Measuring distorting Etymotic earphones with an osciloscope

[RBBVNL9]

This week, I got annoyed by my favourite Etymotic earphones starting to sound distorted in one channel. I decided to use my lab equipment to see to investigate that distortion in mode retail.

In this video you see how I did that both on an R&S RTB and a Siglent SDS2k scope:

After making the video, I realised that I also could have exploited the fact that the SDS can do two FTT analyses at once! To do so, I took two identical microphones (Røde M5) and treated them with the same mic pre-amp settings, then connecting them to channel 1 and channel 3 of the SDS scope (pic attached). The result shows that the 'good' channel (Orange trace, green FFT) has fewer unwanted harmonics than the 'bad' channel (yellow trace, pink FFT).

The dual FFT of the SDS is clearly an advantage over scopes that can do one FFT at the time only ;-)

Next step is to try the audio analyser on the CMU200 to investigate the distortion… And to see whether I can get it repaired, perhaps by compressed air, as someone suggested!

[wasedadoc]

No need for FFTs to see that level of distortion. 

Are the headphones purely passive or do they contain some active electronics?

[RBBVNL9]

No need for FFTs to see that level of distortion. 

Absolutely, the waveform already gives away that the headphones do not reproduce a proper sine wave. But I was interested to learn how the distortion my ears experienced translated into harmonic content. I found it surprising that even the 'good' earpiece had a fair amount of harmonic distortion, but mostly uneven. Want to dig a bit more into that.

Are the headphones purely passive or do they contain some active electronics?

They are passive and have a single driver. After these Etymotic ER4 were introduced in the 1990s, many other brands followed with in-ear models, some single-driver and some multi-driver designs. But for me, as an former live sound engineer, the Etymotics always remained among my favourites! They always ship with individual frequency response graphs for the left and right earpieces you bought. Very nice.

[switchabl]

I found it surprising that even the 'good' earpiece had a fair amount of harmonic distortion, but mostly uneven. Want to dig a bit more into that.

I am not sure that you should draw any quantitative conclusions from this setup. The drivers are tiny and are meant to go into a sealed ear canal. I believe that if you place them in free space, radiation efficiency increases with frequency (something like 6 dB/octave, even though the proximity effect of the cardioid mics may offset this to some degree), so there will be a large emphasis on the harmonics.

That's my understanding at least, I mean, after all acoustics is just a kind of scalar EM. 

[tooki]

This week, I got annoyed by my favourite Etymotic earphones starting to sound distorted in one channel. I decided to use my lab equipment to see to investigate that distortion in mode retail.

Mode retail??

Anyway, do you not see the earwax filter sticking out of the lower one? I’d start there…

[RBBVNL9]

I am not sure that you should draw any quantitative conclusions from this setup. The drivers are tiny and are meant to go into a sealed ear canal. I believe that if you place them in free space, radiation efficiency increases with frequency (something like 6 dB/octave, even though the proximity effect of the cardioid mics may offset this to some degree), so there will be a large emphasis on the harmonics.

Thanks for the reply! Yes, that certainly makes sense. Just in the open air, they are not acoustically loaded as they are designed for and might behave in unintended ways. Unfortunately, I do not have this 'fake head' or whatever they are called for proper tests of in-ear devices.

Still, I find the relative differences interesting: the two earpieces/drivers behave differently, even if tested in free space.

[RBBVNL9]

Code: [Select]

Mode retail??
Oops, it should be 'more detail'. Sorry for the typo...

Anyway, do you not see the earwax filter sticking out of the lower one? I’d start there…

Yes, I now see in the picture one of them has the filter in there; the other not. The one with the filter is the 'good' one; I remember taking it out of the 'bad' one. Still, for a 1:1 comparison, I should have taken them both out. You learn every day and never can be careful enough when experimenting ;-)

[tooki]

Code: [Select]

Mode retail??
Oops, it should be 'more detail'. Sorry for the typo...

Heheh. FYI, that specific type of transposition has a special name, a Spoonerism (after a Mr. Spooner who famously made lots of them, like “well-oiled bicycle” -> “well-boiled icicle”! )

As for characterizing your Ety’s, with these tiny acoustic systems, even the tiniest differences in shape, etc. have significant effects on the frequency response, etc. I don’t think any simple setup is capable of providing any meaningful results.

[bicycleguy]

Not sure about your instruments but with mine the fft's work a lot better if the signal of interest is full screen but not clipping.  Full resolution requires full amplitude.

[RBBVNL9]

Heheh. FYI, that specific type of transposition has a special name, a Spoonerism (after a Mr. Spooner who famously made lots of them, like “well-oiled bicycle” -> “well-boiled icicle”! )

Today I received a new set of Etymotic ER4XR earphones I ordered, and, as always, they include an individual measurement sheet with them. Nice! Interestingly, this sheet also reveals how the manufacturer measures themselves: they use a G.R.A.S. RA0045 Ear Simulator. It's based on the IEC 60318-4 standard.

The  manual (page 5) explains that "The acoustic input impedance [...] closely resembles that of the human ear up to 10 kHz and, as a result, loads a sound source in very much the same way.", and also provides the electrical equivalent diagram. 

Not sure about your instruments but with mine the fft's work a lot better if the signal of interest is full screen but not clipping.  Full resolution requires full amplitude.

Good point. I was in an environment where the mics also picked up my voice (when I was making the video ;-). For more serious testing, I'd need both a proper acoustic load, as well as proper full range settings!