Continued from my previous post...
4) The input signal voltage (coming from a DAC) can have a dramatic impact on SQ with the MG3. When I use my DACmini CX, which has a 2V rms Line Out, the MG3 is at its best, in terms of dynamics, control, detail, and transparency (using 8-Ohm speakers with a 24V power supply (AC adapter or battery). But if I use my Sony PCM-M10 Linear PCM Recorder, with its 1V rms Line Out, the MG3 is still pretty impressive, but queing the same tracks, back to back, I can hear a difference. The MG3 becomes less assertive, more laid-back, and transparency begins to suffer - but I'd say it's still 90% as good as with a 2V rms input signal - and I do have to turn up the volume on the MG3 to achieve the same SPL at the speakers as had with a 2V rms input - so I'm getting less headroom - the very reason it doesn't sound as good.
My Sanza Clip+, which isn't truly a Line Out, but rather a headphone out, at maximum volume, these have been measured to produce only 55mV rms (about one half one twentieth the voltage of my Sony PCM-M10). When I use the Sanza Clip+ with its volume turned all the way up, the MG3 sounds lifeless - it loses a lot of its assertiveness and dynamics, and the transparency falls apart too - sounding quite veiled, as if vocalists are singing in their sleep, under a blanket.
In the owner's manual Steve Deckart (of Decware) wrote for his Decware ZSTAGE (triode tube gain stage), he writes that if you insert his ZSTAGE between a DAC and an amp, then use the gain control on the ZSTAGE to decrease the input gain to the amp (to something less than what's coming from your DAC) while simulataneously increasing the amp's volume control to maintain the same SPL at your speakers or headphones, you can reduce the dynamics and fullness of sound for a given DAC/amp combination. Conversely, he writes that you can use the ZSTAGE gain control to increase the the input gain to the amp (to something more than what's coming from your DAC (up to about a 5dB gain, depending on the type of tube you are using), while simultanously decreasing the amp's volume control, again to maintain a constant SPL at your speakers or headpones - for the purpose of increasing dynamics and fullness of sound for a given DAC/amp combination.
I know I've touched on this subject before, but in an e-mail discussion I had with Jan Meier (Meier-Audio), he said (paraphrasing here) that reducing or increasing the input level to the amp with the purpose of tailoring the dynamics or fullness only works with "some amps." Indeed, I've found that this doesn't work at all with the Emotiva a-100 Mini-x - it couldn't care less what input voltage you hit it with. Once you turn up the volume to compensate a weaker input signal, it pretty much sounds the same. Now that might not be the case if it were having to operate at the limits of its power capacity - with inneficient speakers, for example, that require every bit of the Emotiva's rated power, but it's certainly the case when using headphones.
Not so with the MG3 (into speakers). Even with 24V power, where it's rated at 32Watts into 8-Ohms, I'm finding that, at least with my 8-Ohm speakers, the MG3 is very sensitive to the input signal voltage. It's not nearly as happy with the 55mV rms signal of a Sanza Clip+ as it is with the 1V rms signal of my Sony PCM-M10 or the 2V rms signal of my DACmini CX.
So... If you want to use the MG3 with inefficient headphones (i.e. HE-6) or actual speakers, keep this in mind when considering DACs like the Reconessence Concero, which uses an ESS9023 chip with no gain in its downstream analog section - the Concero puts out only 1V RMS. This rules out lots of other ESS9023 DACs, too - like the Objective DAC, the Audioquest Dragonfly, and the diminuitive Stoner Acoustics UD100. A DAC like the (much more expensive) Wyred4Sound DAC2 would be fun to play with on an MG3, as it offers several choices of output level, but any DAC that puts out 2V rms would be fine, in my opinion, for speakers or inefficient headphones. The MG3 is not at all like the Emotiva a-100 mini-X or even the 1.28W into 50-Ohm Burson Soloist, in this regard. The MG3 wants to see 2V rms (when driving speakers). Please remember that all of my observations at this point are with 8-Ohm speakers that are not as efficient as heapdhones, so it could be that the MG3 has plenty of headroom (and thus less vulnerability to lesser input voltages) with headphones. I just haven't test that yet (waiting for the impedance match).
Update: Desipite the qualification spelled out in my last sentence, above, I've now emphasized that the observations I've made were with speakers, not with more efficient headphones.
Edited by zilch0md - 8/23/13 at 11:44am