[rgonzale]

I just bought a pair of Stax SR-80 electret headphones along with the SRD-4 energizer (SR-84 system) and would like to avoid using them with a full-size (loudspeaker) amplifier if I can avoid it, since I have a couple of decent headphone amps. (Creek OBH-1 and Audio Alchemy class A headphone amp.)

EDIT: some incorrect assumptions here, which are addressed later in this thread

Apparently the SRD-4 unit provides a polarizing voltage via the 5th cable, but it's secondary purpose is to step up the voltage provided by an amplifier to that needed by the high impedance (150 kOhm) driver itself. I believe it basically consists of a transformer with 1 ohm input impedance with a series resistor to pad the input impedance up to 8 ohm for conventional amplifiers -- and this padding allows you to use the volume setting in a normal range without sending too much power to the drivers. Back of the envelope calculation: amplifier delivers 1 watt into 8 ohms with 2.8v signal (P=(V^2)/R). After padding, the transformer sees 0.35v. If the turns ratio is 400:1 (just a guess!), the Stax electret driver will get 140v, which will produce 130mW power and should be plenty loud. (The rated sensitivity is 95dB/100v).

Most headphone amps will of course have trouble delivering a watt into 8 ohms. If you remove the padding resistor, the amp only has to deliver 130mW but is now driving a 1-ohm load, which is very different from the usual 30-300 ohm headphones in terms of damping factor and current requirement.

But it seems easy enough to replace or modify the transformer to have a lower turns ratio. If we reduce the turns ratio by a factor of 20, the headphone amp will see a 20 ohm load and should be able to supply the needed voltage and power.

Does anyone else have insight into this? I'm curious why Stax designed their energizer units for speaker-level signals instead of headphone-level signals in the first place?

 

[labrat]

The Stax SRD-4 energizer is self-biasing, has no power from the grid to set up the bias voltage.
A headphone amplifier does not output the power needed to feed enough for the bias voltage to build up.
A power amplifier does have enough power, but you don't need a big amplifier for that.
Get something like a Dayton Audio class D amplifier , or anything like it.

https://www.parts-express.com/dayto...attery-powered-mini-amplifier-15-wpc--300-380

 

[rgonzale]

Thanks a lot for your clarification @labrat , but I'm still confused.

The SRD-4 circuit is simply: input --> series resistor (actually a thermistor) --> transformer --> output. There is no connection to the bias pin on the output connector (sorry I miss-stated that at the top of this thread). In fact my SR-80 headphones only use a 4-conductor cable, so the self-bias apparently relies on the permanent charge in the electret elements.

 

[labrat]

Sorry that I did not read all from the start, not noticing that these are Electret, not Electrostatic headphones.!
Electret have no bias, but as you say a permanent "bias" set up by magnets.
They still need more input power than normally supplied from a headphone output connector.
You will note a low signal if you try connecting your headphone amp, maybe too low to listen to.
But no damage.

 

[rgonzale]

Thanks again. My fault with the inaccurate comments about polarizing voltage at the top of the thread.

Incidentally I have 2 nice headphone amps, each of which is capable of delivering a lot more power than - say - an iphone.

So back to my original question, has anyone tried removing the resistor/thermistor from the circuit to save power and either reducing the number of windings in the transformer or swapping out the transformer so the HP amp will see a higher impedance?

My guess is that the SRD-4 was designed this way because (1) at the time these units were sold very few people owned dedicated headphone amps, and/or (2) Stax simply re-used the design and parts from the (powered) SRD-5 without bothering to source a new transformer.

 

[rgonzale]

I messaged @spritzer and did some measurements on the SRD-4/SR-80 system, and many of my assumptions at the top of this thread (which I had gleaned from miscellaneous internet sources) are incorrect. The SRD-4 transformers' turns ratio is probably about 25:1. The SR-80 drivers themselves have an impedance on the order of 5000 ohm at 1 kHz, based on some crude measurements. This means they present a load of about 100-200 ohm at the primary side of the transformer. The additional thermistor/resistor doesn't significantly affect this, but helps stabilize the system and protect the amplifier. (The DC resistance measures about 6.7 ohm, a combination of the transformer primary DCR and the thermistor.)

Based on this, it would seem that a good headphone amp designed for 30-300 ohm loads should be able to deliver the drive levels needed to power the system, with no modifications required (other than wiring a stereo 1/4" plug to the SRD-4 input).

I tested this with my Creek OBH-11 amp and was able to achieve suitable listening levels at about a12 o'clock volume setting.

Incidentally here is a frequency response measurement of the system, using a simple instrumentation mic placed as close as possible to the diaphragm. The response is generally smooth and extended, but rolls off prematurely below 150 Hz. (Ignore the 60 Hz peak, caused by grounding issues.) This reflects the common view that Stax electrets are lacking in bass. Furthermore, the channel matching is good but deviates at a few points.

 

[edstrelow]

These are a decent enough phone while you are waiting to move up to full biased Stax. BTW check out my thread on damping in this forum.

 

[arjuna93]

Based on this, it would seem that a good headphone amp designed for 30-300 ohm loads should be able to deliver the drive levels needed to power the system, with no modifications required (other than wiring a stereo 1/4" plug to the SRD-4 input).

Have you succeeded with this in practice?