Radical: Characterized by independence of, or departure from, what is usual or traditional; progressive, unorthodox, or revolutionary (in outlook, conception, design, etc.).

Oxford English Dictionary

Headphones for Radicals
Since I can trust my long-time readers to have reread my Post 604 in anticipation of this post, I will deliver only the briefest recap of my purposed radical recreation of the electrostatic headphone. Let's start with what is scary but not audio related, a concealed-carry handgun tucked into the front of our pants, 9mm or 45 ACP, loaded and cocked, safety off, equipped with a light, nimble trigger, the barrel alarmingly aimed squarely at a man's procreative appendages. Which is the greater potential threat, an armed assailant or a miscalculated misfire?

Another worrying thought: our head placed between two high-voltage plates of metal, our ears just millimeters away from boisterous AC voltages, voltages that swing into the hundreds of volts. Sure, that's what I want to do. No thanks. Yet, that is what I do each time I place my Stax electrostatic headphones upon my head—and I am not alone.

My alternative electrostatic headphone grounds the stators (a perforated thin sheet of metal that sound travels through) nearest the ear, while energetically driving the diaphragms and opposing stators.

As far as the diaphragm is concerned, both arrangements are identical, in terms of the voltage relationships between the diaphragm and its stators, 850V and 350V. At idle, i.e. no music, both electrostatic headphone designs see the diaphragm experiencing a bias voltage of +600V relative to the stators. As the music plays, it is the varying voltage relationships between diaphragm and stators that make the diaphragm move back and forth. Given the same voltage relationships between the diaphragm and its stators, we get the same sonic output.

My Post 642 revealed how smitten I am with AnTek's 15W toroid push-pull output transformer, the MP-15W50 15W, as it not only delivered wide bandwidth but also includes purposely-made airgaps in it construction, making the transformer less susceptible to core saturation due to unidirectional current imbalance in its primary. This post also showed how a conventional electrostatic headphone could be driven by a single-ended amplifier that used the push-pull output transformer to increase the voltage swings.

The secret sauce here is that the transformer core never experiences a net magnetizing DC current flow through the primary and secondary, as the two DC current flows, being in opposition to each other, cancel.

Well, if we transpose the MOSFET and the ground connection, we can drive my radical electrostatic headphones.

Let's start by noticing that the MOSFET's B+ voltage has increased, but its idle current has decreased. The MOSFET's source must undergo much larger voltage swings, but with smaller current swings. What is going on here? When the ultra-linear tap was driven, only 20% of the primary was directly driven. When the center-tap is driven, however, 50% of the primary is directly powered, making for a ratio of 1 to 2.5, so the idle current can be reduced to 25mA/2.5 or 10mA. In contrast, the driving the primary's center-tap requires the MOSFET's output voltage to swing 2.5 times higher, so the MOSFET's B+ voltage must be increased by 2.5 times, i.e. 275Vdc.

Note that ultra-linear taps are not needed in my radical redesign, just a center-tap, which all push-pull output transformers hold (some single-ended output transformers hold an ultra-linear tap, which sometimes is placed at 50% of the primary winding, so these single-ended output transformers could be used in this circuit). The MOSFET experiences voltage swings both positive and negative, so the peak source-to-drain voltage can approach (or indeed exceed) its B+ voltage; therefore, I would use a 700V to 1kV N-channel MOSFET, possibly a SiC type. On the other hand, we could replace the MOSFET with a tube, either a pentode or triode. (Using a pentode could prove tricky, as a floating power supply would be needed to power the screen.)

One possible triode is the 12B4 single-triode tube, which is a low-mu triode designed to work with extra-high voltages.

Its maximum peak-positive pulse voltage is 1kV. It will, however, require a floating power supply for its heater; one per channel.

The B+ voltage need only be upped to +300Vdc, as the triode can draw 20mA with a cathode-to-grid voltage of 0V and cathode-to-plate voltage of 28V. In other words, we should have no problem getting 250V voltage swings from the 12B4, although the 12B4 requires a huge input signal. We can also drive the center-tap with a MOSFET's drain.

Note that the ear-side of the electrostatic headphone is grounded, as is the primary's top termination. By the way, we can invert the transformer, but we must also invert the 24V power supply's polarity.

Which arrangement is better? I prefer the former, as the negative 24V power supply can also power OpAmps.

The OpAmp's noninverting input sees a bias voltage of -0.9Vdc, which is used to auto-bias the MOSFET's idle current flow, as the top half of the 180-ohm primary (i.e. 90 ohms) must experience the 0.9 volt voltage drop, making for a current flow of 0.9/90 or 0.01 or 10mA. One fear I have is that the OpAmp might not like having its inputs so close to it positive rail voltage. One workaround is to impose an additional fixed voltage drop.

The inclusion of the 1N5401 rectifier buys us almost 0.7V more voltage drop. Of course, we could use a zener of voltage reference IC to make an even greater voltage drop, but at the cost of small bias voltage on the not-grounded stator. Okay, it's time to put on our thinking caps. We have the problem with many OpAmps not being happy with input signals near their positive power-supply rail voltage. We have an asset in the 24V power supply that provides the countervailing current flow through the secondary. Who says we cannot use a ± 12Vdc bipolar power supply instead? No one.

The secondary sees the current flow at idle, the OpAmps is happy with input signals near the center of its bipolar power supply voltages (i.e. 0V, making its input stage and output stage happy), and we no longer need to add rectifiers or zener diodes. Ideally, the 332k voltage divider resistor should be 333k, but 332k is dang close. (Of course, I assume a fully regulated bipolar power supply.) No doubt, some tweaking of the negative-feedback loop might be required and finding a high-voltage PNP transistor is difficult, but not impossible. Note the -580Vdc bias voltage for the diaphragm. This is a feature, not a bug, which my last post explains.

I rank this design worthy of half a Musk Prize Medal, but then—every mother's child is beautiful.

12DW7-Based Harmonic-Restoration Circuit
In a recent email exchange, I recommended using a 12DW7/7247/ECC832 dissimilar-triode tube in a harmonic-restoration circuit, a circuit that purposely adds a good-amount of single-ended-like cascade of harmonics to an input signal at its output, thereby softening and warming up otherwise typical brittle and sterile solid-state-produced signals. Another goal was to strive for a design as simple as possible and not to use any internal coupling capacitors. As I saw it, the 12DW7 would prove an ideal choice. The basic idea was that the 12AX7 input triode, configured as a grounded-cathode amplifier, develops a lot of gain, which a following two-resistor voltage divider throws away, leaving only a unity-gain and harmonically-enriched output signal from a 12AU7-based cathode follower.

The 200k and 7.5k resistors form the two-resistor voltage divider. The LM317 functions as a constant-current source that sets the idle current flow for the 12AU7 triode.

In SPICE simulations, this circuit delivered a healthy amount of harmonic-restoration to a 1Vpk output signal.

I overshot my goal of 1% distortion by only a tad. With an input signal of only 0.1Vpk, the distortion falls near to 0.1%, but still exhibits the single-ended cascade of harmonics. What I didn't like, however, was the relatively poor PSRR of only -22dB. Aikido Mojo to the rescue:

The constant-current source is no longer perfectly constant, as we have injected a small sampling of the power-supply noise into its adjustment pin, which induces an anti-phase current variation that nulls the power-supply noise at the cathode follower's output. Here is the SPICE-generated PSRR graph:

A nearly fiftyfold improvement. Nice! Mind you, this PSRR figure is only the base PSRR, which does not include the further PSRR-enhancements wrought from RC filters and LC filters or high-voltage regulation. Of course, the same topology can accept other tube types, such as the 6DJ8, 6SN7 and 12AU7… and many more. The part values and B+ voltage will differ, but the design concept remains the same: generate signal gain with a triode, attenuate, and buffer. By the way in this design, the LM317-HV is not required, but I could be with other tubes and different B+ voltages.

I should point out that this harmonic-restoration circuit inverts the input signal's phase at its output. Is this a big deal? No, as you can simply flip the phase of the loudspeaker cable's connection to the speaker terminals. Of course, some audiophiles will still be bothered by the inversion. (If rational behavior were a strict requirement to be an audiophile, the number of audiophiles would be even more miniscule than it is now.) For such audio folk, here is a non-inverting harmonic-restoration circuit:

The input 12AU7 triode is configured as a grounded-cathode amplifier that gets its input signal at its cathode. It does not invert the input signal at its plate. The cathode follower output stage also preserves the phase of the input signal. The assumption here is that the circuit would follow some modern solid-state audio signal source, such as a CD player, tuner, DAC, or streamer, whose solid-state output stage can easily drive the 4.7k load resistance (almost all can). The signal-flow layout is that the input signal is attenuated by the 4.7k and 620-ohm resistors, and then amplified by the grounded-cathode amplifier and buffered by the cathode follower.

To improve the PSRR, some Aikido Mojo has been included via the 169k and 10k two-resistor voltage divider, which delivers just enough of the power-supply noise to the grounded-cathode amplifier's grid to provoke a ripple null at the circuit's output. The 1N4007 rectifier is a safety device that only engages at startup, when the triodes are still cold and not conducting. Once the triodes are warm, the rectifier falls out of the circuit, as it is no longer forward biased. The input coupling capacitor shown is a nonpolarized electrolytic capacitor, but a film or PIO capacitor could be used instead.

Although this harmonic-restoration circuit does produce some 2nd harmonic distortion, its quantity is low, as the THD approaches 0.01%, with 1Vpk of output, in SPICE simulations. The PSRR at 100Hz came in at a respectable -54dB. Using carbon resistors and PIO coupling capacitors might be enough to further flavor the output signal.

Disposable Suppressor Inserts
As I value my remaining hearing—eagerly—I await the end of legal restrictions on gun suppressors (aka silencers) here in the USA. I look forward to the day when we can as easily buy one at a sporting-goods store (or Amazon) as we can buy a scope or a rifle travel case; no government paperwork, no finger prints or background checks. In addition, I long to be able to design and to make my own suppressor—legally.

In Post 471, I revealed one design idea I had, which made use of a only a metal lathe and standard metal piping.

Basically, a suppressor is tubular enclosure with two openings and internal baffling. They tend to be insanely expensive, often costing more than the gun they intend to quiet. Why this is so is due, in part, to government regulations and restrictions, and the restraints and inefficiencies of small production. More importantly, I believe most suppressors are excessively overbuilt, often using expensive titanium or advanced 3D printing or elaborate machining construction. Why? One reason is that they are usually designed to survive automatic fire of serious ammo. Moreover, like sports cars and racing boats, suppressors are designed by men, the sort of men who naturally incline to showing off and who strive for excellence.

The reality, in contrast, is that few guns are ever used. (A friend has owned his revolver for decades, but never has fired a shot from it. He is not alone in this.) Those few guns that are used are only rarely used, say once a year, and thereby expending only one box of 20 or 50 rounds. These guns do no need a military-grade suppressor. What they do need is a good, but cheap suppressor, preferably one that accepts replaceable inserts, which might only survive 500 rounds—in other words, survive a decade's worth of casual shooting. The insert could be made from porous lava stone or a coil of copper screen rolled into a tube and welded tight. One thought I had was that of a variation on the metal-foil Virvateller expandable garlands, but made from more solid aluminum.

Same principle, a long narrow strip of metal is crinkled and rotated into a disk, but with a center hole drilled through to allow the bullet free travel. We could buy a pack with four disks for let's say $40; we would begin by pulling the disk's ends apart to create a spiral, a spiral slightly longer than the suppressor tube's inner length, which we would insert into the suppressor tube, and screw the end cap in place, replacing the insert as needed. The outer tube could be made from titanium or hardened steel.

John, don't seem to realize that suppressors are used to kill people?

Really? Have they been fired from cannon, dropped from a skyscraper?

No, but they are used with guns.

Okay, by the same logic, scopes, tripods, slings, red-dot sights, gun oil, and ear muffs are also used to kill people, yet we do not outlaw them or restrict their manufacture. Do you think we should? If so, what about the car that was used to bring the gun?

Back in the 1960s, Frankford Arsenal developed a cheap, disposable suppressor:

Caliber . 22 Hi-Standard Pistol / FA Silencer

The experimental silencer shown in Figures 15 through 17 was designed by two Frankford Arsenal employees in 1967. It evolved concurrently with the availability of low cost porous metal machining stock. The porous metal manufacturing techniques, which only recently were refined, consist of casting the molten metal over a salt configuration and dissolving the salt after the metal hardens. Presently, a number of metals can be cast into almost any porosity, density, or shape. The silencer described herein is probably a fair representative of its type.

The caliber .22 Frankford Arsenal silencer vs tested with the same Hi-standard pistol used for evaluation of the French silencer. It is all-aluminum, and measured 1.4 inches in diameter and 6.5 inches in length. The silencer is machined from stock which is partially solid and partially porous. Outside, the porous section of the silencer is wrapped with electrical tape, which limits the propellant gas discharge to only the 0.23 in. diameter projectile exit at the silencer muzzle. Inside, the silencer consists of three chambers of different lengths and diameters. The outstanding characteristics of the caliber .22 Frankford Arsenal silencers are small weight, low manufacturing cost, and relatively quiet acoustical performance, The undesirable features of the silencer are its bulkiness and high erosion rate.

From page 31 of the Frankford Arsenal Report R-1896

Music Recommendation:James Blake's James Blake
Having seen James Blake's eponymously-named album on lists of stereo-demo-worthy albums before, I tried giving it a listen a few years ago. Sadly, I only got through a minute or two of the first track. Well, recently I saw that its sixth track, "Limit To Your Love," was the one to hear.

Wow!

I didn't expect that. I won't ruin the surprise, so you must stream it and find out what all the buzz is about.

//JRB

Did you enjoy my post? Do you want to see me make it to post 1,000? If so, think about supporting me at Patreon.

AI Summary
Adobe's AI's report on this post:

This document discusses innovative designs and modifications for electrostatic headphone amplifiers, focusing on push-pull transformer configurations and power supply considerations.

Innovative Electrostatic Headphone Designs
A new approach to electrostatic headphones aims to improve safety and performance by reconfiguring the stator and diaphragm arrangements. 

• Traditional electrostatic headphones have high-voltage plates near the ears, posing safety concerns. 
• The proposed design grounds the stators nearest the ear, with the diaphragm driven by opposing voltages. 
• Voltage relationships remain the same (850V and 350V), with a bias of +600V at idle. 
• Sonic output depends on voltage variations between diaphragm and stators, which are maintained in the new design. 
• The redesign enhances safety by reducing high-voltage exposure near the ears.

Push-Pull Transformer-Driven Headphone Amplifiers
A novel single-ended electrostatic headphone amplifier uses a push-pull transformer to increase voltage swings safely. 

• Utilizes AnTek's MP-15W50 toroid transformer with airgaps to prevent core saturation. 
• The transformer core cancels net magnetizing DC current, enabling higher voltage swings.
• The circuit can be driven by either tubes or high-voltage MOSFETs, with a preference for SiC MOSFETs (700V-1kV). 
• Alternative tube options include the 12B4 triode, capable of handling peak voltages up to 1kV. 
• The design increases B+ voltage to +580V or +300V depending on the tube or transistor used.
• The circuit features a grounded ear-side and can invert transformer connections for different configurations. 
• The design emphasizes safety and efficiency, with a focus on reducing voltage and current swings for safer operation.

Advanced Harmonic-Restoration Circuits
A tube-based harmonic-restoration circuit enriches audio signals by adding controlled harmonics, warming sound quality. 

• Uses a 12DW7/7247/ECC832 dissimilar-triode tube combination for simplicity and effectiveness. 
• The circuit employs a grounded-cathode amplifier and a cathode follower for gain and buffering. 
• Achieves 0.1% distortion at 1Vpk, with a PSRR of -22dB, improved to -52dB with Aikido Mojo. 
• The design can accept various tubes like 6DJ8, 6SN7, and 12AU7, with different B+ voltages. 
• The circuit inverts the input phase; phase reversal can be corrected by cable wiring. 
• A non-inverting version maintains phase, suitable for modern solid-state sources.
• The circuit offers significant noise rejection and low distortion, making it suitable for high-fidelity audio restoration.

Disposable Suppressor Insert Concepts
Proposes a low-cost, disposable gun suppressor design using porous metal and simple materials. 

• Suppressors are often overbuilt for military-grade durability, which increases cost. 
• Many guns are rarely used, so a cheaper, replaceable suppressor suffices for casual shooting. 
• Design involves a tubular enclosure with internal baffling and replaceable inserts made from porous lava stone or copper screen. 
• The insert is a crinkled metal strip rolled into a spiral, inserted into a titanium or steel outer tube.
• Aims to produce a lightweight, inexpensive suppressor costing around $40 for a pack of four disks. 
• References a 1967 Frankford Arsenal low-cost, porous metal silencer made from aluminum, with small size and high erosion rate. 
• Highlights the potential for affordable, disposable suppressors to reduce costs and ease of manufacturing.

Music Recommendation: James Blake's Album
Recommends listening to James Blake's self-titled album, especially the track "Limit To Your Love," for high-quality stereo demonstration. 

• The album is known for its stereo-demo-worthy sound quality.
• The reviewer initially struggled with the album but found the sixth track compelling.

User Guides for GlassWare Software
Just click on any of the above images to download a PDF of the user guides.