Headphones with tubes...

[Zap]

I was curious to get some comments on the best way to build an amp for headphone use. I've done some reading but I've seen a few approaches out there and was curious what the trade offs were. One site was discussing the use of load resistors with a second resistor tap to drive the headphones. This sounds like a trade off, but what exactly is being traded here? It would be nice to have the option for speakers OR headphones.

Otherwise, is it common to custom order transformers to the impedance of the pair of headphones being used? I'm thinking about a pair of the Audiophile line of Sennheiser headphones.

Thanks!

[DualTriode]

Hello All,
You are honing in on the same thinking that has been going on in my head. The Sennheiser HD600, HD650 and HD800 all are 300 ohm headphones. All have a pretty sharp impedance peaks (high Q) in the bass range. I will come back to this.
There is a standard in the headphone industry that calls for headphone amplifiers to have an output impedance of 120 ohms. This is why you see some people recommending the addition inline resistance for headphone use.
Solid state headphone amplifiers can have near 0 ohms output resistance.
What I have found is that the output impedance has a profound influence on how the headphones perform. The higher the output impedance the greater the bass loudness at the resonance peak of the headphones, (I am speaking of HD600’s) to some extent this is a matter of taste.
Here is the explanation as I see it. The higher the output impedance the closer the amplifier is to a current amplifier. Watts = I^2 * R (R = the impedance of the headphones). The lower the output impedance the closer the amplifier is to a voltage amplifier. Watts = V^2 / R. In short the amplifier with high output impedance puts out more accentuated bass (more watts) some call it flabby.
With the selection of tube rp, transformer impedance ratio, and negative feedback the tube amplifier can be matched to be exelent with the Sennheiser 300 ohm headphones.
I like about 20 ohms optput impedance with the HD600's.
DT
All Just For Fun!

[ddd]

Originally Posted by DualTriode

...pretty sharp impedance peaks (high Q) in the bass range...

Speakers with sharper and higher impedance peaks have lower Q. Why should it be high here?
Regards

[kavermei]

You probably already know this, but there are a lot of tube headphone designs documented on the "headwize" site:

HeadWize - Breaking News

Kenneth

[bigwill]

Originally Posted by ddd

Speakers with sharper and higher impedance peaks have lower Q. Why should it be high here?
Regards

High Q = sharp high peaks

[Arnulf]

Originally Posted by DualTriode

The higher the output impedance the greater the bass loudness at the resonance peak of the headphones, (I am speaking of HD600’s) to some extent this is a matter of taste.
Here is the explanation as I see it. The higher the output impedance the closer the amplifier is to a current amplifier. Watts = I^2 * R (R = the impedance of the headphones). The lower the output impedance the closer the amplifier is to a voltage amplifier. Watts = V^2 / R. In short the amplifier with high output impedance puts out more accentuated bass (more watts) some call it flabby.
With the selection of tube rp, transformer impedance ratio, and negative feedback the tube amplifier can be matched to be exelent with the Sennheiser 300 ohm headphones.
I like about 20 ohms optput impedance with the HD600's.
DT
All Just For Fun!

What ? The output power formulae are two of the same, derived through different (but interlinked) parameter.

Amplifier with higher output impedance will have worse control of transducer's diaphragm. Impedance peak of transducer will help to alleviate this problem, hence amplifier with higher output impedance might sound louder (relatively louder, compared to percieved loudness at higher frequency where transducer's input impedance is lower).

Same result (seemingly louder bass) can be achieved in a normal amplifier with suitable compensation ("equalization") network without having to rely on transdocer to kindly increase its Zinput for this to happen.

There is absolutely no reason why output impedance of an audio amplifier should be artifically inflated with the use of ohmic series resistor. It is only relevant for taming unwanted high frequency response and whatever response transducer might exhibit outside audible band and in this case the extra impedance used is of more complex type, not simply ohmic.

[ddd]

Originally Posted by bigwill

High Q = sharp high peaks

I'm not agree. It's not similar to a LC resonant circuit. Qes describes the electrical damping of the loudspeaker. Higher Bl => higher peak of Z (higher Back-EMF) => lower Qes (Qt). (Qes = K/Bl^2)

[kstagger]

I've built two tube based headphone amplifiers - a triode connected EF86 DC coupled to a 6922 White Cathode Follower. It sounded pretty good, though bass got a little flabby with lower-z headphones.

For a laugh, I built Pete Millet's "Starving Student" with a tube input / Mosfet out. Doing comparisons between the two using Sennheisers HD600s gave the nod to the Starving Student. It had a much more controlled sound and was dirt cheap to make.

[DualTriode]

Originally Posted by Arnulf

What ? The output power formulae are two of the same, derived through different (but interlinked) parameter.


Amplifier with higher output impedance will have worse control of transducer's diaphragm. Impedance peak of transducer will help to alleviate this problem, hence amplifier with higher output impedance might sound louder (relatively louder, compared to percieved loudness at higher frequency where transducer's input impedance is lower).

Same result (seemingly louder bass) can be achieved in a normal amplifier with suitable compensation ("equalization") network without having to rely on transdocer to kindly increase its Zinput for this to happen.

There is absolutely no reason why output impedance of an audio amplifier should be artifically inflated with the use of ohmic series resistor. It is only relevant for taming unwanted high frequency response and whatever response transducer might exhibit outside audible band and in this case the extra impedance used is of more complex type, not simply ohmic.

Hello Arnulf,
Amplifiers with higher output impedance will have higher output power into a speaker with a high Q. The highest power output will be at the impedance peak of the speaker. The lower the amplifier’s output impedance the less pronounced the peak in the power output at the speaker peak impedance will be. This is a real increase in power output. Some call this one note bass, or flabby bass.
I am not advocating adding series resistance. Some do.
In an effort to reduce SET headphone amplifier output impedance I look for an output tube with a low RP and an output transformer with a relative high impedance ratio to the headphones. All this is in an effort to minimize the amplifier output impedance and avoid boomy bass.
DT
All just for Fun!

[MrCurwen]

I have AKG K601 headphones, with which I'm very happy by the way. When I bought these I tried out a lot of phones, the main contenders being AKG and different Sennheiser models. AKG seemed to have a bit more open and airy sound and a bit wider bandwidth, so I went with them.

I have a DC Darling, and I just have a headphone jack on the front, and a switch back (next to the speaker teminals) which turns the speakers on/off. The headphone jack is always on.

Output impedance is 4 ohms. Performance is excellent (except the phones reveal absolutely every single problem in the sound). My Darling has no problem driving any of the following:

1) speakers
2) headphones
3) headphones + speakers
3) two sets of headphones in parallel

But it can't do two headphones and speakers. Oh well.

I think the Darling is a very good candidate for headphones.