Preference For Sealed Cabs? Why?

[will33]

Sometimes in order to provide the perceived necessary correction, the design of the filter suffers from stored energy problems that were either not understood or simply ignored. With a tube amp, this would not be a problem but clearly it can be with other amp topologies.

It's not just minimum impedance, it's the storage of reactive energy. Filter responses contain both capacitive and inductive slopes (leading and lagging current responses), these factors plus stored energy can allow destructive oscillations to occur depending on amp topology and even specific design approaches. They can be especially troublesome on class ab amps using high global feedback and low stability margins, as well as class d topologies.

I don't have a formal background in electronics, so am at risk of being "that guy". I try to keep my impedance curves fairly benign (above as well as within the passband) in both magnitude and phase angle, and use damping resistors on the parallel legs of my filters so they don't show a dead short if a driver blows, but I don't know what specific steps to take to avoid the stored energy oscillations you describe. What should I be looking at? What are the symptoms that might show up in measurements (before the magic smoke escapes)?

If this is beyond the scope of what you're comfortable discussing here, no problem. But I really don't want to be the speaker guy who accidentally kills someone's amp.

Will's ears perk up as he anticipates one of those crucial bits info that keep his triggerfinger perched over the "remember this" button.

 

[DukeLeJeune]

A few years back I toured with an AudioKinesis TC115. I really loved it, but I had about 15 different amps blow up on that tour

Whoa! Fifteen amps??

I took a look at my schematic and all parallel component legs are damped with resistors, and the impedance curve stays between 3.7 and 6.2 ohms from the bass peaks up to 20 kHz, and is inductive from there on up. So on paper it looks benign (I no longer have the phase plots but would have looked at them too). But maybe there's something I didn't take into account.

Did the blowing of amps begin and end with your use of my cab?

 

[dukeorock]

Whoa! Fifteen amps??

I took a look at my schematic and all parallel component legs are damped with resistors, and the impedance curve stays between 3.7 and 6.2 ohms from the bass peaks up to 20 kHz, and is inductive from there on up. So on paper it looks benign (I no longer have the phase plots but would have looked at them too). But maybe there's something I didn't take into account.

Did the blowing of amps begin and end with your use of my cab?

I was joking!!!

 

[ahc]

dukeorock... not to derail the thread, but I'm curious what amp you liked (for electric bass) with the tc115. And did you use any port plugs?
Thx, Al

 

[dukeorock]

dukeorock... not to derail the thread, but I'm curious what amp you liked (for electric bass) with the tc115. And did you use any port plugs?
Thx, Al

Man, I used that cab mainly for my upright gigs...a little electric too. It was a Genz ShuttleMax 9.2. Sounded pretty good on upright with that cab, and not my thing with electric. A buddy tried it with an Aggie TH500 and it was fantastic on both. I'd love to hear my Mesa M9 or a Prodigy through that cab! Dukes crossovers are second to none
EDIT
I meant to say I didn't love the ShuttleMax on electric thru the TC115. I loved the cab!

 

[DukeLeJeune]

I was joking!!!

Oh wow, you totally got me... yeah now I see the smiley, but I glossed over it, thinking you were just being nice about an awful situation that was almost certainly my fault.

When I first read that post, it was a total "sphincter muscles, don't fail me now!!" moment. If having a couple hundred defective and destructive cabs out there isn't my worst nightmare, it's darn close!

Whew!!

 

[ahc]

Man, I used that cab mainly for my upright gigs...a little electric too. It was a Genz ShuttleMax 9.2. Sounded pretty good on upright with that cab, and not my thing with electric. A buddy tried it with an Aggie TH500 and it was fantastic on both. I'd love to hear my Mesa M9 or a Prodigy through that cab! Dukes crossovers are second to none
EDIT
I meant to say I didn't love the ShuttleMax on electric thru the TC115. I loved the cab!

Thank you sir! /derail

 

[agedhorse]

I don't have a formal background in electronics, so am at risk of being "that guy". I try to keep my impedance curves fairly benign (above as well as within the passband) in both magnitude and phase angle, and use damping resistors on the parallel legs of my filters so they don't show a dead short if a driver blows, but I don't know what specific steps to take to avoid the stored energy oscillations you describe. What should I be looking at? What are the symptoms that might show up in measurements (before the magic smoke escapes)?

If this is beyond the scope of what you're comfortable discussing here, no problem. But I really don't want to be the speaker guy who accidentally kills someone's amp.

Unbelievably complicated subject that is about as dry as it gets. Don't have time now but perhaps I can address a little of it during the equally boring sound check & rehearsal tomorrow while twiddling my thumbs at foh.

 

[Arjank]

I don't have a formal background in electronics, so am at risk of being "that guy". I try to keep my impedance curves fairly benign (above as well as within the passband) in both magnitude and phase angle, and use damping resistors on the parallel legs of my filters so they don't show a dead short if a driver blows, but I don't know what specific steps to take to avoid the stored energy oscillations you describe. What should I be looking at? What are the symptoms that might show up in measurements (before the magic smoke escapes)?

If this is beyond the scope of what you're comfortable discussing here, no problem. But I really don't want to be the speaker guy who accidentally kills someone's amp.

I'm curious too

 

[AstroSonic]

Unfortunately, the potential benefit this brings often includes the issue of stored energy in the form of a tank circuit that can interact with class d amps where the PWM carrier oscillates in damped within the zobel correction networks. I have encountered this on several poorly designed (though the designers felt otherwise of course!) speaker products.

The result is amp shutdown in amps that contain appropriate protection, and magic smoke in those that don't.

Just curious if anyone has bothered to simply test for runaway oscillation of a class d amp using an oscilloscope and pulse generator. A 100 MHz bandwidth is probably sufficient.

The other point of view questions whether the juice is worth the squeeze: are such complex corrective and 'linearizing' networks audibly beneficial in a gig setting? If they are , then amp makers will need to make their products more robust.

 

[agedhorse]

It's not the presence of signal but the reactive energy stored that then interacts with the PWM carrier in a grossly under-damped manner causing runaway. Most PWM oscillators change frequency somewhat with load and/or audio frequency so a powerful outside influence can have dramatic effect.

With regards to complex frequency linearization of impedances, IME, designs and driver choices that require such networks to be useable in the real world are usually better addressed within the basic design rather than with multiple corrections.

 

[DukeLeJeune]

Unbelievably complicated subject that is about as dry as it gets. Don't have time now but perhaps I can address a little of it during the equally boring sound check & rehearsal tomorrow while twiddling my thumbs at foh.

I really appreciate your being willing to toss out whatever you reasonably can. It wouldn't be fair for me to expect you to teach me what a college degree would have, but any pointers in the right direction will be very helpful.

With regards to complex frequency linearization of impedances, IME, designs and driver choices that require such networks to be useable in the real world are usually better addressed within the basic design rather than with multiple corrections.

I assume you're talking about Zobels and/or resonant peak filters like Arjank mentioned a couple pages back. Can you walk through an example (doesn't have to be in great detail)? Only if you want to and at your convenience of course... your coming to this site is supposed to be fun, and I don't want to ruin that for you!

 

[AstroSonic]

It's not the presence of signal but the reactive energy stored that then interacts with the PWM carrier in a grossly under-damped manner causing runaway. Most PWM oscillators change frequency somewhat with load and/or audio frequency so a powerful outside influence can have dramatic effect.

With regards to complex frequency linearization of impedances, IME, designs and driver choices that require such networks to be useable in the real world are usually better addressed within the basic design rather than with multiple corrections.

I'm thinking the reactive energy is derived from the signal and stored in reactive components of the network. If not, the impulse generator is not needed, but the oscilloscope could still be used to actually observe (or check for) a runaway condition. A person could do this to gather real (and not just theoretical) observational data. And, it would be useful even for someone who wasn't sure exactly what was going on. Observations help people gain knowledge and experience.

Totally agree that the problems addressed by complex corrective and linearization networks are most often best addressed in the basic design phase (driver choice, crossover frequency and type, bass alignment).

 

[DukeLeJeune]

I'm thinking the reactive energy is derived from the signal and stored in reactive components of the network.

On the crossover network design side, how would oscillating energy storage be prevented? Seems to me it's potentially there every time we have an inductor and capacitor in series, and that happens all the time.

On the circuit complexity issue, is it the complexity in and of itself that is detrimental, or are there specific resonant interactions that are detrimental, and those are just more likely to arise as the component count goes up? And if so, how do we check for them? I don't need the whole derivation - just a shove in the right direction.

 

[Luke the Fluke]

I'm new around here - but Talkbass is really great - you can ask questions like I have on here and people who really know their stuff answer
Sorry back to your

 

[Arjank]

On the crossover network design side, how would oscillating energy storage be prevented? Seems to me it's potentially there every time we have an inductor and capacitor in series, and that happens all the time.

On the circuit complexity issue, is it the complexity in and of itself that is detrimental, or are there specific resonant interactions that are detrimental, and those are just more likely to arise as the component count goes up? And if so, how do we check for them? I don't need the whole derivation - just a shove in the right direction.

same thoughts here

 

[agedhorse]

They don't teach such details in college engineering programs, this is the important stuff that lies between the lines that you learn through experience and general theory based troubleshooting analysis.

I will think of some practical examples, one that comes to mind in the civil engineering field is the Tacoma Narrows Bridge failure that was a result of store energy under damped runway oscillation. I can't post a link to the video on my iDevice here at foh, but I'm sure there are YouTube videos documenting the tragic failure. Absolutely worth watching.

Since the equations governing oscillation are similar across engineering fields, this is a good mechanical example of an electrical phenom

 

[agedhorse]

Anytime L's and C's are used, it's worth keeping in the back of your mind the POTENTIAL got resonant interactions. There are usually ways to mitigate the side effects but not always, and sometimes the interactions are not at all obvious. Sometimes they are very sneeeeeky!

 

[AstroSonic]

On the crossover network design side, how would oscillating energy storage be prevented? Seems to me it's potentially there every time we have an inductor and capacitor in series, and that happens all the time.

On the circuit complexity issue, is it the complexity in and of itself that is detrimental, or are there specific resonant interactions that are detrimental, and those are just more likely to arise as the component count goes up? And if so, how do we check for them? I don't need the whole derivation - just a shove in the right direction.

I think that the oscillation is set in motion by an ac signal (bass guitar or PWM carrier). DC will behave like a finger that pulls a bass guitar string, but does not let go. Energy is stored, but there is no oscillation. In ac, the string is let go when the wave direction changes. In an oscillating circuit, energy to tossed back and forth between reactive components (caps and coils). The resonant frequency is that at which the cap and coil store (and return) the same amount of energy (their impedance is equal). In crossover networks, the energy is dissipated in the load (the drivers). The dissipative (primarily driver) resistance needs to be more or less equal to the impedance of the reactive components at the crossover frequency (-3db point). In a simple 1st order case for ideal 8 ohm drivers, the coil and cap each have an impedance of 8 ohms at the crossover frequency and are essentially oscillation-free if connected to 8 ohm drivers. As networks get more complicated, it's easiest just to send the speaker system a signal and watch what happens with an oscilloscope.

Here is the 'shove in the right direction': As complexity goes up, the chance for a resonant interaction to occur increases. With complex networks, the potential problem occurs when the network interacts with the class D amplifier circuitry and oscillates with the PWM carrier. A well designed amplifier incorporates protective circuitry. Such an amplifier would be a good 'test bed' for checking for oscillation. First, simply connect the speaker system to the amp, turn it on and watch for oscillation. If nothing unusual is seen, send the speaker either a burst of white noise or a sharp pulse and watch for oscillation. I suppose the 'kids' among us will want to load an o-scope and signal generator app on their smart phone, but I prefer a real signal generator and oscilloscope with knobs and dials. I suspect such an app would be bandwidth limited by digital sampling rate, and might not pick up the PWM carrier. Early digital oscilloscopes had this problem, while their older analog relatives did not.

Hope this is helpful, or at least entertaining…

 

[bftbassman]

I like 2 15's for my rig. What if i used one sealed and one ported ? would i get the best or worst of both worlds in my sound ?