Let me state at the outset: This may not work. I am going to try it and see, and record the results for posterity on this thread.
What it is:
A unity gain headphone driver for 16 ohm in-ear monitors (aka headphones) powered by two AA batteries.
Why it may not work:
Well, this is an exercise in minimalism, not the "less is more" minimalism of high end audio, but instead the "is this possible?" minimalism of trying to run a very inefficient circuit from batteries by cutting down the operating points down to the bare necessity.
Why bother?
Curiosity. Amusement. Fame and glory.
Okay, lets get started.
- purchased some leaded single cell AA battery holders on eBay. (link)
- grabbed two b-board pcbs from the in-house stack.
- assembled a power switch (DPDT), headphone jacks, a reasonably large case so I have plenty of room to work
- grabbed some Nichicon FG 6.3V 4700uF caps from Mouser. I don't believe in battery power unaided. The AA cells will feed a 4x4700 uF energy reserve.
- the b-board circuit will be modded as R1-R6 : 47 ohms (kamaya 1/2W carbon comp) R7-R9 : 1 ohm (kamaya 1/2W carbon comp) R10-R13 : unpopulated. The board will be powered from the "regulated input" V+, V-, COM pads. The Z-reg stage is left unbuilt.
- readied the test phones: Denon AH-C700 (link) 16 ohm 104 dB/mW.
And off we go!
What it is:
A unity gain headphone driver for 16 ohm in-ear monitors (aka headphones) powered by two AA batteries.
Why it may not work:
Well, this is an exercise in minimalism, not the "less is more" minimalism of high end audio, but instead the "is this possible?" minimalism of trying to run a very inefficient circuit from batteries by cutting down the operating points down to the bare necessity.
Why bother?
Curiosity. Amusement. Fame and glory.
Okay, lets get started.
- purchased some leaded single cell AA battery holders on eBay. (link)
- grabbed two b-board pcbs from the in-house stack.
- assembled a power switch (DPDT), headphone jacks, a reasonably large case so I have plenty of room to work
- grabbed some Nichicon FG 6.3V 4700uF caps from Mouser. I don't believe in battery power unaided. The AA cells will feed a 4x4700 uF energy reserve.
- the b-board circuit will be modded as R1-R6 : 47 ohms (kamaya 1/2W carbon comp) R7-R9 : 1 ohm (kamaya 1/2W carbon comp) R10-R13 : unpopulated. The board will be powered from the "regulated input" V+, V-, COM pads. The Z-reg stage is left unbuilt.
- readied the test phones: Denon AH-C700 (link) 16 ohm 104 dB/mW.
And off we go!
LTSPICE file
This is a model for simulation, not the actual circuit. I'm using it to get a feel for what the distortion looks like vs. output power.
The circuit manages about 0.5 mW class A into 16 ohms. 10 mW class B before voltage clipping. In class A the distortion is extremely low (<0.01%) but it rises a little with some prominent 3rd harmonic as you move into class B operation.
All very expected and textbook-y. The point is, though, that as long as you are listening at less than 0.5 mW, it's really clean and you've good a fairly massive class-B power reserve before things get really nasty. Even on transients real-world output power is highly unlikely to reach 10 mW for >100 dB/mW-class headphones
This is a model for simulation, not the actual circuit. I'm using it to get a feel for what the distortion looks like vs. output power.
The circuit manages about 0.5 mW class A into 16 ohms. 10 mW class B before voltage clipping. In class A the distortion is extremely low (<0.01%) but it rises a little with some prominent 3rd harmonic as you move into class B operation.
All very expected and textbook-y. The point is, though, that as long as you are listening at less than 0.5 mW, it's really clean and you've good a fairly massive class-B power reserve before things get really nasty. Even on transients real-world output power is highly unlikely to reach 10 mW for >100 dB/mW-class headphones
Attachments
Sorry for the wait, been too busy this last couple of months to deal with this - finally managed to case the B-board Boxer up this morning.
+/-1.26 V rails, dual mono, 4700 uF filter capacitance per rail.
1k ohm input resistance, 47 ohm 0.1 Zobel on outputs.
9 mA bias, <20 mV offsets.
+/-1.26 V rails, dual mono, 4700 uF filter capacitance per rail.
1k ohm input resistance, 47 ohm 0.1 Zobel on outputs.
9 mA bias, <20 mV offsets.
Attachments
For my reference:
Sennheiser HD600: impedance 300 ohms, sensitivity 97 dB/mW => 102 dB/V.
Denon AH-C700: impedance 16 ohms, sensitivity 104 dB/mW => 122 dB/V.
So to get the HD600s to the same loudness as the AH-C700s, I should apply an additional 19 dB (10x) voltage gain.
It seems instead that its closer to 10 dB (3x). Either someone is playing hard and fast with the published sensitivities (Denon model seems closer to 94 dB/mW than 104 dB/mW), or I've managed to mess up a pretty trivial calculation...
Test notes:
Sanyo eneloop AA voltage is 1.17 V on turn on, but rises to 1.26 V after a few minutes, falling very gradually thereafter.
47 ohm emitter resister measures 0.63 V (13 mA)
1 ohm emitter resistor measures 0.008 V (8 mA)
~ 75 mA per channel power draw.
~ 0.4 mW class A output power into 16 ohms
~ 10 mW class B output power into 16 ohms
Adjusting the simulation to lower the voltages to from 1.4 to 1.2 V, the currents displayed match the physical measurements exactly. LT spice is impressive! Hopefully the distortion calculations are as reliable...
Sennheiser HD600: impedance 300 ohms, sensitivity 97 dB/mW => 102 dB/V.
Denon AH-C700: impedance 16 ohms, sensitivity 104 dB/mW => 122 dB/V.
So to get the HD600s to the same loudness as the AH-C700s, I should apply an additional 19 dB (10x) voltage gain.
It seems instead that its closer to 10 dB (3x). Either someone is playing hard and fast with the published sensitivities (Denon model seems closer to 94 dB/mW than 104 dB/mW), or I've managed to mess up a pretty trivial calculation...
Test notes:
Sanyo eneloop AA voltage is 1.17 V on turn on, but rises to 1.26 V after a few minutes, falling very gradually thereafter.
47 ohm emitter resister measures 0.63 V (13 mA)
1 ohm emitter resistor measures 0.008 V (8 mA)
~ 75 mA per channel power draw.
~ 0.4 mW class A output power into 16 ohms
~ 10 mW class B output power into 16 ohms
Adjusting the simulation to lower the voltages to from 1.4 to 1.2 V, the currents displayed match the physical measurements exactly. LT spice is impressive! Hopefully the distortion calculations are as reliable...
Last edited:
So how does it sound?
Um ... fine?
With the Denon AH-C700 IEH and my Zune HD as the source, I compared with/without the Boxer between player and headphones.
By adding the Boxer, the Zune headphone output becomes a line out, driving a relatively benign 1k load. The only improvement that can be expected is to the degree that the Boxer is superior to the Zune output stage at driving headphones. If the Zune is a lousy line stage as well as a lousy headphone amp, then adding an external headphone amp won't help matters. Likewise if the Zune is a perfectly capable headphone amp already, then there isn't much to gain.
Given that it's not a really high end setup, it shouldn't come as too much of a surprise to learn that there wan't a huge difference to be heard. That, in itself, though, is a "pass": no obvious distortion, certainly no noise. Can you listen to 16 ohm headphones with only 1.2 V peak-to-peak available output swing from the amp? The answer is emphatically "yes".
Although there wasn't much difference, I did prefer listening through the B-board Boxer over the Zune directly. Remarkably lifelike reproduction of percussion, quite exceptional. Tons of ambient info and subtle detail. Deep, controlled bass. A bit too much: both the headphones and the player tend to be bass-heavy, each designed to compensate for poor low frequency response in the other. The Zune output coupling cap (I'm assuming it has one, most PMPs do) has a natural LF roll off driving 16 ohms, this is lost when driving 1k, and the B-board is DC coupled so no roll off there. End result is it all sounds a bit like an old discman with the "MegaBASS(TM)" switch engaged. It's not the amps fault, but an occupational hazard of using an external buffer amp in these sorts of applications.
Realistically I don't think it is really worth the effort for this setup. It was an improvement, though, so if you are interested in playing around I would not let this discourage you.
Me, I'm more interested in dropping the batteries for a low voltage transformer, and settling for a non-portable, desktop amp design. Possibly with a gain stage of some sort, and a volume control. Yeah, that's back to where I started with this, but there is room for a couple more variations on that theme, and I'm encouraged by the results to go on and see how good I can get the Denon IEHs to sound.
Um ... fine?
With the Denon AH-C700 IEH and my Zune HD as the source, I compared with/without the Boxer between player and headphones.
By adding the Boxer, the Zune headphone output becomes a line out, driving a relatively benign 1k load. The only improvement that can be expected is to the degree that the Boxer is superior to the Zune output stage at driving headphones. If the Zune is a lousy line stage as well as a lousy headphone amp, then adding an external headphone amp won't help matters. Likewise if the Zune is a perfectly capable headphone amp already, then there isn't much to gain.
Given that it's not a really high end setup, it shouldn't come as too much of a surprise to learn that there wan't a huge difference to be heard. That, in itself, though, is a "pass": no obvious distortion, certainly no noise. Can you listen to 16 ohm headphones with only 1.2 V peak-to-peak available output swing from the amp? The answer is emphatically "yes".
Although there wasn't much difference, I did prefer listening through the B-board Boxer over the Zune directly. Remarkably lifelike reproduction of percussion, quite exceptional. Tons of ambient info and subtle detail. Deep, controlled bass. A bit too much: both the headphones and the player tend to be bass-heavy, each designed to compensate for poor low frequency response in the other. The Zune output coupling cap (I'm assuming it has one, most PMPs do) has a natural LF roll off driving 16 ohms, this is lost when driving 1k, and the B-board is DC coupled so no roll off there. End result is it all sounds a bit like an old discman with the "MegaBASS(TM)" switch engaged. It's not the amps fault, but an occupational hazard of using an external buffer amp in these sorts of applications.
Realistically I don't think it is really worth the effort for this setup. It was an improvement, though, so if you are interested in playing around I would not let this discourage you.
Me, I'm more interested in dropping the batteries for a low voltage transformer, and settling for a non-portable, desktop amp design. Possibly with a gain stage of some sort, and a volume control. Yeah, that's back to where I started with this, but there is room for a couple more variations on that theme, and I'm encouraged by the results to go on and see how good I can get the Denon IEHs to sound.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.