RME ADI-2 FS Version 2 DAC and Headphone Amp Review

[JIW]

How does the load impedance get halved?

By using balanced connection each terminal of an individual driver is connected to identical amplifiers operating as a differential pair across the driver. Thus, each amplifier has half the load compared to being single ended where only one terminal is connected to an amplifier and the other terminal is connected to ground.

For example, the Pro in balanced mode has a maximum power of around 3 W at 64 Ohm. The current for that is around sqrt(3/64) = 0.2165 A = 216.5 mA which is the maximum current at 32 Ohm single ended. See section 34.18 of Pro manual v.3.8 and MC's corrected figure above.

 

[sonitus mirus]

Did you try the single ended output and lacked volume? Based on the manual, at 23 Ohm, I can only see a level difference on the 2/4 of around 20*log10(2*4.1/6.9) = 1.5 dB. Compared to the Pro FS and the DAC in single ended, the difference is around 20*log10(2*4.1/5.6) = 3.31 dB and around 20*log10(2*4.1/(2*3.2)) = 2.15 dB compared to the Pro FS balanced.

Thank you for explaining all that. I got the 2/4 around the same time as the Stealth and chose a 4.4 mm balanced cable for the Pentaconn connection to get as much power as possible. I was coming from an ADI 2 DAC FS, which probably would have been fine, but the 2/4 is my end game setup for headphones, so I wanted to give it every advantage. I have not tried single ended with an adapter because I am worried about damaging the headphones, so I cannot say whether single ended volume was an issue. Reading a few comments in the review thread made me cautious, and your insights here are very helpful. Greatly appreciate you for sharing your knowledge.

 

[JIW]

Thank you for explaining all that. I got the 2/4 around the same time as the Stealth and chose a 4.4 mm balanced cable for the Pentaconn connection to get as much power as possible. I was coming from an ADI 2 DAC FS, which probably would have been fine, but the 2/4 is my end game setup for headphones, so I wanted to give it every advantage. I have not tried single ended with an adapter because I am worried about damaging the headphones, so I cannot say whether single ended volume was an issue. Reading a few comments in the review thread made me cautious, and your insights here are very helpful. Greatly appreciate you for sharing your knowledge.

There is no problem using a balanced cable with a single ended amp as long as the negative terminals are connected to ground in the adapter. The other way around using a symmetrical amp with a cable with combined ground for both drivers is asking for trouble as the amps for the inverted signals are directly connected through the cable.

What volume setting is loud enough for you even with the most demanding recordings?

 

[sonitus mirus]

There is no problem using a balanced cable with a single ended amp as long as the negative terminals are connected to ground in the adapter. The other way around using a symmetrical amp with a cable with combined ground for both drivers is asking for trouble as the amps for the inverted signals are directly connected through the cable.

What volume setting is loud enough for you even with the most demanding recordings?

I casually listen to music at -44 dB to -40 dB in high gain, mostly with music from the 60s to today, and quite a bit of techno/drum n bass. This volume level is well within the capability of just about any amp. I can push it higher with classical and 192 kHz tracks, since I usually stream Apple Music at their highest resolution. For typical 44.1 kHz rock, I rarely go beyond -18 dB for more than five to ten seconds, though I have jammed at around -24 dB in high gain with a balanced connection. I’m nearly 60, so my hearing isn’t the best.

 

[JIW]

I casually listen to music at -44 dB to -40 dB in high gain, mostly with music from the 60s to today, and quite a bit of techno/drum n bass. This volume level is well within the capability of just about any amp. I can push it higher with classical and 192 kHz tracks, since I usually stream Apple Music at their highest resolution. For typical 44.1 kHz rock, I rarely go beyond -18 dB for more than five to ten seconds, though I have jammed at around -24 dB in high gain with a balanced connection. I’m nearly 60, so my hearing isn’t the best.

Volume -18 dB in high gain is +7 dBu at 0 dBFS. Amir measured the DCA Stealth's sensitivity as 94 dB SPL at 0.549 V RMS which is equivalent to 94 - 20*log10(.549) + 10*log10(0.6) = 97 dB SPL at 0 dBu. At -18 dB in high gain, max peak SPL is thus 97+7+3 = 107 dB SPL. At maximum voltage for 23 Ohm, 8.2 V RMS, maximum peak SPL is 94 - 20*log10(.549) + 20*log10(8.2) + 10*log10(2) = 120.49 dB SPL.

You can use Auto Ref Level, which should put you into IEM for most of your listening.

 

[sonitus mirus]

Volume -18 dB in high gain is +7 dBu at 0 dBFS. Amir measured the DCA Stealth's sensitivity as 94 dB SPL at 0.549 V RMS which is equivalent to 94 - 20*log10(.549) + 10*log10(0.6) = 97 dB SPL at 0 dBu. At -18 dB in high gain, max peak SPL is thus 97+7+3 = 107 dB SPL. At maximum voltage for 23 Ohm, 8.2 V RMS, maximum peak SPL is 94 - 20*log10(.549) + 20*log10(8.2) + 10*log10(2) = 120.49 dB SPL.

You can use Auto Ref Level, which should put you into IEM for most of your listening.

Yes! Thanks a bunch. Using Auto Ref Level now. This is what I use with my speakers.

 

[sonitus mirus]

I set my auto loudness on the RME 2/4 Pro balanced headphone (Phones Out 3/4) to be an equivalent match to what I enjoy with my speakers (Line Output 1/2) and I immediately found that I no longer had the urge to crank up the volume with certain music. For some reason I confused Auto Ref Level with the Loudness feature and I was missing out on my favorite feature with the RME DACs.

 

[unpluggged]

By using balanced connection each terminal of an individual driver is connected to identical amplifiers operating as a differential pair across the driver. Thus, each amplifier has half the load compared to being single ended where only one terminal is connected to an amplifier and the other terminal is connected to ground.

For example, the Pro in balanced mode has a maximum power of around 3 W at 64 Ohm. The current for that is around sqrt(3/64) = 0.2165 A = 216.5 mA which is the maximum current at 32 Ohm single ended. See section 34.18 of Pro manual v.3.8 and MC's corrected figure above.
View attachment 423737

Yes, I figured this out shortly after posting, but it's good to have it clarified here for everyone else

 

[rcstevensonaz]

By using balanced connection each terminal of an individual driver is connected to identical amplifiers operating as a differential pair across the driver. Thus, each amplifier has half the load compared to being single ended where only one terminal is connected to an amplifier and the other terminal is connected to ground.

For example, the Pro in balanced mode has a maximum power of around 3 W at 64 Ohm. The current for that is around sqrt(3/64) = 0.2165 A = 216.5 mA which is the maximum current at 32 Ohm single ended. See section 34.18 of Pro manual v.3.8 and MC's corrected figure above.

Update: I have a section 34.3 34.6 "Balanced Phones Mode" in my manual that shows the block diagrams for inverting the signal on both the left and right channels in balanced headphone mode, so I assume this was the section @JIW was referring to.

Can you explain what "differential pair" means in this context? The word "differential" does not even appear anywhere in my ADI-2 Pro FS manual.

Could you kindly share the name of section 34.18 from the v.3.8 manual? In my version (v2.5), that section is "ADI-2 Pro as Hardware I/O for Measurements" and does not seem to say anything about balanced vs. unbalanced headphone power.

 

[JIW]

Can you explain what "differential pair" means in this context? The word "differential" does not even appear anywhere in my ADI-2 Pro FS manual.

Could you kindly share the name of section 34.18 from the v.3.8 manual? In my version (v2.5), that section is "ADI-2 Pro as Hardware I/O for Measurements" and does not seem to say anything about balanced vs. unbalanced headphone power.

One amplifier is amplifying the inverse of the signal amplified by the other. Symmetrical is another term that could be used.

It is called "Extreme Power Charts". You can download the latest version of the manual from RME's product site under the downloads tab.

 

[rcstevensonaz]

One amplifier is amplifying the inverse of the signal amplified by the other. Symmetrical is another term that could be used.

It is called "Extreme Power Charts". You can download the latest version of the manual from RME's product site under the downloads tab.

Thanks! What I was needing to see was the block diagram for the balanced headphone mode. And yea, the term used by the RME manual in that section is "inversion".

 

[JIW]

Thanks! What I was needing to see was the block diagram for the balanced headphone mode. And yea, the term used by the RME manual in that section is "inversion".

That section is curious. Most DAC chips these days - certainly the AK4493 in the Pro - have symmetrical output and so inversion is not required even with one chip. So for a combined DAC/amp, balanced headphone out does not require analog inversion. Of course that doubles the number of analog low-pass filters and gain amps so might be more expensive and hence not preferred.

 

[MC_RME]

That section is curious. Most DAC chips these days - certainly the AK4493 in the Pro - have symmetrical output and so inversion is not required even with one chip. So for a combined DAC/amp, balanced headphone out does not require analog inversion. Of course that doubles the number of analog low-pass filters and gain amps so might be more expensive and hence not preferred.

See chapter 34.6 in the ADI-2/4 Pro SE manual. The 2 Pro's way to do it is the best way to let unbalanced outputs also offer a balanced mode. The 2/4 had more emphasis on that mode during the design phase, and also keeps outputs 1/2 working independently.

 

[KSTR]

That section is curious. Most DAC chips these days - certainly the AK4493 in the Pro - have symmetrical output and so inversion is not required even with one chip.

Please not that the symmetrical output of basically all DAC chips is there for a reason, to be able to subtract out the common-mode noise/distortion.... and you have to do that early in the signal chain.
That is, you must subtract the chip's output signals and re-reference the difference to the local ground. If you need/want a balanced output from that stage, you need an additional inversion.
This (filtering, subtraction and inversion) can be done in one simple clever circuit and that's what RME is doing.

 

[sonitus mirus]

There is no problem using a balanced cable with a single ended amp as long as the negative terminals are connected to ground in the adapter. The other way around using a symmetrical amp with a cable with combined ground for both drivers is asking for trouble as the amps for the inverted signals are directly connected through the cable.

What volume setting is loud enough for you even with the most demanding recordings?

While cleaning out my office closet, I noticed that I had this adapter stuck in my old ADI-2 DAC FS 6.3mm TS jack. I will try and set up the Auto Ref and loudness profiles to be similar to my 2/4 Pro SE and see how my DCA Stealth headphones perform with regards to volume levels. I'm sure it will provide plenty of volume for me after looking at the numbers.

https://www.amazon.com/gp/product/B0BG5TZZ42

 

[ObjectiveSubjectivist]

I know I'm asking for a quirk, but maybe
@MC_RME can we please get an update so the spectrum analyzer in ADI2 would look like the one in old Winamp?
I mean the colors and those bars falling from the top.

I'm missing the old days

I assume it's technically possible but it cost time for basically a gimmick.

 

[Oldson]

in the opinion of RME the 5 bands (+2)of peq is enough in the majority of cases for headphone correction!
would you guys agree?

 

[Matias]

If it needs more than 5 bands, change your headphone lol.

 

[Oldson]

If it needs more than 5 bands, change your headphone lol.

tbh, the main headphone is hd800 and i know that will be fine.
just dont know what i may get in future.
was just curious on opinions as i am likely to order the dac later tonight.

 

[Rednaxela]

in the opinion of RME the 5 bands (+2)of peq is enough in the majority of cases for headphone correction!
would you guys agree?

Don't forget that the RME PEQ is not only limited in number of bands but not all bands are equally flexible either.

I haven't tried this myself, but maybe you can get a feel for the RME's PEQ (in)capabilities through https://autoeq.app/ and its Custom Parametric EQ.