Want to build a headphone amp.

[Zad]

Build it and see. Anything is better than the hundreds of theoretical designs discussed in Internet forums!

Personally, for portable use I wouldn't bother with external transistors, I'd stick to a pair of op-amps, or a TPA6120 (current feedback amp, needs reasonable PCB design), but like I say, the best headphone amp is the one you actually build, not the one sat on a drawing board.

[Lightages]

Yup, Zad, I totally agree. Even trying out a 5532 op amp, or some in parallel, would be a good and easy start.

Like this simple one: https://easyeda.com/DiyGuy/NE5532_Based_Headphone_Amplifier_Circuit-53XfxrJ1i

[f5r5e5d]

wanting low Z output over audio brings another bit of EE into play with feedback amplifiers and the capacitance of the headphone cable potentially resulting in loop oscillations

"Zobel" series RC to gnd on a amplifier output to give damping is often seen

power amps for audio often use a series isolating Z such as inductor or L||R

op amp based headphone amps sometimes just use a R, at least for higher Z headphones 10-25 Ohms added in series doesn't give big FR error with typical single driver headphones

[SeanB]

My first PC amplified speakers was literally a car audio amplifier, driving 2 4R 10W speakers in plastic enclosures. Added a LC filter to the 12V supply taken from the PC, random choke from the scrapped radio it was sold with in the auction, and a 2200uF 16V capacitor from the same radio. Did the job for around a decade, stayed there with motherboard upgrades and was eventually scrapped when I changed case to fit an ATX supply in to power the then new Pentium Pro processor I bought.

[madires]

Why, WHY, would you put a 120 ohm resistor in series? High impedance outputs are a guarantee that you will have frequency anomalies.

When I built the amp I've found several references claiming 120 Ohms to be the standard impedance for headphones. Another interesting thing is that the impression of the sound level is quite similar to my normal speakers. So I can just plug in the headphones and don't have to adjust the volume. Extra point for convenience

[suicidaleggroll]

When I built the amp I've found several references claiming 120 Ohms to be the standard impedance for headphones.

That still doesn't explain anything.  This is not RF, you don't want to match the impedance.  Headphones, like all speakers, have a very dynamic impedance.  At one frequency they might be 120 ohms, while at a different frequency they might be 50, or 200.  By sticking a a 120 ohm resistor in series, you've created a frequency-dependent voltage divider, so you frequency response is going to ride the headphone's impedance curve.  That's not a good thing.

Another interesting thing is that the impression of the sound level is quite similar to my normal speakers. So I can just plug in the headphones and don't have to adjust the volume. Extra point for convenience

There are a hundred other ways of accomplishing that without screwing up the frequency response of the headphones.

[mich41]

Headphones already have substantial voicecoil resistance in series with the "actual load" (i.e. moving diaphragm around).  As I understand, getting reasonable FR involves matching this series resistance with any case resonance, damping, etc.

The 120R standard was meant to give headphone designers some way to predict the total series resistance for given voicecoil resistance, which is something they control.

120R fails miserably with multiple armature IEMs which contain highly reactive crossover networks.

It isn't clear whether manufacturers really target 120R amps or not in actual products, I have never seen solid data on this.

[f5r5e5d]

120 ohms is dead as a standard - not much use with 1Vrms DAP and 16-32 Ohm IEM

matching Z is the wrong concept in driving speakers and headphones - in audio the long accepted assumption is that amplifiers are nearly Zero output Z Voltage Sources and most loudspeakers and headphones are built with that expectation, give their flattest frequency response when driven from a Vsource or at least most are most accurate to whatever FR they were designed for when driven from a Vsource

but you can find exceptions too - some laud "current drive", others think some particular output Z may complement a specific loudspeaker or headphone - Sennheiser makes a amplifier for their 300 Ohm HD800 headphones that has 40 Ohms output Z

but matching Z really isn't a thing in audio amplifier/loudspeaker or headphones

[dom0]

When I built the amp I've found several references claiming 120 Ohms to be the standard impedance for headphones.

That still doesn't explain anything.  This is not RF, you don't want to match the impedance.  Headphones, like all speakers, have a very dynamic impedance.  At one frequency they might be 120 ohms, while at a different frequency they might be 50, or 200.  By sticking a a 120 ohm resistor in series, you've created a frequency-dependent voltage divider, so you frequency response is going to ride the headphone's impedance curve.  That's not a good thing.

I think the idea originated somewhere in maybe the 70s where it was common practice to just use a voltage divider from the main output (of the power amplifier for speakers) to drive the headphone jack on the front-panel. I guess transistors were expensive.

[madires]

I think the idea originated somewhere in maybe the 70s where it was common practice to just use a voltage divider from the main output (of the power amplifier for speakers) to drive the headphone jack on the front-panel. I guess transistors were expensive.

I think you're right. The series resistor is a simple and inexpensive solution to whack in a headphone jack. If you want to add it the proper way you would add a small amp around 1W and maybe a dedicated volume control.

For me the 120 Ohms resistors are fine since I don't wear headphones regularly and I can't say that the non-linear (regarding frequency) voltage divider causes any major issues for me. Maybe it's a perfect match for the frequency response and impedance of my inexpensive headphones

[retrolefty]

I think the idea originated somewhere in maybe the 70s where it was common practice to just use a voltage divider from the main output (of the power amplifier for speakers) to drive the headphone jack on the front-panel. I guess transistors were expensive.

I think you're right. The series resistor is a simple and inexpensive solution to whack in a headphone jack. If you want to add it the proper way you would add a small amp around 1W and maybe a dedicated volume control.

For me the 120 Ohms resistors are fine since I don't wear headphones regularly and I can't say that the non-linear (regarding frequency) voltage divider causes any major issues for me. Maybe it's a perfect match for the frequency response and impedance of my inexpensive headphones

 Even in the 70s some manufactures had models that did it right. I have A Sony 2000F (high voltage FET stereo preamp) that had a dedicated amp for headphones with their own dedicated volume control.

http://www.thevintageknob.org/sony-TA-2000F.html

 Many, (at least me lol), still think the 70s was the high point in consumer stereo hi-fi equipment. This was pre PC computer days when a persons hi-fi system was the first to get disposable income spent for quality and performance.
There was lots of competition from the Japanese manufactures that drove quite a horse-power race in receivers and amps. I have a Pioneer SX-1980 receiver that has 270 WPC rms and weighs like 90 pounds!

 Anyway your correct in that most stereo receivers just used resistor drop on output to drive the headphone jack(s).
But a few higher end pre-amps included a separate headphone drive.

[AF6LJ]

I think the idea originated somewhere in maybe the 70s where it was common practice to just use a voltage divider from the main output (of the power amplifier for speakers) to drive the headphone jack on the front-panel. I guess transistors were expensive.

I think you're right. The series resistor is a simple and inexpensive solution to whack in a headphone jack. If you want to add it the proper way you would add a small amp around 1W and maybe a dedicated volume control.

For me the 120 Ohms resistors are fine since I don't wear headphones regularly and I can't say that the non-linear (regarding frequency) voltage divider causes any major issues for me. Maybe it's a perfect match for the frequency response and impedance of my inexpensive headphones

 Even in the 70s some manufactures had models that did it right. I have A Sony 2000F (high voltage FET stereo preamp) that had a dedicated amp for headphones with their own dedicated volume control.

http://www.thevintageknob.org/sony-TA-2000F.html

 Many, (at least me lol), still think the 70s was the high point in consumer stereo hi-fi equipment. This was pre PC computer days when a persons hi-fi system was the first to get disposable income spent for quality and performance.
There was lots of competition from the Japanese manufactures that drove quite a horse-power race in receivers and amps. I have a Pioneer SX-1980 receiver that has 270 WPC rms and weighs like 90 pounds!

 Anyway your correct in that most stereo receivers just used resistor drop on output to drive the headphone jack(s).
But a few higher end pre-amps included a separate headphone drive.

I would agree once you got past the consumer junk and poorly designed power amplifiers that had too much stage gain and required massive amounts of negative feedback.