dnmun said:
the currents flowing in the parallel link are not measurable. the parallel link is there to allow the cans to balance to each other. the parallel link does not carry current.
the current is carried in the serial connection from one channel to the next in series. for that you need almost nothing for these little cans. at 2.7Ah then 4C is about 10A and it only takes an 18AWG wire to carry 10A.
the 1mmx4mm strip to connect the cans in series can carry up to 40A. if the connector is made so that it only contacts one edge of the case on the bottom then the amount of heat needed to reflow the solder could be kept minimal and the sponge would soak it up in a flash as soon as you hit it.
a soldered connection is more reliable and has the minimal contact resistance of all forms of connection.
certainly soldered or welded contacts are known reliable and used ~100%. With soldering the question is about critical heat impact on the plus pole - repeatedly when reconfiguring. Soldering on the negative pole is not critical at all because of the big heat capacity of the can: With strong short heat - just rub (with sandpaper or so) before soldering to minimize oxide coating.
Has anybody investigated cells via 4-point cell resistance & capacity measurements before and after soldering on the plus pule?
So welding is professional, used 99% and is clean and flat - but hard to reconfigure/repair (and problematic DIY for some).
The point and question of this thread yet is research about a fast removable contact, if reliability and low resistance is good enough for practical use.
The big link surface is needed here for the NIB contact.
A magnet contact and some positioning taping just on the plus pole and solder on negative pole would still retain much of the flexibilty and charm.
The parallel links also carry substantial current unless one makes p serial connections near each of the parallel cells. But the links in the very small milliohm range are not critical regarding resistance - cooper or nickel. So nickel is much much better here (where gold is out of question) because its a typical cheap contact material with little oxidation, and the pole and magnet cover material is also nickel, so one does not create a
corrosion cell (-> humidity). Cooper is a no-go for such contacts in many regards.
I think the problem of small flexing and micro movements is not big. some taping in whatever form. Some fixation of the magnets to the strips (welding, small screws or whatever) would improve the coherence and/or allow direct magnet-cell contact which is better than indirect magnetism through the strip (The nickel-steel strips are better for the latter as well).
The micromovements even would repeatedly rub oxide away from the moveable cell contact: I have made many contact measurements in another context: totally motionless pressure contacts tend to increase electrical resistance over time, any small movements decrease again.
Interesting would be some first systematic measurements of the magnitude and reliability of the resistance of a magnet-cell contact: Change during one day .. week of total immobility, changes during micromovments/flexing, with heat of ~55°C (hair dryer), in sunlight, when placing into the fridge, with longer lasting substantial currents of 2A ...
If the contact resistance stays consistently below 5 mOhm with that magnet force, this subject may turn interesting.