Question - A simple compressive non-linearity (x - ax**3) comes with a not inconsiderable suppression of the fundamental. Since the tracing distortion would be both level and frequency dependent how does this factor in or out. If I remember my expansions correctly -30dB thirds gives -0.8dB on the fundamental.
Last edited:
Yes it represents losses in the pickups. Hallgren also experimented to see if RLT was affected by mechanical forces, and found it was not.
Yes I was considering the Permalloy pole pieces. If the permeability falls above a few kHz, depending on lamination thickness amongst other things, the inductance L is far from constant, which in turn is going to make the shunt loss resistance also variable with frequency. The only constants are the dc resistance and load capacitance.
This could easily explain the 3~5 kHz dip seen on so many MM cartridges and the unexpected phase shifts that some of us have seen when measuring cartridges
This could easily explain the 3~5 kHz dip seen on so many MM cartridges and the unexpected phase shifts that some of us have seen when measuring cartridges
The shunt loss resistance cannot vary with frequency. If it does, then it's not resistance, it's reactive (either L or C). It's shunting an inductor, which impedance will change with frequency. The capacitive load (the impedance component that is capacitive reactance) is not constant with frequency either. Only resistance is constant.
However, not to beat this up too much, there's no dynamic component here specifically, at least not so far. And if there were, the question would be how that dynamic component would be different, and how the normal methods of dealing with cartridge impedance effects would somehow not be effective.
Also, you may want to read the Holman papers. There are three cartridge impedance effects to consider. Over-thinking the cartridge model is pointless when you consider it varies with cartridge model, as do the effects of cart impedance, but the way of working around them does not. Again, Holman's work should be reviewed.
No, it's a non-linear resistor, and as such it's resistance can depend on frequency and/or level. It represents losses, so whatever they depend upon.
Yes, but Davidsrsb's point is that if permeability changes with frequency, so does inductance itself. And so does efficiency/transfer function of the generator.
No, if loss mechanism depends on flux slew rate, frequency and level are correlated variables. At a constant level, impedance would be frequency sensitive, and at a constant frequency, impedance would be level sensitive.
LD
It is relevant here, not least because test sweeps tend to be toward the warm end of modulation levels. And so invoke notable hd, esp in upper mid bands. As you say, this must show up as fundamental suppression.
LD
I just remembered the Telarc Omnidisk has several cuts of music recorded in 4 2dB stepped tracks. It can't hurt to try and apply some modern signal processing to these, just as a datapoint.
I need to post a photo of the 1812 cannon at max level from this while I'm at it, talk about cutting something that nothing can play.
Not sure of the omnidisk, but there are apparantly 3 masters of the telarc 1812 out there of various trackability levels.
Attachments
That is the thing that I would like to get to the bottom of. The Ortofon eddy current stuff on solid poles and the Van Raalte LA paper give a compelling story of early roll off propped up by cantilever resonance. A model that explains this that can lead to experiments to measure it would I think set us off in a good direction.
Yes, there is one big reason. By definition, resistance cannot vary with frequency. If it does, it is not pure resistance, and would exhibit a reactive component. If it varies with level, it is not pure resistance either, but rather a nonlinear component (like a diode, for example).
The skin effect is absolutely not involved as all conductors are smaller than skin depth at the frequencies in question.
What other effect might be at work?
Level dependence means a non-linear system and superposition does not apply, you simply can not apply normal small signal analysis to the problem. I don't want to buy any.
Thanks for the Tomlinson papers!
Thank you for those - I think I have previously seen the first but not the second and third. In the third ("New Factors in Phonograph Preamplifier Design", October 1975) Tomlinson cites Hallgren on page 3, as follows:
Taking it that Holman has correctly represented Hallgren's finding, I would expect RLT to be specified as a function of frequency [RLT = f(ω)] though not amplitude, with the function taking a form derived from measurements of the cartridge in question.
Have I read that correctly?
Thank you for those - I think I have previously seen the first but not the second and third. In the third ("New Factors in Phonograph Preamplifier Design", October 1975) Tomlinson cites Hallgren on page 3, as follows:
Taking it that Holman has correctly represented Hallgren's finding, I would expect RLT to be specified as a function of frequency [RLT = f(ω)] though not amplitude, with the function taking a form derived from measurements of the cartridge in question.
Have I read that correctly?
More clues about Permalloy behaviour:
https://aip.scitation.org/doi/abs/10.1063/1.1710305?journalCode=jap
I have never heard of "magnetic viscosity"
https://aip.scitation.org/doi/abs/10.1063/1.1710305?journalCode=jap
I have never heard of "magnetic viscosity"
I understand non-linear systems. I'm asking for at least some metric re: how "non-linear" it is.
And I don't want to make that typo again. But may anyway.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.