My concern was that the contact resistance might be significantly higher than for a soldered joint leading to a significant voltage drop across them (they will be carrying a fair amount of current from a 3.3V power supply). Time to do some testing.
I started out with four pieces of wire of (approximately) equal length. Three of them I cut in half and then joined back together using one of those butt-splice connectors, a solder joint or a simple twist. The fourth piece of wire I didn't cut: it will serve as the control.
So what is the resistance of each of these? It is going to be pretty low no matter what, so a four-wire measurement is going to be needed. Happily, I have a 5.5 digit multimeter that does have four-wire resistance measurement capability.
This is what the test setup looks like. One pair of multimeter leads provides the test current and the other pair senses the voltage across the joint.
The results were pretty good:-
|Measured Resistance (mΩ)
Treating the uncut wire as a control, that means that there is only around 2-4mΩ of contact resistance regardless of what method is used. I have to admit, that's a good bit less than I was expecting. There is also less difference between the solder joint and the butt-splice joint than I had expected.
As a secondary check, I also measured the voltage drop across the joint while running a decently-high current through it. My bench power supply can supply a current-limited 5.8A. Knowing this and measuring the voltage drop across the joint is a secondary way of measuring the contact resistance. Here are the results:-
|Measured Voltage Drop (mV)
|Calculated Resistance (mΩ)
Curiously, the uncut wire measures a little higher than the solder joint. My suspicion is that the current-limiting on my (el-cheapo) power-supply isn't that rigid and it was letting a little more current through as it warmed up. This is borne out by the fact that the last digit on the ammeter showed the current climbing just a little as the tests went on (just a few tens of mA over the course of the tests) and I did the test with the uncut wire first. I could have controlled for this by monitoring current with a external - more accurate - ammeter, but I didn't and I think the discrepancy is small enough to ignore (for my purposes, anyway). If anyone is interested in seeing the test repeated a little more rigorously, I do have enough equipment to do it: leave a comment below and I'll redo the test.
I think the takeaway is that a solder joint is a little better than a crimped butt-splice joint, but not by a whole lot. There are certainly factors that I haven't taken into consideration here: will a solder joint age better than a crimped-on connector (I rather suspect it will)? Which one is mechanically more stable (again, my money is on the solder joint)? These don't matter in my particular application so - happily - the result is that I should be able to use the much-easier-to-make butt-joint method for joining the cables I need to join. A win !!