• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

NEW GaN Amplifier : Dual ACM8816 GaN amplifier chips + AK4118AEQ [YJCO D816]

I tried to understand your explanation. I only partly succeeded. Thanks anyway. So I ask as a technical amateur:
If a so-called full digital amplifier has a digital input, at what point in such a device does the conversion to an analog waveform take place, and how exactly is this achieved, with which components? I'm asking for the following reason:
Most readers know the digital-to-analog converter chips that are built into D/A converters, for example from manufacturers like ESS or AKM. So in your opinion, how does this conversion happen in a full digital amp?
If you are asking about Lyngdorf, they did not disclose exactly what they do. My guess is that they are not doing anything unusual. Basically converting the incoming digital audio into full strength PWM and send it to the speakers, and without feedback.

Here are some sample measurements of their TDAI-2210 from SoundStage!

Output frequency responses into different loads indicated that there is no post filter feedback (PFFB). Which is consistent with what was said in the Audioholics video that Lyngdorf does not use feedback in their digital amplifiers.
1788650806073.png


5 W SINAD at <80 dB is, frankly, not competitive these days, especially when the asking price is ~$5k US.
1788651296019.png
 
If you are asking about Lyngdorf, they did not disclose exactly what they do. My guess is that they are not doing anything unusual. Basically converting the incoming digital audio into full strength PWM and send it to the speakers, and without feedback.

Here are some sample measurements of their TDAI-2210 from SoundStage!

Output frequency responses into different loads indicated that there is no post filter feedback (PFFB). Which is consistent with what was said in the Audioholics video that Lyngdorf does not use feedback in their digital amplifiers.
View attachment 556828

5 W SINAD at <80 dB is, frankly, not competitive these days, especially when the asking price is ~$5k US.
View attachment 556830

Thank you very much for your reply. You yourself write:

My guess is that they are not doing anything unusual. Basically converting the incoming digital audio into full strength PWM and send it to the speakers, and without feedback.
Well, what do you think: Should we call a PWM data stream digital or analog? I googled what PWM means. It means:
Pulse-width modulation is a method in which a continuous square wave signal is generated. This means that the signal switches between two values ('1' and '0', or 'on' and 'off') within a certain period and at a consistent duty cycle. By deliberately turning devices on and off, the average power can thus be directly influenced.

PWM is mainly used in control and regulation technology. A concrete use case is controlling the dimming of LEDs. Due to their characteristics as semiconductors, the brightness cannot be changed via voltage. The LED basically only knows the state 'on' or 'off'. Through PWM and corresponding control, the LED can be turned on and off in a targeted and rapid sequence. This reduces the average power. The resulting flickering of the light is not noticeable to the human eye. The light appears dimmed to us humans.

So, I claim, if THIS isn't digital, then what is supposed to be digital?

Have you ever actually heard a full digital amp yourself? I don’t think so. I’m pretty sure. Because if you had heard one, especially a Lyngdorf, which doesn’t measure particularly well, you would have noticed this:
These amps sound clearer than ANY others. They sound more direct. They have unrestrained dynamics. The resolution is better and more distinct. The unique character of these amps is less pronounced than in conventional Class D amps, which means, in a positive sense, they sound more like “nothing”.

I know that for sure because I’ve been listening to music with these kinds of amplifiers since 2012: Lyngdorf TDAI 2200, Lyngdorf TDAI 2170, SMSL VMV A2, Sabaj a30a, Aiyima D05, and currently the brand-new YJCO D816. I’ve listened to all of these amplifiers extensively, and they all have better clarity compared to 'normal' Class D amplifiers, at least when you compare them to similarly priced devices. EVEN THOUGH they actually measure worse. Try it yourself. You’ll agree with me. For sure.
 
You’ll agree with me. For sure.
Nope. Sighted listening has no value if it is not supported by measurements. I've had a TDAI2200 for 15 years and a TacT2150 before that. I just sold it and am using a Fosi V3 stereo for now. I can't recognize any of what you write. It's an amplifier, and it sounds like any other amplifier.
 
Nope. Sighted listening has no value if it is not supported by measurements. I've had a TDAI2200 for 15 years and a TacT2150 before that. I just sold it and am using a Fosi V3 stereo for now. I can't recognize any of what you write. It's an amplifier, and it sounds like any other amplifier.

"It's an amplifier, and it sounds like any other amplifier."

Thanks a lot. So, if that's the case: why would anyone even buy a new amplifier?
 
Thanks a lot. So, if that's the case: why would anyone even buy a new amplifier?
Marketing. Reviewers and manufacturers depend on people buying new stuff when in reality there hasn't been any improvement in the last 40 years. Look at the review of the NAD 2200. It has only been marginally improved by newer amplifiers, but none of it is audible.

 
 
Marketing. Reviewers and manufacturers depend on people buying new stuff when in reality there hasn't been any improvement in the last 40 years. Look at the review of the NAD 2200. It has only been marginally improved by newer amplifiers, but none of it is audible.


Thanks for your apparently honest impression. I’ll reply, my impression is different. I’m sure that with good, high-resolution speakers, it’s possible to hear amplifier sound differences even in a blind test. Especially a tube amp is pretty easy to distinguish from a modern amplifier. Honestly, no problem.

Significantly smaller, but still easily audible to my ears are the sound differences between Class D and Full Digital Amplifiers. They sound pretty similar, but to me, the FDA amplifiers are clearly crisper, more direct, clearer, and more dynamic. At completely normal room volume.

Actually, distinguishing between FDA amplifiers is pretty difficult. In my ears, the Sabaj a30a or the SMSL VMV A2 are the best FDAs, sound-wise. But I admit that the differences I hear compared to a Lyngdorf or even the new YJCO D816 are so small that I wouldn't trust myself to recognize them with a good hit rate in a blind test.

But maybe my speakers help me with that too: I’m listening on Martin Logan Ethos electrostatics.
 
Thank you very much for your reply. You yourself write:

My guess is that they are not doing anything unusual. Basically converting the incoming digital audio into full strength PWM and send it to the speakers, and without feedback.
Well, what do you think: Should we call a PWM data stream digital or analog? I googled what PWM means. It means:
Pulse-width modulation is a method in which a continuous square wave signal is generated. This means that the signal switches between two values ('1' and '0', or 'on' and 'off') within a certain period and at a consistent duty cycle. By deliberately turning devices on and off, the average power can thus be directly influenced.

PWM is mainly used in control and regulation technology. A concrete use case is controlling the dimming of LEDs. Due to their characteristics as semiconductors, the brightness cannot be changed via voltage. The LED basically only knows the state 'on' or 'off'. Through PWM and corresponding control, the LED can be turned on and off in a targeted and rapid sequence. This reduces the average power. The resulting flickering of the light is not noticeable to the human eye. The light appears dimmed to us humans.

So, I claim, if THIS isn't digital, then what is supposed to be digital?

Have you ever actually heard a full digital amp yourself? I don’t think so. I’m pretty sure. Because if you had heard one, especially a Lyngdorf, which doesn’t measure particularly well, you would have noticed this:
These amps sound clearer than ANY others. They sound more direct. They have unrestrained dynamics. The resolution is better and more distinct. The unique character of these amps is less pronounced than in conventional Class D amps, which means, in a positive sense, they sound more like “nothing”.

I know that for sure because I’ve been listening to music with these kinds of amplifiers since 2012: Lyngdorf TDAI 2200, Lyngdorf TDAI 2170, SMSL VMV A2, Sabaj a30a, Aiyima D05, and currently the brand-new YJCO D816. I’ve listened to all of these amplifiers extensively, and they all have better clarity compared to 'normal' Class D amplifiers, at least when you compare them to similarly priced devices. EVEN THOUGH they actually measure worse. Try it yourself. You’ll agree with me. For sure.
All electrical signals are analog, period. Digital is a representation that is useful only in signal transmissions (that requires back and forth conversions), storage, and computations.

Here is, again, an example of a typical eye-diagram (for 2-levels binary, i.e. NRZ modulation. There are modulation schemes that use more than 2 levels, such as PAM4). It shows the electrical signal that represents some digital signal being transmitted.
index.php


This is an example of what the electrical signal may represent in digital form.
DigitalWaveform-1708036433.png

When translating the electrical signal to its digital representation, the absolute voltage levels do not matter a lot. For example, for 5V TTL logical levels, an input level of >2 V is considered logic high; and an input level of <0.8V is considered logical low. In-between high and low levels is considered to be an error when encountered. 3V is the same 4 V is the same as 5 V. They are all logic high. Timing is similar. Jitters in the signal timing are unimportant when they are within the window of timing tolerance. That why the eye doesn't need to look anywhere close to perfect square waves. If your eye-diagram "looks really good", you have left a lot of bandwidth on the table.

But when you want to convert the PWM pulses back to the analog waveform by passing it through a low pass filter, then everything matters. The exact timing, the exact levels, all will affect the outcome. That's why PWM are not digital. It is no longer just two discrete levels of 0 and 1. 0.0 is different from 0.1 is different from 0.2, etc. Timing is again similar. A pulse train with rise time of 10 ns will give you different analog results after filtering than a pulse train with a rise time of 11 ns. For digital, they are identical as long as they are within spec.
1788699911593.png


Therefore, you cannot treat PWM as anything other than analog.

Back to your earlier question (post #40) regarding where the DAC process in the chain is optimal. High power amplifiers driving difficult loads will require feedback for high performance. The feedback signal, optimally, has to come from the amplifier output, which is an analog signal. Therefore, it is most practical to operate the power amplifier fully in the analog domain, regardless of it being class A, AB, B, D, or something else.

The actual conversion from digital to analog inside a DAC require high precision voltage or current sources. (The highest performance DAC's are usually current output DAC's.) It is extremely difficult to construct high precision sources that can drive difficult loads at high power levels. It is much easier (= practical) to make high precision sources that are only required to drive a benign and stable load such as that of an op-amp input. Therefore, the optimal arrangement (with the technology available at this time) is to have DAC's with line level outputs to drive conventional analog amplifiers.
 
Last edited:
Thanks, NTK!
It’s hard to shake off certain beliefs...
You’ve shed light on the "100% digital" myth in an accessible way. :)
With all due respect to some...
 
  • Like
Reactions: NTK
Actually, distinguishing between FDA amplifiers is pretty difficult. In my ears, the Sabaj a30a or the SMSL VMV A2 are the best FDAs, sound-wise. But I admit that the differences I hear compared to a Lyngdorf or even the new YJCO D816 are so small that I wouldn't trust myself to recognize them with a good hit rate in a blind test.
You have to test blind and level matched. Then the differences disappear.

 
Marketing. Reviewers and manufacturers depend on people buying new stuff when in reality there hasn't been any improvement in the last 40 years. Look at the review of the NAD 2200. It has only been marginally improved by newer amplifiers, but none of it is audible.

Okay. I got it. The technology of audio amplifiers has improved over the last 40 years, but it was already so good in the 80s that you can measure these improvements, but not hear them.

So if that's true, I want to ask everyone, all readers: What are you actually doing here?
 
Okay. I got it. The technology of audio amplifiers has improved over the last 40 years, but it was already so good in the 80s that you can measure these improvements, but not hear them.

So if that's true, I want to ask everyone, all readers: What are you actually doing here?
Loudspeakers are an exception. I think tangible improvements are being made there. Also, whole new categories of equipment are coming along, like streamers.
 
I want to ask everyone, all readers: What are you actually doing here?
There are still poor designs. Hissing, bad frequency response, bad with complex loads, too sensitive protective circuits etc
 
You have to test blind and level matched. Then the differences disappear.

The differences disappear when the volume while listening is exactly the same. Ok. Do you mean the ones between FDA amplifiers or the ones between ALL good amplifier types?
 
All electrical signals are analog, period. Digital is a representation that is useful only in signal transmissions (that requires back and forth conversions), storage, and computations.

Here is, again, an example of a typical eye-diagram (for 2-levels binary, i.e. NRZ modulation. There are modulation schemes that use more than 2 levels, such as PAM4). It shows the electrical signal that represents some digital signal being transmitted.
index.php


This is an example of what the electrical signal may represent in digital form.
View attachment 556886
When translating the electrical signal to its digital representation, the absolute voltage levels do not matter a lot. For example, for 5V TTL logical levels, an input level of >2 V is considered logic high; and an input level of <0.8V is considered logical low. In-between high and low levels is considered to be an error when encountered. 3V is the same 4 V is the same as 5 V. They are all logic high. Timing is similar. Jitters in the signal timing are unimportant when they are within the window of timing tolerance. That why the eye doesn't need to look anywhere close to perfect square waves. If your eye-diagram "looks really good", you have left a lot of bandwidth on the table.

But when you want to convert the PWM pulses back to the analog waveform by passing it through a low pass filter, then everything matters. The exact timing, the exact levels, all will affect the outcome. That's why PWM are not digital. It is no longer just two discrete levels of 0 and 1. 0.0 is different from 0.1 is different from 0.2, etc. Timing is again similar. A pulse train with rise time of 10 ns will give you different analog results after filtering than a pulse train with a rise time of 11 ns. For digital, they are identical as long as they are within spec.
View attachment 556888

Therefore, you cannot treat PWM as anything other than analog.

Back to your earlier question (post #40) regarding where the DAC process in the chain is optimal. High power amplifiers driving difficult loads will require feedback for high performance. The feedback signal, optimally, has to come from the amplifier output, which is an analog signal. Therefore, it is most practical to operate the power amplifier fully in the analog domain, regardless of it being class A, AB, B, D, or something else.

The actual conversion from digital to analog inside a DAC require high precision voltage or current sources. (The highest performance DAC's are usually current output DAC's.) It is extremely difficult to construct high precision sources that can drive difficult loads at high power levels. It is much easier (= practical) to make high precision sources that are only required to drive a benign and stable load such as that of an op-amp input. Therefore, the optimal arrangement (with the technology available at this time) is to have DAC's with line level outputs to drive conventional analog amplifiers.
Hello NTK, I honestly thank you for your explanation. My problem with it, however, is that I understand at most half of it. Maybe even less than half. Now here’s the thing: Google clearly says, in my opinion, that PWM is a digital signal form. You, on the other hand, say PWM is analog. And Lyngdorf says their amplifiers are fully digital—they even cheekily write it on the devices themselves: TDAI, Total Digital Amplifier Integrated. You say that’s nonsense, that such a thing doesn’t exist. Who should I believe?

The thing gets even weirder because apparently I'm the only one here who dares to say i can hear a difference between, let's say as an example, a 40-year-old NAD 2200 Class A/B amplifier and a current Sabaj A30A full digital amp. At exactly the same volume, in a blind test. No problem, I can do that. I'm TOTALLY sure about that. The only thing I need are very high-quality speakers, let's say a pair of passive Canton Reference 9s, and a really good classical recording, let's say from Hessischer Rundfunk. I'm 100% sure that if you're not deaf, you can do it too. And so can many others here. Bet on it.
 
You're going to end up annoying him...
Get yourself some fully digital hearing implants...
If you manage to decode the bits... you're Terminator...:rolleyes:
 
Nope. Sighted listening has no value if it is not supported by measurements. I've had a TDAI2200 for 15 years and a TacT2150 before that. I just sold it and am using a Fosi V3 stereo for now. I can't recognize any of what you write. It's an amplifier, and it sounds like any other a
I'm amazed. Honestly. You write that you had a Lyngdorf TDAI 2200 for 15 years. Before that, you had a Tact 2150. Both devices are VERY expensive, especially compared to a Fosi Audio amp. How does someone who says it’s just an amplifier and sounds like any other end up with a device like that? How does that add up?
 
I'm amazed. Honestly. You write that you had a Lyngdorf TDAI 2200 for 15 years. Before that, you had a Tact 2150. Both devices are VERY expensive, especially compared to a Fosi Audio amp. How does someone who says it’s just an amplifier and sounds like any other end up with a device like that? How does that add up?
One of the first things you learn on ASR is that price has very little to do with sound quality for electronics. Have you read the thread pma linked to?

And read a good chunk of this as well.

 
One of the first things you learn on ASR is that price has very little to do with sound quality for electronics. Have you read the thread pma linked to?

And read a good chunk of this as well.

I agree with you there. That's how it is. All the more, I wonder how someone like you, with that attitude, ends up with not just one but two very expensive fully digital amplifiers in a row? Your words: It's an amplifier, and it sounds like any other amplifier
 
Hello NTK, I honestly thank you for your explanation. My problem with it, however, is that I understand at most half of it. Maybe even less than half. Now here’s the thing: Google clearly says, in my opinion, that PWM is a digital signal form. You, on the other hand, say PWM is analog. And Lyngdorf says their amplifiers are fully digital—they even cheekily write it on the devices themselves: TDAI, Total Digital Amplifier Integrated. You say that’s nonsense, that such a thing doesn’t exist. Who should I believe?
For a "Fully Digital Amplifier", how do the speakers decode the digital signal output of the amplifier?
 
Back
Top Bottom