Class D "digital" amps. Opinions on performance

[will33]

Geez, why is it that threads full of great tech info end up like this? For those of us who aren't professional engineers, this thing is a goldmine.

 

[projectMalamute]

No.

It's a different topology but the same class.

Like a Fender Champ and Nelson Pass Zen are both class A, but completely different topologies. The Fender Champ uses a 6V6 to drive a single ended transformer and the Pass Zen uses a mosfet to directly drive the speaker... but they're both class-A amps.

A CA9 is a class H amp. It is a fundamentally different kind of amplifier. It is related to but not the same as a class A/B amp. It is absolutely not a class A amp.

Just like how a class I amp is related to but fundamentally different than a class D amp.

This whole debate about whether the differences warrant a separate definition is smoke. There is different tech under the hood. Period.

When you get a gazillion watts in a 3 pound package for 200 bucks you are not getting the same kind of amp used by a touring sound company.

 

[jnewmark]

Geez, why is it that threads full of great tech info end up like this? For those of us who aren't professional engineers, this thing is a goldmine.

Maybe it's because great minds CAN'T think alike ? (just a joke )

 

[Bob Lee (QSC)]

A CA9 is a class H amp. It is a fundamentally different kind of amplifier. It is related to but not the same as a class A/B amp. It is absolutely not a class A amp.

Just like how a class I amp is related to but fundamentally different than a class D amp.

This whole debate about whether the differences warrant a separate definition is smoke. There is different tech under the hood. Period.

When you get a gazillion watts in a 3 pound package for 200 bucks you are not getting the same kind of amp used by a touring sound company.

Uh, no. This post is a big step backwards.

 

[Melvin7822]

A CA9 is a class H amp. It is a fundamentally different kind of amplifier. It is related to but not the same as a class A/B amp. It is absolutely not a class A amp.

Just like how a class I amp is related to but fundamentally different than a class D amp.

This whole debate about whether the differences warrant a separate definition is smoke. There is different tech under the hood. Period.

When you get a gazillion watts in a 3 pound package for 200 bucks you are not getting the same kind of amp used by a touring sound company.

 

[projectMalamute]

Uh, no. This post is a big step backwards.

so a CA9 is not a class H amp?

I suppose they have it just as wrong as the guys at Crown?

 

[Bob Lee (QSC)]

so a CA9 is not a class H amp?

I suppose they have it just as wrong as the guys at Crown?

Where do you get this from?

 

[greenboy]

Geez, why is it that threads full of great tech info end up like this? For those of us who aren't professional engineers, this thing is a goldmine.

Here, Will. Maybe reading through some of this will prove revealing and interesting. Among other things, you'll likely run across a couple of revelations about a holy grail for true digital power amps and not what currently is being done, as well as seeing how pervading active enclosures and DSP are in high-end installation and touring.

Invalid Link Removed

Enjoy!

 

[seamonkey]

yup... not picking on the original comment but i wish everyone could keep this in mind. and the best scientists will agree with this statement. true science is just a method and nothing is ever even "proven" 100% by it.

Sure - there's no proofs in science. It's all theory. Like the electron theory. We have it down enough to build amps.

I happen to work for the IEEE. It's nice to see us mentioned. [Invalid or Expired Link Removed]

This is a great paper! Needs to be FAQed.


Why does technology get smaller better and lighter - beats me:

"Where a calculator on the ENIAC is equipped with 18,000 vacuum tubes and weighs 30 tons, computers in the future may have only 1,000 vaccuum tubes and perhaps weigh 1.5 tons."
Popular Mechanics, March 1949

 

[Febs]

When you get a gazillion watts in a 3 pound package for 200 bucks you are not getting the same kind of amp used by a touring sound company.

As far as I can tell, no one in this thread is claiming otherwise.

 

[will33]

Here, Will. Maybe reading through some of this will prove revealing and interesting. Among other things, you'll likely run across a couple of revelations about a holy grail for true digital power amps and not what currently is being done, as well as seeing how pervading active enclosures and DSP are in high-end installation and touring.

Invalid Link Removed

Enjoy!

Thank you. Both for the link, and for trying to bring some sanity in here.

 

[BurningSkies]

Here, Will. Maybe reading through some of this will prove revealing and interesting. Among other things, you'll likely run across a couple of revelations about a holy grail for true digital power amps and not what currently is being done, as well as seeing how pervading active enclosures and DSP are in high-end installation and touring.

Invalid Link Removed

Enjoy!

Apparently every other group is using LA-8's...

Invalid Link Removed

 

[thudfromafar]

Sure - there's no proofs in science. It's all theory. Like the electron theory. We have it down enough to build amps.

yep, science is good for stuff like that.
it's nice to work with what we have, but we gotta stay open minded anyways. fact is, "science" doesn't know anything for sure. technically it's a method, not a body of knowledge. but yes, we have enough to be functional and progressive in a multitude of ways.

 

[R Baer]

I happen to work for the IEEE. It's nice to see us mentioned. [Invalid or Expired Link Removed]

Interesting article. We had many conversations with Bruno during the course of the R&D on the Valkyrie. Of all the class D modules we tried, I liked the sound of the Hypex the best. Our engineer also felt it was the most well designed module of the bunch. Very nice stuff, but considerably more expensive than the IcePower modules, especially at the time, with the not so favorable exchange rates.

 

[will33]

Apparently every other group is using LA-8's...

Invalid Link Removed

Thanks to you too. You see folks, if you can get past the "my dad can beat up your dad" stuff, when discussions like this take place, just bookmark links, copy-n-paste phrases, terms, etc. for later research, self-teaching, etc. It's an incredible wealth of information that can give you months worth of reading, a better understanding, and an insight on how to apply this stuff to your own personal situation. Even if that just means setting up your band a little different, training your ears better, or selecting a driver other than the one you thought was best at first glance, rather than running an entire engineering division, developing the latest technology, or planning FOH support for a big-name tour, you'll be much better of in the wnd if you sit back, soak it up, and use it rather than throw mud. Some of this stuff is akin to arguing with your other half.....after a while, you don't even know what it is you're arguing about, just that you must argue.

 

[Chef]

Exactly.

There's no need to defend any viewpoint "to the death," confrontationally.
One can choose to make one's point and move on.
Or, read up.

Be nice to each other, first and foremost.

Big thanks to those who have hung in, been polite, and contributed helpful tech here.

Thanks to you too. You see folks, if you can get past the "my dad can beat up your dad" stuff, when discussions like this take place, just bookmark links, copy-n-paste phrases, terms, etc. for later research, self-teaching, etc. It's an incredible wealth of information.....after a while, you don't even know what it is you're arguing about, just that you must argue.

 

[agedhorse]

For those interested in learning and expanding their technical horizons, here are the basic classes and some related topology comments that go along with the classes. Other engineering types feel free to jump in and add to this where I have missed something or did a less than clear description...

CLASS A: the conduction angle of the current flows through the output device(s) is for the entire 360 degrees.

CLASS B: the conduction angle of the current flows through the output device(s) is for 180 degrees. Topologies can include single ended push-pull, bipolar push-pull, capacitor coupled, transformer coupled, direct coupled.

CLASS AB: the conduction angle of the current flows through the output device(s) is for slightly greater than 180 degrees.

CLASS G: the basic amplifier is either class A, AB or B, the difference is that the power supply rails are tiered and the upper tiers are driven by a voltage offset of the audio waveform and bring in the higher rail(s) as needed to keep the supply voltage ~10V above what is needed by the lower rail class A/AB/B amp. It’s a form of linear rail modulation that serves to increase efficiency.

Class H: the basic amplifier is either class A, AB or B, the difference is that the power supply rails are tiered and the upper tiers are driven by a voltage commutater to switch in the higher rail(s) as needed to keep the supply voltage 1 tier above what is needed by the lower rail class A/AB/B amp. It’s a form of non-linear rail modulation that serves to increase efficiency.

CLASS TD: the basic amplifier is either class A, AB or B, the difference is that the power supply rails are tiered (generally 2 tiers) and the lower voltage rail is driven by a class D amplifier to modulate the voltage of the lower rail to keep the supply voltage about 5-10V above what is needed by the lower rail class A/AB/B amp. It’s a hybrid form of linear/non-linear rail modulation that serves to increase efficiency. This is what is at the core of Lab Gruppens (larger) amplifiers. It’s also used in variations in products by Carver (Sunfire) and Bash Audio.

CLASS D: the basic process is that the analog audio waveform is converted to a PWM representation (an amplitude to time transform) at a logic level, then it’s level shifted to the voltage necessary to achieve the desired output power, then the time to amplitude transform is made converting the high voltage, high current PWM signal back into a linear analog signal using a reconstructing low pass filter. All of the class D amps follow this basic process. Differences between them include things like feedback topologies (pre-filter, post-filter or combination), oscillating types (clocked, self oscillating, etc), switching bridge configuration, and deadtime management.

CLASS I: based off of the class D topology above, Crown switches the bridge in quadrature which increases the effective switching speed. This brings some added efficiency as switching losses are non-linear with respect to frequency. By slowing down each switch by 50%, switching losses in each switch may decrease by a factor of 4 or more. This yields a savings of 50% minimum even with more devices required. It’s was even more significant when device speeds were slower than they are today.

Class T: Based on the class D topology above, it is based Tripath’s method of output stage drive and modulation.

CLASS C: the conduction angle of the current flows through the output device(s) is for less than 180 degrees. This class is rarely used in audio amplifiers but is common in RF amplifiers driving heavy tuned circuits.

 

[Chef]

AgedHorse:

Thanks, that's the first time I've seen that all in one place, and I like it

 

[projectMalamute]

For those interested in learning and expanding their technical horizons, here are the basic classes and some related topology comments that go along with the classes. Other engineering types feel free to jump in and add to this where I have missed something or did a less than clear description...

CLASS A: the conduction angle of the current flows through the output device(s) is for the entire 360 degrees.

CLASS B: the conduction angle of the current flows through the output device(s) is for 180 degrees. Topologies can include single ended push-pull, bipolar push-pull, capacitor coupled, transformer coupled, direct coupled.

CLASS AB: the conduction angle of the current flows through the output device(s) is for slightly greater than 180 degrees.

CLASS G: the basic amplifier is either class A, AB or B, the difference is that the power supply rails are tiered and the upper tiers are driven by a voltage offset of the audio waveform and bring in the higher rail(s) as needed to keep the supply voltage ~10V above what is needed by the lower rail class A/AB/B amp. It’s a form of linear rail modulation that serves to increase efficiency.

Class H: the basic amplifier is either class A, AB or B, the difference is that the power supply rails are tiered and the upper tiers are driven by a voltage commutater to switch in the higher rail(s) as needed to keep the supply voltage 1 tier above what is needed by the lower rail class A/AB/B amp. It’s a form of non-linear rail modulation that serves to increase efficiency.

CLASS TD: the basic amplifier is either class A, AB or B, the difference is that the power supply rails are tiered (generally 2 tiers) and the lower voltage rail is driven by a class D amplifier to modulate the voltage of the lower rail to keep the supply voltage about 5-10V above what is needed by the lower rail class A/AB/B amp. It’s a hybrid form of linear/non-linear rail modulation that serves to increase efficiency. This is what is at the core of Lab Gruppens (larger) amplifiers. It’s also used in variations in products by Carver (Sunfire) and Bash Audio.

CLASS D: the basic process is that the analog audio waveform is converted to a PWM representation (an amplitude to time transform) at a logic level, then it’s level shifted to the voltage necessary to achieve the desired output power, then the time to amplitude transform is made converting the high voltage, high current PWM signal back into a linear analog signal using a reconstructing low pass filter. All of the class D amps follow this basic process. Differences between them include things like feedback topologies (pre-filter, post-filter or combination), oscillating types (clocked, self oscillating, etc), switching bridge configuration, and deadtime management.

CLASS I: based off of the class D topology above, Crown switches the bridge in quadrature which increases the effective switching speed. This brings some added efficiency as switching losses are non-linear with respect to frequency. By slowing down each switch by 50%, switching losses in each switch may decrease by a factor of 4 or more. This yields a savings of 50% minimum even with more devices required. It’s was even more significant when device speeds were slower than they are today.

Class T: Based on the class D topology above, it is based Tripath’s method of output stage drive and modulation.

CLASS C: the conduction angle of the current flows through the output device(s) is for less than 180 degrees. This class is rarely used in audio amplifiers but is common in RF amplifiers driving heavy tuned circuits.

I would just like to point out that class D, class I, and class TD are listed here as different things.

Which is exactly what I have been saying this entire time.

 

[will33]

AgedHorse:

Thanks, that's the first time I've seen that all in one place, and I like it

+1

I'd never even heard of class T or TD,....seems as if we've skipped some letters.