Topping A70 Pro Preamp to control volume - SNR Compared to direct digital and passive volume control

[Rja4000]

I measured the
Topping A70 Pro as a preamplifier

My goal was to assess its
performance in terms of noise versus level, compared to other methods.
(I was inspired by this post from @nagster , where he's done the same with the Topping Pre90)

The Topping A70 pro is said to be using stepped relays attenuator,
which is a performant way to control the volume in the analog domain.

How does it compare on that level against the DAC digital volume control and a cheap (but very performant) passive attenuator ?

The contenders:
Topping D90LE as a DAC, with volume controlled from the measurement software. I purchased it through Audiophonics for 799€ (now discontinued). It was measured with 4V range.
Topping A70 Pro as an active preamp with stepped relays volume control. I purchased it through Audiophonics, where its now at 499€. It was measured on low gain.
Behringer Monitor1, a cheap passive volume attenuator. I paid mine 49€.


Measurement chain is as follows:
Topping D90LE as a source (both output summed, 4V range, balanced output)
DUT to control volume (if any).
RME 12Mic-D as ADC, Autoset gain (gain is automatically set to avoid clipping and maximize performance/SNR)
Measurement software Virtins MultiInstrument 3.9.8.1

I limited at 1Vrms, since what's interesting for me is the noise with some attenuation.

Obviously, the passive performs better on that aspect, except at highest level.
The A70 Pro shows some benefits when the level is between 0.03V and 1V.
I don't know what the D90LE is doing internally, but there seems to be also some kind of analog change around 0.1V or so, allowing it to catch-up at very low level. Impressive, actually.

This is very much in line with the measurements performed by @nagster in 2021 (copied below for convenience).
The Topping A70 Pro performs similarly to the Topping Pre90 on that aspect.
(The Topping D90LE is said by Topping to be identical to the D90SE, if you except MQA support. Measurements support that claim.)

PS: I own 2 Behringer Monitor 1. That gives me 4 channels.
The one measured here is the one with best SNR performance.
There is up to 2-3dB difference between th extremes.
Note that the Behringer is... a Behringer: cheap. You see it when you try to unplug your genuine Neutrik XLR connectors and struggle. It's also too light and can be pulled by heavy cables.
I also own the SPL Volume 2, but I know it will not be able to compete.
I measured mine here. (That was in 2019. One of my first measurements.)

EDIT: I changed colors in my graph, to allow easier comparison with AP plot

 

[Rja4000]

Here the measurement extended to 4V

Note that the A70 Pro has a gain >0dB in low gain balanced mode.
Horizontal scale is limited at 5V.

The A70 Pro gives the best noise performance of the 3 methods between 0.6V and 2V approx. (-16.5dB to -6dB gain approx.)

Measurements here are combined results from 12Mic-D ADC (see above) and RME ADI-2/4 Pro SE with E1DA APU notch (above approx 0.36V)

 

[antcollinet]

So here is the thing. At 0.01V output, you are -52dBFS (12 dB from 4 down to 1V, then 40dB from 1V down to 0.01V


Even with the digital volume control, at -52dBFS, you still have an SNR of 83dB

So even if full volume is set at 100dB spl (Too loud), -52dB will be close to inaudible - or at least very quiet. And then the noise floor is 83dB below that. These are not levels we need to be worrying about IMO

 

[DVDdoug]

If you aren't hearing noise there is no problem either way.

SNR doesn't tell the whole story... With digital attenuation you lose SNR and resolution but that doesn't matter if you can't hear the noise. At low levels you aren't using all of the bits and when you're down to a few bits the quality is terrible but you can't hear ANYTHING anyway. The same thing happens as a song fades-out... You are using fewer bits as the volume goes-down. The absolute level of the quantization noise isn't worse even though the ratio is worse...

Where you get into trouble is if you re-amplify. Something similar happens with analog (depending on where the noise is coming from). When you re-amplify you are amplifying the signal and noise together, plus more (analog) noise will be introduced by the amplifier and the noise may become audible.

 

[antcollinet]

Where you get into trouble is if you re-amplify. Something similar happens with analog (depending on where the noise is coming from). When you re-amplify you are amplifying the signal and noise together, plus more (analog) noise will be introduced by the amplifier and the noise may become audible.

That is maybe going to be an issue if you have a high gain power amplifier. Lets have a look at that.

30dB would be pretty high gain for most amps. Lets say you typically want to listen at 85dB with 87db/W 4ohm speakers, sitting 3m from the speaker.

In that situation you (according to this calculator) need about 6W. With 4ohm speakers that means about 5V out of the amp. (all figures rounded).
https://www.extron.com/calculators/amplifier-power/?tab=tools

With a 30dB gain, you'd only want to be putting in around 0.16V. From the chart above, you'd still have around 95 to 100dB SNR : About the same as redbook. In many cases, I suspect the amplifier self noise would be a bigger problem.

If you want more than that, use an amp with lower or switchable gain, or put a passive attenuator on the input.

 

[Rja4000]

I agree: there is no issue with properly implemented digital attenuation for real use.

My purpose was more to optimize SNR for measurements. Noise is quite often the bottleneck there.

 

[Rja4000]

I updated the colors to match the AP plots
And added values up to 4V+

 

[Rja4000]

Now, we observe an interesting thing :
We have better results with the A70 Pro headphones out.

Look at the SNR measurement for 50mV I published in the A70 Pro review thread :

If you compare with the plots above, it seems that the A70 Pro is performing several dB better through its headphones out.
Actually, its SNR matches now the SNR achieved with the Behringer Monitor1 for 50mV.

Why is that ?
Is it because the 50mV headphones noise measurement is done on the 3.5mm unbalanced headphones output ?
And, as Topping specifies it, the noise should be 3x lower on unbalanced out.
Which is what we see here, I suppose.
Or is that because the output and load impedances are very different ?
(The load I used for headphones output noise measurement is 300 Ohm)

(More to come)

 

[Rja4000]

So here I added a plot with Level controled by the A70 Pro, using its unbalanced Headphones out (with a 300 ohm load)

Interesting result, ins't it ?
It looks like the SNR remains the same for a voltage divided by 2.
Like if all was shifted left by 6dB.

Even more interesting: since the A70 has a remote control, I may automate the level setting.
(Which I'm not able to do with the Passive Monitor1, of course)


@nagster

NB: I used the Topping D90LE since it was the quickest way to get a 4V signal
But I could have been using the RME ADI/4 pro SE as well.
With the A70 Pro, I measured the same SNR performance with it at 10mV or 50mV (+13dBu range +1.5dB, for 4.1V output).

Also: The RME 12Mic-D is more performant than the E1DA APU linear amp for this, since it offers a variable gain (which I may also automate through Midi, I suppose.).
And even for the same gain.
Of course, it's not exactly the same price. But since I have one...

 

[Wisehybred]

Any noise or issues in using the a70 pro in preamp mode. I am wanting to use one primarily as a preamp to my B&W speakers and Cary Audio monoblocks. I don’t want to spend a small fortune on a preamp with a bunch of inputs as I only use my streamer and dac to steam hi-res files.
I’ve read a bunch about digital volume versus analog, but I still think having a variable relay volume control would be best. Plus you get the benefit of a headphone amplifier and preamp in 1.
I wouldn’t mind spending the money on a better preamp but from all the research I’ve done is that the topping pre90, a70 pro and RME ADI 2, make for very neutral/transparent preamps. I need a remote and would really like a Home Theater bypass but I could make do without that feature.