|
Post by ms on Jan 6, 2019 12:33:51 GMT -5
Computing the frequency response of a pickup from impedance requires good measurements. (Part of the eddy current effect is not included from these measurements and must be obtained from exciter coil measurements.) Here are a couple of plots to demonstrating the accuracy of these impedance measurements. First, let's look at the impedance of a resistor across frequency. The amplitude is constant across frequency to better than 2 parts in 10,000, which is better than required. The noise on the amplitude measurements is a small fraction of 1 part in 10,000, which is good for an measurement time lasting about one second. The measurements are not as good at low frequencies because the preamps of the Apogee Element 24 are AC coupled. Thes phase varies by about .03 degrees, again with some deterioration in performance at low frequencies, primarily limited by the match between the capacitors in the Element 24 preamps. Art least, I think this is the problem. Now we need to see how well large changes in impedance with frequency are reproduced. For this we use film capacitors, since their behavior is so close to ideal. The following plot shows the impedances of two capacitors, plotted on log scales, which should give straight lines for an ideal variation. Two capacitors are measured so that more than the full range required for a pickup measurement is covered. No setting was changed between the two measurements. There is no apparent deviation from straight lines over the full range from just under 800 ohms to just over 3 Mohms. This is better than required.
|
|
|
Post by aquin43 on Jan 7, 2019 7:37:20 GMT -5
I must say that the capabilities of your test system seem very satisfactory indeed. I am particularly interested at the moment because my Audio Precision test set has just died with an amazing stink of burning electronics. It was, in its day, state of the art but it is very inconvenient because it was made before USB so it has its own proprietary interface which needs a Windows tower for the interface card.
You seem to make your measurements by voltage driving the pickup and measuring the current. That might cause difficulties in measuring a pickup mounted in a guitar, or any cased pickup where you would have to drive the case. Presumably a similar measurement technique could be used with the current driven approach.
Arthur
|
|
|
Post by ms on Jan 8, 2019 7:00:03 GMT -5
I must say that the capabilities of your test system seem very satisfactory indeed. I am particularly interested at the moment because my Audio Precision test set has just died with an amazing stink of burning electronics. It was, in its day, state of the art but it is very inconvenient because it was made before USB so it has its own proprietary interface which needs a Windows tower for the interface card.
You seem to make your measurements by voltage driving the pickup and measuring the current. That might cause difficulties in measuring a pickup mounted in a guitar, or any cased pickup where you would have to drive the case. Presumably a similar measurement technique could be used with the current driven approach.
Arthur
One side of the pickup is connected to the driving signal, which is a low impedance to ground. The other side is connected to a 1K resistor; the other side of it is connected to ground. There do not seem to be any serious issues with driving it if the pickup case or guitar ground is connected to the first rather than the second. That said, it is important to realize that if you want to measure the complete set of parameters of any pickup with a metal base plate and two coils, you need to remove it from the guitar so that the connection to the baseplate can be removed. This is to avoid effects from the secondary resonance that is seen above the main one. Unless this secondary resonance is suppressed, it is much more difficult to get an accurate C value, so that the effect of the C can be removed, revealing the extent of the eddy current effects over the useful range of the pickup. (For those not familiar with this effect: each coil of a two coil pickup has a resonance with its L and C. Ideally the series connection becomes like a single L and C. But the effect of a grounded baseplate is to add more C to one than the other, resulting in a more complicated circuit. This is not much of an issue with the ;pickup sound in actual use because the cable capacitance lowers the main resonance, and thus the effect from the secondary resonance occurs at a frequency range where the pickup has little response. But it matters for measurements because we want the resonance high and the circuit simple to best determine the response where it matters.) Not sure how this would work with a current source drive. I prefer to stick with the voltage source because you make best use of the way the sampler works without external circuitry.
|
|