A wire causing distortion? Maybe.

[Bacon, WA3WDR]

I was trying to test some switching power supply coils.  The real spec is X number of turns on such-and-such core, because that works.  There is a stated inductance, maybe it's right and maybe not. But having wound the coils several times, I know that the number of turns on such-and-such core works, so I want to make sure I duplicate that as I proceed to build these things.

The trouble is, it's too easy to lose count of the turns while winding the darn things. So I thought, let me carefully count the turns on a sample unit, and measure its inductance, and then ensure that all coils perform the same in that test.  Since I didn't have a decent inductance meter, and I didn't want to go buy one, I thought I would instead put a ten turn test winding on the coil core, and measure the voltage ratio of the windings, to make sure at least that I had the right number of turns on the winding in question.

I took some #30 teflon-insulated solid tinned jumper wire, and I used it to wind a ten turn test winding on the core, and I put some AC on it.  I got very strange results, and I tried a bunch of different frequencies and the results continued to be strange.  I thought gee, the core must be really strange.  The voltages were all wrong, and looking with a scope, I saw a big mess of crossover-like distortion.  The generator was not producing the distortion, so I thought that the core must be really nonlinear, and I wondered how I was ever going to measure this, etc.

To make a long story short, the problem turned out to be the #30 test winding wire.  IT was causing the crossover-like distortion!  I changed the test setup to a resistor in series with the winding, applying AC at a fixed frequency, and measuring the voltages across the resistor and the winding, etc.  THAT works.  I can see the change that one turn more or less than the correct number makes, and correct the error.

OK, but why did the wire cause distortion?  It didn't cause very much distortion, but the sharp resonance of the coil emphasized the distortion that was present.  What caused the distortion?  My thought is that it is the junction of dissimilar metals - the copper in the wire, and the tin(?) plating.  My thought is that the #30 conductor is so thin that a significant amount of current flows through the copper-tin(?) barrier, and some distortion results - enough to screw up my measurements.

I took the wire and put signal on it through a resistor with the core not present, and I monitored the voltage across the wire with the scope, and I saw distortion on the scope.  I cut the wire in half and tried again with the resistor - and I saw about half as much distortion.  The whole thing is strange and confusing, but I concluded that the wire was causing the distortion, and I got back to work.

I'll play with this some more one of these days, but I am wondering - has anyone else ever noticed this?  I know this much: I'll think twice before dismissing audiophiles who claim to hear a difference between various wires.  I have seen evidence that they may be correct.

[WA1GFZ]

Bacon
I have run into a similar problem with inductors. The series resistor will give you a good clue so you can sweep the inductor and measure the reactance across the sweep by comparing the inductor voltage to the voltage across the resistor. We found these little switching regulators with surface mounted iron case inductors lose reactance fast above about 6 MHz. This is due to C between turns. We would simulate a filter but performance above 10 MHz was way off. The more wire you put on the worse it will get. I really like J material Ui 5000 for common mode chokes. You don't need many turns to get a lot of L

[k4kyv]

Maybe you should have used oxygen-free copper.  Audiophools insist that the type of wire used for speaker cable makes a tremendous difference in the sound - but only after the cable wire has been "broken in" for several hours.   

[Tom WA3KLR]

Hey Bacon,

I wonder if you had copper-coated steel wire?

Did you do a 4-wire resistance test on a length of wire and compare to the copper tables?

[WA1QHQ]

Sounds like your generator didn't like working into the low impedance caused by the ten turns of wire, it might also object to running into a DC short circuit. Adding the series resistor would have fixed the low impedance situation.

[Bacon, WA3WDR]

Interesting.  I did not measure the resistance of the wire per unit length; I will do that and see what it is.  It is fairly stiff, so it may not be copper.  As for the load presented to the generator, the resistor was 100 ohms 5% (measured 98.8 ohms) in series with the generator output, and the generator has 50 ohm output Z, so I was well above the load Z where the generator would be unhappy. Also, the scope showed a good waveform from the generator under load.  It was the voltage between the ends of the wire that showed the distortion.  I made sure that the connections to the wire were not loose, etc.

I just experimented with a magnet, and the wire did not respond to the magnet, as far as I could tell.  If it was steel wire, it would have; many component leads and even 316 stainless steel hardware respond slightly to this magnet.  I cut off a 1" length of the wire and tried that with the magnet, no effect.  Hmmm.  Later I will take some measured length of the wire, and run an amp through it, and measure the voltage drop across it.  Somebody bought this stuff at a hamfest over ten  years ago, and we have been using it as jumper wire stock ever since.  The spool label says MIL87816  MFG PN 191292-9 and is dated 4/93, and somebody wrote 2841-1/1 on the label with a marker pen.

Don, I was thinking about oxygen-free copper...  I sure hope it doesn't come to that!  What kind of tin would I need for the foil hat I would need to make for myself???

[Bill, KD0HG]

Bacon, I'm thinking that for whatever reason, you were causing the core to be nonlinear.

I can't see distortion happening in a conductor itself, no matter what.

Take that 5 feet (or whatever length) of wire off the core make an air-core coil out of it, and apply the AC to it looking for a distorted waveform. If there isn't any (most likely), then the issue is the core. Try winding it on a big steel nail and check for distortion.. If you're using a magnetic wire (steel, nickel alloy) it's going to affect the core's response. If the wire is aluminum, there will also be an effect.
Just some irrelevant thoughts..