TOPIC:

Hi,

I finally fired up the EARS I got a few month back in anger.

I had a few false starts, trying to run this up and validate results against known headphone samples that had been tested independently on GRAS43.

With minor simple mods (that would ideally be applied at the factory) and some effort to get around the calibration, the EARS are a VERY useful piece of hardware. Once you got your setup done, just connect, measure, 10 minute job. At the price this is a no-brainer even if some time must be put in.

Here is what I found and struggled with...

First I found that the silicone ear/cheek pieces had poor seal against the metal and it did not relate purely to the fixing screws being overtightened (or rather tightened to the mechanical stop). I fixed that seal issue using clear silicone from a gun, applied sparingly and carefully on the metal plates before re-attaching the ear/cheek simulator silicone plates.

After this I got good seal for in ears, sometimes I got decent seal and sometimes not for over ears, .

I found that the screw heads even if fully tightened (as received, which also caused issues with seal) protruded enough to break the seal easily. So I replaced the screws with countersunk screws carefully tightened "just so" to leave the silicone surface flat without indention or protruding bits but flat.

Finally, reliable seal for over ears.

Over Ear headphones can be measured fairly repeatably, as well as with alternative solutions I'd say (care is always needed when placing headphones on ear simulators).

In Ear Monitors show good seal and good repeatability of test results as well.

I got around the calibration issue (I cannot really "get" what the process is supposed to do and it gave me wrong results to what I expected) as well (see my other thread asking for actual sensitivity).

Suggestions for improvements:

1) Apply silicone sealant during assembly to ensure reliable seal (needed for in ear / over ear).
2) Use countersunk screws of the correct length during assembly to ensure reliable seal (over ear only).
3) Provide a nominal (uncalibrated) sensitivity as "XXXdBFS @ 1Pa / 94dB and 0dB gain somewhere in the (on line) documentation.

These should have minimal (if any) impact on cost and would make a product with good intend and generally good hardware but flaws out of thge box into something simply excellent.

Thor

PS, as source I use an ifi-audio iDSD micro "Black Label" which outputs 2.2V @ 0dBFS in "eco" setting, volume at max and with a 0dBFS signal. Channel difference is low enough to not show up as a material difference

Thus a -7dBFS signal is very accurately 1V and with the EARS calibrated to absolute SPL it is easy to get SPL referenced as XXdB/1V (as Sennheiser specifies and as I prefer to use) or -27dBFS is 0.1V and appropriate for IEM measurements (Knowles uses 0.1V as reference signal).