Topping A90 Discrete Review (Headphone Amp & Preamp)

[escalibur]

Please note that the noise performance of the XLR output is always gonna be slightly worse than the 6.35mm output same for the A90.
The 6.35mm output performance is essentially the same as A90.

It's great to see you back!

Interesting release, but I was expecting some improvements at volume knob as it was pretty much the only and extremely picky downside of A90.

 

[AndreaT]

A state of the art headphone amplifier at a relatively affordable price. Thank you Amir!

 

[nagster]

This graph is automated only for the DAC plot.This graph is automated only for the DAC plot.
The other ones are 12 manually changed steps, as far as I understood.
Maybe @nagster can confirm that ?

I'm not confident in the explanation in English, but I'll try it anyway.

The D90SE is in DAC mode from start to finish.
Only the graph of D90SE alone is an automatic sweep of 80 points. The AP's digital generator output level is variable.

When measuring D90SE + MONITOR1, D90SE + Volume2 and D90SE + Pre90, the AP digital generator output level is always 0dBFS. I turned each volume knob to attenuate in the analog range.

From here, we will return to the Stone Age.
To change the measurement level, turn the volume knob while staring at the dBrA meter on the dashboard. The pot is difficult.
I want to record the values for the plot, but I can't follow the THD+N meter with my eyes, so I'll take a screenshot. Rewrite the file name to the model and dBrA value for later identification.

Open the screenshot image from the edge and make the dashboard meter value text.

Change them into a form that can be read by the AP, and import them into the graph of the D90SE alone that was first automatically swept.
There is still some work to be done, but the above is a rough procedure.
I was a primitive man.

It took more than an hour from the first automatic sweep to the completion of the posted graph image.
I do not recommend it to others. Choose a more efficient method.

edit
Now I noticed the button.
The CSV file is generated the first time, and the meter values are added to the same file from the second time onward...

 

[kopczas]

The measurement is very beautiful. But Topping needs to learn Singxer and Gustard, using higher quality components for their products to increase reliability and make it more premium.

Which components are you referring to? Looking at graphs components used for this unit are adequate quality.

 

[Pumpkin.0114]

Why not?

I asked to topping and they told me that gain can shift only by remote ...... it was sad

 

[Snoopy]

SA-2 in the second half of this year.

You keep saying that.. links?

 

[Ra1zel]

I was reading the user manual and definitely the Gain can't be adjusted with the front volume knob.

Eh, very sad.

 

[Billy Budapest]

Which components are you referring to? Looking at graphs components used for this unit are adequate quality.

I think they probably mean bigger, fancier cases.

 

[pseudoid]

In the recent Luxman SQ-N150 Review; an appropriate comment was made by

Marketing-driven engineering...

In the case of this Topping unit, the results of @amirm testing prove that both the marketing and engineering brain-trusts can work in lock-step!
I am also betting that the Topping A90 was specifically designed to also please the hardest critic of them all... even able to comfortably drive his favorite headphones.

 

[BenF]

@amirm When testing SMSL SP400 you wrote:

I did hit a usability issue though that is common with R2R volume controls: they are slow to respond. You have to crank and crank the volume control to go from loud to average level. Wish they would add some kind of acceleration to the volume control and jump a bunch of steps when you rotate it fast.

Is this amp behaving the same way?

 

[Gremlins]

Hello everyone

Is there any advantages to use this preamp A90 discrete instead of a Topping dx30 pro used as a preamp in front of a class D amp?

I start to think that a dac preamp offer lifeless music , i'm looking for a preamp that would bring life back

Cheers

 

[RosalieTheDog]

Do I understand correctly that this is more or less a Pre90 with headphone outs? What is the purpose of dedicated preamplifiers if headphone/preamps do the same job but are generally cheaper and have additional functionality?

 

[respice finem]

Do I understand correctly that this is more or less a Pre90 with headphone outs? What is the purpose of dedicated preamplifiers if headphone/preamps do the same job but are generally cheaper and have additional functionality?

Good question. I think, as with all hobbies, it's for the joy of playing with stuff, more than anything else. And everybody has their own mind and preferences, some will prefer matching different components, instead of buying a "Jack of all trades", even knowing that really hearing a difference is unlikely.
To each their own.

 

[ShiZo]

Sure. I plan to do a much more extensive study of volume control in different places.

Note that this volume control involves me manually turning the knob. it is NOT an automated sweep where input value changes as you are showing in above graph.

Did you happen to measure crosstalk?

 

[pseudoid]

Do I understand correctly that this is more or less a Pre90 with headphone outs?

What would make this Topping A90 (as w/most tested hw) more usable is to either have:
* Both rear XLR/RCA outputs simultaneously available, or at least
* One of the rear (XLR or RCA) outputs simultaneously available w/one of the front headphone outputs.

 

[Labjr]

Does this display which input is selected when the EXT is connected?

 

[Veri]

Did you happen to measure crosstalk?

WolfX at l7audiolab did, in any case. Not quite on Benchmark level but "very good".

 

[pseudoid]

Does this display which input is selected when the EXT is connected?

It seems like that is all you get and the Ext90 is of no help either!

 

[Labjr]

WolfX at l7audiolab did, in any case. Not quite on Benchmark level but "very good".

About 5dB worse than the PRE90. I wonder if crosstalk is measured between different inputs or left and right channels? I remember when stereo separation of a good phono cartridge was 25db.

 

[Rja4000]

I'm not confident in the explanation in English, but I'll try it anyway.

The D90SE is in DAC mode from start to finish.
Only the graph of D90SE alone is an automatic sweep of 80 points. The AP's digital generator output level is variable.

When measuring D90SE + MONITOR1, D90SE + Volume2 and D90SE + Pre90, the AP digital generator output level is always 0dBFS. I turned each volume knob to attenuate in the analog range.

From here, we will return to the Stone Age.
To change the measurement level, turn the volume knob while staring at the dBrA meter on the dashboard. The pot is difficult.
I want to record the values for the plot, but I can't follow the THD+N meter with my eyes, so I'll take a screenshot. Rewrite the file name to the model and dBrA value for later identification.
View attachment 213937

Open the screenshot image from the edge and make the dashboard meter value text.
View attachment 213938

Change them into a form that can be read by the AP, and import them into the graph of the D90SE alone that was first automatically swept.
There is still some work to be done, but the above is a rough procedure.
I was a primitive man.

It took more than an hour from the first automatic sweep to the completion of the posted graph image.
I do not recommend it to others. Choose a more efficient method.

edit
Now I noticed the button.
The CSV file is generated the first time, and the meter values are added to the same file from the second time onward...
View attachment 213939

I follow almost the same methodology.

I can't afford an AP, so I use Virtins Multi Instrument.
This has a .Net interface, so I can control it from a .Net App.
I wrote a "step test", where it stops and ask me for a value before any measurement. All is stored in a SQL.

So I'm able to do this as follows:
I set the volume to max, record "100" as a value and Mi records the SINAD.
Then to 75% and record "75". And so on.
My program measures level (ADC dBFS or Vrms with a DMM) as well.
I'm then able to plot multiple measurements.
Or to produce "reference" files to load in MI.