Building better High sensitivity headphones

Here's a page with info using piezoelectric elements, specifically Kyocera KBT-33-RB-2CN. Not a lot of info on the build but does seem to say these would be quite a bit more sensitive than even the best sound powered balanced-armature headphones.

and the piezo element page. MikeK

The Kyocera data book is

KBT-33-RB-2CN data is found on page 20. It's about 35mm in diameter.

The impedance is capacitive, stated as 3k at 1 kHz, and less at higher frequencies, being about 1 k at 3 kHz.

Where is the 107dB SPL measured? Perhaps not at a distance of 1m. Perhaps at the transducer surface?

I have not seen reports of crystal receiver rectifier (CR) output voltages... perhaps millivolts?

It's still practical to test CR circuits with the radiation from local AM broadcast stations.

My last venture with a CR such worked OK on WOSU AM radiation, that transmitter antenna being about two miles east of me. [I used an old 2k-ohm magnetic headphone set, which also produces enough sound from my old B&K

2203 SLM ac-out port signal to show me whether that SLM was working OK in the field...]

Ange

ALL of those devices have narrowband resonances all over the place. The response looks like a hedgehog. And they have to be driven with a high-Z source or it gets worse.

Probably, if the transducer is not necessarily intended to couple into air anyway.

So what's wrong with using a JFET and a low-efficiency moving coil driver? It's hard to get efficiency and flat response at the same time.

--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Typically the output would be 86 dB at three feet 1 watt.. Add a horn and add

9-10 dB. They are driving with equivalent 8 ohm .125 watt or 1 VRMS and who knows the distance. or complete seal. 1 VRMS draws .00036 watt at 1 kHz roughly. At 300 Hz very efficient headphone drawing .0001 watt.

I guess its not efficient. The orginal post I think is about crystal radios.

greg

True.

BUT the critical point is to pay attention to the part of the audio spectrum that purveys the maximum speech intelligibility. We find that it is in the vicinity of 1,000-2,000 Hz. (The exact metric is to be found somewhere in ANSI lore where the contribution to speech intelligibility of each 1/3 octave band from maybe 100 Hz to 10kHz is catalogued).

Most practical audio piezoelectric resonators show a resonance in the

2,000-3,000 Hz range, which suits... more or less.

The original telephone systems resonances are to be found in the 1500 Hz area, which is a crude balance between the the chief frequency range of the sound of the vowels (500-800Hz) and the consonants (1,000-3,000Hz). Kryter (also a ham), in a 1960's paper showed that if the dedicated speech purveyance bandwidth had to be only 500 Hz, that the best place was that it be centered at 1500 Hz (old telephones). If two 500 Hz bands were available, then they should be centered at about 1,000 Hz and 2,000 Hz respectively. If three such 500 Hz bands were available, he suggested about 800 Hz, 1500 Hz and, I think 2500 Hz as the center frequencies. Such info was handy to design crystal SSB filters in days-gone-by. So I think we should use these criteria for selecting earpieces for CR...

Absolutely, but that may require an external power supply, unless the DC from the crystal rectifier is sufficient to feed the drain (or source) of a JFET....

Ange

Yes, headphones for crystal radios. The object is to use no external power, then hear as many stations as possible with a crystal radio. So we want the most sensitive headphones we can get, also want high impeadance to prevent loading the resonant circuit. There are those that receive a strong signal and use that energy to power an amplifier. The amp is then used to amplify the audio from a weaker station. MikeK

I wonder what the highest Z might be. I think 600 ohm nowdays. I don't know if it would drive a transformer. ??

Yes, a transformer is often put at the detector output and this drives the headphones. I see the datasheet for the KBT-33-RB-2CN says 2.8 kohms at 1 khz. These are often put two in series (one for each ear) for 5.8 kohms. MikeK

That's true... the voice intelligibility from those things is actually pretty good... if anything you could argue that some of the clanginess might even improve the ability to make words out.

Maybe, unless your goal is the highest possible fidelity with that Canadian big band station on 860 KC!

So this is cheating? Even if I use a lemon and some zinc and copper plates for power?

--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

Well we'd have to define two classes; one with air RF only; the other using a single 1.5v cell (AA, AAA, etc)

Ange

"Angelo Campanella" wrote in news:i8hse4$an7$ snipped-for-privacy@news.eternal-september.org:

Neither. The KBT-33-RB-2CN transducer is designed for use as a telephone headset receiver. The specified response is measured by a microphone in a IEC-318 coupler.

Hey answerman, I noted that test on the spec. page and the coupler used in the measurement. Now I wonder, to get the good output do I need there specific coupler to mount the transducer and then mount that assembly into the headset?

And if that is so, Why don't the sell it installed in the special designed loading coupler? I haven't found a seller that has them in stock with a price, only RFQ pages. MikeK

"amdx" wrote in news:6415d$4cafc194$18ec6dd7$ snipped-for-privacy@KNOLOGY.NET:

No. You install the transducer in a headset and measure the response of the headset on an IEC-318 coupler.

The IEC-318 coupler is a device that simulates the average acoustic characteristis of the human ear. The microphone in the IEC-318 coupler measures the pressure that is expected to exist at the eardrum when the transucer is tightly coupled to a real ear. In order to get the specified response and acoustic output, the transducer needs to be coupled tightly to either the ear simulator or to a real ear.

Because everyone has one. It's your ear.

The IEC-318 coupler is a design/measurement tool. It's very expensive.

MikeK

"amdx" wrote in news:4a781$4ca92512$18ec6dd7$ snipped-for-privacy@KNOLOGY.NET:

If you want to build a high sensitivity, high acoustic output headphone, a piezoelectric transducer isn't the way to do it. The sensitivity of a piezoelectric transducer is poor compared to that of a dynamic transducer, and distortion is high, especially at high levels. A high sensitivity dynamic transducer such as that which is used in the Sennheiser HDA-200 audiometric headphone is what is needed. The HDA-200 produces 114dB SPL at the eardrum at 1kHz for 25mW electrical input (1 Vrms). At maximum 500mW input, the acoustic output would be 127dB SPL. If you want even higher sensitivity and associated acoustic output, there's the VeriPro headphone that is used for testing the real-ear attenuation of insert hearing protectors. The VeriPro headphone is capable of producing 120-125 dB SPL for 25mW electrical input (1Vrms), At maximum 500mW input, the acoustic output of the VeriPro is 133-138dB SPL as meeasured in an IEC-318 coupler. That's some serious acoustic output.

We're looking to get maximum sound from a sliver of driving energy. The output from a crystal radio can be down in the microwatts, and maybe even nanowatts if I'm understanding this page;

Which can be found as a link here.

This is the Ben Tongue from Blonder and Tongue Labs, he did quite a lot of fine research on crystal radio design. MikeK

They don't care about high acoustic output. They care about high efficiency but they _mostly_ care about highest possible load impedance.

That's nice, but now you need a substantial step-down in front of it, and the consequent transformer losses. On top of which it's really hard to make a transformer with much higher than 50k input impedance and it would be _nice_ to have an order of magnitude higher.

--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

"amdx" wrote in news:7efee$4cb125ed$18ec6dd7$ snipped-for-privacy@KNOLOGY.NET:

Then YOU need to be precise in defining your terms. If YOU define transducer sensitivy as acoustic output per applied unit power, as opposed to acoustic output per applied volt, then a piezo transducer will almost always win because, except at its resonant frequency, its impedance is almost purely reactive (capacitive) and virtually NO power is consumed.

Thank you so very much for setting me straight.