SMSL VMV A2 Review (Stereo Amplifier & DAC)

[Francis Vaughan]

What is there is not capable of the feature set of VMV A2:

"Conversion between 44.1, 48, 88.2, 96, 176.4 and 192KHz input and output sample rates"

VMV A2 goes to 768 kHz sampling, and DSD. CPU rate scales with sample rate so you need massively more optimization or computation power than the stock library.

The API Readme isn't all that up to date. XMOS have a DSD implementation, and it is possible to set the limits of the resampler by changing a couple of constants in the code. This code base works across a range of their chips, and the x200 series have enough grunt to do what is needed out of the box. This amp uses the 16 core version so it easily meets the processing requirements. There are a few application notes that describe how to modify the code to do what is needed.
The point is that XMOS have all the needed code available, and any customer can use it without having to write stuff from scratch. I doubt any of the audio manufactures write any of their own audio processing code.
Kardas took years to work out how to remove XMOS from the USB ident and put their own name in. They clearly just recompiled the firmware from base source.

 

[Mulder]

What is really interesting is that they sell the amplifier as a module. No price or datasheet, but looks interesting. But difficult for them to compete with Hypex's dedicated Malaysian operation. You are on your own for a power supply.

There is a price actually. The price is 599 Euros. The Hypex NC400 is 275 Euros, but the later is excl. tax. And you need two Hypex modules. So the True-Blue Box COBALT modules seems to be less expensive I think.

It would be very interesting to see the Blue-Box measured. As it has a phono amp built in, maybe they have put some extra effort into the design of the analogue inputs. For those who really need a phono input, I think it is possible that The Blue Box possible hold it´s own against SMSL VMV A2 what price matters.

 

[Stokdoof]

Interesting. Very much a boutique operation. Amplifiers are custom made to order. Pricey compared to the SMSL, a bit hard to compare as their prices almost certainly include VAT, but even then higher. All amps include phono. They look a bit hobbybox, but the CNC cut wood panels are cute.

What is really interesting is that they sell the amplifier as a module. No price or datasheet, but looks interesting. But difficult for them to compete with Hypex's dedicated Malaysian operation. You are on your own for a power supply.

The price is below the info click the + sign

Yet another alternative with Axign inside:

https://www.harmankardon.nl/citation-series/CITATION+AMP-.html

 

[Francis Vaughan]

The price is below the info click the + sign

Yet another alternative with Axign inside:

https://www.harmankardon.nl/citation-series/CITATION+AMP-.html

Ah, I never really though the do that. I just assumed the site was incomplete.

Yeah, the Harman Citation is the really interesting one. Mainstream brand acceptance will make a big difference to Axgin. The competition becomes mainstream, and we benefit.

 

[jokan]

NOTE: this is a review of a new product. My recommendation is strictly based on its performance in the limited time I had with the unit. You are responsible to research issues of reliability, functionality, support, long term stability, etc. on your own. Please don't be an early adopter if are risk averse in these regards.

I particularly agree with this statement!
I'm looking forward to more independent owners reviews on this amplifier.
I too am a bit bummed by the fan, but it is otherwise a sealed box so I can understand the need for ventilation.
I'm going to follow this thread now.

 

[jokan]

The performance differences between the SA400 and the VMW A2 remains a puzzle. All indications are that the power amp is identical. In both cases it seems the analog inputs have been messed up. Given that is all the SA400 has, it seems quite a shame, and makes the SA400 a product without a reason. There could and should be a middle ground product, similar price to the SA400, digital inputs only.

In total agreement with you on this statement here. I don't get it at all.

 

[SylphAudio]

Additional testing is always a good thing.

If there is a watchdog system in the chip, yes. But I would not take this as a given, some additional testing as suggested by @restorer-john might be interesting.
It would anyhow be a sign of safety, which is good, but not of stability.
Having ADCs with both high speed and high accuracy is not an easy task. For the moment analog loops still have the lead for high quality class D amplification.

 

[SMSL-Mandy]

Hi Mandy, can you comment on the high SINAD observed by Amir when using Analog In?

The analog input uses ADC, according to the TI official circuit, there is a 20Hz high pass filter. That is normal.

and A2 only uses ST in the Post-level and uses STA311 only as an EQ processor, and the DSP used for feedback is not ST.

 

[Willem]

What alternatives would you suggest with this kind of performance and a built in DAC?

Yamaha AS701/801 At least over here these Yamahas are even cheaper (and were a lot cheaper until recent price increases).

 

[Toku]

The second puzzle was the frequency response. Here I am driving the amplifier using Coax at 192 kHz sampling so we should have flat response to half of that but we don't:

View attachment 138311

Indeed, you can set the sample rate to anything you want, change to USB, etc. and you always get what seems to be 44.1 kHz sampling! SMSL advertises the "high-res" capability heavily in this amp but while the DAC is likely capable of delivering that, the amplifier happily shuts any extra spectrum above audible band.

The graph of the frequency response of amirm shows the characteristics like cutting 20kHz or more with a knife.
I don't think that such characteristics will be created unless the sampling rate is 44.1kHz, but where do you create such characteristics?
The signal processing configuration and the IC chip used are unknown, and I have no idea how this strange characteristic is created. I would like to know more about VMV A2, but do you have any information?

 

[PeterOo]

I tried to get the datasheet for AX5689 even using a non personal, big engineering email address and didn't even get a reply which is a bit of a turn off.

They have a vacancy for Product marketing manager. That might explain the lack of response.
You did find the product brief?
https://www.axign.nl/wp-content/uploads/AX5689_ProductBrief_v1.0.pdf

 

[SMSL-Mandy]

They have a vacancy for Product marketing manager. That might explain the lack of response.
You did find the product brief?
https://www.axign.nl/wp-content/uploads/AX5689_ProductBrief_v1.0.pdf

The vacancy for the product marketing manager. The found is so fun.

 

[Francis Vaughan]

The signal processing configuration and the IC chip used are unknown, and I have no idea how this strange characteristic is created. I would like to know more about VMV A2, but do you have any information?

We are assuming the Axgin AX5689 is the core of this.There isn't a lot of information to be had. The product info sheet is:
https://www.axign.nl/wp-content/uploads/AX5689_ProductBrief_v1.0.pdf
The article from Audioxpress is:
https://audioxpress.com/files/attachment/2659
and the patent is:
https://patents.justia.com/patent/10659028

None of these tell you much (if anything) about the control loop parameters, and it is here that the response is created. So you can see a general overview of operation. The more terms in the loop the faster the roll-off can be.

 

[Tks]

Thats an amazing brick wall if ive ever seen one. I just dont understand how they did it. Is this done right after the DAC chip or is the amp somehow doing this cut in thr analogue stage? Or can the chip be programmed with a custom filter or two and SMSL just made a filter as good as the insanely costing M-Scaler?

I know its pretty trivial doing this sort of cut in software through the use of EQ (heck even Windows' audio stack itself has a proper brickwall). Its just a curious thing wondering what SMSL did here, since most brickwalls on AKM and ESS arent this sharp and strong.

 

[Francis Vaughan]

is the amp somehow doing this cut in the analogue stage

It would seem that the amp's PWM modulator is doing it in the digital loop filter part of the modulator. So all in the digital domain, although obviously the output filter removes other high frequency artefacts.

 

[JohnYang1997]

Thats an amazing brick wall if ive ever seen one. I just dont understand how they did it. Is this done right after the DAC chip or is the amp somehow doing this cut in thr analogue stage? Or can the chip be programmed with a custom filter or two and SMSL just made a filter as good as the insanely costing M-Scaler?

I know its pretty trivial doing this sort of cut in software through the use of EQ (heck even Windows' audio stack itself has a proper brickwall). Its just a curious thing wondering what SMSL did here, since most brickwalls on AKM and ESS arent this sharp and strong.

They are completely different tests.
One is sine sweep one is white noise.
One is +-5dB scaled one is 0dB to -120dB scaled.
It also may very well be that the output is simply turned off after certain frequency.
This is certainly not something to care too much about imo.

 

[SylphAudio]

I tried to get the datasheet for AX5689 even using a non personal, big engineering email address and didn't even get a reply which is a bit of a turn off.

I also emailed them few days ago, they opened the mail on the same day but I didn't get any reply.

 

[21qz]

Who is the a1 for, vs a2 in a simple write up?

 

[SMSL-Mandy]

Who is the a1 for, vs a2 in a simple write up?

VMV A1 is a class-A amp, it is different modle.
Thanks

 

[21qz]

Right, but who is a class a amp for? Oppose to class d. What type of audiophile