Output impedance as compensation?

[Marine Boat]

In a UPS truck far far away, there are a pair of in-ear monitors known as the Ortofon e-q5's. The main source I'll be using with these will be the creative x-fi pro USB DAC. According to measurement's done by nwavguy this DAC seems to be pretty respectable aside from a fairly low output voltage and high-ish output impedance. Before that UPS truck barrels through my front door (let's hope not), I'd like to have a decent headamp built. However, considering that they're IEMs and have a rated efficiency of 117 db/mw the output voltage of the x-fi will be plenty. So I'm left to build a simple current buffer. Easy, and if that were all I wouldn't be posting this.

But according to measurements done by goldenears, the impedance quoted by Ortofon (40 ohms) isn't quite right. I'm taking it for granted based on what I've read that I'll need a lower output impedance than the x-fi can provide if I want a decent bass response, but the impedance shoots up like a bottle rocket above around 3khz. The frequency response drops as well past about 7. This is no doubt partly due to the fact that they were measured with the foam tips, but the drop will likely still exist, based on the comparisons they did between the eq7s with silicon and foam.


I've looked through Rod Elliott's work on the subject of output impedance and it's got me thinking:

I know current drive will cause muddy-sounding under-damped bass, but are there any negative effects for the treble where I plan to maybe use it?

Would it be feasible to design an amp that has a higher zout above a certain frequency so that by the time I hit that rising impedance I'm running in current drive? Perhaps by doing something akin to bi-amping but with both amps going to the single driver I could achieve a good damping factor (via low output impedance) for the bass frequencies yet also get an adequate amount of current into the load by using a high zout for the treble.

 

[jcx]

Reference earphones - suggests finding the EQ curve with a test signal source and your own ears

for sensitive IEM a audio step down transformer could be a better choice than an amp

 

[Marine Boat]

Reference earphones - suggests finding the EQ curve with a test signal source and your own ears

But the target with IEMs is not a "uniform perceived amplitude response to sinewaves" as Mr. Linkwitz says. Unless you're listening to binaural recording, of course. Where an ideal microphone is flat, the ear has a varying response corresponding to the target curves you can read about on etymotic's website. I will keep that article in mind though should I chose to explore the world of binaural recordings. Nevertheless, the drop in HF response on the Ortofons seems to be (at least partly) because of the rising impedance.

for sensitive IEM a audio step down transformer could be a better choice than an amp

What's your rationale for this? Aside from noise performance, a real current amp tends to be a lot closer to an ideal current amp than a real transformer is to an ideal transformer in my experience. Even noise performance can be made a non-issue with careful design.

 

[jcx]

Likwitz shows how to do passive series EQ - you can decide your target curve - I just thought a directly relevant link might be appreciated

current outputs are going to come down to a sense R and a signal Vsource

so you get is the sense R Voltage divider relation with the IEM Z - and several added noise sources from your active componets boosting the compliance V to make a controlled current source

with a transformer you get the turns ratio division for the command V+noise - but fewer added noise sources, and the turns squared relation for source and primary Z reducing their effects