Matrix M-Stage HPA-3B mods

[maxxevv]

Here is the single-ended version of HPA-3B: https://www.audiosciencereview.com/...-dac-and-amp.5571/#lg=attachment18517&slide=0. See that noise difference is big.

That's a non-SMT version of the HPA-3U. Not the same circuitry as the 3B.

 

[trl]

Audeze has the same schematic as 3U. Deckard is an OEM headamp made by Matrix.

 

[trl]

That's a non-SMT version of the HPA-3U. Not the same circuitry as the 3B.

I bet HPA-3B = 2 x HPA-3U inside.

 

[maxxevv]

I bet HPA-3B = 2 x HPA-3U inside.

If you look at the Deckard versus HPa-3B specs, they are pretty different when it comes to the SE output specifications. The output impedance differences alone suggests something quite different internally between them.

Deckard:

SNR: -106dB A-weighted
Output Impedance : 3 Ohms
Power : 4W at 20 Ohms load

https://www.audeze.com/products/amplifiers/deckard

HPA-3B: (the one on the Matrix site is the 3U+, so no longer directly comparable).

Single ended:

SNR: -108dB A-weighted
Output Impedance : 0.4 Ohms
Power : 2.1W at 33 Ohms load

Balanced output:

SNR: -114dB A-weighted
Output Impedance : 0.6 Ohms
Power : 3.8W at 33 Ohms load

https://drop.com/buy/matrix-m-stage-hpa-3b-balanced-headphone-amplifier/details#details

 

[trl]

There are still 4 amplifiers inside DeckardHPA-3U, but they're paralleled/bridged for more power. You can compare PCB and internal components and their placement on the PCB on the Head-Fi - Deckard thread.

The HPA-3B has a bit different connections of the four internal amplifiers (channels) and of course an additional dual-opamp to make everything inside to be fully balanced from head to tail.

What I'm trying to say is that differences are small between these amps and all are designed and implemented by the same manufacturer.

 

[Veri]

Looks like new version is out, M-stage ⁺
https://www.shenzhenaudio.com/catalogsearch/result/?q=m-stage++

THD+N＜0.0003%
Full-band of THD + N less than 0.0003%, SNR up to -112dB, bandwidth not less than 150k at -3dB gain, 0.2Ω ultra low output impedance, three gain level options to fit for the different impedance and sensitivity headphone, all of these excellence performance improvements to be born the brand new HPA-3U⁺!

Also some measurement included but it's a little silly without any commentary of what they're measuring

 

[JohnYang1997]

Looks like new version is out, M-stage ⁺
https://www.shenzhenaudio.com/catalogsearch/result/?q=m-stage++

THD+N＜0.0003%
Full-band of THD + N less than 0.0003%, SNR up to -112dB, bandwidth not less than 150k at -3dB gain, 0.2Ω ultra low output impedance, three gain level options to fit for the different impedance and sensitivity headphone, all of these excellence performance improvements to be born the brand new HPA-3U⁺!

Also some measurement included but it's a little silly without any commentary of what they're measuring

I think it's 33ohm load. So, good but not impressive.
"THD+N＜0.0003% at 33ohms 90mW 1kHz A-Weighting"

 

[maxxevv]

That's approximately -110 dB ?

Pretty good, but not SOTA yet.

 

[trl]

Same THD+N has the older HPA-3U and HPA-3B in their specs, but at least my HPA-3B only got 0.003% instead of 0.0003%. Maybe is it me and my primitive measuring tools, don't know, but I really have doubts these devices really measures -110dB of THD+N (0.0003%).

However, HPA-3U+ was lunched in 2018 2017 if I remember well and it has a different DAC/transport card, but same powerful single-ended amplifier is inside. I think it's decent combo and worth trying, but it should not measure so well like Khadas/D50 + Atom or similar setups.

 

[maxxevv]

So far, Matrix have shown (with the Element X) that their measurements are not "plucked from the sky" so, I would think that their published specs would most likely be close to what Amir would get from his measurements.

 

[trl]

I do respect MATRIX as a company too, but I own HPA-3B and measured specs are below the published specs. However, audible transparent for sure, so nothing to worry about.

 

[NoNameNPC]

Crosstalk -0.1db is this bug of measurements?

 

[trl]

That's a really solid performance if the measurements are correct !

I wonder how does the unmodified version measure though ?

Unmodified version was measured by Amir, so here's the graph:

And here's the modified one below, measured with my brand new E1DA Cosmos ADC:

A SINAD of 113.56 dB @ 4.25V RMS, unity gain (lowest gain)
​

So, about 10 dB improvement in SINAD and in background noise on the modded version of HPA-3B and a even more than 10 dB decrease in mains hum. Quite a decent modded headamp and with an inaudible noise on all of my sensitive IEMs.