Tuesday, September 4, 2012

AKG K3003 part3: In-depth analysis


Introduction

AKG K3003 is such an innovated product. The manufacturer, AKG, put a lot of engineering effort in fine-tuning this IEM. As mentioned before, this IEM has so much room -  for better or for worse - for an user oriented acoustic tweaking. However, unless a careful attention is given to the tweak itself to accurately estimate what the effect would be, the effort might end up with negative consequences.

This in-depth analysis will demonstrate the consequences resulting from such tweaks, and hopefully provide some helpful insights to the general public. Below are main topics, and respective topics will be discussed separately.

First, The effect of various insertion depths using different ear sleeves

Second, the effect of under-damping, using a serial resistor adapter.

Third, the effect of acoustic dampers

Fourth, the effect of a pressure equalization port


Various insertion depths with different ear sleeves

It seems the manufacturer was well aware of K3003's bulkiness, and it is virtually impossible for the users to insert this giant into their ear canals. So AKG designed the product in such way that the frequency response measured at a shallow insertion depth to be the reference. The picture on the left is K3003 inserted in a proprietary external ear canal adapter, with the other end of the adapter is at which the reference plane of an occluded ear simulator is located. As seen on the picture, the insertion does not go any deeper due to its enormousness in size. The distance from the tip of the IEM to the reference plane is approximately 3 mm, and that should be about where this IEM would end up at in a real ear canal in practice, regardless of the sleeve size.

And what's seen on the right is the left and right channel responses averaged, with different ear sleeves attached. The sleeves differ in size to a great degree, and consequently, they end up with different insertion depths. The smallest one is inserted the deepest, and vice versa.









Damping AKG K3003 electrically

Impressively, although K3003 is a multi-driver IEM, its impedance characteristic is quite linear with increasing impedance as the frequency rises. This tendency is also shown on Etymotic Research's ER-4 series, of which the manufacturer increased the high frequency output by simply adding a resistor to a Knowles ED-9689. The same electric damping principle can also be applied to K3003, and it will further extend the high frequency bandwidth.

With a simple frequency-dependent attenuation calculation borrowed from the Audioholics.com,

According to the formula, a resistor bigger than 48 Ω will not do much for AKG K3003, since it is the value at which the output becomes fully velocity-driven. Anyway, the result can be estimated to be around 4 dB increase in output at 10 kHz, and 5 dB at 20 kHz, with the DC remaining a zero relative level as a reference.  And in practice, with the amplifier output impedance added, the result is shown below:


Damping K3003 acoustically

AKG K3003 comes with three different acoustic filters, which have various acoustic impedances. The black filter is 3000 Ω, the gray filter is 1500 Ω, and finally, the white one has only a protective steel mesh.


And the relative effect of each dampers are shown on the right. With the gray being a reference, Left and right channels are averaged together to further improve accuracy of the data.

Please keep in mind the data are for reference only. DO NOT EVER USE K3003 with the filters REMOVED unless you exactly know what you're trying to do.







A vent-slit pressure equalization 

Again, here is the another proof AKG meticulously thought out the overall design of K3003. Not only this port effectively avoids the over-pressurization in the ear canal, which causes an excursion issue of a dynamic transducer suggested by Ambrose, it also prevents the low frequency response to be overly slow & bassy. This port is actually an attenuator.






When this port is blocked, the ear canal becomes over-pressurized, resulting in even more bass, up to 6 dB @ 20 Hz. As seen previously, K3003 is quite bass-heavy even to begin with. Unless the user is an hardcore bass-head, this is not a very desirable tweak, considering the ultimate goal of an audio reproduction is hi-fidelity.





In conclusion

Everything has been covered in regards to AKG K3003's technical potentiality, and it shall be up to the individual users to determine what kind of tweaks would suit them the best. For my personal preference, an under-damped with the reference filters yields the most linear performance.

AKG K3003 is a great IEM, and there's so much fun to delve into. One of my blog readers commented on the part 1 of this series:
 "But... Can it beat the Etymotic ER-4?
Well, the answer to his/her question really depends on what you're looking for from your IEM. If you're looking for a frequency response linearity in accordance to the diffuse-field reference, NO. Actually, no IEM can beat ER-4 in that league. However, for everything else, including high frequency extension, tweakability, and distortion figure, definitely YES.


Appendix

Inks mentioned the discrepancy of my data with other websites. In order to explain the issue, and to advocate my good friend, ソノベ-san's honor in data accuracy, I super-imposed my data, forcibly inserted up to the reference plane, to ソノベ-san's data.



As you can see, I tried to match the resonant frequencies as close as possible. I suspect the deviation here is caused by the acoustic output impedance mismatch between a 2-cc coupler and an occluded ear simulator, as shown below. Is ソノベ-san's 2-cc data accurate? YOU BET! There is a reason why he is my all-time mentor. (my own self-claimed, that is. LOL)
plots acquired from B&K and ACO Japan, respectively

1 comment:

  1. Once again, and always...

    Great Work!

    Thank You

    ReplyDelete