Manufacturers like Infineon add to this myth unfortunately. Of course the marketing cowboys have a way with such out of context facts.
https://www.infineon.com/cms/en/pro...egrated-class-d-audio-amplifier-ics/ma5332ms/
I have devices with the chip and it sure needs a (quite small) heatsink. However, if I would like to waste energy to useless heat and desired a 2 x 20W class A amplifier only the needed heatsink and a PSU would be more expensive than that complete class D device. When comparing side by side playing at the same output level most people would perceive the 2 x 160W MS5332MS device to be "cool" and "small" and the 2 x 20W class A device to be "very hot" and "large".
All off topic as the B100 don't seem to be class D but good old class AB possibly cleverly combined with other classic topologies resulting in best of class measurements. For 299$ per channel.
https://www.infineon.com/cms/en/pro...egrated-class-d-audio-amplifier-ics/ma5332ms/
I have devices with the chip and it sure needs a (quite small) heatsink. However, if I would like to waste energy to useless heat and desired a 2 x 20W class A amplifier only the needed heatsink and a PSU would be more expensive than that complete class D device. When comparing side by side playing at the same output level most people would perceive the 2 x 160W MS5332MS device to be "cool" and "small" and the 2 x 20W class A device to be "very hot" and "large".
All off topic as the B100 don't seem to be class D but good old class AB possibly cleverly combined with other classic topologies resulting in best of class measurements. For 299$ per channel.
Last edited:
Hypex power ratings are peak power.
Continuous power is typically 1/5th of peak power, and cooling must be good enough for that continuous power rating.
Recently I assembled an NCx500; what struck me was that semiconductors (also SMPS) kept relatively cool, but the output inductor got really hot; actually the hottest component in the amplifier.
I reported back to Hypex and they confirmed to be aware.
Continuous power is typically 1/5th of peak power, and cooling must be good enough for that continuous power rating.
Recently I assembled an NCx500; what struck me was that semiconductors (also SMPS) kept relatively cool, but the output inductor got really hot; actually the hottest component in the amplifier.
I reported back to Hypex and they confirmed to be aware.
This is a DIY community and there is important question regarding the present development; both technologically and economically.
Let me be speciffic: there is a signifficant market here, running in background, of "ultra-low distortion amplifiers". Obviously, Topping devices beat them dearly, in both compartments: the prices of both Topping devices are indeed tempting whilst they offer unbeatable set of parameters.
Let me be speciffic: there is a signifficant market here, running in background, of "ultra-low distortion amplifiers". Obviously, Topping devices beat them dearly, in both compartments: the prices of both Topping devices are indeed tempting whilst they offer unbeatable set of parameters.
This amplifier is tempting, even for DIY purposes. I’m planning ‘monoblocks’ for active loudspeakers, so at least two different amplifiers in a single case (class D for LF range). Looks like well suited, distortion and power wise, for mid and HF range. Adding a decent linear power supply could improve things (or not).
Alternative, for me, is class A with close enough distortion and noise levels, but it would cost more. We still don’t know enough how it behaves with real loudspeaker loads and conditions.
Alternative, for me, is class A with close enough distortion and noise levels, but it would cost more. We still don’t know enough how it behaves with real loudspeaker loads and conditions.
What I don't get here with regards to a textbook Class B are 3 things:
1) mitigation of crossover distortion
2) idle power vs. max power
3) amount of feedback (no figures anywhere just "nested" and other fancy buzzwords)..
An amplifier class - by default - is defined as: in how much of the 360° input signal's cycle does an amplifier's output device(s) conduct.
- For Class A, this is 100%, output device is all the time conducting and ON. Least efficient design: with no input signal applied, there's still a ton of heat generated (idle current).
- For Class B, this is 50% for one device and 50% for the other device, so each of them are in 0-180° and 181-360° conducting (ON state). In theory.
In practice, there's crossover distortion which comes from the fact that at the crossover point one device isn't conducting anymore while the other hasn't started conducting yet. Efficiency is way better than for Class A since idle current is only half (or somewhat less) then in case of Class A.
- For Class AB, this is more than 50% for both devices, with a slight overlap to mitigate crossover distortion - a tiny bit less efficient then Class B (depending on the width of the overlapping region, more overlapping = more Class A-like operation). Still more efficient than pure Class A however the first couple of watts are behaving (can be described) as Class A.
- Class D - highly efficient design with a somewhat more complex signal processing (chopper, comparator, PWM/PDM and a filter circuit at the very end to get rid of ultrasonic content, nowadays with a PFNF - post filter negative feedback for real frequency-independent behavior). Energy efficiency is very high, idle current is very low.
- Class G, H - different implementation of Class AB actually, the trick is in different / switching power rails.
When I see a B100 and look at the specs of 9.1W idling, amplifier ON and no signal, I instantly think "no way it's a Class AB.. if this is Class AB with 100W output power capability, idle power must be at around another 100-ish Watts or so instead of 9.1W".
So in general I agree with some previous comments before me that there's a trick inside this amp which the Chinese will never tell of course and for me this Class B statement alone is a weird thing already, not to mention the sheer ease and simplicity they state this is Class B. No, it isn't. Or at least it sounds like you're looking at a Formula-1 car and state "it runs on processed mineral oil". Yeah, sure, but .....
You get it ?
Crossover distortion: a separate amplifier circuit maybe working together with the enclosed real Class B design ? But still, if this is class B with 1.9W idle current and 100W output power (max), it's simply not Class B but any kind of clever hybrid I don't really care, just makes me think how Topping (or anybody) is making such bold statements it's Class B. Of course there's something else going on under the hood. I can imagine complementary circuits, nested feedback, even feed-forward (look at Hegel's diagrams) and all that kind of stuff.
For a REAL textbook Class B amp you simply cannot increase feedback (even when nested) so far that you get 100W power remaining while retaining the amp's stability - especially not with such a small footprint, because increase in feedback pushes any amplifier to the direction of instability and less power and these both have a limit where an amp easily gets a self-oscillator and boom.
Or maybe it's a non-linear amplifier, consuming very little idle current when idle while still being capable of 'enormous' output power in a Class B fashion and the huge nonlinearity (and also crossover distortion) are mitigated via nested feedback and built-in auxiliary amplifiers and aforementioned techniques to produce such great numbers at the end.
It would be very interesting to look at the block diagram of this amplifier but I think this will never come. 🙂
So, these were my 5 cents only.. just imagining the unimaginable.
Class B ? Maybe, but then it's not the whole picture I think.
Far from it, actually.
Same as you would state "Class A" for a Hypex Class D just because the input buffers' op-amps are in Class A operation.
Cheers 😉
1) mitigation of crossover distortion
2) idle power vs. max power
3) amount of feedback (no figures anywhere just "nested" and other fancy buzzwords)..
An amplifier class - by default - is defined as: in how much of the 360° input signal's cycle does an amplifier's output device(s) conduct.
- For Class A, this is 100%, output device is all the time conducting and ON. Least efficient design: with no input signal applied, there's still a ton of heat generated (idle current).
- For Class B, this is 50% for one device and 50% for the other device, so each of them are in 0-180° and 181-360° conducting (ON state). In theory.
In practice, there's crossover distortion which comes from the fact that at the crossover point one device isn't conducting anymore while the other hasn't started conducting yet. Efficiency is way better than for Class A since idle current is only half (or somewhat less) then in case of Class A.
- For Class AB, this is more than 50% for both devices, with a slight overlap to mitigate crossover distortion - a tiny bit less efficient then Class B (depending on the width of the overlapping region, more overlapping = more Class A-like operation). Still more efficient than pure Class A however the first couple of watts are behaving (can be described) as Class A.
- Class D - highly efficient design with a somewhat more complex signal processing (chopper, comparator, PWM/PDM and a filter circuit at the very end to get rid of ultrasonic content, nowadays with a PFNF - post filter negative feedback for real frequency-independent behavior). Energy efficiency is very high, idle current is very low.
- Class G, H - different implementation of Class AB actually, the trick is in different / switching power rails.
When I see a B100 and look at the specs of 9.1W idling, amplifier ON and no signal, I instantly think "no way it's a Class AB.. if this is Class AB with 100W output power capability, idle power must be at around another 100-ish Watts or so instead of 9.1W".
So in general I agree with some previous comments before me that there's a trick inside this amp which the Chinese will never tell of course and for me this Class B statement alone is a weird thing already, not to mention the sheer ease and simplicity they state this is Class B. No, it isn't. Or at least it sounds like you're looking at a Formula-1 car and state "it runs on processed mineral oil". Yeah, sure, but .....
You get it ?
Crossover distortion: a separate amplifier circuit maybe working together with the enclosed real Class B design ? But still, if this is class B with 1.9W idle current and 100W output power (max), it's simply not Class B but any kind of clever hybrid I don't really care, just makes me think how Topping (or anybody) is making such bold statements it's Class B. Of course there's something else going on under the hood. I can imagine complementary circuits, nested feedback, even feed-forward (look at Hegel's diagrams) and all that kind of stuff.
For a REAL textbook Class B amp you simply cannot increase feedback (even when nested) so far that you get 100W power remaining while retaining the amp's stability - especially not with such a small footprint, because increase in feedback pushes any amplifier to the direction of instability and less power and these both have a limit where an amp easily gets a self-oscillator and boom.
Or maybe it's a non-linear amplifier, consuming very little idle current when idle while still being capable of 'enormous' output power in a Class B fashion and the huge nonlinearity (and also crossover distortion) are mitigated via nested feedback and built-in auxiliary amplifiers and aforementioned techniques to produce such great numbers at the end.
It would be very interesting to look at the block diagram of this amplifier but I think this will never come. 🙂
So, these were my 5 cents only.. just imagining the unimaginable.
Class B ? Maybe, but then it's not the whole picture I think.
Far from it, actually.
Same as you would state "Class A" for a Hypex Class D just because the input buffers' op-amps are in Class A operation.
Cheers 😉
wouldn’t this be that of Class A for that level of high idle consumption?
To me B100 looking more like Class AB for that sort of low idle power,.
Also, the Class of operation refers to the power output stages and not the preceding buffers, VAS etc that will be low power Class A for best linearity
Last edited:
Looks something similar to Benchmark AHB2, using nested feedback and feedforward correction to achieve ultra low distortion figures. Keeping SMPS external reduces noise (cheaply) to better AHB2 (psu built-in) in SINAD - though seems a bit cumbersome with dual mono arrangement for the B100.
Interesting to see how these compare in terms of actual performance - and then the price difference $600 vs $3,500 😱
Interesting to see how these compare in terms of actual performance - and then the price difference $600 vs $3,500 😱
Surprisingly none of the wealthy diyaudio.com guys reports the results of the purchase of 2 x B100. These B100 cost peanuts compared to the average device displayed here. Or do many have similar objections? Two amplifiers, two adapters and one preamp so a five device setup just for amplification...Personally I say NEVER.
Now what about having all these in 1 x 19" casing? The return of the "DIY" part of activities! Out of the box thinking 😀
Now what about having all these in 1 x 19" casing? The return of the "DIY" part of activities! Out of the box thinking 😀
Last edited:
How would the noise figures look like, especially with B100s smps all inside the box?
I don't care, I am allergic to the number 5. And multichannel but you knew that 🙂 Not even if I got either for free I would consider to use them.
Maybe those SMPS can be separated/shielded with metal walls inside the casing. Anything in a single casing is better than having to use 5 devices and all the necessary cabling for amplification.
Maybe those SMPS can be separated/shielded with metal walls inside the casing. Anything in a single casing is better than having to use 5 devices and all the necessary cabling for amplification.
Last edited:
Don’t forget to add to the costs re: the final build for all extra work on the B100s
Even thogh B100's spec is impressive, I have no need for them. If the goals is objective performance, I can build a PGP amplifier by @syn08 and @Edmond Stuart with comparable or even better performance but can provide 100W/8 Ohms-200W/4 ohms for a bit more money. If the goals is subjective performace, then First Watt amps (for low power) or trying to clone an Accuphase ones (P7100, M6000) is a better choice, and they are more satisfied for me as diyer.
And max 83W/4 Ohms and 50W/8 ohms is hardly enough for a main amplifier to drive a low sensitivity passive speaker.
I can see them an excellent device for horn section but it is a story for next six month when I get an DFM2535
And max 83W/4 Ohms and 50W/8 ohms is hardly enough for a main amplifier to drive a low sensitivity passive speaker.
I can see them an excellent device for horn section but it is a story for next six month when I get an DFM2535
@acko: could still be less than the average serious DIY amplifier project.
Chances are that they may sound and perform better than amplifiers built or bought on emotion or brand name (which are really just biased assumptions). That would be good to know. Objective user reports and open minded real life comparisons would be nice. No one seems to be able to beat the chinese as they have just started to design good to very good products. That is in the end only good for us as to be competitive the western companies have to improve products too.
Prediction:"if you can't beat them join them". That will likely be the near future in audio. Just look in the average audio device, they are already rarely designed and produced in the west. Since we educate chinese students at our universtities what do we expect?
Chances are that they may sound and perform better than amplifiers built or bought on emotion or brand name (which are really just biased assumptions). That would be good to know. Objective user reports and open minded real life comparisons would be nice. No one seems to be able to beat the chinese as they have just started to design good to very good products. That is in the end only good for us as to be competitive the western companies have to improve products too.
Prediction:"if you can't beat them join them". That will likely be the near future in audio. Just look in the average audio device, they are already rarely designed and produced in the west. Since we educate chinese students at our universtities what do we expect?
Last edited:
Interesting to see actually so just playing around from somewhere to start 🙂
I have got Benchmark AHB2 (bought 5yrs ago) driving Pear Sibelius speakers. Not really easy speakers to drive but AHB2 has got the authority to bring out something nice. So I don’t have a requirement to upgrade to anything else for now otherwise I would have gambled on the B100s 😊 Happy to do a shootout if anyone else could loan the B100 (pair)
I have got Benchmark AHB2 (bought 5yrs ago) driving Pear Sibelius speakers. Not really easy speakers to drive but AHB2 has got the authority to bring out something nice. So I don’t have a requirement to upgrade to anything else for now otherwise I would have gambled on the B100s 😊 Happy to do a shootout if anyone else could loan the B100 (pair)
Sure those B100's would have been only about 30% of the cost of the AHB2. Not really an honest comparison but possible. Then discover that they are better....
Lower cost is driving the audio market for many enthusiasts (probably the majority) . If it would not be like that we would have Acko DACs. Explains partly why ASR is so popular.
vs AHB2 : has smps built in and stereo. Elegant solution and engineering feat to keep noise so low. Compact unit and runs cool. No mods needed to suit
If it would need mods it would not be worth 3999 Euro. It better has elegant engineering and an excellent SMPS in it too. Mods are usually done to devices that have design imperfections due to cost cutting. The 13th Law of DIYaudio.com says "the lower the price the higher the chance things have been done too cheaply". Now rumor is that the 13th Law is about to be refuted by chinese designers.
Completely different segment of audio to most. This thread is about 299$ mono blocks so 598$ for a stereo setup plus preamp. Let's say 300$ for a Topping Pre90 preamp with them so 900$ in total. That is what used to be called mid class sector but now as fully separate mono blocks in balanced configuration and relay based volume control.
Completely different segment of audio to most. This thread is about 299$ mono blocks so 598$ for a stereo setup plus preamp. Let's say 300$ for a Topping Pre90 preamp with them so 900$ in total. That is what used to be called mid class sector but now as fully separate mono blocks in balanced configuration and relay based volume control.
Last edited: