• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. 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!

HIRESFI AMPER502 Full Digital Amplifier Test

The concept of "power DAC" has always been appealing in the way it seems to bypass a number of audio stages. But so far, I have not seen any implementation that performs very well. Some things are not intuitive and this is one of them!
There are some good ones. The first I was aware of was the Sharp SM-SX100 , but that was a while ago and things have moved on.

I currently use a Sony TA-DA9100ES, which is a multi channel power DAC. I used to debate with myself whether it was a DAC with no amplifier, or an amplifier with no DAC. I think it was the latter.

Anyway it sounds very good, and saw off a lot of good competition. I'd love you to measure it, but its very heavy and its a long way. I recently got an E1DA ADC, so will try to satisfy my curiosity myself.

Nick

 
I currently use a Sony TA-DA9100ES, which is a multi channel power DAC. I used to debate with myself whether it was a DAC with no amplifier, or an amplifier with no DAC. I think it was the latter.
It's neither, it's just an AVR. Nor is the Sharp any good, at least not by today's standards:


We do have some modern examples though, particularly the ones employing Axign technology. One of them was tested here:


These newer generation devices have a big advantage over the old Sharp: they properly employ feedback (post-filer).
 
Last edited:
It's neither, it's just an AVR. Nor is the Sharp any good, at least not by today's standards:
It is not just an AVR.

With a digital input, there's no active analogue processing or amplification at all, just the passive output filter.

It is 17 years old though, and I'm looking forwards to comparing it with contemporary equipment.
 
With a digital input, there's no active analogue processing or amplification at all, just the passive output filter.
Ah, I see now. There is also no post-filter feedback, though. Then it's an AVR with a power DAC ;)
 
Only way it could be more broken would be if someone threw it out the window of a 25 story building.
 
From this, one might imagine what Topping and SMSL routinely manage to do is actually difficult.
I'll give Topping and SMSL credit for using measurements to guide their design, but with ample feedback, it's not that difficult to achieve 100 dB or more of SINAD.

FDA is an idea that looks good on paper, but has a major flaw in practice: there is no analog reference to create a control loop. Whether it's negative feedback (what most well designed amps do) or feed-forward (what THX and Benchmark amps also do), you have a control loop that keeps the output signal close to the input signal.

There are ways around that with FDA, by using an ADC on the output (for feedback) and/or by using a DAC after the input (for feed-forward), but this adds cost and complexity that kills the point of FDA in the first place. Another approach is to take a look at the feedback that happens in the digital domain, within the noise shaper, and to use some complex DSP to try to tame the output. This is arguably the least complex way, but requires intimate understanding of the mathematics behind PWM, which isn't all that common unfortunately.
 
I really appreciate Amir's work, but I don't quite understand the point of publishing a review when you don't know if the tested device is broken or not.
 
It's neither, it's just an AVR. Nor is the Sharp any good, at least not by today's standards:


We do have some modern examples though, particularly the ones employing Axign technology. One of them was tested here:


These newer generation devices have a big advantage over the old Sharp: they properly employ feedback (post-filer).

I thought of the SMSL VMV A2 and the Sabaj A30a immediately when I read Amir's comment.

I know the Sabaj A30a hasn't been reviewed here - but I thought Amir recommended the VMV A2 - ?
 
The audiophile concept of avoiding stages is just idiocy. Ten correctly designed stages will produce clean sound, while one incorrectly designed stage will produce garbage, and it takes a few stages to actually produce audio correctly. It's like skipping the tires and suspension in a car to get a more direct feel. Well, you'll certainly feel every bump in the road.
 
It's supposed to be 2x50W so something's wrong and hence I'm not sure what the measurements really mean.
A lot of my measurements are not at max power. So that information is useful. In addition, if there is a design issue, then folks need to know about that.
 
The concept of "power DAC" has always been appealing in the way it seems to bypass a number of audio stages. But so far, I have not seen any implementation that performs very well. Some things are not intuitive and this is one of them!
I used to be dead keen on the idea too, but I seem to recall an EE mentioning that the digital electronics can be done more cleanly at line level voltages than at high power voltages, so, on the basis of using each technology for what it is best at, it will be optimal to do A/D/A process at line level and amplification to power levels in the analog domain.

I used to have a signature on audio discussion forums that went something like, "I am agnostic to technology and agnostic to complexity, but not agnostic to performance."
 
There are ways around that with FDA, by using an ADC on the output (for feedback) and/or by using a DAC after the input (for feed-forward), but this adds cost and complexity that kills the point of FDA in the first place.
If you can't make a decently performing amp, the question is if there is a point to this at all ;)
Another approach is to take a look at the feedback that happens in the digital domain, within the noise shaper, and to use some complex DSP to try to tame the output. This is arguably the least complex way, but requires intimate understanding of the mathematics behind PWM, which isn't all that common unfortunately.
That is what the aging Sharp does that @welwynnick mentioned:
1726815390156.png

But it doesn't properly solve the speaker impedance dependency, among other things. Relative to using an ADC, this is actually an elegantly simple, but ultimately incomplete, solution ;)
 
It's neither, it's just an AVR. Nor is the Sharp any good, at least not by today's standards:...

We do have some modern examples though, particularly the ones employing Axign technology.

These newer generation devices have a big advantage over the old Sharp: they properly employ feedback (post-filer).
Does the new Marantz M1 qualify?
 
I used to be dead keen on the idea too, but I seem to recall an EE mentioning that the digital electronics can be done more cleanly at line level voltages than at high power voltages, so, on the basis of using each technology for what it is best at, it will be optimal to do A/D/A process at line level and amplification to power levels in the analog domain.
Must have been Bruno Putzeys, who, in this very interesting lecture, concluded "Keep in mind that no converter designer in their right mind would ever consider building a DAC using power FETs" (from 35'54''):

 
Back
Top Bottom