Icon-YouTube(250x93px).png
Icon-LinkedIn(250x93px).png

© 2017 Ian Watts

To Solder ? ...or... To Spot-weld ?

I've been lucky; I've had some experience with soldering and (to me at least) it seems pretty simple. Soldering is all down to technique... and, once some experience is gained, soldering, really is, pretty simple. The 'problem' here is that soldering directly onto individual cells is NOT as simple as simply soldering some wires together or soldering a component onto a circuit board - generally, a LOT more heat is required.

Whilst this is NOT a 'How to solder' lesson, we do need to, briefly, consider what soldering is and the basics of how to solder, in order to appreciate what's required to solder 18650 cells and why, after some reflection, you may prefer not to !

First; in order to achieve a good solder connection, both metals need to be heated to a temperature that allows the solder, that's going to be added, to flow and fill any gaps thereby making an electrical 'joint' and, when cooled, a semi-permanent connection. Sounds easy but, in reality, it's not so... There are various factors to consider here:

  • Solder does NOT 'stick' well to oxidised surfaces. That's one of the reasons for the flux (either built-in to the solder or added as extra whilst soldering).

  • Cell caps (+ve / -ve) have a large nickel content and nickel oxidises quickly - that's the shiny surface we all see.

  • Solder (depending upon its precise composition) melts at around 200' C / 392'F. For it to 'run' smoothly and 'stick' to the metal surfaces to be joined, all surfaces need to come to this (minimum) temperature.

  • A Li-Po cell has a recommended maximum, no long-term damage, temperature of 70'. Above this temperature, irreversible chemical reaction begins to occur to the electrolyte within the cell... followed by material expansion... followed by cell venting... followed by eventual cell destruction and possibly even fire or explosion.

  • Raising the temperature of an object (or objects surface) above ambient involves applying heat. The time it takes to heat the object(s) will depend upon the rate at which the heat can be applied AND the rate at which that heat is dissipated throughout the rest of the object(s) being heated.

Taking the above points in order :

  • In order to solder to a Li-Po cell cap (high nickel content) the oxidised outer layer must first be removed. This can be achieved by the flux integral to the solder or added, however, quite often, the flux integral to the solder is insufficient so localised 'scrubbing', sanding  or 'roughing-up' of the nickel surfaces to remove the oxidation in way of the joint is necessary (time consuming at best).

  • Heating the cell cap (to around 200'C) needs to be achieved as QUICKLY as possible in order to minimise localised heating of the electrolyte. This requires a high power & capacity soldering iron; 150W - 200W should do the trick... and a larger that normal chisel tip (although, a 200W iron is large and comes with a pretty big chisel tip). This is not your 'usual' PCB soldering iron (typically 15W - 30W) with a needle tip. The latter will, quite likely, get the job done BUT the limited available heat transfer results in a longer contact time being required which results in more heat being transferred to the surrounding material which results (in this case) of localised (at best) electrolyte heating and (at worst) venting and destruction of the cell. 

So, with the above in mind, I prefer to spot-weld - heat time is less that 1second and EXTREMELY localised. Indeed, I have never seen a cell manufacturer (Panasonic/Sanyo, LG, Samsung...) soldering tabs to any of their cells - all connections are made with spot welding.

Having said all the above, there is no doubt that direct-to-cell solder connections can be achieved - the challenge in so doing is ensuring that the cell is not damaged in the process and, unfortunately, (ignoring extreme cases) it is nigh-on impossible to detect any damage / cell degradation until much further on down the road when it's even more difficult to attribute shortened cell-life or lower-than expected capacity / performance to initial solder process or general life-time mis-use.

Anyways, if soldering is to be your 'method of choice' then, prepare everything in advance, arm yourself with a high-powered solder-iron, a quality solder, a pot of additional flux (can help a lot)  and make each connection as quickly as you can in order to localise heat transfer / minimise heat transfer to the cell itself. And away you go...

There are plenty of people who DO solder, even swear by it - although I can't help but wonder if is this simply because they don't have a spot-welder ?!

Spot-Welding...

Having decided that spot-welding was the way forward for me, I needed to get myself a spot-welder. I could have bought one (there are plenty on eBay and Amazon) but I preferred to try and make one ! I've covered this little 'nugget' in my Spot-Welder project <HERE> so I won't go over that again.

Suffice to say, it took maybe 15 hours of work (a lot more waiting for parts to arrive or be printed on the 3D printer) to build a suitable unit and put it to work. My biggest challenge was how to spot-weld fuse wire to the cell caps. That follows, below...

Spot welding nickel to nickel and nickel to cell-caps proved to be pretty simple. I did find a couple of things though (bearing in mind, my home-made spot welder does not allow me to adjust the power of a spot-weld pulse, rather the pulse duration) ;

1: Using the dual sprung-electrodes to spot-weld nickel strip to the cell-caps, I found got a MUCH better spot-weld if one electrode was placed on the nickel strip and the other on the cell-cap. After welding, I then moved the electrode on the nickel strip to the next location (leaving the one placed directly on the cell-cap) and initiated the next weld. Normally, I have four spot-welds per cell cap.

I figured this was because by placing one electrode directly on the cell cap and the other on the nickel strip I was effectively 'forcing' the pulse through the nickel (at the point of electrode contact) and into the cell cap, then out of the cell-cap to the second electrode and back home to the spot-welder. If, on the other hand BOTH electrodes remained on the nickel strip, it appeared to me that a lot of the pulse energy was simply passing through the strip and not across the joint between strip and cell-cap. The end result was I needed longer pulse durations to effect the same spot-weld. Just my findings...

Version '1'

'Sprung-pair' Electrodes

2: Anyways, the result is that I designed, 3D printed and made up a pair of electrode pens. Each pen has a momentary switch to initiate the firing pulse.

The 'problem; I had with the pen electrode pair was, when spot-welding fuse wire.

I found it difficult and time-consuming to keep the pointed tip of the 'pen' electrode on the fuse-wire and stop it from rolling off / stop the fuse-wire rolling from under the tip of the 'pen' electrode (whichever)... Either way I quite often found myself spot-welding electrode to cell-cap... not particularly helpful or  effective !

I also found that more heat was needed to spot-weld fuse-wire to cell-caps than nickel strip to cell-caps and that the copper electrode dissipated the heat through itself far too quickly. The result was longer pulse durations, hotter electrode tips = shorter continuous welding time (more breaks required to let tings cool down) and generally a not-so-good spot-weld.

Version '2'

'Pen' Electrodes

3: Enter Version '3' - the ferrite chisel-tipped electrode.

Probably the simplest of the three to make (I was getting anxious to move on / make some progress) and started to look at other materials in order to concentrate the heat at the electrode tip and thereby create a better and faster spot-weld with fuse-wire.

I 'played' with steel (no good at all) and tungsten (rather too inflexible / brittle) and finally ferrite (easily filed and a poor conductor of heat).

Finally, I simply cut a 5mm copper tube to length (100mm), directly soldered the inter-connect cable (from the spot-welder) into it and screwed a ferrite rod up into the other end of the tube. The remaining end of the ferrite rod was (easily) filed to a chisel tip and an insulating 'handle', made on the 3D printer, was fitted to the copper tube.

Coupling this electrode with a pen-electrode provided a fast and simple way to spot-weld fuse-wire onto the cell caps - the pen electrode providing the switch with which to initiate the weld and the chisel tip electrode resting firmly on the fuse-wire and not rolling around ! 

Version '3'

Ferrite 'Chisel-Tip'  Electrode