What would be the most transparent DIY headphone amp to build?

Misread

The answer is probably a AMB Labs Beta22.

The Wire

Try a Dynalo.

sensitivity, Ohms? - otherwise no one can know how much I, V needed - the dynalo in particular has low V swing may not drive "studio monitor" headphones to acceptable SPL

"balanced" bridged output amps double the Vout for a given type - may be useful for op amp + buffer in the loop circuits when you need > ~ 8 Vrms

I would recommend an O2 or something similar, but if you want extra drive capability for really difficult headphones (As in you want to listen to the most inefficient of the orthodynamic models at very high volumes) you couldn't really go wrong with the Wire.

I would advise against amplifiers that haven't got a decent set of measurements. Especially really expensive ones. Especially really expensive ones with bizarre designs.

O2?? Are you serious? 

HiFi1972: I would definetely build DisPre II JFET (balanced if necessary), because that is an real amp.

http://web.telecom.cz/macura/dispre2-jfet_en.htm

Quote:

Originally Posted by AmarokCZ 

O2?? Are you serious? 

HiFi1972: I would definetely build DisPre II JFET (balanced if necessary), because that is an real amp.

http://web.telecom.cz/macura/dispre2-jfet_en.htm

What's wrong with an Objective2?  It can drive anything from 16 ohms to 600 ohms and was designed on transparency.  

Well, you can say that about (almost) every amp.

O2 is based only on OA, OAs are not primaly designed to drive loads and if they do, their performance is not very good. On the other hand DisPre is without any compromises.

Quote:

Originally Posted by AmarokCZ 

Well, you can say that about (almost) every amp.

O2 is based only on OA, OAs are not primaly designed to drive loads and if they do, their performance is not very good. On the other hand DisPre is without any compromises.

The designer of the O2 has done some very extensive measurements that would tend to disagree with with you.. Of course I would agree with you that a Discrete amplifier may have a performance advantage, but I take issue with blanket statements like:  "OA(opamp)s are not primaly designed to drive loads and if they do, their performance is not very good"  

Quote:

Originally Posted by Willakan 

Especially really expensive ones. Especially really expensive ones with bizarre designs.

What about really cheap ones with bizarre designs?

The Wire uses op amps and gets somewhat better performance than that...also it has slightly higher voltage swing (9.2V rms), for those 600 ohms vintage AKG models and so on.  I don't think like a couple dB of difference is make-or-break level though.

Quote:

Originally Posted by AmarokCZ 

Well, you can say that about (almost) every amp.

O2 is based only on OA, OAs are not primaly designed to drive loads and if they do, their performance is not very good. On the other hand DisPre is without any compromises.

every engineering design is a trade off of some sort, Dispre clearly is compromised in all of PSRR, output Vswing, output Z at audio, gain could be higher too for hi Z load - a TPA6120 will match it for noise at low source Z, match its Vswing, likely beat any distortion measure - a multiloop using OPA627 around a TPA6120 would be even better

whether any of that makes for "more transparent" audio seems unlikely - if I were trying to make the case for a discrete amp over monlolithic op amp options I would at least use higher supply V, say +/-24 V, some "studio monitor" headphones like even more V 

Quote:

Originally Posted by nikongod 

What about really cheap ones with bizarre designs?

 

Run like the wind from the first person who tries to sell you his $30 single-ended triode-ified monstrosity.

Also @Amarok: What?! There are opamps which are fine with low impedance loads. There are opamps which are not. I think you've been taking Burson's PR sheets a little too seriously.