A theoretical question for someone WAY smarter than myself...
1) We know that the output of a typical transistor power amplifier is low impedance (able to drive a low impedance device such as an 8 Ohm voice coil loudspeaker without the need for any impedance matching transformers).
2) We also know that the input of a typical electrostatic loudspeaker is very high (~10K Ohms), therefore a separate impedance-matching transformer is required to couple the low impedance output of the typical amplifier to the high impedance electrostatic.
3) Finally, we also know that the anode plate of a vacuum tube (the output point prior to the output transformer) is very high, around 8-10K ohms or thereabouts- a seemingly good impedance match to drive an electrostatic speaker.
So my question is, "Why can't we simply drive the electrostatic directly from the plate of the electron tube (through a decoupling capacitor) as the impedance at that point is already very high (7-10K Ohms) and bypass all the impedance matching transfomers downstream in the signal line?
1) We know that the output of a typical transistor power amplifier is low impedance (able to drive a low impedance device such as an 8 Ohm voice coil loudspeaker without the need for any impedance matching transformers).
2) We also know that the input of a typical electrostatic loudspeaker is very high (~10K Ohms), therefore a separate impedance-matching transformer is required to couple the low impedance output of the typical amplifier to the high impedance electrostatic.
3) Finally, we also know that the anode plate of a vacuum tube (the output point prior to the output transformer) is very high, around 8-10K ohms or thereabouts- a seemingly good impedance match to drive an electrostatic speaker.
So my question is, "Why can't we simply drive the electrostatic directly from the plate of the electron tube (through a decoupling capacitor) as the impedance at that point is already very high (7-10K Ohms) and bypass all the impedance matching transfomers downstream in the signal line?
Broskie explores this on his site...
Here's one of his posts on electrostatic tube headphone amplifiers.
Electrostatic Headphone Amplifiers
Here's one of his posts on electrostatic tube headphone amplifiers.
Electrostatic Headphone Amplifiers
Think more in terms of voltage.
An example:
A 50 watt @ 8 ohm (tube) amplifier will have 20 VRMS output (56 Vpk-pk).
Pk-pk plate-plate voltage will be around 1000 V, depending on type of amplifier and impedance/voltage ratio of the output transformer.
Electrostatics need high AC voltages.
The Quad ESL63 needs ~ 10000 Vpk-pk for full power (membrane bias voltage is ~5 kV).
The Quad has a 1:150 step up transformer, so some 67 Vpk-pk is required at the output of the amplifier.
Driving the Quad with a tube amplifier will still need a transformer.
The step up ratio however can be much lower, in the order of 1:6 to 1:10, depending on several factors.
Without any transformer you would need very high voltage tubes.
It has been done in the past, but mostly failed due to reliability problems.
An example:
A 50 watt @ 8 ohm (tube) amplifier will have 20 VRMS output (56 Vpk-pk).
Pk-pk plate-plate voltage will be around 1000 V, depending on type of amplifier and impedance/voltage ratio of the output transformer.
Electrostatics need high AC voltages.
The Quad ESL63 needs ~ 10000 Vpk-pk for full power (membrane bias voltage is ~5 kV).
The Quad has a 1:150 step up transformer, so some 67 Vpk-pk is required at the output of the amplifier.
Driving the Quad with a tube amplifier will still need a transformer.
The step up ratio however can be much lower, in the order of 1:6 to 1:10, depending on several factors.
Without any transformer you would need very high voltage tubes.
It has been done in the past, but mostly failed due to reliability problems.
Last edited:
I know that Beverage Audio did that in the 1970's with their model X-1, but I haven't looked into them lately.
Model X-1
The Model One was the first pair of speakers my father made utilizing his invention of an acoustic lens.
Each speaker measured six feet tall by three feet wide by some two feet deep. It was a single source, full range, wide dispersion cylindrical wavefront, monopole electrostatic loudspeaker. It was also an active speaker with its own dedicated integral OTL tube amplifier, developed by my father. The design's frequency range is unknown, but I believe it could go well into the 20 Hz range.
They performed beautifully in the very difficult room at our home on Franceschi Rd., on the hilltop overlooking Santa Barbara. I believe that it was this room that caused my father to conceive the idea of a vertical, wide dispersion cylindrical
wavefront as well as the concept of the acoustic lens necessary to accomplish such a wavefront. This room was about 60 feet long by 20 feet wide (ending in a small L-shaped extrusion). At each end there was a solid wall with no windows. What made this room so difficult to work with was the fact that along each of the long walls there was about 30-40 feet of floor-to-ceiling interlocking sliding glass doors. These glass walls were parallel to each other and had a polished concrete floor between them for most of their lengths. There were curtains, but they were almost always pulled back into the corners. Very few speaker systems, then or now, can work with such a difficult room, but the Model One handled it beautifully.
http://bevaudio.com/
Model X-1
The Model One was the first pair of speakers my father made utilizing his invention of an acoustic lens.
Each speaker measured six feet tall by three feet wide by some two feet deep. It was a single source, full range, wide dispersion cylindrical wavefront, monopole electrostatic loudspeaker. It was also an active speaker with its own dedicated integral OTL tube amplifier, developed by my father. The design's frequency range is unknown, but I believe it could go well into the 20 Hz range.
They performed beautifully in the very difficult room at our home on Franceschi Rd., on the hilltop overlooking Santa Barbara. I believe that it was this room that caused my father to conceive the idea of a vertical, wide dispersion cylindrical
wavefront as well as the concept of the acoustic lens necessary to accomplish such a wavefront. This room was about 60 feet long by 20 feet wide (ending in a small L-shaped extrusion). At each end there was a solid wall with no windows. What made this room so difficult to work with was the fact that along each of the long walls there was about 30-40 feet of floor-to-ceiling interlocking sliding glass doors. These glass walls were parallel to each other and had a polished concrete floor between them for most of their lengths. There were curtains, but they were almost always pulled back into the corners. Very few speaker systems, then or now, can work with such a difficult room, but the Model One handled it beautifully.http://bevaudio.com/
Attachments
Last edited:
daanve,
I am aware that some electrostatics require 10,000v, but that is just excessive and pretty much an arbitrary value. As a point of interest (and a bit off topic), I just bought a $2.95 electronic flyswatter from Harbor Freight, pulled the electronics of the handle and hooked it up to the electrostatic speaker as the polarizing DC supply. With only the 9v battery and cheap electronics, the flyswatter produced 400-600v and made one hell of a fine sounding 1' x 3' electrostatic speaker! Really surprising is that when I pulled the drain-down safety resistor off the circuit board, I found that the charge remained on the plates for a considerable amount of time without any need for a recharge. I could press the momentary "power button" on the flyswatter handle, charge the electrostatic diaphragm, and that charge would bring the speaker alive and last 5 FULL MINUTES without the need to recharge the plates again! I could listen to the electrostatic speaker for quite some time without any power supply or degredation to the sound. Apparently, there is negligible current drain from the plates and as long as you maintain a moderate parastitic voltage on the plate, sound can be quite good even over a large surface.
The question I raise on the original post does not have anything to do with the voltage on the electrostatic diaphragm, rather it relates to a more simple design that would allow a direct coupling to the amplifier plates in order to bypass the need for the impedance matching transformers in the AC signal line. I would be removing the DC using a decoupling cap so that I only have high impedance AC, and use an external harbor freight DC power supply anyway.
I am aware that some electrostatics require 10,000v, but that is just excessive and pretty much an arbitrary value. As a point of interest (and a bit off topic), I just bought a $2.95 electronic flyswatter from Harbor Freight, pulled the electronics of the handle and hooked it up to the electrostatic speaker as the polarizing DC supply. With only the 9v battery and cheap electronics, the flyswatter produced 400-600v and made one hell of a fine sounding 1' x 3' electrostatic speaker! Really surprising is that when I pulled the drain-down safety resistor off the circuit board, I found that the charge remained on the plates for a considerable amount of time without any need for a recharge. I could press the momentary "power button" on the flyswatter handle, charge the electrostatic diaphragm, and that charge would bring the speaker alive and last 5 FULL MINUTES without the need to recharge the plates again! I could listen to the electrostatic speaker for quite some time without any power supply or degredation to the sound. Apparently, there is negligible current drain from the plates and as long as you maintain a moderate parastitic voltage on the plate, sound can be quite good even over a large surface.
The question I raise on the original post does not have anything to do with the voltage on the electrostatic diaphragm, rather it relates to a more simple design that would allow a direct coupling to the amplifier plates in order to bypass the need for the impedance matching transformers in the AC signal line. I would be removing the DC using a decoupling cap so that I only have high impedance AC, and use an external harbor freight DC power supply anyway.
Find some very high voltage power triodes and use them to build an Aikido amplifier. You would need two Aikido amps for each ESL because there are two stators in one ESL. No output transformers, I guess that was the goal.
It's all physics.
To reach enough SPL you need high voltages.
To reach full bandwidth (bass!) you need enough excursion of the membrane.
Larger membrane - stator distance means high AC voltages to get enough electrostatic drive.
You can make an electrostatic work at lower voltages, but you will not get a grown-up loudspeaker that way.
The lower voltages are good for electrostatic headphones.
To reach enough SPL you need high voltages.
To reach full bandwidth (bass!) you need enough excursion of the membrane.
Larger membrane - stator distance means high AC voltages to get enough electrostatic drive.
You can make an electrostatic work at lower voltages, but you will not get a grown-up loudspeaker that way.
The lower voltages are good for electrostatic headphones.
Perhaps you could make an amplifier with these...
6S20S / 6C20C high voltage tube → Single triode → Tubes-Store.com
They have a plate voltage rating of 25kV and a gain of 2500!
6S20S / 6C20C high voltage tube → Single triode → Tubes-Store.com
They have a plate voltage rating of 25kV and a gain of 2500!
Long ago I was reading an old issue of The Audio Armature where they had a tube electrostatic amp project to drive the electrostatic speakers in an earlier project I believe it was using 3 or 4kv on the output tubes. These articles might have been by Sanders.
On a lesser note I heard about someone enjoying there Quads by simply running them from the plates of the output tubes in a St70.
On a lesser note I heard about someone enjoying there Quads by simply running them from the plates of the output tubes in a St70.
An electrostatic loudspeaker needs not only high audio voltage, but also some current, to deliver acoustic power. In addition, they are of capacitive nature, meaning that the voltage and current on the output tube is out of phase by 90°. But the main reason I would not use direct drive is that it is extremely dangerous. It is a real challenge to make a construction with proper isolation (and still acoustically transparent) that is absolutely safe.
Typically output transformers of tube amplifiers decrease voltage and impedance suitable to a usual loudspeaker load. But in case of ESL, then there is another transformer to increase voltage and impedance. Transformers also isolate circuits and prevent DC. But sufficient capacitors can also block DC.
A triode parafeed amplifier for ESL would be interesting. Especially the chokes would be extraordinary.
A triode parafeed amplifier for ESL would be interesting. Especially the chokes would be extraordinary.
The Acoustat model X is where I would start. I had a look at one some years back, the heater transformer that heated the upper triode in the output stage had pretty impressive insulation!
An old TV HV shunt regulator tube? Don’t those things create X-rays and require shielding?
The tube plate needs a DC path. This can be a choke (transformer primary was often used), a resistor (terrible power/energy efficiency), or another tube (probably working FAR above ground and thus awkward).
If you have a happy tube+transformer amp, connect to the primary (don't kill the cat). Efficiency is as good as it gets. But the L+C load (with only a hint of resistance) is a tough one.
Sanders did indeed use huge load resistors. If you could stand full voltage at 6kHz(?) the nominal efficiency is maybe 6%; at 100Hz more like 0.01%.
Totem-pole tubes are easy to sketch and tough to implement. If you have to ask, enjoy the trip.
The X-Rays are about the voltage, not the tube.
Triode parafeed amplifier with series connected chokes
I must add: Series connecting smaller chokes is also an option.
I must add: Series connecting smaller chokes is also an option.
This is a fun discussion. Thanks for the responses folks! Soon I will be trying out a few tricks suggested herein.
I was trying to get away from a transformer and connect directly through the plates, but perhaps I could compromise and use a 10,000 Ohm 1:1 center-tapped transformer (connected to the electrostatic device) and see how that works.
I bet using a single transformer would be an improvement from having to go the "usual" route from a 10 KOhm down to 8 Ohm transformer (the output stage of a typical tube amp) and then turn around and go back up from 8 Ohms to 10K Ohms again using another transformer... I don't know how much efficiency we all lose in this process, but I bet quite a bit, and also wonder that adverse effects we get by using two transformers become in terms of damping factor, transient and frequency response. Me thinks that a single 1:1 10K Ohm transformer would be a lot more efficient than two transformers with large winding ratio differences...but that is all supposition.
I was trying to get away from a transformer and connect directly through the plates, but perhaps I could compromise and use a 10,000 Ohm 1:1 center-tapped transformer (connected to the electrostatic device) and see how that works.
I bet using a single transformer would be an improvement from having to go the "usual" route from a 10 KOhm down to 8 Ohm transformer (the output stage of a typical tube amp) and then turn around and go back up from 8 Ohms to 10K Ohms again using another transformer... I don't know how much efficiency we all lose in this process, but I bet quite a bit, and also wonder that adverse effects we get by using two transformers become in terms of damping factor, transient and frequency response. Me thinks that a single 1:1 10K Ohm transformer would be a lot more efficient than two transformers with large winding ratio differences...but that is all supposition.
Last edited:
Did you look at the Acoustat amp I mentioned? It's exactly what you asked for and the service manual has the schematics.
Hi,
at purist we tested a KR 845 amp to drive my ESL panels directly by raising the plate supply to ~2kV and simply taking off the signal from the 845's anode, using the original transformer as anode choke load.
Worked very well, since my panels are highly efficient panels requiring only a couple of hundreds of volts to sing loud.
The transformer could still have been used to drive the mids and lows.
The latter trick was not uncommon in old tube radios featuring se ESL tweeters afair.
jauu
Calvin
at purist we tested a KR 845 amp to drive my ESL panels directly by raising the plate supply to ~2kV and simply taking off the signal from the 845's anode, using the original transformer as anode choke load.
Worked very well, since my panels are highly efficient panels requiring only a couple of hundreds of volts to sing loud.
The transformer could still have been used to drive the mids and lows.
The latter trick was not uncommon in old tube radios featuring se ESL tweeters afair.
jauu
Calvin
Audiowize...
I looked up the Acoustat X as you suggested. Looks like the exact answer to the question I was seeking. Direct drive tube system driving an electrostatic. Very nice device in both theory and execution. Thanks for the heads-up.
I still think it might be interesting to see how my suggestion of a single transformer + an external DC power supply (my post above) would stack up against a direct drive. Me thinks it might be an easy design with only a small compromise in performance...
I looked up the Acoustat X as you suggested. Looks like the exact answer to the question I was seeking. Direct drive tube system driving an electrostatic. Very nice device in both theory and execution. Thanks for the heads-up.
I still think it might be interesting to see how my suggestion of a single transformer + an external DC power supply (my post above) would stack up against a direct drive. Me thinks it might be an easy design with only a small compromise in performance...