"1) Let's first clear the air about "0-ohm." We all know there is no such thing as there is going to be some resistance in the circuit. So when I say "0-ohm" or "bypass the 22-ohm resistors," please make sure you know I really mean "less than 1 ohm."
I have tested the RWAK100 to be totally stable after bypassing the 22-ohm resistors. After bypassing of the 22-ohm resistors, there isn't a safety issue even if the output was shorted. How do I know this? I tested it! The output stage does not get damaged. Sure - this is not recommended and the output FETs get hot, but they don't fail.
How many headphone amp designers do you know who are putting 22-ohm resistors on the output of their amps for "safety?" "
I apologize for making improper short abbreviation term like 0 ohm as to shorten < 1 ohm output impedance. If you read my earlier posts, I noted Dr. Jan Meier's products for using < 1 ohm output impedance as well. For FET you said tested and it's OK. Sure it was hotter but not fail like you said, did you test this at highest current drainage level for long run like 200hrs continuous? Even if you put some power hungry phones and run at 100% volume for days and it didn't fail. It doesn't mean there's no damage to FET at all. If FET were to be worn out for x days, I'm sure doing this won't exactly last for the same time if you keep using it at its extreme condition having high current surge frequently with high dynamic swing track.
"2) Do you really believe this is true? I do not see any evidence of this being true when tracing out the circuit. The 22-ohm output resistors are unnecessary and bypassing them only IMPROVES the presentation. With the resistors in the circuit, the dynamics and PRaT you speak of are inferior (especially with lower impedance headphones or IEMs). You can easily hear this if you put in a DPST switch that simply shorts the resistors (bypassing them), or leaves them alone. Do this and listen as you switch on the fly - it is very obvious that bypassing the 22-ohms is an improvement all around - nothing about the music suffers. The sound opens up with them bypassed and there is more life to the sound.
Have you tried this and listened to compare? "
I listened to and compared. I didn't say this is a bad mod not even once. How can you be sure that they only add 22 ohm resistor to reduce load of impedance? by-passing it has pros and its cons. Overall sound performance surely is improved having lower output impedance but its stability doesn't improve from what I heard myself. I found tons of good ears people including me complaining they sound at the same level, no depth, poor imaging and you can't feel much of the slam and impact. It's more revealing and linear with cleaner sound so fixing its cons would make this perfect.
"3) But is already is safe, and bypassing is the cleanest way. Adding capacitors in series to the output (as your picture shows) when they are not needed will only add their coloration to the sound (for better or worse). Anyone can play with adding caps in series with the output, but it doesn't make it a "complete mod." The mod is already complete as is.
And you say:
I setup tons of decent sources from Esoteric/Emm Labs/others with proper PRaT fine-tuning so I daresay that mod screwed this up and I'm trying to fix it here.
and then you say:
Well, I'm not experts in this field and can't read much from seeing bare board so I'm not daring enough to say what is actually completed.
Unless I am not understanding you, these are somewhat conflicting statements. You are daring so say the mod screws up the sound and you are trying to fix it, but you are not daring enough to say what you are doing instead and how it is better?"
Adding output coupling capacitor also has its pros and cons just like resistor by-passing mod. You have your own design and philosophy believing this is the best way why others like Jan Meier here may have different approach from you. caps has its character for colorization and practically make it less rock solid for fastest speed. But your current mod is way too fast for device to handle causing imaging and pace caused by too fast PRaT speed and balance. Add c output can help balancing signal current flow speed and prevent surge from high current drain making FET works less extreme and prolong its lifetime. C output coupling also have other pros for reducing background noise/THD effects/and vibration influence.
For last part I'm saying to admit that I don't know enough of engineering to design and make my own like that so I won't say what will be better between Vinnie and Meier's approach. But I also have listening experiences in this field first hand. I heard many world-class super highend equipments and they're generous enough to show me what's inside and how things work internally. Why < 1 ohm impedance is good and what's difficult to make a very good ones.
Let me say this clear again that it's not a bad mod but there're parts working better and some parts having problems. Even original design has some good reason to do that and bad reason why it shouldn't be like that.