[Joe Bloggs]

One effect that would be common to all headphones is the change in overall damping factor--the sound would supposedly be 'looser' than a 0 ohm impedance output, independent of changes in frequency response.

The other change, the change in frequency response, depends on the impedance response curves of the phones.

The power available to the phones at a particular frequency= V^2 / Z. The voltage V depends on the ratio of output impedance : headphone impedance at the frequency. Specifically,
V = Z(headphone) / Z(headphone) + Z(output).

The implication is that with 0 ohm output, the voltage across the headphones would be 100% of the output for all frequencies, while for higher ohm outputs, the higher the impedance of the headphones at a particular frequency, the higher the proportion of voltage available across the headphones. As output impedance tends to infinity, ratio of voltage available across the headphones for 2 frequencies = ratio of headphone impedance for those two frequencies.

I'll write more about this, but quick skinny is that headphones with an impedance hump in the low frequencies will end up with a low frequency response hump with high output impedance, while headphones with an impedance hump in the high frequencies will end up with a high frequency response hump. So the midbass hump of the HD580s will get worse with higher output impedance thanks to the midbass impedance hump, while the K501 may sound more shrill with higher impedance. Do the etymotics driver's impedance response look more like the K501 than the HD580?

 

[ServinginEcuador]

Joe,

So it sounds like what you're saying is that the impedance curve of a low outpu timpedance source is basically the same as the impedance curve readings?? IOW, where there's a bump in the frequency curve, it's because of the high impedance at the frequency since the low output impedance drops very little voltage compared to the high impedance presented at that frequency??

So, when you more closely match the output impedance of the amp to that of the headphone, you should end up with a flatter curve as the variations in impedance is much less when the two are equal at a static dc state??

 

[gdahl]

In a nutshell:

An amp with a very low source impedance will be relatively unaffected by the peaks and dips in the headphone's impedance curve.

An amp with a very high source impedance will cause the frequency response to vary according to the shape of the impedance curve.

Right?

 

[fredpb]

Let's look at the impedence curve above.

Your little amp has 50mw output at 8 ohms for an 8 ohm set of headphones.

How many mw output will it supply to the Senn 600's at 100hz????


Given that low freqs need lotsa power.

Duh.

So how will the Senns sound with that little amp with 50mw rating?????

 

[dd3mon]

fred, according to your profile you just turned 3 a month ago.. hm.

 

[gdahl]

fredpb,

Actually, the 600's impedance peak at 100 Hz will make it easier to drive at that frequency, because it will draw less current.

But if the amp's output impedance is too high, there will also be a response peak at that frequency, resulting in a bloated, poorly controlled bass response.

 

[Joe Bloggs]

Quote:

Originally posted by gdahl In a nutshell: An amp with a very low source impedance will be relatively unaffected by the peaks and dips in the headphone's impedance curve. An amp with a very high source impedance will cause the frequency response to vary according to the shape of the impedance curve. Right?

Yeah, that's just about what I'm saying

An amp with a very high source impedance will cause the frequency response to vary according to the shape of the impedance curve, in addition to the variations in frequency response that are already present at 0 ohms output impedance. Given that amp power is not a problem, that is, as fredpb mentions...

Another example...

The DT931 is considered to be on the thin, overbright side at 0 ohm output impedance on e.g. the Corda...

At higher output impedances the bass is boosted by the impedance response bass hump.

 

[Gariver]

Joe Bloggs:

I am glad you are discussing this subject. Output Impedance is important. However, it is easy to get confused when talking about it.

Transcendent Sound--On Output Impedance:
Here's a page that discusses OI. Notice how negative feedback is used to drive down a tube amp's output impedance. I found that part to be very, very interesting.

http://www.transcendentsound.com/amp..._impedance.htm

Good luck with this info!

 

[gdahl]

Gariver,

The link provides good information about the use and purpose of negative feedback. But part of the story is left out.

Pentodes require negative feedback for two purposes: bringing distortion down to reasonable levels and lowering output impedance to a usable value.

Unfortunately, the results are not as good with a real speaker or headphone as they are with a test resistor. In the real world, feedback is always filled with errors when music signals are played into transducers. As the author of the link wrote, the speaker is a motor. It is also a generator, and its motion creates a junk signal directed back at the amplifier. This junk signal becomes part of the amplifier's negative feedback! You can add to this problem the timing errors. Negative feedback is always late. It is always trying to correct for something that has already reached the outputs.

If we consider triodes, the most linear amplifying devices available, they have impedance much lower than pentodes, but still not low enough to prevent some problems. Many people have found that they could get excellent results with triodes as long as they could match them with transducers having flat enough impedance curves to keep the trouble to a minimum. The payoff is being able to avoid negative feedback, with its tendency to suck the life and three-dimensionality out of the music.

I believe it is sonically preferable to bring down a triode's output impedance using a transformer, or in the case of a headphone amp, an autotransformer. I looked at the performance specifications of some OTL headphone amps, and they are not nearly as good as a triode/autotransformer amp. Also, the OTL designs I have looked at have a large-value electrolytic capacitor at the output. I have never heard an electrolytic capacitor outperform a good transformer!

I do think that it is worthwhile working to understand feedback and its applications, but pentodes and OTL's have plenty of proponents. I wanted to take the opportunity to wave the banner for triodes and iron!

 

[Gariver]

Hi Gary:

After your post, I went to Headwize's Projects and found a nice SRPP headphone amp. This is the type of amp you are advocating, right?

http://headwize2.powerpill.org/proje...bridge_prj.htm

*

 

[gdahl]

Hi Gariver,

You're teasing me, right?

You know I advocate for the Espressivo approach to headphone amps!

Anyway, I looked over Simon's project. Very interesting, and nicely done! The SRPP is used as an input/gain stage, and the output tube, with very low gain, is used to drive the transformer.

What I can't figure out is why he didn't just use a single 417A/5842 stage, which is what his transformer was designed for. As shown, Simon has three triode sections per channel, not particularly low-distortion ones at that. The SRPP configuration does not have very good power supply rejection ratio either. A 417A/5842 would have provided enough gain at lower distortion. I would expect greater transparency from fewer stages, especially when the SRPP can be avoided.

But it is a cool project, and I'll bet it sounds quite good!

 

[Gariver]

He, he! I'm not teasing you. I'm just trying to learn from you.

I focused on the Busbridge amp due to this quote: "The output power will depend on the impedance of the headphones; the amp gives about 10 V output before clipping. The gain of the output stage is 0.33, so quite a bit of amplification is needed before it. I am a fan of SRPP because it improves linearity, lowers output impedance and allows for a large voltage swing."

That last sentence on SRPP is interesting...
--SRPP improves linearity.
--SRPP lowers output impedance (Busbridge gets 2.5 Ohms).
--SRPP allows a large voltage swing.

 

[gdahl]

Hi Gariver,

It is true that the SRPP circuit improves linearity, reduces output impedance and allows a greater voltage swing. But it's all relative. The 6072A used in the SRPP circuit doesn't have good linearity or low output impedance in the first place. Using it as an SRPP improves it but still doesn't make it as good as some other options. Think about it--if these factors were sufficiently improved, Simon could have just driven the transformer with the SRPP. But the SRPP isn't up to that task, so he uses it to drive a second tube, which has low gain but higher current capability.

Even with the SRPP, the 6072A still has higher distortion than the 3A/167M in the Espressivo. The 3A/167M has enough gain and low enough output impedance, all in one stage, to drive the TL-404 autoformer.

The frequency response is also better using the 3A/167M and the autoformer--the minus 3dB points are 7 Hz and 120 kHz!

 

[fredpb]

Quote:

Originally posted by gdahl fredpb, Actually, the 600's impedance peak at 100 Hz will make it easier to drive at that frequency, because it will draw less current. But if the amp's output impedance is too high, there will also be a response peak at that frequency, resulting in a bloated, poorly controlled bass response.

Actually, heavy bass draws mucho current. The lack of current at 100hz makes the phones sound thin at that frequency unless you have a good amp.

They do NOT get easier to drive they get worse, by far.

Took me quite a while to get decent performance with my Senns.

But they still don't have the low end they should. But Grados and AKG's are worse, so I live with it.

 

[Joe Bloggs]

Quote:

Originally posted by fredpb Actually, heavy bass draws mucho current. The lack of current at 100hz makes the phones sound thin at that frequency unless you have a good amp. They do NOT get easier to drive they get worse, by far. Took me quite a while to get decent performance with my Senns. But they still don't have the low end they should. But Grados and AKG's are worse, so I live with it.

Wrong... it's 3am 1/1/2003 here and I can't think of anything more to say / explain

Happy new year