How about some very robust high voltage electrostatic elements.
Then power up a pair of 833 thoriated filament transmitting triodes per channel, some very high voltage chokes, and B+ of up to 3000V.
Complex, dangerous, but it might just sound real good.
One issue with either choke or transformers and electrostatics, is the resonance of the electrostatic capacitance and the amplifier's chosen transformer or choke inductance.
Then power up a pair of 833 thoriated filament transmitting triodes per channel, some very high voltage chokes, and B+ of up to 3000V.
Complex, dangerous, but it might just sound real good.
One issue with either choke or transformers and electrostatics, is the resonance of the electrostatic capacitance and the amplifier's chosen transformer or choke inductance.
Last edited:
Winding a transformer
Winding a 1:1 transformer would be simple using a bifilar winding. Inductive coupling would be near perfection. Parasitic capacitance is one thing to worry about. It depends on how you fill the bobbin. There will be parasitic capacitances series connected. If you had many small capacitances in series that would be better than small number of big capacitances series connected. Multi section bobbins may help. Wind one section full, then continue to the next, etc.
Probably you need 1:1:1 winding if you are going to use a push-pull amplifier.
But just think how simple and elegant it would be with one triode and one 1:1 transformer.
Winding a 1:1 transformer would be simple using a bifilar winding. Inductive coupling would be near perfection. Parasitic capacitance is one thing to worry about. It depends on how you fill the bobbin. There will be parasitic capacitances series connected. If you had many small capacitances in series that would be better than small number of big capacitances series connected. Multi section bobbins may help. Wind one section full, then continue to the next, etc.
Probably you need 1:1:1 winding if you are going to use a push-pull amplifier.
But just think how simple and elegant it would be with one triode and one 1:1 transformer.
Jack Elliano of Electra Print Audio, wound me a special pair of 1:1 interstage transformers for use with ECC99 drivers, and 300B output tubes.
I measured the frequency and phase response of those interstage transformers.
I used a $50,000 Vector Network Analyzer 10Hz to 4GHz) with special optional input and output measurement ports, $10,000 Precision Call Kit, speacial measurement software, through measurements, and a special DC to 4 GHz return loss bridge.
The interstage gave excellent performance from 20Hz all the way to several hundred kHz.
A 1:1:1 output transformer should be able to work quite well, but the voltages involved make insulation and wire spacing something to solve.
My guess is that perhaps a 1:1 + 1:1 might work as good or better.
I built a special case "push pull" amp out of 2 single ended amplifiers, driving the inputs differentially, and connected the output secondaries in parallel, but reversed common and 8 taps (in anti-phase). That worked, because each se amp was opposite phase of the other amp (differential signal of 1 versus 2 inputs) There was very good coupling of each primary to secondary, low leakage inductance, so the parallel secondary connection provided good 2nd order cancellation, similar to push pull.
I measured the frequency and phase response of those interstage transformers.
I used a $50,000 Vector Network Analyzer 10Hz to 4GHz) with special optional input and output measurement ports, $10,000 Precision Call Kit, speacial measurement software, through measurements, and a special DC to 4 GHz return loss bridge.
The interstage gave excellent performance from 20Hz all the way to several hundred kHz.
A 1:1:1 output transformer should be able to work quite well, but the voltages involved make insulation and wire spacing something to solve.
My guess is that perhaps a 1:1 + 1:1 might work as good or better.
I built a special case "push pull" amp out of 2 single ended amplifiers, driving the inputs differentially, and connected the output secondaries in parallel, but reversed common and 8 taps (in anti-phase). That worked, because each se amp was opposite phase of the other amp (differential signal of 1 versus 2 inputs) There was very good coupling of each primary to secondary, low leakage inductance, so the parallel secondary connection provided good 2nd order cancellation, similar to push pull.
Last edited:
I once built a direct-drive amplifier for a DIY electrostatic speaker built by a then colleague. With its 2660 V supply capable of supplying more than enough current to cause cardiac problems, it was by far the most dangerous circuit I ever built. Its peak output voltage was 2 kV, which turned out to be insufficient to play anything but background music. See https://home.kpn.nl/verwa255/esl/amp.zip and Elektrostatic Loudspeakers
I know Lampie519 from this forum has a more efficient and more successful direct-drive amplifier. He uses SRPP output stages and a custom-wound transformer for the high-voltage supply. That's far more efficient (and even more dangerous) than what I did.
I know Lampie519 from this forum has a more efficient and more successful direct-drive amplifier. He uses SRPP output stages and a custom-wound transformer for the high-voltage supply. That's far more efficient (and even more dangerous) than what I did.
High voltage rock 'n' roll
With these voltages, I wonder, would a triple insulated wire (Furukawa TEX-E or some other comparable) used on both primary and secondary be safe enough?
If the power tube was used as a cathode follower, then the transformer primary winding could be connected to earth, making the high voltage issue somewhat easier. Both primary and secondary would swing around the same potential.
With these voltages, I wonder, would a triple insulated wire (Furukawa TEX-E or some other comparable) used on both primary and secondary be safe enough?
If the power tube was used as a cathode follower, then the transformer primary winding could be connected to earth, making the high voltage issue somewhat easier. Both primary and secondary would swing around the same potential.
No.
Inductive coupling is through the transformer core; bifilar winding does not change that.
Parasitic capacitance is typically not something to worry about when winding bifilar; on the other hand: it's your friend!
Good high frequency bandwidth is possible because of capacitive coupling which bifilar winding provides. But it only works that well with 1:1 ratio.
Bumped into this thread looking for tube amp for STAX headphones.
I believe there is an easy solution, at least for earlier models, like SR-3, that have higher sensitivity. SR-3 is 107 dB at 100 VRMS.
A regular push-pull 45 amplifier biased to AB1 (Ua=300 V, Ia=20 mA per tube) has peak output of 150 VRMS, if unloaded. STAX stators connected directly to plates. Without speaker at the secondary, output transformer acts as plate choke, and the whole PP stage as choke-loaded voltage amplifier.
This wouldn't require any modification of the amp, which, with headphone unplugged, can be used with its intended speakers.
Anything wrong with this?
I believe there is an easy solution, at least for earlier models, like SR-3, that have higher sensitivity. SR-3 is 107 dB at 100 VRMS.
A regular push-pull 45 amplifier biased to AB1 (Ua=300 V, Ia=20 mA per tube) has peak output of 150 VRMS, if unloaded. STAX stators connected directly to plates. Without speaker at the secondary, output transformer acts as plate choke, and the whole PP stage as choke-loaded voltage amplifier.
This wouldn't require any modification of the amp, which, with headphone unplugged, can be used with its intended speakers.
Anything wrong with this?
sser 2,
I use exacly this way of doing it in my STAX amp I have been using for some 15 years or so. I have some signal relays in there so that I can just change between speakers and headphones with a switch on the front. Works well. Better to bias the cathode near -300 and have a low voltage at the anodes. For the HV bias to the headphone you can use a series resistor in the megaohm range. In mine I use 6B4G DHTs as output tubes.
For only the headphone, a tube like ECC99 or 6N6p will be more than enough to sound good and play loud.
I use exacly this way of doing it in my STAX amp I have been using for some 15 years or so. I have some signal relays in there so that I can just change between speakers and headphones with a switch on the front. Works well. Better to bias the cathode near -300 and have a low voltage at the anodes. For the HV bias to the headphone you can use a series resistor in the megaohm range. In mine I use 6B4G DHTs as output tubes.
For only the headphone, a tube like ECC99 or 6N6p will be more than enough to sound good and play loud.
Many years ago I built a pair of small , slim HF ESL panels . Ultralinear push pull output transformers were used as step-up autoformers (secondary disconnected) with the stators taken from the anode taps and the UL taps driven by push pull 7236 . I ran the diaphragm with a negative polarising supply . Not direct drive and potentially lethal with the stators being at HT but a simple , elegant solution if you don't need to swing tens of kilovolts and have suitable transformers in your parts stash
cheers
316a
cheers
316a