Tubes or Solid State

[Leporello]

Since the current setup uses a step-up transformer to boost the voltages, don't we already have very high voltages on the stators?
We see this when a stator seal cracks or peels away and arcs to the diaphragm.

The main difference vs direct drive is the risk of having that high voltage in the cables between the amp and the speaker; that's where the real problem lies.
I dealt with this conundrum in my design for the Monoray line by having a custom Kevlar-wrapped 'umbilical' cord to carry the signals (and any LV power) needed by the speakers. Using sealed locking connectors, an amp output relay that only closes once the connections are validated as safe. My direct-drive amp is a power-DAC like some of the NAD DDFA amps but with beefy modern power FET devices.

See the brochure PDF I wrote for more details: https://cdn.imagearchive.com/martinloganowners/data/attach/15/15199-MonorayBrochureFlat.pdf

The stepped up audio signal from a normal ESL driver transformer system could be harmful or lethal under unlucky circumstances, and you should always use caution when working inside the electronics, especially if feeding an audio signal. All the direct coupled tube amps I have seen use plate voltage for the DC bias, which is very dangerous. That could be mitigated somewhat by using a grounded anode arrangement with the cathode floating at -2000V or so, but I have not seen this. The Hermeyer 'stats, referenced in the Sanders white paper, had bare metal rods as stators, at transmitting tube plate voltage. No infants or pets should ever be in the room with them!

The Acoustat panels, when they went with direct coupled tubes, used insulated wires as stators inside plastic supports (fluorescent light diffusors were the material of choice for home made versions). So quite a bit of isolation between the user and high voltage. ML panels use perforated metal with propitiatory insulation--probably safe but I'd be nervous about touching both sators at once with a signal applied, let alone with both floating at transmitting tube plate voltages.

I have long believed the ideal setup for driving ESL's with tubes would be a single, low turns ratio transformer between the tube plates and ESL stators. Low turns ratio transformers are easier to build with less flux loss. Stepping the signal down to an 8 ohm nominal impedance then back up to high voltage to drive the 'stats just doesn't make engineering sense to me.

 

[JonFo]

ML panels use perforated metal with propitiatory insulation--probably safe

I saw the process at the Kansas factory during our MLO tour in 2007. It starts with flat perforated metal sheets, then runs through a curving press to give them the 30-degree arc, and then hangs them on a charged line that goes through a powder coating booth. Then they go into a baking oven to set it. I think it was a fairly standard silica-based powder coat material. It is able to insulate the high voltages as long as it's not cracked or peeled away.
From my efforts at separating the old panels for my sidewall installation, the coating is pretty resilient and can handle rough treatment without (visual) harm.

 

[Ercana]

Quite afraid to speak after persons speaking of OTL amps directly coupled to ESL stators. This could obviously be the best technical solution to eliminate many parasite capacitors and inductances, but in a world made for electromagnetic transducers, this is not a path paved with roses. After all, I decided to use the 4 ohm output of all amps with my refurbished Aerius which goes down very quietly to 2 ohm and less at 20kHz. The new Peerless woofer with its big 82uF PP cap accepts this drive and gives everything needed to sustain the electrostatic counterpart. At the 4ohm output, The Aerius (not-i) are giving a great sound with any of the tube or silicium amp that I have tested (Audio-Research, Yamaha and even the tiny Jolida Jd-202a). Fifty watt per channel are more than sufficient in those conditions, and no 6550 or EL34 were harmed.

 

[Bodyslam]

I have long believed the ideal setup for driving ESL's with tubes would be a single, low turns ratio transformer between the tube plates and ESL stators. Low turns ratio transformers are easier to build with less flux loss. Stepping the signal down to an 8 ohm nominal impedance then back up to high voltage to drive the 'stats just doesn't make engineering sense to me.

I agree with this, and have thought so for a long time. But I have never heard of anyone actually trying it.

Of course it would require a custom designed transformer. There's nothing you can buy off the shelf that would cover all the requirements. I once asked Tim de Paravicini whether he would make a pair for me so I could use his amps to direct-drive my panels. That was neither the first time nor the last time he called me an idiot.

In any case, I have been driving my Monolith panels with Tim's 549 250-watt tube amps for decades. Out of all the gear that has come and gone from my system, it's the only gear that my wife has forbidden me to sell. I'm pretty happy with it too.