Axe-FX III Dyna-Cabs

[FractalAudio]

If you’re in the Ft. Lauderdale area, A) you’re crazy because it’s Spring Break and B) shoot me a PM if you want any recommendations!

Naples.

 

[jay mitchell]

Speaker resonance is a function of displacement.

No. Speaker resonance is a function of the parameters I listed above. Whenever speaker impedance is a function of displacement, there will be either measurement error or decidedly nonlinear behavior.

Let's simplify: at low frequencies (below, say, 1 kHz), a cone transducer is a second order damped oscillator. Its resonant frequency is the frequency at which it will naturally oscillate in response to a negative step (a displacement from equilibrium followed by a release). Reread that statement. The resonant frequency of a cone driver is a mechanical parameter. It can be determined with a high degree of accuracy with no electrical input and is not affected by displacement within the linear range of the suspension.

Having read Quantasylum's page on speaker impedance measurements, I'll say that their methodology, while potentially useful for production QC, poses many possibilities for substantial error (including, but not limited to, the ones they disclose). The page itself contains several obvious errors. If the system output levels and amplifier gain shown in the above graph are accurate, the highest voltage level applied to the speaker produced less than .4 watts (note the decimal) nominal power into an 8 ohm load. That is too low to drive even the wimpiest guitar speaker suspension into nonlinear behavior. That data is therefore highly suspect.

 

[FractalAudio]

No. Speaker resonance is a function of the parameters I listed above. Whenever speaker impedance is a function of displacement, there will be either measurement error or decidedly nonlinear behavior.

Let's simplify: at low frequencies (below, say, 1 kHz), a cone transducer is a second order damped oscillator. Its resonant frequency is the frequency at which it will naturally oscillate in response to a negative step (a displacement from equilibrium followed by a release). Reread that statement. The resonant frequency of a cone driver is a mechanical parameter. It can be determined with a high degree of accuracy with no electrical input and is not affected by displacement within the linear range of the suspension.

Having read Quantasylum's page on speaker impedance measurements, I'll say that their methodology, while potentially useful for production QC, poses many possibilities for substantial error (including, but not limited to, the ones they disclose). The page itself contains several obvious errors. If the system output levels and amplifier gain shown in the above graph are accurate, the highest voltage level applied to the speaker produced than .4 watts (note the decimal) nominal power into an 8 ohm load. That is too low to drive even the wimpiest guitar speaker suspension into nonlinear behavior. That data is therefore highly suspect.

I'm talking about the electrical resonance of the speaker impedance. That's what matters as that is what a tube amp "sees". The electrical resonance varies with drive level. There are multiple studies showing this and I've verified it with my own tests. Klippel's papers are the best reference on the subject.

 

[imperator]

I'm talking about the electrical resonance of the speaker impedance. That's what matters as that is what a tube amp "sees". The electrical resonance varies with drive level. There are multiple studies showing this and I've verified it with my own tests. Klippel's papers are the best reference on the subject.

Is the "sees" due to VSWR or something else, any references?

 

[FractalAudio]

Is the "sees" due to VSWR or something else, any references?

A tube amp is a high impedance output in most cases. Basically a current drive.

Thiele-Small is an outdated, linear view of things. The modern way to analyze things is to realize that nothing is linear. The various components of a speaker (cone, surround, suspension, spider, etc.) are quite nonlinear. Thiele-Small treats a speaker as a *linear* second-order system. In reality the "spring" (compliance) is nonlinear.

This is unsurprising because springs are almost always nonlinear. Even a simple coil spring is only linear over a small range.

When you apply power to a loudspeaker the diaphragm moves. As the diaphragm moves from the rest position the compliance changes. Therefore the resonance changes because the resonance is a function of the compliance.

 

[FractalAudio]

To elaborate further. This is similar to the electrical model we use for the speaker impedance:

Our voice coil inductance is more complex though.

Note the inductor represents the speaker's compliance. In the old Thiele-Small view of things Cm is static. However Cm in the "real world" is Cm(x) and is a function of displacement. Therefore the electrical impedance is also a function of displacement.

The voice coil inductance is also a function of displacement. But for a different reason. The BL product changes with displacement which changes the effective inductance. Basically the same thing as a solenoid. As you change the amount of metal inside the windings the inductance changes.

 

[Will Chen]

To elaborate further. This is similar to the electrical model we use for the speaker impedance:

Our voice coil inductance is more complex though.

Note the inductor represents the speaker's compliance. In the old Thiele-Small view of things Cm is static. However Cm in the "real world" is Cm(x) and is a function of displacement. Therefore the electrical impedance is also a function of displacement.

The voice coil inductance is also a function of displacement. But for a different reason. The BL product changes with displacement which changes the effective inductance. Basically the same thing as a solenoid. As you change the amount of metal inside the windings the inductance changes.

I love reading this stuff...but aren't you supposed to be on vacation? Take a well deserved break.

 

[jay mitchell]

Even a simple coil spring is only linear over a small range.

I'm glad you bring that up. What happens when a "simple coil spring" is extended beyond its linear range? Its rate increases (i.e., it gets stiffer). What effect does increasing spring rate have on the resonant frequency of a mechanical oscillator? The resonant frequency increases. What does the chart you posted show? It shows the resonant frequency decreasing.

Your argument does not support the data. In fact, the data contradicts your argument.

FYI, Newtonian physics have not gone "out of date."

The BL product changes with displacement which changes the effective inductance.

The degree to which that happens depends heavily on the motor design. However, let's do a simple thought experiment. Assume an underhung voice coil (the full length of the windings sits within the magnet gap). In that circumstance, the VC inductance will be at its maximum in the equilibrium position. As the cone moves forward, at some point part of the VC leaves the gap. What effect does that have on VC inductance? It decreases. What effect will decreasing inductance have on resonant frequency? It will increase. The chart you show contradicts that argument as well.

 

[Mongillo19]

I'm glad you bring that up. What happens when a "simple coil spring" is extended beyond its linear range? Its rate increases (i.e., it gets stiffer). What effect does increasing spring rate have on the resonant frequency of a mechanical oscillator? The resonant frequency increases. What does the chart you posted show? It shows the resonant frequency decreasing.

Your argument does not support the data. In fact, the data contradicts your argument.

FYI, Newtonian physics have not gone "out of date."

The degree to which that happens depends heavily on the motor design. However, let's do a simple thought experiment. Assume an underhung voice coil (the full length of the windings sits within the magnet gap). In that circumstance, the VC inductance will be at its maximum in the equilibrium position. As the cone moves forward, at some point part of the VC leaves the gap. What effect does that have on VC inductance? It decreases. What effect will decreasing inductance have on resonant frequency? It will increase. The chart you show contradicts that argumment as well.

I for one can't wait to see the Jay Mitchell amp modeler at NAMM

 

[Jarick]

Cliff is working harder on vacation than I do most days at the office.

 

[jay mitchell]

I for one can't wait to see the Jay Mitchell amp modeler at NAMM

I, OTOH, look forward to evaluating Cliff's loudspeaker designs.

 

[Gearzilla]

I, OTOH, look forward to evaluating Cliff's loudspeaker designs.

Speaking of loudspeakers…

Whatever happened to/with your bookshelf "FRFR" design?

 

[fretworn]

 

[jay mitchell]

Speaking of loudspeakers…

Whatever happened to/with your bookshelf "FRFR" design?

Which one? I've been manufacturing speakers that fit that description since 1986.

 

[FractalAudio]

I'm glad you bring that up. What happens when a "simple coil spring" is extended beyond its linear range? Its rate increases (i.e., it gets stiffer). What effect does increasing spring rate have on the resonant frequency of a mechanical oscillator? The resonant frequency increases. What does the chart you posted show? It shows the resonant frequency decreasing.

Your argument does not support the data. In fact, the data contradicts your argument.

FYI, Newtonian physics have not gone "out of date."

The degree to which that happens depends heavily on the motor design. However, let's do a simple thought experiment. Assume an underhung voice coil (the full length of the windings sits within the magnet gap). In that circumstance, the VC inductance will be at its maximum in the equilibrium position. As the cone moves forward, at some point part of the VC leaves the gap. What effect does that have on VC inductance? It decreases. What effect will decreasing inductance have on resonant frequency? It will increase. The chart you show contradicts that argumment as well.

Agreed but you're looking at things from a lumped element view. A real speaker is a distributed, nonlinear system. It's not a simple spring/mass/dashpot. Nothing is. Everything is distributed and nonlinear. We approximate systems with lumped, linear versions so that we can analyze them easier.

It seemed counterintuitive to me too but multiple papers and my own tests show the frequency decreases with increasing drive. Up to a point. It then increases, but this is at very high drive levels.

You should read some of Klippel's papers on the subject. Quite fascinating.

 

[Gearzilla]

Which one? I've been manufacturing speakers that fit that description since 1986.

I can’t say specifically, however you mentioned something without elaborating related to applying with guitar modeling. Smaller than the CLR.

 

[jay mitchell]

I can’t say specifically, however you mentioned something without elaborating related to applying with guitar modeling. Smaller than the CLR.

I have a couple of those that I use, but I've not offered them for sale. My customer base would have no use for them, and I've refrained from selling to end users for several decades.

 

[jay mitchell]

Agreed but you're looking at things from a lumped element view.

You mean like the lumped element schematic in your post 186? Yup.

 

[TheOrangeChannel]

I’m definitely not looking at this to revolutionize the way I go about dialing in tones, but more of an area I’ll have that much more control over. There are times I’m picking IR’s and the difference between IR #2 and IR #3 isn’t huge, but neither are exactly what I’m going for because I want what’s between 2 and 3. Or sometimes I’m thinking “Man, if I could just angle this mic a pinch it’d be there!”

Just some finer detail that wasn’t available before.

Definitely anxious to see what the 3rd party market does with this. If York rents that robot I’ll have no problem re-purchasing some cab packs I already own.

I generally go with an IR that more or less mirrors a cab i previously owned and loved. Either the 2x12 Recto or a Straight Front 4x12 Traditional Recto. Then I can get into minutia like V30's vs G12H's or blends of speakers. It tends to be a case where that was the wheelhouse I enjoyed standing in front of so if I can approximate the feel and there's one that sounds great without tweaking to much I'm golden. I honestly don't even like to move mics in these units if i don't have to.

 

[TheOrangeChannel]

• 2017, maybe earlier: Two Notes releases their cab sim thing with movable mic/position/angle. How it worked under the hood I do not know.
• 2019: ML Sound Lab MIKKO is the first to offer a cab sim that lets you adjust not only the mic position/distance/angle but also adjust it all around the cone. It also supports mixing up to 9 different cab/mic combinations together and you can export it in mono, stereo, different sample rates and more. They have just released MIKKO 2 that has a ton of cabs and an updated UI. I think it's still the most advanced cab sim plugin on the market.
• 2020: Quad Cortex 2 is the first hardware unit to my knowledge to offer movable mic UI on the device itself.
• 2022: Line6 brings a new cab sim system with Helix 3.5 firmware.
• 2023: Tonex has the VIR cabs feature in their pedal.
• 2023: Fractal brings their own cab sim system.

Don't forget the GetGood stuff... they've got a bunch of cab IR generators. It's like Mikko and Nolly more or less perfected shooting IRs over the years.