The idea of making my own housings for Stax drivers was born out of my experience with the Lambda and the Sigma Pro. Whereas the Lambda sounded too shrill (with the SRM-1 [?]) -- and almost literally caused toothache --, the Sigma's sound was too boomy, diffuse, muddy and lacked midrange (with the SRM-T1). I wanted to strike a balance between both, conveying some of the Sigma's spatiality and bass fullness, but preserving the electrostatic virtues such as resolution, accuracy, detail and refinement -- which were mostly lost with the Sigma -- and at the same time reduce the Lambda's treble glare.
Edited by JaZZ - 5/15/10 at 11:14am
So I decided to go the middle course between the two designs, i.e. angling the drivers, but clearly less than the Sigma's ~80°: I chose 30°, out of my belly. The prototype consisted of a wire grid, lined with double-layer velour -- I wanted to try a non-reflective housing and hence accepted a just decent isolation against phase cancellation due to rear-/front-wave interaction. So I slaughtered my Sigma and implanted its drivers into the prototype. And I was disappointed about the weak bass and an over-all muddiness, apparently resulting from the inappropriate acoustic isolation of the housing.
Nevertheless, the sonic balance and particularly the soundstage were as I had hoped for. So I built some more stable housings made of «airplane (birch) plywood» -- the same wood I used for making my boomerangs. To prevent them from being overly reflective, I carefully lined them with pieces of carpet, wrapped in black velvet. Unfortunately removing the drivers from the prototypes was almost impossible, and both got torn apart! So I had to order a pair of Lambda drivers from the local distributor -- they are the same as the Sigma's, just with different soldering flaps.
This design was a full success. The larger air volume between driver and ear (compared to the Lambda's) provided a fuller, more impactful bass, and the angling made for the optimal compromise between spatiality/airiness and accuracy/refinement. The bass has exactly the right proportion to the rest of the spectrum, reaches farther down than all other headphones I've heard then, the midrange was beautiful, colorful and smooth, and the treble had all the clarity of the original Lambda, just a much ear-friendlier loudness level. And finally the soundstage... it was incredibly realistic.
A while later I decided to make a similar design around the Lambda Signature Pro drivers. I chose a narrower angle -- 22° -- just for the sake of variation.
This design fulfilled all my expectations, too. This «Electrostat 2» sounded even faster and had even more detail than «Electrostat 1», in fact it strongly reminded me of the Sennheiser Orpheus. (A few years later -- 2007 -- I had the opportunity to compare both extensively. There was definitely a similarity, but the HE 90 had still better detail and resolution. Nevertheless, in the end I preferred my own electrostats to it, and it wasn't even a close call. With the HE 90 I missed blackness, fullness and impact in the bass as well as some smoothness and liquidness in the treble. Over-all it sounded too dry and not lively enough to me. It was the first time the term «PRAT» seemed adequate to me, although I generally don't like it.)
In the meantime I have a pair of Lambda Nova drivers laying on a bookshelf since many months, waiting for some nice wooden housings. The Lambda Nova I now own since about a year makes me think it actually wasn't the worst decision to be lazy.