[eightbitpotion]

OPA2134

or

OPA2132

What's the big difference? I've already ordered a 2132.... did I mess up?

 

[majkel]

OPA2132 is sonicallly slighlty better IMHO, but they're the same regarding the specs.

 

[fordgtlover]

I have used both and I can't actually tell the difference in sound. Tangent disusses the details of each at his website linked below. Note that Tangent reviews the single versions of each, which are called OPA134 and OPA132 respectively.

http://tangentsoft.net/audio/opamps.html

 

[Joshatdot]

OPA2227

 

[MrMajestic2]

Quote:

Originally Posted by Joshatdot /img/forum/go_quote.gif OPA2227

I agree, great op-amp for a cmoy.

 

[majkel]

OPA2227 is horribly dull and laid back - big bore guaranteed. Maybe good for Grados but terrible for Senns. It's uncompensated version OPA2228 is waaaaaaaay better, I'd even say, pleasant sounding.

 

[balou]

has nobody yet mentioned the AD8066

?
yeah, confusing poor newbies is fun

 

[majkel]

balou, I understand we'll go one by one?

TLE2142 is my answer.

 

[balou]

um... are you serious? looks like a rather strange opamp

one chip which imho blows the ad8066 in a cmoy (not refinement-wise, but in almost every other respect) if you use grados is the njm4556/jrc4556, the same as in the grado ra1 amp.

edit: tle2142 has an output current of 80mA. I now understand the 'why'

 

[kristleifur]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif OPA2227 is horribly dull and laid back - big bore guaranteed. Maybe good for Grados but terrible for Senns. It's uncompensated version OPA2228 is waaaaaaaay better, I'd even say, pleasant sounding.

I tried the 2227 in a PIMETA and I agree with you, majkel - and I've got Grados. (The 2227 is too laid back for my tastes, that is, even with Grados.)

 

[Filburt]

TLE2142 is a basic BJT input op-amp with decent output current. Load drive capability is...okay I guess. Distortion is fairly high, open loop bandwidth is very small (looks like around 10Hz before -3db), and it's not really designed for very low impedance loads like a headphone. You could use it but it's no AD8397.

OPA2227 is ill-suited for this task. Distortion is high, especially after a few KHz, load drive capability is low, open loop bandwidth is tiny, and settling characteristics aren't very good on it. The 2227 is a 'precision' op-amp designed to work at or around DC; it is not designed to work for the audio band, and it is not designed to push a low impedance reactive load.

 

[mminutel]

The best thing to do would be to use a DIP-8 socket and roll them around until you find one suitable for your needs. I have found that the 2227 is better than the 2032 because it relaxes the bass. The 2032 is very tubby in the bass. If you need more bass in your phones then you might go that route, but if you find that the bass is not lacking then you might want to go with something else. The AD8066 is a really good chip, but you will have to solder it onto an adapter because it only comes in SOIC-8 and MSOP which are not compatible with the layout of the CMOY without an adapter. You may be able to find somebody that will sell you one that is premounted.

 

[majkel]

Guys, I just mentioned one op-amp for fun.

I agree the best is AD8397 and mentioned it many times in similar topics. Its biggest advantage is current efficiency - 310mA peak. NJM4556 is around 70mA per channel AFAIK, still better than a regular op-amp. TLE2142 is quite good sounding, IMHO better than OPA2132, 2134, 2227. 2228 is a different story. I also like AD8066 and TLE2082 but they are just regular op-amps, so no significant current amplification expected. AD8620 also very nice. And, my favourite as a driving stage op-amp is... AD825. It is single, so you need two but it excels even OPA637 IMHO. Instead of a Cmoy it's better to build a PINT - low noise version.

 

[balou]

but it's not really suited for a cmoy - it also has high quiescent current draw, so you almost need a battery charger, and it needs decoupling, also something which is beyond a simple cmoy design (especially a first one)

the lmh6654 and lm6171 might also be something to consider in a cmoy design. but at least the lm6171 needs decoupling too - but seemingly it's not as cranky as the ad8397, and also got lower quiescent current

 

[majkel]

The only decoupling you need for AD8397 is a 100nF ceramic cap between V+ and V-. No decoupling to ground is recommended, especially when you're going to build virtual ground channel on op-amps. AD8397 is unity gain stable, so it won't oscillate like a LM6172 without compensation. The problem with AD8397 is its high bias and offset current, that's why there's coupling via resistor in the low noise PINT design.