Minimal Headphones Amp

[ruffrecords]

I am working on a very simple headphones amp for a mixer. The attached pdf shows my thinking so far. There is a level control before the input in each case and appropriate coupling and decoupling components are not shown

I started with a standard hi-fi type phones amp using a pair of 5532 op amps shown in the top figure. The two op amps work in parallel with 47 ohm build out/ current sharing resistors. The NFB is taken from the output so the output Z is very low and the gain is about 10dB.

The next version is one I found in the Mackie 1202 VLZ mixer sown in the middle figure. The negative feedback is now taken from the output of the first op amp and the build out resistors have been increased to 120 ohms. In this configuration the output impedance is essentially the two 120 ohm resistor in parallel which presumably gives it some advantage when working with a wide range of headphone impedances. The gain is still about 10dB.

I only need unity gain as my tube mixer internal level is +4dBu. So the third and I hope final version simply removes the two gain setting resistors and turns the first op amp into another unity gain follower.

I have built and tested the first version and as expected it has too much gain. I plan to modify it to match the other two and retest.

Comments?

Cheers

Ian

 

[JohnRoberts]

I am working on a very simple headphones amp for a mixer. The attached pdf shows my thinking so far. There is a level control before the input in each case and appropriate coupling and decoupling components are not shown

Headphone amplifier design is quite mature ranging from simple dedicated one chip headphone amps, to well over engineered with discrete transistors hanging off an opamp output stage. I have done both, and more.... Back in the 70s I used a national semi chip inside a disco mixer kit. Last century at Peavey I made a GP accessory headphone amp with discrete medium power transistors cobbled up to an opamp output that would easily drive loudspeakers. I had to current limit the PS because it would take out the thermal fuse in a 1A wall wart driving speakers loudly.

I started with a standard hi-fi type phones amp using a pair of 5532 op amps shown in the top figure. The two op amps work in parallel with 47 ohm build out/ current sharing resistors. The NFB is taken from the output so the output Z is very low and the gain is about 10dB.

Paralleling op amps is a strategy sometimes uses when a mixer maker is already using lots of a common op amp.

The next version is one I found in the Mackie 1202 VLZ mixer sown in the middle figure. The negative feedback is now taken from the output of the first op amp and the build out resistors have been increased to 120 ohms. In this configuration the output impedance is essentially the two 120 ohm resistor in parallel which presumably gives it some advantage when working with a wide range of headphone impedances. The gain is still about 10dB.

The only time I doubled up opamps in a design was situational. The design I was working on had a spare TL074 section and I wanted to drive a less than 2k feedback resistor. Using degeneration resistors in series with each opamp output nicely forced them to share.

I only need unity gain as my tube mixer internal level is +4dBu. So the third and I hope final version simply removes the two gain setting resistors and turns the first op amp into another unity gain follower.

The dedicated headphone amp ICs are not routinely unity gain stable so they might not be happy at unity gain, while you can trick them into behaving at low gain.

I have built and tested the first version and as expected it has too much gain. I plan to modify it to match the other two and retest.

Comments?

Is this intended for GP "everything for everybody" use? In console headphone amps it is common to use a serious build out resistor (51 ohms or so) to accommodate the full possible range of cans (3.2-600 ohms).

Cheers

Ian

Good luck...

JR

 

[ruffrecords]

Thanks for the comments JR. I started a general headphones topology thread some time ago from which I eventually selected the one in the first post.

I seem to remember PRR once wrote something about why headphones should be driven from a 120 ohm source. Does this mean damping factor in headphones is not an issue?

Cheers

Ian

 

[RuudNL]

Keep in mind that most (32 ohm) dynamic headphones are inductors.
This means that the frequency response will be infuenced if there is a series resistor at the output.
Also the damping factor will suffer.
(I tried an amplifier with and without a series resistor and the difference is audible.)

I have designed and delivered headphone amplifiers for Dutch national radio.
There was a shootout to find the 'best' headphone amplifier.
The people with the 'golden ears' thought my design sounded the best...
In fact it consisted of nothing more than an OpAmp followed by a PNP/NPN transistor pair for more output current.

 

[MicUlli]

Your super simple approach may not work as expected. It is simply an amplifier with an output impedance of 60 Ohm

 

[ruffrecords]

Your super simple approach may not work as expected. It is simply an amplifier with an output impedance of 60 Ohm

Please explain/expand

Cheers

Ian

 

[MicUlli]

I would expect that your design should have a very low output impedance. Isnt it better to connect the - input of U5 directly to OUT ?

 

[Tubetec]

Does an op amp with extra transistors at the output mean you need the usual 2x15 volt rails , plus another seperate dual rail supply for the output stage ?

For general control room usage I find the internal op amp headphone amps in mixers sufficient mostly ,
For vocals , drums and electric guitars extra headroom always comes in handy to help prevent the most objectionable distortions reaching the talents ears .

As we know Op amps tend to hit distortion very abruptly , maybe a slight concession on simplicity is worth the extra effort in this case .
In terms of the extra cost of the parts ,compared to the cost of an ITB mixer itself , isnt it worth having a HP circuit that can supply the goods no matter what headphones are connected .

Back in my studio days we had RANE HP amps , was fine for moderate levels ,
but did not cut it when someone had a set of Sennheiser DT-100s and playing a Marshall 100w cranked or a drum kit tuned to maximum volume was trying to monitor .

The other way of dealing with headphone amp running out of steam is to move to a higher senistivity headphone , Sony MDRV-6 is my go to in that situation

 

[JohnRoberts]

Thanks for the comments JR. I started a general headphones topology thread some time ago from which I eventually selected the one in the first post.

I seem to remember PRR once wrote something about why headphones should be driven from a 120 ohm source. Does this mean damping factor in headphones is not an issue?

Cheers

Ian

Built-in headphone amps involve a couple design compromises. We want the headphone output to perform serviceably (sound good) over a wide range of headphone impedances, and keep it simple to not consume too much PCB real estate or BOM cost.

The common compromise to incorporate a series build out resistance (like PRR's 120 ohm). Adding any series impedance there degrades performance but is considered a lesser evil to accommodate driving outlier low impedance cans without releasing smoke. The last time I did a survey (last century) of what kinds of headphones I had to drive, I came up with a range of 3.2 Ohms to 600 ohms. There were one our two outliers even higher impedance but the majority were in the range of 32-600 ohm. When designing for the mass market I learned to expect extremes. A moderate buildout impedance (50-150 ohms) protects against drawing too much current, while still being able to drive 600 ohms to modest output levels.

For one-off headphone amps, design for your personal headphone choice. For a high end/premium mixer customer you can probably make assumptions that your customers will use popular headphones (32-600). Thinking out loud you might be able to provide a low impedance headphone output and incorporate something like a PTC resettable current limiter. That way the head phone amp will deliver good fidelity and harmlessly shut down if overloaded with some cheap 3.2 ohm cans.

A common newbie mistake when incorporating a headphone amp into a console master section is to not properly account for the headphone current return path in the master section PCB. The headphone return current can easily corrupt clean signal grounds inside a complex master section layout causing crosstalk, a lesson typically learned the hard way, but hopefully only once.

I haven't reviewed modern offerings of dedicate headphone amp chips, so there may be a better canned solution.

For TMI perhaps search discussions about my old HB-1, while that was a stand alone headphone amp that could drive loudspeakers.

JR

 

[JohnRoberts]

Does an op amp with extra transistors at the output mean you need the usual 2x15 volt rails , plus another seperate dual rail supply for the output stage ?

you can often just power the emitter follower devices from the unregulated rails easing the current draw from the regulators.

For general control room usage I find the internal op amp headphone amps in mixers sufficient mostly ,
For vocals , drums and electric guitars extra headroom always comes in handy to help prevent the most objectionable distortions reaching the talents ears .

My wake up call was hearing how load a drummer's headphone mix was playing between takes during a session back in the 80s. The HB-1 was my answer to inadequate headphone level. I could hear his headphones from across the room and his signal was heavily clipped.

Many small studios repurpose and old audio power amp to drive headphones. Those studio headphone distribution systems routinely use something like 22 ohms in series with every drop.

As we know Op amps tend to hit distortion very abruptly , maybe a slight concession on simplicity is worth the extra effort in this case .
In terms of the extra cost of the parts ,compared to the cost of an ITB mixer itself , isnt it worth having a HP circuit that can supply the goods no matter what headphones are connected .

YMMV. Small amounts of signal clipping are generally well tolerated as long as the active stages recover cleanly and quickly without sticking or generating audible artifacts. Discrete transistor stages are generally pretty easy to do, but not impossible to screw up.

JR

Back in my studio days we had RANE HP amps , was fine for moderate levels ,
but did not cut it when someone had a set of Sennheiser DT-100s and playing a Marshall 100w cranked or a drum kit tuned to maximum volume was trying to monitor .

The other way of dealing with headphone amp running out of steam is to move to a higher senistivity headphone , Sony MDRV-6 is my go to in that situation

 

[fazer]

Many small studios repurpose and old audio power amp to drive headphones. Those studio headphone distribution systems routinely use something like 22 ohms in series with every drop

That’s what the studio I started in used. 27 ohm build outs, 120w amp stereo and AKG 240 phones. I think you need to differentiate between tracking sessions and mixing sessions. What worked well for one is the wrong choice for the other. We always had 3 to 8 players for tracking sessions in those days. To me worrying about the series resistance in a smallish room with a drummer at 110 DB is a waste of time. Mixing on the other hand with open back phones requires precision handling of the signal and the space (your in mixing) with an NC20 noise curve is a totally different set of requirements. You don’t blast headphones when mixing except to check for noises to remove from the mix.

 

[ruffrecords]

I would expect that your design should have a very low output impedance. Isnt it better to connect the - input of U5 directly to OUT ?

If you want a really low output impedance then yes. However, from the point of view of operating with headphones of different impedances and for the protection of the op amps and for the protection of peoples ears a build out resistor can be a good thing.

Cheers

Ian

 

[MidnightArrakis]

Keep in mind that most (32 ohm) dynamic headphones are inductors.
This means that the frequency response will be infuenced if there is a series resistor at the output.
Also the damping factor will suffer.
(I tried an amplifier with and without a series resistor and the difference is audible.)

I have designed and delivered headphone amplifiers for Dutch national radio.
There was a shootout to find the 'best' headphone amplifier.
The people with the 'golden ears' thought my design sounded the best...
In fact it consisted of nothing more than an OpAmp followed by a PNP/NPN transistor pair for more output current.

[I have designed and delivered headphone amplifiers for Dutch national radio] -- How about a schematic of said "Dutch National Radio" headphone amplifier? Sounds interesting!!! Gotta -- LOVE -- the capability for more output current!!!

/

 

[Matador]

From PRR:

This over-high level can be done over most headphone impedances with 7V behind 29 ohms. 300r phones get 6.4V (136mW), 32r gets 3.5V (380mW), the ancient 4r phones would get 0.8V (180mW). With thrifty output stage, this can run on +/-12V power.

 

[Tubetec]

A drum kit in a room ,
with or without headphones the player and any other musicians in the area are exposed to dangerous SPL as it is , distortion on the peaks in the cans does the ears involved no favours .

Where the Rane Hp amp didnt cut it or when higher head room was required typically an integrated hifi amp with the usual build out resistances in the HP circuit was used to supply the talents lugholes .

 

[ruffrecords]

From PRR:

I wonder if this was before or after the advent of rare earth magnets in headphones which significantly increased their sensitivity?

Cheers

Ian

 

[C H]

I haven't darkened the doors of a pro studio this century, but when I did they always seemed to have beyer dt100s driven from a big amplifier. Is that still the default? Sounded ghastly, I think the amp was the least of the problem.
Anyway. Q: How many headphones does a French recording studio need?

A: 15

 

[Michael Tibes]

In fact it consisted of nothing more than an OpAmp followed by a PNP/NPN transistor pair for more output current.

This was my first thought as well.

In my experience the audible differences between headphone amps are huge. I once couldn't finish a soundcheck because the band in the recording room was totally unhappy with the headphone sound. After a long and painful time of trial and error I replaced the Mackie mixers which they used as headphone amps in that studio with dedicated Behringer headphone amps and everything went smooth from there on. After that I wanted to buy more of these amps but they weren't available in the local store. So I bought a different brand, highly recommended by the shop, tried it, it sounded like shit, returned it, got the Behringer... Long story short for: this topic is so important but doesn't nearly get the attention it needs.

I believe there should be plenty of schematics for current boosted opamp based headphone amps. I didn't do really extensive research on the topic, but the best headphone amp I tried so far was a discrete amp I bought from china, certainly a clone of something expensive. If you have the space available (ca. 15 x 15 cm pcb) it might be an easy fix, it wasn't expensive. If you are interested I can try to find it for you.

Cheers,

Michael

 

[mhelin]

A recent thread "Best opamp for simple headphone buffer?" in diyaudio might be worth checking...

 

[JohnRoberts]

This was my first thought as well.

In my experience the audible differences between headphone amps are huge. I once couldn't finish a soundcheck because the band in the recording room was totally unhappy with the headphone sound. After a long and painful time of trial and error I replaced the Mackie mixers which they used as headphone amps in that studio with dedicated Behringer headphone amps and everything went smooth from there on. After that I wanted to buy more of these amps but they weren't available in the local store. So I bought a different brand, highly recommended by the shop, tried it, it sounded like shit, returned it, got the Behringer... Long story short for: this topic is so important but doesn't nearly get the attention it needs.

I believe there should be plenty of schematics for current boosted opamp based headphone amps. I didn't do really extensive research on the topic, but the best headphone amp I tried so far was a discrete amp I bought from china, certainly a clone of something expensive. If you have the space available (ca. 15 x 15 cm pcb) it might be an easy fix, it wasn't expensive. If you are interested I can try to find it for you.

Cheers,

Michael

Over the decades a great deal of work has gone into making premium stand alone headphone amps. Generally low output impedance and high enough voltage swing to deliver decent power into hi-Z cans is useful. IIRC Wayne has a premium headphone amp thread over on his forum.

These are easier said than done for GP built-in mixer headphone amps, that are expected to work with anything customers can plug in (i.e. 3.2ohm to >600 ohm).

JR