Software for performing pickup analysis with a recording interface.

[Joe Gwinn]

Very interesting! The randomized multi-sine looks very much like Gaussian random noise, as it should, but since it is a finite number of discrete sine waves, it is like repeating the same random noise over and over.

Yes. One common name for the multisine signal is "periodic noise".

That does look like a good technique; you can make a flatter spectrum than, say, code105, but the waveform has less average power per unit time because of its dynamic range.

I dug a bit deeper into the impedance spectroscopy literature, and the multisine approach is the clear winner in that world. The reason is that the multisine approach doesn't waste any test signal energy on frequencies that will not be measured. They spend significant effort on choosing phases to minimize the crest factor of the resulting periodic noise sequence, the purpose being to preserve the dynamic range.

But it should be understood that they use very weak test signals, about 10 millivolts, to avoid various interfering electrochemical effects, and yet they achieve 50 and 60 db SNR. Low frequencies are a problem for them too. Their solution is to make the low-frequency sines stronger than the high-frequency sines, to compensate for the highr loss at low frequencies. But we can also use very strong test signals, and simply overpower the problem.

Back when I was looking for nonlinearities in pickups, specifically in the magnetic materials, I was driving the test pickup hard enough that there were 70 volts rms across the coil. It would take quite the guitar player to approach that. I'm not sure I want to meet such a player.

For multisines, five or ten frequencies per decade suffices in the electrochemical world. There are various rules, but one rule is to choose from the odd harmonics of 10.07 Hertz (don't yet know why). The chosen test frequencies can be logarithmically spaced. It may be possible to also arrange things so the test sines land cleanly in single FFT bins, without straddling.

Might be worth a try after fully analyzing what we have.

I don't see any mention of the impedance spectroscopy folk using barker codes, but they did do something similar, but using a far longer pseudorandom sequence. A sequence of 105 elements will have sidelobes 20 log(105)= 40.4 db down, while they are looking for 60 db. This may be the reason.

[Joe Gwinn]

This drawing ([ATTACH]27861[/ATTACH]) shows the hardware connections needed to measure the pickup frequency response. It is pretty simple.

So we are measuring drive coil current versus pickup coil output voltage.

Joe Gwinn brought up an interesting question as to how much integration is necessary when using code105. This figure ([ATTACH]27862[/ATTACH]) shows two measurements, 300 shots underneath 1 shot. You really cannot see much difference. However, if you look more closely at the lower frequencies ([ATTACH]27863[/ATTACH]), Then you can see differences, and depending upon what you are doing, you might want the more accurate measurement. If I were measuring a bass pickup, I would use a coil that pumps out more flux and also integrate longer. There is no point in having much uncertainty at frequencies that matter.

The SNR may already be very high except at low frequencies, where the induced voltage is very low. This would explain the small advantage from added integration - it's already good enough (except at low frequencies).

As discussed in the thread now on impedance spectroscopy, one can counter this by pre-emphasis on the low frequencies. And by brute force.