On direct drive vs interface. "Our recent auditions with the new Solaja 300B² showed that even with our transformer Interface after its output transformer, it had more life, detail and breath than the direct-drive VM-1a in pentode mode with only one transformer in the signal path. The price you pay with transformers relates to their transformation ratio. If you need a 100:1 ratio, it really doesn't matter if you combine two 10:1 transformers or use a single 100:1. Actually, you can better optimize each trafo if you divide the ratio between the two. The 10:1 ratio of our interface is a very relaxed design with about 1/10th the turns or windings you'll find in a typical tube output transformer. Our simpler requirements really drive down the parasitic effects which are a square function of the number of turns. We use a quality core and properly anneal the toroid after winding it. The VM-1a wasn't superior because it was a 'direct-drive' design where other amps need our interface. It's because Dragan Solaja knows a thing or two about designing great valve amps. His VM's output transformer designed by Menno van Der Ven was a toroid which required a 0.3Ω load at its secondary to present 4kΩ to the anodes. With a single toroid, you can actually sort out the parasitic inductance even at 100:1 ratios. Other core types not fully covered in windings make that far more complex. The solid-state HSA amps are designed for 1Ω loads and have a 0.68Ω ballast resistor to not rely on the 0.4Ω/1.8µH cable resistance like Jotunheim R does. My job was to make a good interface/cable system that won't compromise excellent amplifiers by other designers. Our interface really doesn't dumb things down to create more usage scenarios for ribbon headphones. That'd be anathema to always moving forward not back. If the interface creates more application scenarios, I'm all for it. But that's entirely a secondary benefit. My primary focus is advancing the state of the ribbon headphone art." On cost of manufacture, "after counting our work hours, increased wages following inflation, growing costs for services and materials, we can't price these any lower. They take an extraordinarily long time to build. Including control and measurements, it takes one experienced worker a full day to just make 2½ ribbons; and we have two guys making them. Still, the final sound makes it all worthwhile."

On my desktop the Enleum AMP-23R drives either headphones or speakers, currently Final D-8000 and sound|kaos Vox3awf with upfiring Raal ribbon tweeters.

On Magna/Immanis efficiency vs Susvara. "Magna's apparent loudness equals Susvara or on vocals even feels a touch louder. Immanis goes about 2dB louder again so any amplifier with sufficient gain to play Susvara at the desired SPL will work on our new models through their 32Ω interface." In the head-fi.org Raal 1995 thread, a user of the Feliks Audio Euforia reports brilliant results with the SR1a through the interface. Apparently this amp's original gain setting has been changed to an impedance adjuster since Alex's brief show trial to now work splendidly. Likewise for misunderstanding the Bakoon AMP-21R and Enleum AMP-23R. In default low gain they lack ribbon drive. Without visible gain adjustment, users trying default mode will write them off. Set to high [option mode ⇒ ±vol ⇒ V1 LED = lo, V2 LED = hi], the circuits' full 25wpc/8Ω speaker power routes to their 6.3mm outputs. Those drive Susvara with copious headroom to spare; and will power Magna and Immanis at +6wpc into 32Ω so with aplomb.

From Herb Reichert's review of the Feliks Envy 300B SET: "… With its 83dB/mW sensitivity, HiFiMan's Susvara headphone needs plenty of voltage to produce sufficient volume – hence, high gain. And because its impedance is on the low side at 62Ω, it needs sufficient power to support that voltage… To accommodate difficult loads like the Susvara, the Envy's power is rated at 8wpc and a front-panel control knob allows gain selection in three steps: Lo, Med, Hi. When the Susvara was driven at the Envy's lowest gain setting (also a low-impedance setting), music sounded clipped, dull, dark and distorted. Even with the volume knob set at 5 o'clock, the music was insufficiently loud. At the medium gain setting, the volume level was comfortable with the control at about 12 o'clock; the sound brightened up and got nice. At the Envy's highest gain setting, at the same volume, it's like someone threw a knife switch and the stage lights went on. It made the famously hard-to-drive Susvara light up and sing with great purity and feeling."

On advancing the SR1a/b concept: "I gathered certain insights while working on Magna and Immanis which eventually I'd like to incorporate into new open-baffle headphones. I just don't yet know how, exactly. I think I could introduce them in perhaps two years? Most probably their 'wings' would sit at a fixed angle with some additional padding at the top and bottom for less wiggly wear and deeper bass. Future open-baffle models would also focus on just front not back leakage. First I'll have to decide though whether they should rely on dedicated amplifiers; become independent of them; or be a mix of both depending on model tier. Still, for now all that's far in the future." Why the SR1a/b themselves won't change: "Our new approach stopped exchangeable ribbons. We still slide in the frames which stretch the ribbons but the smaller gaps between frame and free ribbon edges are now so tight that we can't allow for the earlier slightly arced frames which provided friction to hold them in place. The new frames are straight and sit so loose between the magnets that they'd rattle if we didn't cement them in place. Our new ribbon tuning adjustments don't allow for the frames to be under any tension. Tension would change the tuning differences we want between adjacent ribbons inside one driver. In short, these new developments wouldn't really benefit our original SR1a/b whose single ribbon already settled at its resonant frequency." From Alex's earlier drawing of a ribbon and a planarmagnetic, we remember the latter's multiple drum-resonant modes whilst a ribbon only exhibits one easily suppressed longitudinal standing wave. Does the offset tuning of adjacent diaphragms involve different resonant frequencies based on slightly different ribbon lengths? Without lifting trade secrets, there's clearly more to the MO of these models than meets the eye; or can be duplicated without hard-won experience and highly specialized labour skills.