Topping B100 Amplifier Review

[Deleted member 72219]

I hope one day they make a new strong allinone amp like the MX3s.

Topping MX7s

Can't seem to find a single thing on this MX7 or MX7s. Does it even exist?

-Ed

 

[Roland68]

Can't seem to find a single thing on this MX7 or MX7s. Does it even exist?

-Ed

I hope one day they make a new strong allinone amp like the MX3s.

Topping MX7s

This is exactly how rumors start...

Next, an influencer asks Topping for an MX7s test sample

 

[Mike_Trueman]

Sorry for the mistake, I just want an MX7s. People can dream after all.

 

[Mort]

Topping MX7!!!! Take my money!

 

[Bleib]

This is exactly how rumors start...

Next, an influencer asks Topping for an MX7s test sample

And a month later they've made it

 

[ryanmh1]

@ you three distortion and noise adding experts (nice discussion btw.): If the power supply would have had less distortion at 60 Hz, 120 Hz, and 180 Hz - would the SINAD of the B100 be even better than 120 dB? I know, it’s only academically, because everything over 116 dB is fully transparent anyway, but if the power supply would have had the LA 90 48 Volt Power Supply-Qualities of minus 145 dB at 60 Hz:
the B100 would have killed the LA90 discrete by even more than 0.5 dB margin?

SNR is at 130dB. Since SINAD is significantly "worse" at 120dB, I think that would suggest SINAD is distortion limited, since SINAD is just THD+N. Typically, THD+N/SINAD tends to be only about 3dB worse than SNR at low powers. Rare instance where the SNR is so high that the amp is already largely THD constrained at 1W/5W.

And I don't want to belabor the point, but while "everything over 116dB is fully transparent" is true, the threshold is vastly lower than that for a power amplifier. Vastly. Use Benchmark's calculator which is pretty accurate: https://benchmarkmedia.com/blogs/calculators/thd-n-absolute-inaudibility-calculator. I'm reluctant to point to it since there's some questionable commentary designed to sell Benchmark products, but here we go. The problem is conflating "audibility" into a single SINAD/THD+N figure. To get it right and make it work, you first need to ignore SINAD at full power and first use SNR instead. (If you want to get into the nitty gritty of it, the calculator is "false" in claiming a 0dB level for inaudibility, in that an amp would have to have -110dB THD+N at full power, which is ludicrous. That's .0003% distortion at full power. Way beyond remotely audible. So do noise first. The level for 0dB noise on a 200W amp on 87dB/2.83V/1M speakers is around -105dB. My Denon receiver is rated 100dB A-weighted, but appearently measured -110db(!). Either way this is totally inaudible.

Once you've checked on the noise issue, you can then safely add harmonic distortion on top of that up to at least .1% at full power according to virtually every study ever done, and be absolutely positively inaudible unless you have the goldenest ears around. The issue with the .1% THD threshold is that an amp with that level might have IMD problems, which are more audible. So let's dial this back to .05% THD (-66dB) from 20Hz to 20kHz. You're pretty safe there. Dial it further to .02% (-74dB) and you're probably completely safe (10kHz/15kHZ/ideally 20kHz THD figures will give us a good measure of an amps propensity for IMD, since it is a linearity test). Again, this assumes correct gain staging and feeding it with something perhaps 10dB cleaner than the amplifier.

In the end, on the 87dB speakers you just need about -105dB SNR and distortion under .02% from 20Hz to 20kHz and you're pretty much guaranteed total transparency (on the typical distortion and noise metrics) unless you clip the amp or have some other issue with capacitive or inductive loads. Note that this does not necessarily apply quite the same to Class D amplifiers, for reasons that are a whole article.

If you're drawing the conclusion that you do not need the performance this amplifier offers and cannot possible come close to hearing it, you're absolutely right. I love seeing numbers like this, but I want to be careful we don't start fetishizing them as some sort of benchmark of necessary or even desirable performance, from an audibility perspective. It's a bit like looking at a 4K 60" TV from 15 feet away. 1080p will appear identical. As I and other have already said, if you're a dyed in the wool numbers guy, outside of some oddball cases like highly sensitive horn speakers, you're unlike to be able to blind test the difference between this (or a Benchmark or whatever) and a $300 Denon or Onkyo receiver (or equivalent) with music playing. You don't need 116dB SINAD for transparency. Your number is probably high by 40dB, for amplifier distortion. For amplifier noise, a bit high, but not as much. 10db? But at full power. Always remember to use full power SNR when doing the calculation, not 5W.

I think I've gotten this mostly right, but anyone feel free to jump in and correct it, if they think my numbers are way off the mark here.

 

[amirm]

@ you three distortion and noise adding experts (nice discussion btw.): If the power supply would have had less distortion at 60 Hz, 120 Hz, and 180 Hz - would the SINAD of the B100 be even better than 120 dB?

No. The spikes are at -140 dB which is not an impact on SINAD of 120 dB. Broadband noise is some 30 dB higher than what FFT shows which would swamp those spikes.

 

[HYQID]

@ you three distortion and noise adding experts (nice discussion btw.): If the power supply would have had less distortion at 60 Hz, 120 Hz, and 180 Hz - would the SINAD of the B100 be even better than 120 dB? I know, it’s only academically, because everything over 116 dB is fully transparent anyway, but if the power supply would have had the LA 90 48 Volt Power Supply-Qualities of minus 145 dB at 60 Hz:
the B100 would have killed the LA90 discrete by even more than 0.5 dB margin?
View attachment 394063

You can see what Topping post in the product page.Power amplifier test is vulnerable to interferences such as radiation the power supply.With all things put in the right place to minimize interferences,it will achieve near 123dB SINAD just like a HPamp from topping because of the same low noise and distortion.

 

[HYQID]

I hope one day they make a new strong allinone amp like the MX3s.

Topping MX7s

There will be MX7 first then the s.

 

[terryforsythe]

I think I've gotten this mostly right, but anyone feel free to jump in and correct it, if they think my numbers are way off the mark here.

Unless I am wrong (I could be, but I don't think so), depending on what sources are being used (e,g., DAC, phono stage, etc.), I think the bigger issue with this amplifier will be the noise from the input sources. It is amplified by the amplifier, the amount of amplification depending on the gain of the amplifier.

 

[ryanmh1]

No. The spikes are at -140 dB which is not an impact on SINAD of 120 dB. Broadband noise is some 30 dB higher than what FFT shows which would swamp those spikes.

Then what explains SNR being where it is? If medium gain SNR is 128.5db, mgSINAD can only be at 119.52 if mgTHD is -120.1. The three are all mathematically related and solve for one another. You got me unless one of the measurements hit an instrumentation limit.

Edit: I looked at the numbers again this morning. According to AP's datasheets and Topping, the low gain measurement is basically measuring the analyzer. But SINAD is apparently THD limited. Topping measured even higher SNR than APx was able to do by using a booster. So it's -135dB+. Distortion, if not just the analyzer residual, is -127dB at 1kHz, which would still cause SINAD to be distortion limited. You probably need to rely on high gain to have any sense of a potentially reliable measurement. Take the high gain for example. Visible Distortion components of -145, -148, -144, -148, etc will sum up (despite being down in the noise), eventually (since the equation says they must), to around -114dB. Then you have the SNR at 122dB, and end up with overall SINAD around 113. But since you have noise there, which is also distortion (at least from a visible perspective), you're adding any "noise" at the remaining harmonics, so you still end up being distortion limited despite the remaining harmonics being "invisible" in the noise. At least, that's what the math says must be going on. Utterly pointless academic exercise, but interesting nonetheless. And how reliable is any of this when you're so close to instrument limits? Who knows.

 

[Bleib]

Put any speaker into the calculation and see the number drop drastically. Now add to it recording quality as well, a lot of them not exactly superb

 

[amirm]

Put any speaker into the calculation and see the number drop drastically.

Noise level/dynamic range won't change as speakers are totally silent when not driven.

 

[totti1965]

You can see what Topping post in the product page.Power amplifier test is vulnerable to interferences such as radiation the power supply.With all things put in the right place to minimize interferences,it will achieve near 123dB SINAD just like a HPamp from topping because of the same low noise and distortion.

No. The Topping Datasheed shows A-weighted data. @amirm is a little stricter about it!

But I have another question:
The frequency respose of the Topping Datasheed shows fantastic bandwith with less than 1 dB loss at 100 kHz at 4 Ohm and 8 Ohm
@amirm measurements are showing about minus 3 dB at 100 kHz

Is this also related to the A-weighting of the Fosi-guys?
Or is there another explanation? Perhaps two different samples (One for the Chinese market and one for export (EU and USA) with coils in the output for slightly higher interference immunity but lower bandwith [worse rise time and slew rate]?) I am just speculating.

 

[pjug]

Use Benchmark's calculator which is pretty accurate: https://benchmarkmedia.com/blogs/calculators/thd-n-absolute-inaudibility-calculator.

I don't get that calculator. It asks for SNR of noisiest component. But then it wants to force including 25dB amplifier gain. But usually the power amplifier is the noisiest component. Well I can do it in my head anyway. And I don't care about listening position, I want to know if I can hear it when I walk by one of my speakers. Easy rule of thumb is if your amplifier SNR at 2.83V is about the same or better than your speaker sensitivity you won't hear anything in a normal room. In my case I don't hear any noise with my ears within a couple inches of 85dB speakers, even though my amplifier SNR at 2.83V is only about 82dB.

 

[somebodyelse]

But then it wants to force including 25dB amplifier gain.

That is calculated from the amplifier power you entered a few lines above - somewhere round 300W would give that.

 

[pjug]

That is calculated from the amplifier power you entered a few lines above - somewhere round 300W would give that.

So then for entering THD+N (I prefer using just SNR) use the value at full power? When I do that, use single speaker, and use 0.1 meters distance I get SPL in the 20s so seems reasonable. Sorry for my misunderstanding!

 

[ryanmh1]

I don't get that calculator. It asks for SNR of noisiest component. But then it wants to force including 25dB amplifier gain. But usually the power amplifier is the noisiest component. Well I can do it in my head anyway. And I don't care about listening position, I want to know if I can hear it when I walk by one of my speakers. Easy rule of thumb is if your amplifier SNR at 2.83V is about the same or better than your speaker sensitivity you won't hear anything in a normal room. In my case I don't hear any noise with my ears within a couple inches of 85dB speakers, even though my amplifier SNR at 2.83V is only about 82dB.

In the real world, yes. Listening for noise probably works fine. Using the calculator is just for numbers junkies who think they can hear whatever they can see on a chart... No one would ever do that here, right? It's a good pre-check for unusual situations too, I suppose.

So then for entering THD+N (I prefer using just SNR) use the value at full power? When I do that, use single speaker, and use 0.1 meters distance I get SPL in the 20s so seems reasonable. Sorry for my misunderstanding!

Use SNR at the same power level it was measured, and it should work, mostly. SNR measured at full power? Input the full power. Measured at 1W? Input 1W. Use both speakers. Use distance of 2.5M or however far away you sit. Benchmark's calculator is for guaranteed inaudibility at listening position of noise, when used in this fashion. In the real world, 10dB probably does it. Dedicated listening rooms you might need the full 0dB, or close to it if you're terribly concerned about absolutely zero hiss. One you hit "play" though the noise floor on most recordings is going to undo that quickly, although digitally mastered stuff can be almost dead silent.

This is all quite fascinating when you have decades of reading audio reviews under the belt. The whole "inky blackness" and "low level detail" stuff starts to make sense when you look at the SNR of various pieces of audiophile equipment. Not all of the florid language is or was wrong. Some old megabuck stuff was quite noisy. Sometimes worse than 70 SNR at 1W, which might equate to 10dB of potentially audible noise at the listening position. Put that amp into a quiet, dedicated listening room, and different amplifiers could sound different when it comes to "detail retrieval" on then-new digitally mastered recordings. Huh.

 

[firestart3r]

THX standard suggests that the gain should at 23db for balanced or 29db for unbalanced.

Wouldn't be better to have the standard gain instead of 20 and 23db?

For example, if your AVR preouts are a bit low in voltage, a bit more gain could be handy.

I feel Topping wants impressive SINAD numbers so they bring the gain down...

 

[dpuopolo]

OK, I found some information on this. Perhaps the best explanation is this one from Keysight support:
_________________________________

A feedforward amplifier is a type of power amplifier that utilizes a feedforward technique to improve linearity.
Here is a step-by-step explanation of how a feedforward amplifier works:

1. The input signal is split into two paths: the main path and the error path.
2. The main path signal is amplified by the power amplifier.
3. The error path signal is delayed to match the propagation delay of the main path signal.
4. The delayed error path signal is then amplified by a separate amplifier called the error amplifier.
5. The amplified error path signal is subtracted from the main path signal.
6. The resulting difference signal, known as the error signal, represents the distortion introduced by the power amplifier.
7. The error signal is then fed back to the input of the power amplifier.
8. By subtracting the error signal from the input signal, the power amplifier is forced to follow the input signal more closely, reducing distortion.
9. The combination of the power amplifier and the error amplifier effectively cancels out the distortion introduced by the power amplifier, resulting in improved linearity.

It is important to note that feedforward amplifiers require precise component manufacturing and knowledge of the signal path to achieve optimal performance. Additionally, the loop in a feedforward amplifier is not self-stabilizing, so careful monitoring is necessary to maintain stability. Feedforward amplifiers are typically used in high-frequency (HF) and lower very high frequency (VHF) applications due to the requirement of small delays for stability.
________________________________

It looks like Topping have used their experience with radio frequency design concepts to apply them to an audio amplifier. That's very interesting stuff. I can see how -- if implemented successfully -- a feed-forward network could be used to remove the crossover notch from a class B output stage. But this is the kicker...

"...feedforward amplifiers require precise component manufacturing and knowledge of the signal path to achieve optimal performance. Additionally, the loop in a feedforward amplifier is not self-stabilizing, so careful monitoring is necessary to maintain stability."

Is this Topping amplifier unconditionally stable, under all conditions of operation? Yes, it appears to be so with sine waves applied to its input, but what about a 10kHz square wave, or other more difficult tests? Is the 'monitoring' constant enough for the desired results at low frequencies such as those in the audio band?

It could be that this amplifier accomplishes all the goals, including stability under all signal conditions. That would make it a true breakthrough product. The smallish power output is probably a product of the complexity of this kind of design compared to the usual class AB or class D designs.

I wonder how this amp sounds in operation, driving today's relatively difficult speakers with very wavy impedance curves dipping down to under 4 ohms with steep phase reversal angles, etc.

This is called PRE DISTORTION where you add in an inverse distortion signal to make the output better. It's been used in Tube TV transmitters for just about FOREVER! I published an article about it six years ago: https://www.radioworld.com/tech-and-gear/put-synergies-to-work-at-your-station