Portable headphone amp w. batteries

I´m testing several discrete OpAmp´s for my DIY portable headphone amplifier (with first generation NiMh 16VDC/2Ah battery pack), all with good results. And the batteries are drained to about 7VDC, then the volume will be a lot lower and if going below the OpAmp´s working voltage, it will also be distorted. Now, I wonder about some things (please see attached schematics):

1. As the input power is directly connected to the +/-15VDC output at the voltage regulators LM317/LM337 to minimize the power draining from the battery, so also the voltage reg´s and the two big power power filter cap´s and of course also the rectifier diodes, will not be used - may the voltage reg´s draw any mAh current backwards, altough there are no draining component at this direction other than the unused power stage. Maybe it´s best to remove the whole power section (by removing the voltage reg´s or cut the PCB´s copper conductor´s for the output pin´s)?

2. Should the two big filter cap´s (with original value 4700uF swapped to 10000uF) be used altough the amp is running by batteries, or will the OpAmp be satisfied without them powered in realtime directly from the batteries?

3. Can I safely remove all the three (or any of them) LED´s for the amp, to minimize the battery draining?

4. As those discrete OpAmp´s will accept 24VDC, will it be fine if raising the batteries voltage to something between 18 and 24VDC?

5. With above in mind, should the amp be better if replacing the NiMh with Lithium, such as dynamic and transient speed etc?

mikaelmark said:

I´m testing several discrete OpAmp´s for my DIY portable headphone amplifier (with first generation NiMh 16VDC/2Ah battery pack), all with good results. And the batteries are drained to about 7VDC, then the volume will be a lot lower and if going below the OpAmp´s working voltage, it will also be distorted. Now, I wonder about some things (please see attached schematics):

1. As the input power is directly connected to the +/-15VDC output at the voltage regulators LM317/LM337 to minimize the power draining from the battery, so also the voltage reg´s and the two big power power filter cap´s and of course also the rectifier diodes, will not be used - may the voltage reg´s draw any mAh current backwards, altough there are no draining component at this direction other than the unused power stage. Maybe it´s best to remove the whole power section (by removing the voltage reg´s or cut the PCB´s copper conductor´s for the output pin´s)?

2. Should the two big filter cap´s (with original value 4700uF swapped to 10000uF) be used altough the amp is running by batteries, or will the OpAmp be satisfied without them powered in realtime directly from the batteries?

3. Can I safely remove all the three (or any of them) LED´s for the amp, to minimize the battery draining?

Click to expand...

Of course, they don't do anything - they also don't waste a great deal of power either.

For a battery operated device you don't need all the complicated mains power supply, and do you really need such a high power amplifier just to feed headphones?.

4. As those discrete OpAmp´s will accept 24VDC, will it be fine if raising the batteries voltage to something between 18 and 24VDC?

5. With above in mind, should the amp be better if replacing the NiMh with Lithium, such as dynamic and transient speed etc?

Click to expand...

No different - battery type should have no effect, but Li-Ion are vastly superior to NiMh as a battery (and far easier to charge).

The supply to the amplifier needs to be approximately symmetrical and not significantly exceed + and - 15V to 0V
(The absolute limit for the opamp, at which it may be damaged, is +/- 18V so 36V across the power connections - try to stick to no more than 30).

You do not need the rectifier, regulators and all the extra input smoothing.

You need either two batteries to give + and - supplies, or add an extra circuit to create an artificial "zero volts" at half the battery voltage.

Thank´s for your quick replies!

I´m using a TLE2426 from Texas Instr. paired with a MKP 1uF cap. to create a virtual signal ground, and NOT using the negative voltage from the battery for ground. The datasheet say´s the Supply current for the TLE2426 are 170 micro Ampere. Is this what this component part will consume/drain? If so, then it will not be much as the diskrete OpAmp itself will consume between 14 and 28mA.


Some other components used for this amp: Dale 24 step ladder attenuator, Jantzen Superior Z/Russian Teflon/polystyrene caps, Holco resistors, Elna Silmic II, Panasonic FR and as opamp. Burson V6 Vivid/NewClassD/SS3602/Orange.

The reason I´ve lithium-batteries for replacing in mind, are they´re may have better draining factor with a much higher C drain - think about the RC-racing; 25 years ago they used NiCd, then they were replaced with NiMh (that rather had worser C drain) and last years they´re using Lithium with a lot higer C drain factor. But I´m not shure this advantage will have an affect for use in audio as there are a very fast current need.

And of course it´s a bit of overkill with such an extreme high amp. power for portable, but as I´m using it connected to a FiiO X5 with an Burr Brown PCM1792 directly from the low level line out to the amp. and using it with some really high grade phones, such as K702, K812, HD800 and SE-A1000, I get an audio quality that´s higher than most peoples home audio equipment (not to mention the crap they have in their ears today!). The downside is that I have to carry everything in a quite heave back pack, with thick cables and a big lamp-switch hanging out - has get some suspicious gazes, so not a good idea to carry this at the airport custom!

I notice that the amplifier design uses NO Negative Feedback from the output. Then the driver and output transistors will produce uncontrolled distortion if the headphones have a fairly low impedance and the loudness level is high. Also the damping factor will be low.

Loudness level? It will be vast. What will prevent deafness?? Oh, maybe you are deaf already and need the damaging loudness.

audioguru; Are you saying this amp will have a volume gain in bass and treble section (as this is loudness)?

My headphones are all between 64 and 300 Ohm impedance and I definitely not use any loudness or any cheap/crappy audio setting!

And the volume setting are just as high as needed to not damage my ears. When listening in my living room, I normally listen at about 70db. When I one year ago did an hearing test, it was as close to perfect it can be - and I´m 42 years old.

Oh, I almost forgot to tell this amp is a clone of a now selling amp with MSRP at about US$1200 (altough with a toroidal and lower grade components)!

I assume you are using the opamp & cap to produce a virtual ground for the input side.
Do you have the outputs capacitively coupled to that, or to negative? Coupling to the input ground could cause distortion, it would be giving some positive feedback at low frequencies.

Otherwise, as long as the single supply is adequate for the output you need, it should work OK - though you will have to adjust some resistor values in the transistor buffer sections to allow for the lower overall voltage.
The 1.5K resistors will need reducing to restore a similar bias leves between the output transistor bases, to what would have been present with 30V total supply. Possibly 680 ohms is worth a try.


Re. the $1200 amp - I'd put that in the "idiot bait" class, as with a lot of so-called audiophile gear. It's a simple design with not a great value of components and for most stuff like that you can find equal (or better) quality gear at a fraction the price.

Example - a "Headphone DAC" someone on another forum was advising a person to buy.

The cost was over £4000 but looking at the technical specs, the figures were not as good as with some common audio interfaces used for home studios - the headphone section of a Focusrite Scarlett 2i2, which you can get for about £100, has far better specifications & enough output level to go beyond painful.
Pair that with some Sony MDR-5706 headphones (less than 100-) and I'd challenge anybody to find anything that is _actually_ better in any functional sense, rather than just claimed to somehow be magically better.


I have an extreme dislike of places that push ludicrously overpriced stuff & rip people off..

rjenkinsgb said:

Re. the $1200 amp - I'd put that in the "idiot bait" class, as with a lot of so-called audiophile gear. It's a simple design with not a great value of components and for most stuff like that you can find equal (or better) quality gear at a fraction the price.

Click to expand...

Yeah - idiot bait

However, as usual AG is against any audio system that he doesn't agree with the design of - and particuarly OCD about damping factors, something which is of FAR less importance than he always claims, and depending on exact speaker types may be of no consequence at all (such as any speakers used on valve amplifiers).

And while I do agree with him about the stupid output levels from the design (it's hardly a 'headphone' amplifier), he seems to have ignored the local feedback round the output stages. It has been fashionable at various times to reduce feedback to a minimum, and to have separate feedback in different parts of the amplifier.

I said "loudness" to mean the audio output level, not the bass and treble boost used years ago with a "loudness" switch.
Every audio power amplifier and opamp have seen has the output stages included in the overall negative feedback to reduce distortion. Not this one.

mikaelmark said:

headphone amplifier (with first generation NiMh 16VDC/2Ah battery pack

Click to expand...

I think this amplifier will produce more signal with less power from the battery. "Stereo 20W Class D Audio Amplifier - MAX9744 "

• Power from 4.5V-14V DC voltage
• Up to 93% efficient (88-93% typical)
• 20mA quiescent current (or put into shutdown for 1uA quiescent)
• Up to 29.5dB max gain
• Use DC or AC coupled line-level input, up to 3Vpp
• Filterless Spread-Spectrum Modulation Lowers
  Radiated RF Emissions from Speaker Cables
• 20W Stereo Output (4Ω, VDD = 12V, THD+N = 10%)
• Low 0.04% THD+N
• Integrated Click-and-Pop Suppression
• Short-Circuit and Thermal-Overload Protection
• Link

rjenkinsgb said:

I assume you are using the opamp & cap to produce a virtual ground for the input side.
Do you have the outputs capacitively coupled to that, or to negative? Coupling to the input ground could cause distortion, it would be giving some positive feedback at low frequencies.

Otherwise, as long as the single supply is adequate for the output you need, it should work OK - though you will have to adjust some resistor values in the transistor buffer sections to allow for the lower overall voltage.
The 1.5K resistors will need reducing to restore a similar bias leves between the output transistor bases, to what would have been present with 30V total supply. Possibly 680 ohms is worth a try.
Re. the $1200 amp - I'd put that in the "idiot bait" class, as with a lot of so-called audiophile gear. It's a simple design with not a great value of components and for most stuff like that you can find equal (or better) quality gear at a fraction the price.

Example - a "Headphone DAC" someone on another forum was advising a person to buy.

The cost was over £4000 but looking at the technical specs, the figures were not as good as with some common audio interfaces used for home studios - the headphone section of a Focusrite Scarlett 2i2, which you can get for about £100, has far better specifications & enough output level to go beyond painful.
Pair that with some Sony MDR-5706 headphones (less than 100-) and I'd challenge anybody to find anything that is _actually_ better in any functional sense, rather than just claimed to somehow be magically better.


I have an extreme dislike of places that push ludicrously overpriced stuff & rip people off..

Click to expand...

rjenkinsgb said:

I assume you are using the opamp & cap to produce a virtual ground for the input side.
Do you have the outputs capacitively coupled to that, or to negative? Coupling to the input ground could cause distortion, it would be giving some positive feedback at low frequencies.

Otherwise, as long as the single supply is adequate for the output you need, it should work OK - though you will have to adjust some resistor values in the transistor buffer sections to allow for the lower overall voltage.
The 1.5K resistors will need reducing to restore a similar bias leves between the output transistor bases, to what would have been present with 30V total supply. Possibly 680 ohms is worth a try.


Re. the $1200 amp - I'd put that in the "idiot bait" class, as with a lot of so-called audiophile gear. It's a simple design with not a great value of components and for most stuff like that you can find equal (or better) quality gear at a fraction the price.

Example - a "Headphone DAC" someone on another forum was advising a person to buy.

The cost was over £4000 but looking at the technical specs, the figures were not as good as with some common audio interfaces used for home studios - the headphone section of a Focusrite Scarlett 2i2, which you can get for about £100, has far better specifications & enough output level to go beyond painful.
Pair that with some Sony MDR-5706 headphones (less than 100-) and I'd challenge anybody to find anything that is _actually_ better in any functional sense, rather than just claimed to somehow be magically better.


I have an extreme dislike of places that push ludicrously overpriced stuff & rip people off..

Click to expand...

As the standard version with a toroid will have 15 - 0 - 15 VAC, with 0 to signal ground, I missing this when using batteries that will only have + 16 - 16VDC, and that´s why I used the TLE2426 rail splitter with a 1uF cap to be connected from the batteries to the signal ground.

Yes, refer to the Grado RA-1, with a MSRP of ~USD600, that has a VERY slimmed component list/schematic and are very popular as a clone with a component cost of ~USD30 and can be assembled in a coffee break - and almost at the same level as a CMOI!

That´s why I upgraded almost every component I putted in. When buying this clone as a DIY kit from China, it will cost ~USD50, but this will be rather crappy. If buying all components separately with about the same component grade as original, it´s ~USD200. I´ve payed about ~USD400 (about 1/3 or the original retail price) and I´m really satisfied with the result!

Also remember that Sennheiser choosed this amp (with an OPA2134) when they demoed the HD800 at exhibitions.

A few years after the release of this amp, the german manufacturer released a "special edition" with some higher grade components and replaced the aluminium enclosure with a wood box, and the MSRP was raised ~USD300.

audioguru said:

I said "loudness" to mean the audio output level, not the bass and treble boost used years ago with a "loudness" switch.
Every audio power amplifier and opamp have seen has the output stages included in the overall negative feedback to reduce distortion. Not this one.

Click to expand...

Not everyone, as I've already mentioned there was a spate of low feedback and local feedback only audio amplifier over the last 20 years or so.

Any thoughts how to get negative feedback for this amp; should i solder some resistors or caps between the output and input pins of the OpAmp? Remember I´m using discrete OpAmp with pure Class A.

The opamp has feedback - that's what the 4k7 resistor between output and negative in does.

Only the output buffer is not included in the feedback loop, but as it is just emitter followers with high bias, that should be reasonably linear over a moderate range of output voltages, more than enough for headphones.

That is very dependent on the bias resistors being matched to the supply voltage, as I said previously.

Note that the original operates on 30V with a centre ground.
You supply is not plus and minus 16V, it is 16V total - with the extra circuit to produce a mid supply zero, the supply relative to that (which is the critical bit) is +/-8V

The bias resistors need to be changed appropriately.

As it's also unregulated you have to allow for the full range of battery voltage, from 1.5V per cell the instant they are taken off charge down to 1V per cell, which is when they are considered totally flat. Most of the time it will be working at around 1.2V per cell & downward from that.

mikaelmark said:

Any thoughts how to get negative feedback for this amp; should i solder some resistors or caps between the output and input pins of the OpAmp? Remember I´m using discrete OpAmp with pure Class A.

Click to expand...

Why would you want to? - I thought the entire point of the design was that you liked the sound it gives?, so why would you want to alter how it sounds?.

But unless you're posted circuits are completely wrong, you're NOT using discrete opamps anyway, as is obvious by the opamps in the circuit. It's just an opamp with a current buffer on the output.

rjenkinsgb said:

The opamp has feedback - that's what the 4k7 resistor between output and negative in does.

Only the output buffer is not included in the feedback loop, but as it is just emitter followers with high bias, that should be reasonably linear over a moderate range of output voltages, more than enough for headphones.

That is very dependent on the bias resistors being matched to the supply voltage, as I said previously.

Note that the original operates on 30V with a centre ground.
You supply is not plus and minus 16V, it is 16V total - with the extra circuit to produce a mid supply zero, the supply relative to that (which is the critical bit) is +/-8V

The bias resistors need to be changed appropriately.

As it's also unregulated you have to allow for the full range of battery voltage, from 1.5V per cell the instant they are taken off charge down to 1V per cell, which is when they are considered totally flat. Most of the time it will be working at around 1.2V per cell & downward from that.

Click to expand...

Should it be best to raise the battery voltage to about 30Volt? Then it should be 15VDC to the OpAmp!?

I´m also worried about the rail splitter TLE2426 for virtual ground, as it may not be able to deliver the current needed as it can deliver up to 20mA! Is that maybe enough for this purpose? The OpAmp themselves will drab between 14 and 20mAh. Replacing it with two resistors from + and - are maybe not ideally, but what if I remove the virtual ground generator completely? But as I told before, the sound are remarkably good.

mikaelmark said:

Any thoughts how to get negative feedback for this amp; should i solder some resistors or caps between the output and input pins of the OpAmp? Remember I´m using discrete OpAmp with pure Class A.

Click to expand...

The driver and output transistors will probably add some phase shift which will cause the opamp to oscillate at a high frequency requiring careful compensation if negative feedback from the driver and output transistors is added.

Does anybody have detailed spec's for this amplifier so we can see its distortion levels? Since its output power is not much then the massive driver and output transistors might be linear enough.

64 ohm headphones with a very loud output of 100mW use an RMS current of only 40mA. Then the buffer transistors do not do much more than make heat.
With 600 ohm headphones then 100mW needs a peak voltage of 11V and an RMS current of only 13mA which the OPA2134 opamps can easily provide without the output buffer transistors.