Matrix Audio HPA-3B Balanced Headphone Amp Review

[maxxevv]

That's because 3B is a Class A amplifier, class A amplifier will keep a certain temperature when it works for a while. This is determined by working characteristics of class A amplifiers.

I do understand that. However, as with all electronic devices that generate heat, better heat dissipation measures will definitely help in lowering its overall temperature at equilibrium. Like how different heatsink designs can alter the stable temperatures of computer CPU chips by as much as 15 degrees.

 

[FireLion]

@amirm I bought another one (my second) that had sparkos voltage regulators already installed. I think that was to allow certain oamps.

I did fit two opa1656's which were recommended to me by a TI engineer last year. It took a while but the sound grew on me. I have the 789 running alongside it and they are very close but the HPA-3B has a little body to it and sounds fuller. It's very marginal but might be something to do with class A vs A/B.

Admittedly I lean more to the Matrix over the 789 and it pairs excellently with my mdr-z7m2 which are also upgraded with lambskin. DBS tech talk reviewed my modded set.

It's also worth noting that while the 789 has 6 watts vs 3.8 the HPA-3B @300 and 600 ohms has about double the power of the 789 which is why I think it's a great amp to have.

I wish they would redo this amp with the 789 form factor or just re-release it as it probably didn't get the recognition it deserved.

To anyone who reads this If you ever get the chance to pick one up I highly recommend it.

@MatrixAudio I had one of these and sold it, managed to get another and love it. Amazing amplifier especially with opa1656 opamps.

 

[trl]

I did fit two opa1656's which were recommended to me by a TI engineer last year.

Worth measuring the output DC with a true-RMS multimeter. Of corse, comparing the value with OPA1652 or other opamps might worth the challenge.

 

[nhatlam96]

Where can I find the SINAD chart for headphone amps?

 

[trl]

The 50mV one is probably this one:: https://www.audiosciencereview.com/...ts/most-quiet-headphone-amp-review-png.55847/, but it's continuously updating, so whatch latest reviews posted by Amir.

 

[FireLion]

God I love this amp! Would love to see Matrix bring something like this out again.

 

[Human Bass]

God I love this amp! Would love to see Matrix bring something like this out again.

They just need to upgrade the psu.

 

[trl]

God I love this amp! Would love to see Matrix bring something like this out again.

They won't, I asked them the same thing few years ago.
Design is great, otherwise the Audeze guys wouldn't ask Matrix to build the rebranded Deckard; just few minor modifications and a new Matrix HPA can be done in no time, but they don't want to continue this trend.

They just need to upgrade the psu.

Ya, well...at least to build it in a bigger case or to shield the transformer, because the most of the mains hum is caused by the transformer being too close to the audio PCB traces.

 

[FireLion]

They won't, I asked them the same thing few years ago.
Design is great, otherwise the Audeze guys wouldn't ask Matrix to build the rebranded Deckard; just few minor modifications and a new Matrix HPA can be done in no time, but they don't want to continue this trend.


Ya, well...at least to build it in a bigger case or to shield the transformer, because the most of the mains hum is caused by the transformer being too close to the audio PCB traces.

I don't really understand why not, it sounds great and more and more people are spending in the $500 price bracket. All the more reason to never sell this second one I acquired.

It also works brilliant with iems even sensitive ones. It has good measurements and for me much better sounding than the likes of a 789.

Some people might not like the long shape or the class A heat.

 

[maxxevv]

They won't, I asked them the same thing few years ago.
Design is great, otherwise the Audeze guys wouldn't ask Matrix to build the rebranded Deckard; just few minor modifications and a new Matrix HPA can be done in no time, but they don't want to continue this trend.


Ya, well...at least to build it in a bigger case or to shield the transformer, because the most of the mains hum is caused by the transformer being too close to the audio PCB traces.

I just did a little experimenting with the shielding of the toroid transformer using 0.5mm Mu-metal sheet. Which I then wrapped in double layer of Kapton tape to make sure that there were no accidental short circuits. Don't have a means to measure the resultant system level noise for any difference but I do have Gauss meter that measures for magnetic leakages. Before and after shielding difference in case measurements of magnetic flux was about 25%. So not sure if it will make a difference in the overall audible quality if at all.

I honestly couldn't pick up any significant differences with what I used, a HD-6XX and a HE4XX planar. Maybe I need something really sensitive like a very efficient IEM ?

I saw that you did a partial shield with audible results ?

 

[trl]

Difference is at least 10 dB, just do a loop test with a decent ADC/soundcard or do a FFT. You will notice how 50/60 Hz and it's 2nd and 3rd harmonics will severely decrease in amplitude after installing the shield.

The headphones you used make no sense for such a test; get some 16 Ohms IEMs instead. You need sensitive headphones able to hear the mains hum noise, not planars.

 

[FireLion]

Difference is at least 10 dB, just do a loop test with a decent ADC/soundcard or do a FFT. You will notice how 50/60 Hz and it's 2nd and 3rd harmonics will severely decrease in amplitude after installing the shield.

The headphones you used make no sense for such a test; get some 16 Ohms IEMs instead. You need sensitive headphones able to hear the mains hum noise, not planars.

I sometimes get noise using burson opamps, what materials do you use to make the shield?

Try a set of opa1656's, same ones in the A30 pro. I have had then in my HPA-3B for a long time.

 

[trl]

I sometimes get noise using burson opamps, what materials do you use to make the shield?

What kind of noise? Why only "sometimes" and not all the time?

What exact opamps from Burson? V6 Classic were the best I've tested. V5-i were incompatible, so not recommended.

The shield was made out of G.O.S.S., double sheet, but this was only to lower the mains hum, not for any other improvements. More in-depth about GOSS here: https://www.nipponsteel.com/en/tech/report/nsc/pdf/8110.pdf.

 

[Veri]

What exact opamps from Burson? V6 Classic were the best I've tested.

Really? I thought the Vivid were the more accurate ones.

 

[trl]

Really? I thought the Vivid were the more accurate ones.

I'm a Classic lover, what do you know...

 

[FireLion]

Buy two, thank me later!

USA
https://www.ebay.com/itm/333850298075?hash=item4dbb0746db:g:yv0AAOSw7tBgTNIs

Hungary
https://www.ebay.com/itm/233994108948?hash=item367b22c014:g:tJgAAOSwLIZeD4j8

 

[Veri]

Buy two, thank me later!

Hungary
https://www.ebay.com/itm/233994108948?hash=item367b22c014:g:tJgAAOSwLIZeD4j8

The Hungary one looks higher quality

 

[FireLion]

The Hungary one looks higher quality

One on eBay Germany now, if you want class A amp that should push a Susvara to the moon. Strongly recommend the opa1656, there is a reason topping are using it.

 

[JohnYang1997]

One on eBay Germany now, if you want class A amp that should push a Susvara to the moon. Strongly recommend the opa1656, there is a reason topping are using it.

Opa1656 is used because:
It is relatively reasonably priced.
It has enough bandwidth for NFCA to work.
It has rail to rail output.
It has relatively high output current, still needed 10 opamp parallel to get what we want.
It behaves well near and at clipping.
It doesn't break itself at current limit.

In the application of 3B, all those merits are basically invalid or not needed. You might like it for all the different reasons.

 

[trl]

I'm using OPA1652 in my HPA-3B due to:
- lower output-DC
- less-to-none "thump" sound when switching between gain stages and power On/Off

Also tested with OPA1688 and OPA1656, still OPA1652 will remain my favorite in this application.