Cmoy with lowest noise possible

Yes, use R5 to raise the impedance seen by the op amp.

You will need a little extra gain to compensate for R5.

Keep the parallel value of the gain and feedback resistors

to a value of 300 to 500. If you etch your own board, use

a ground plain design. A volume control of 10k will not

add much noise. A gain of five should be low enough to

keep the hiss down but high enough for good stability

and compensation for R5. An R5 value of 51 ohms should

be enough. Keep R5 as close to the op amp as possible.

R5 also isolates the op amp from cable capacitance.

You can use the same resistor type for all resistors.

an amp is simply the wrong approach for sensitive iem

http://www.head-fi.org/forum/thread/553094/continued-sidetrack-discussion-from-tiniest-portable-amp-i-can-build-nikongod-microtransformer-based-impedance-step-down-box 

The windings of the transformer would appear as a very low impedance to DC.

This could cause problems for your iMod. On the other hand, the transformer

will not pass DC.

Transformers will color your sound. Whether or not that is a bad thing

is up to your ears.

Quote:

Originally Posted by juswyq 

thank you for the very informative posts. 

 

I intend to build a high quality cmoy for some cans as well, so having information to build a low noise cmoy would be very helpful.

 

for the transformers:

I understand that it's meant to make your low-impedance IEMs appear as high-impedance so that the source can drive them better. however, can the cheap transformers cause any distortion or add in any unwanted colourations to the sound?

 

a separate question: imods have a dc voltage offset if you don't use capacitors. would running them through this transformer set up also get rid of the dc voltage offset? since transformers don't work with dc circuits. that would eliminate the need for an output capacitor, by far my greatest problem with the imod (i don't like running my audio signal through a capacitor - was looking for a way to remove it from the cmoy but it seemed like the potential cost of ruining my headphones far outweighed the sound quality gain)

 

I apologise if I'm making any mistakes in my electronics theory, I'm not EE trained so I'm seeking enlightenment. Tried to search for answers to my questions but I couldn't find any.

RE distortion - 

EVERYTHING distorts. At best you can pick where & how. Transformers do have some inherent distortions, but at low levels they are pretty small. The key point for IEMs is that a transformer adds basically zero noise to the signal, which is hardly true for active circuits. 

Compared to a Cmoy (some of the newer ultra-high current op amps designed for low impedance loads dont suffer as badly) you have to account for "overloading" the op amp. 

RE DC offset:

The transformer will block DC but I'm not sure how well the little guys in that thread deal with DC current. I'm also not sure that the line out on the ipod will deal well with sourcing output current. If its any consolation you could probably get away with a 10-22uf cap :)

Something I have been thinking about for a while is an active voltage shifter to replace the output caps in the ipod. 

IEMs rarely do need any voltage amplification, even with the .5V output of an iPod, so transformers, as Ari and jcx suggests, or an active buffer with good current performance should fit the bill.

NwAvGuy just reviewed a Cmoy designed to these goals, based on a JRC4556 and dual 9V power supply. Maybe some of the more tech savvy folks here can find ways to further improve on the design to drive IEMs, but already it is looking very promising.

I built a bridged/balanced "JRC4556 cmoy" for my Grado RS-1 headphones. Its a fun little amp. I built it as a joke about "balanced amps are inherently better" (I honestly wanted it to suck) but it turned out better than I hoped. Grumbles.

As a tangent, does anyone know what the DC offset of the ipod line out is before the coupling caps? 

series R can change the headphone frequency response if the headphone impedance varies

single driver headphones may have a hump in their impedance curve at low frequency which givies a small bass boost with a hi output resistance amp (or with a normal amp and added series resistor)

by making a resistor divider you can reduce the output level and have a smaller equivalent added series resistance  - see the "hiss buster" thread

  http://www.head-fi.org/forum/thread/198828/the-hissbuster-for-sensitive-headphones

some multidriver iem have very "lumpy" impedance curves and series R would be expected to give audible frequency response coloring

but a transformer is the most "efficient", literally and in terms of giving lower noise and can be wound for exremely low added series R 

Thanks for the info on the DiyImod voltage offset. Im going to think about this one for a while. I have some ideas though 

Regarding the ER4 P-S resistor. Yes, adding a resistor does reduce the load on the amp in a kind-of-sort-of similar way to the transformer BUT the sensitivity of the ER4S is much lower than the P so the amp has to swing more voltage. It has to swing more voltage into the transformer too, so I cant say for sure which is the easier load to drive. Id like to guess its the transformer as 100ohms is still in a kind of tricky range for a lot of op amps and >1Kohms is pretty easy but the voltage swing could do the opamp in.

Something nice about the transformer is that it has applications with the multi driver IEMs where a resistor cable definitely does not. Resistor cables like the Ety P-S cable interact very poorly with the crossover networks in the multi-driver IEM's and generally mess up the sound. The transformer has a low output impedance and knocks hiss down a bunch.