by an M3 bolts on cell tabs sandwiched inbetween two copper bus bars should be well enough to provide a strong connection, at least as good as what a XT90 connector is capable of doing, yet I just need to pass 5A, not 90A. What's your opinion on that?
We know bolted connections work,..most packs have a few somewhere in the build..
The level of current is not important assuming the connector is sized correctly, a bad connection is just as likely at any power level
....but the difference between a single fully developed and tested connector system such as XT90 , and say 50-100+ Multiples of unproven small bolted connections, lies in the increased probability of a bad contact, loose nut, corrosion, etc.
Reliability is improved by keeping things simple with temporary connections and part count at a minimum
Example... Toyota hybrid drives in their Prius, Camry, etc,are ultra reliable and use a NmHi pack with multiple prismatic cells connected by screw connections.
The most common failure on those packs is loss of pack performance /capacity due to corrosion and bad contact at the screw connectors .
 
And a pack that is easily atomized for inspection / testing of the individual component cells

would reveal any such problems allowing them to be easily fixed.

I completely agree that the spotwelding approach is ideal wrt that one factor, and also cheap cost for professional / commercial pack builders.

But members here spending their time and money to advance our science,

come up with a cheap and reliable alternative that allows for that top cited goal

is something to be encouraged and applauded.

And I am confident will ultimately be successful.

Not at all a waste of time, from a community hive-mind POV.

Of course those making a living selling spot-welded packs will oppose these efforts.
 
Hillhater said:
...50-100+ Multiples of unproven small bolted connections, lies in the increased probability of a bad contact, loose nut, corrosion, etc.
Reliability is improved by keeping things simple with temporary connections and part count at a minimum

A spot-welded or soldered connection can also fail, especially when done by a hobbyist with suboptimal tools, especially in a high vibration environment such as an ebike. And there are also 50-100+ of these connections in a battery pack...not to mention that it would be extremely difficult to test and fix such connections if they do fail, unlike a simple bolt/nut system. So the question is not professionally built vs DIY packs (in which case professional spot welded pack wins hands down, except costwise maybe) but which method is the simplest, safest and most reliable for DIY battery packs. In that respect, bolted connections still have some place IMO. But it is really motivating to read arguments from all camps to progress and think further.
 
If you use cells with tabs, the quickest, simplest, easyest, and proven method to assemble would be to solder the tabs to each other and the buss bars.
Solder connections can be undone quicker than bolts and provide a better electrical connection without the risk of loose parts rolling about inside the pack should a connection come undone
I have assembled various ways , solder to the cell, solder tabs , welded, and a few small screw connected cells.
Personally for me, the welds are too permanent, and i do not believe the electrical conductivity of tiny spot welds is the best, i have spent too much of my life hunting for dry/corroded connections on various screw/bolt connections, to want to go that way on a 50+ cell pack,.....so soldering is the lesser of evils for me.
Again, i have had no problems from soldering direct to the cell base ..( the top cap has never been considered a risk).. so i do not bother with welded tabs.
But, everyone has an opinion and preference, together with individual requirements, ..so you should decide for yourself after you have assembled a pack or two using those M3 nuts and bolts (+washers ?)
. PS dont forget the threadlock !
 
After reconsidering @Hillhater suggestions, I may also consider soldering the U tabs on the copper bus bars instead of using brass bolts but this is still to be decided. Anyway, here are a few picture of the layout for a 10S6P battery:

top view:
r40b5UB.png


bottom view:
tfJZfNY.png


Each cell is bought using U tabs spot-welded on it (grey) and placed in a regular 18650 cell holder (light blue). A 3mm hole is drilled in each tab at the location of each yellow dot. Then copper bus bars (orange) are used for the parallel and series connections. These bus bars are also drilled at these locations and M3 brass bolts are used to hold the tabs from 2 cells on the bus bar (may also solder the tabs on the bus bar instead of using bolts). The small wires show the locations of the balance leads. I plan on using probably 8mm wide and 0.5mm/1mm thick copper bus bars and eventually tin plate them (electrolessly) by dipping them in a stannous chloride solution with thiourea.

Now you'll see I'm considering only one serial connection between parallel groups. Obvisouly it would be better to use two here for better load sharing. This is also possible here but I thought one serial connection would be enough for several reasons:

1) I'd be using 0.5mm/1mm thick copper bus bars which are able to carry much more current than nickel strips anyways
2) if you look at the serial connection on the far right of the bottom picture, you'll see this has to be a single piece. So does it make sense to have 2 serial connections between most parallel groups when this last serial connection cannot be doubled?
3) If I use a single serial connection as displayed, I can use some rubber feet on both side of the battery (see dark blue squares) so that the battery rests on these feet and not on the connections.

What do you think of this design? Do you think I should use two serial connections instead of one anyway despite the fact that one of the serial connection will remain single anyway? What copper thickness would you consider? I think 0.5/1mm is already overkill probably but I'd need that to have some structural rigidity I guess (for the bolts or for soldering). Would be glad to know what you think of this and/or have some suggestions!

PS: I'm actually more and more heading towards soldering the tabs onto the bus bars than using brass bolts/nuts. A nice update would be to drill the bus bars at the location of the yellow dots and stick a copper nail in it. Then the drilled U tabs could be hooked to the nail and soldered to it (I guess it would be easier to solder the tabs on the nail rather than on the bus bar since the nail would heat-up quickly). Finally the excess nail could be cut and this would result in a low profile reliable electrical connection without transferring heat to the cells.
 
kilou said:
Hi,

I'm investigating a semi-solderless design
Behold, completely solderless ghetto design!
I somewhat screwed up my spot welder built, now waiting for the Pickit programmer to fix it, meanwhile my new BMS was already here.
I've decided to make temp 24S1P pack to test it.
Since A123 26650 are aluminum, I came up with the following idea.
Cells are connected with 0.15 nickel strip, 40 mm long, and on each side I used 1 cent coin to apply enough pressure on the strip, with 2 6" zip ties to hold it all together.
Well, it worked well.
Maybe even too well :)
2 nights ago, after testing new BMS, everything worked fine, I've disconnected pack from BMS, and later on, somehow, managed to shorten it.
It was barely charged, 3.4 volt on average. But it def woke me and my GF up. Sparks were flying, negative end caught fire, these nickel strips got heated so fast, all the balancing wires became detached.
This is no 18650 LG batteries, from which I made my first battery 6 years ago.

eArfnF5.jpg


Needless to say, Im getting 2 fast blow, 100 Amp, 125 VDC fuses for my 24S6P battery now.
 
Back
Top