Congratulations! Please share how they sound to you.
1. 800Vpp so did you apply +-400V to the membranes? 1.2mm with such voltage will sound quiet. But if you double these 400 to +-800, you will see a huge difference. Going to 0.6mm though even with +- 400v will stick membrane to the stator especially with such low stretching (108Hrz). Remember it's not like normal design with one membrane which experience equal force from both stators in idle. These 2 membranes are attracted to each other with 800V electrical force, so they are always bended in the middle and closer to the stator, hence I used in my latest design dual middle stator bended like this )( . With a straight stator my experience with 0.6 mm spacers, 140 Hrz and +- 630V is that one of the membranes sticks after 5 mins to the stator. So 1.2mm is good if you want stability with +-800. And by the way distance is definitely not a linear factor - in fact it is quadratic to the e-force - (k x Q1 x Q2) / d^2. The reason you did not hear a difference is in my opinion, that membranes are stuck to the stator, discharged and of course very quiet. With 1.2mm distance and 400V it cannot be loud. Even normal design with 1.2mm spacers will be quiet (see p.3)
2. 2u mylar is in my opinion too thick for this configuration - that's one of the reason for the big bass - heavier membranes. You can try to stretch the membrane a bit more, to remove the bass, but with me this did not work - my current stretching is 170 Hrz and bass is almost the same as with 120Hrz. My opinion is that ear sealing plays much more role for the bass than stretching. I guess the only solution is to make the area smaller (which is good from esthetic point of view) and to order 1um Mylar. I love bass, so it's not that bad for me, I made some equalization to make it flat and it became a bit poor.
3. As for the loudness, ignoring the smaller space (because of the constant attraction in my design), +-400V (800V) and 800Vpp (single) should be equal to single 400V and 400Vpp (800 differential) in normal design. Assume stator is at +400V then one membrane sees 0V the other 800V. Now with normal design one stator is at -400 the other is +400, membrane is at +400. Again - one side is 0V, the other is 800, so it's the same es-force. And this is the main flow of single stator design - you have to double either the output of the amp or both biases to get the same level. I can't explain how you got such huge 20dB difference (my guess is the spacers are not the same thickness), I don't have much experience with a single membrane, but as I mentioned, I'm with +-630V now and about 580Vpp (single ended) and it's quite loud, my normal listening is at half volume. It's still quieter than a dynamic phones at the same source level, but still loud enough. If I go to high volumes my wife starts complaining - she can hear the phones pretty loud standing 2m away
. I see now how you are going in my steps - I used to have +-300V biases and yes it was quiet, but +-630 made a huge difference. Stretching is also a nightmare - you have to stretch 4 membranes in about the same frequency (better - slightly different) and glue them and check after that again and match them by pairs. The smaller space between center of membrane and stator also plays some role in the loudness. And if this is not enough you can always go for a higher biases (increasing amp output is not easy), by doubling the power rails voltage of your amp.
4. Highs - something is wrong with your setup - I have measured 30kHz there without any problem. This design also has theoretically almost 0 capacitance (only cable capacitance), so it shouldn't have any problems with high frequencies. May be you have somewhere high load capacitance to the output of the amp? By the way hum and crosstalk of channels is removed with small capacitors (NO MORE than 10nF, considering about 3MOhm biasing resistors) AFTER the biasing resistors on both +- biases to gnd. PLEASE be careful - biases become now very unpleasant - 10nF although not lethal is not pleasant to discharge through you. If you can - put 20Mohm and 1nF. And before anyone starts to blame me - I saw original STAX amp schematic with 10nF capacitors after the resistors. Just make sure you never touch naked wires, membranes is ok to touch you won't even feel the voltage. The best is to put the caps inside the phone frames, but that would require a gnd wire to run to the phone, which is inconvenient and adds additional capacitance.
5. The highs will surely go much higher when you close the back of the phones, this should also happen with normal design or any open baffle design, but that's not a solution at all. Electrostats don't tolerate closed volume (except at your ear), they must be opened at the back. You can put a very transparent grill at the back, that would not disturb the sound waves in any way.
PS. Just remembered something. I have an ellipse design - 100x70mm AA. Used the same amp and biases. The spacers are 0.6 and membranes stuck to the stator all the time. I made some additional spacers on a 3D printer - 0.2mm (single layer) and now membranes are stable with about 0.8-0.85 mm. I think, that the shape affects also stability, because the shorter 70mm distance of the ellipse equals to a circle with 70mm diameter, but the area is much bigger. I can tell these ellipses sound the same as my 90mm circles - may be a bit louder in the mids-highs. This could be also because my circles have these thin foam inside the earpads.