[pmillett]

I'm curious - all you builders out there: what series resistor, if any, do you use on the output of your headphone amps? This would be the resistor between the headphone (jack) and the output buffer of your amp.

This really applies only to amps with a solid-state output buffer (opamp, BUF634, diamond buffer, etc.)

The reason I ask - this R (or if you prefer, the amp's Zo or damping factor) affects the sound of many headphones dramatically. So it's hard to settle on a single "best" number.

IEC61938, which I suspect at least the German headphone makers follow, stipulates 120 ohms. But not many amp builders follow this, myself included.

I seem to prefer ~20-50 ohms, usually, though it does seem to depend on the headphones.

Thanks for your feedback!

Pete

 

[Andrea]

My Xenos 0HA-REP uses 6.8 ohm series output resistors, which seems perfect to me with all the headphones I have.

 

[AndrewFischer]

I built My PPA with no series resistor. Yes this does put the buffer chips at risk.

Even with HD-580s I wouldn't use a resistor. Part of the reason I DIY is to get away from compromises like an output resistor.

Andrew

 

[kklee]

I've settled on 4R7 as being the best compromise for the various headphones I have.

 

[rickcr42]

10 ohm for low Z cans,120 ohm for the Euro cans which seem to respond well to the added resistance,both outside the feedback loop for more RFI load isolation (isolates the feedback loop)
If you look at Jan Meiers amps he usually has both a high-Z and a low-Z headphone out on the panel but no reason it can't be a DPDT "center off" toggle switch for choice of Low/Off/High to satisy all possible headphone requirements

 

[xstephanx]

why nto just put a rheostat there, and you can change your damping factor to whatever headphone you cant relatively quickly. You could ever (correct me if im wrong) slowly adjust it while listening to more accurately find your sweet spot?

though im assuming ther must be something wrong in my thinking or people would already be doing it.

 

[rickcr42]

I suggested that very thing in another thread almost two years ago but be aware that is EQ thus changes the sonics depending on rotation/load capacitcane/load inductance whereas this is more about getting the output into some kind of standard so there is a known benchmark.

If it is to be assumed the Euro cans are all designed and tested to be played from a standard impedance,and there is much evidence to support this,then it follows that any amp that wanders from the styandard is providing its OWN EQ profile to the end sonic profile much as you would by adding series resistance to a speaker amp driving an 8 ohm load.This is a way to claim "pure signal path" and maintain tone controls suck while back door-ing the consumer with the very thing you claim to despise.

Where it gets a bit into unknown territory is not so much the Euro amps but the U.S. and Asian amps that seemingly follow their own in house standards from manufacturer to manufacturer and that leaves the door open for a damn lot of mismatches/accidental synergy that would be better for the consumer if there WERE imposed standards so at the minimumn two headphones bearing the same standard could be measured then compared head to head against a common benchmark,as in the loudspeaker arena.

just way too many variables the way things stand and again makes it a crap shoot on getting it right unless you KNOW what the manufacturer was using as his/her model when dialling in the final profile.How else to ever know what exactly the headphone is MEANT to sound by the designer of if you have no insight into the design/test parameters ?

I think Jan has it mostly right in the optional load that the consumer can choose to use or not with the cost to add it so small as to be inconsequential at best.

When it comes to low Z cans or "off standard" loads the resistance may do more harm than good when you consider it will change the damping factor though if contemplated or actually used by the original designer may actually be dead on,unknown territory but territory that really needs to be explored if only because is is a good idea to totally isolate the output with a resistive load in a feedback amp,something that has already been established for damn near 50 years because in a feedback amp any signal appearing at the output that does not appear also at the input (RFI/EMI) WILL be treated as a signal instead of cancelled out since only a common mode signal can be canceled by a differential input amplifier creating a problem that you either fix at the source-the output/feedbak loop input-or in the loop itself with compensation.

Open loop amps do not suffer this problem so another "plus" for open loop single ended class-a operation making the resistor a purely optional tone control/bass peak tamer

 

[tomb]

Rhoestats are particularly noisy, and none are wound with technology that is adequate for high-fidelity audio, I think.

A jumpered output resistor is best, I think - except for some tube amps with lots of gain on a low impedance phone (Grado's). In my limited experience, 20 - 50 ohms makes no difference in sound compared to a jumpered output - at least on my Milletts with various gains (7-14), various buffers, and anything from Grado's to Koss's to Senn's.

Once you get into gains like 18-20 (SOHA, for instance) with something such as Grado's, then 100-150 ohms at the output can make a worthwhile difference on the volume control.

Except for the Intersil's, there's probably not much danger of frying them due to a short. Most of the DB's built around here are robust enough to take the intermittent shorts of plugging/unplugging of headphones, and many other incidental shorts - slipped meter probes, etc. The BUF634's have output current protection, anyway.

 

[nikongod]

the place where i find a resistor to have the "best" value is BETWEEN an output coupling cap, and a low impedance headphone.

adding someting like 22 ohms to a grado does little to the sound on a direct coupled/transformer coupled amp, but it extends the bass response soooo far when there is a cap in the output.

please take my comments a "fwiw" because i dont have any beyers that seem to love the 120 ohm outputs...

 

[rickcr42]

Quote:

Rhoestats are particularly noisy, and none are wound with technology that is adequate for high-fidelity audio, I think.

any potentiometer can be used as a rheostat if you simply attach one end of the fixed resistor portaion to the slider which makes a variable resistor out of it instead of a voltage divider which a volume control is.


Quote:

Except for the Intersil's, there's probably not much danger of frying them due to a short. Most of the DB's built around here are robust enough to take the intermittent shorts of plugging/unplugging of headphones, and many other incidental shorts - slipped meter probes, etc. The BUF634's have output current protection, anyway

See above.Most modern devices are current limited so the old "series resistor current limiting" is not the main reason you would use a resistor on the output unless you are trying to dial in the damping to get to a flat response.Where they are needed is to isolate the feedback loop in any amps that use opamps,multloop opamp/buffer combinations and any amp that uses feedback from the output stage to the input stage (and a lot of buffers having internal feedback

) or you risk cantaminating the sound with additional signals not part of the input source.

For the measly under a buck it adds to the cost should be a no brainer during the design stage and if it changes the SQ for the worse then the design is not adequate to begin with and you need to beef it up to compensate

 

[tomb]

Quote:

Originally Posted by rickcr42 /img/forum/go_quote.gif any potentiometer can be used as a rheostat if you simply attach one end of the fixed resistor portaion to the slider which makes a variable resistor out of it instead of a voltage divider which a volume control is.

Yes, well a "rheostat" is a particular kind of variable resistor usually used in a power application - implying motor control or similar applications.


Quote:

See above.Most modern devices are current limited so the old "series resistor current limiting" is not the main reason you would use a resistor on the output unless you are trying to dial in the damping to get to a flat response.Where they are needed is to isolate the feedback loop in any amps that use opamps,multloop opamp/buffer combinations and any amp that uses feedback from the output stage to the input stage (and a lot of buffers having internal feedback ) or you risk cantaminating the sound with additional signals not part of the input source.

Well, since it was Pete asking the question, I deduced that current limiting protection might be part of the reason for the question. His own Audio Xpress article for the original Millett Hybrid stated,
Quote:

Originally Posted by Pete Millett, Audio Xpress, 11/02 "The output of the BUF634 is connected through a 22Ω resistor, which is needed only to help protect the BUF634 in case of a short circuit of the output, but it also affects how different headphones sound."

Of course after years of using it, we now know that the BUF634 has its own protection. Nevertheless, it's a prudent measure that is still used in many cases today, re: the Intersils, recent of PPAV1 and revMillett Hybrid, the Millett design itself, and the SOHA (150 ohms due to the large gain). Most discrete DB design/construction - most often preferred these days - at least gives it consideration at some point along the way.

If fact, after reviewing his original article, either topic (current limiting/impedance matching) may be the reason for his question above:
Quote:

Originally Posted by Pete Millett, continued I usually recommend a series resistor of between 10% and 50% the impedance of your headphones— e.g., if you have 200Ω headphones, use a resistor between 20 and 100Ω. If you don’t know what impedance your headphones are, or are going to use several different headphones, stick with a smaller resistor (such as 22Ω). Again, you can experiment with this resistor value to see what differences you hear without worrying about hurting anything."[/i]

 

[tomb]

BTW, PETE: Many thanks for your wonderful Millett Hybrid design. I'm not up to insane build numbers yet, but I do have 4 of them so far. It's a wonderful, versatile and nostalgic little amp (sounds really good, too)!!

 

[joe_cool]

I'm convinced that many sound variations in headphone and interconnect cables have to do with reflections in the transmission system (source, cable, and load). When the impedances are matched, the reflections are minimized.

Let's consider an amp with an output impedance of <1 ohm, a ~75-ohm cable, and a headphone load of 16-300 ohms. Reflections from the (inductive?) load back to the amp will affect the feedback loop and the amplifier response. What does this sound like?

Theory suggests that the source should have a non-zero impedance to balance the system. This may not be practical in a loudspeaker system due to power losses but in the headphone system we have plenty of power work with.

 

[rickcr42]

Quote:

Yes, well a "rheostat" is a particular kind of variable resistor usually used in a power application - implying motor control or similar applications.

since you seem intent on splitting hairs

http://en.wikipedia.org/wiki/Potentiometer

and especially this

"A rheostat is essentially a potentiometer, but is usually much larger, designed to handle much higher voltage and current."

notice there is a "usually" added and not an "always".If there were then this would make no sense at all

http://www.allaboutcircuits.com/vol_6/chpt_3/7.html


Quote:

Well, since it was Pete asking the question, I deduced that current limiting protection might be part of the reason for the question. His own Audio Xpress article for the original Millett Hybrid stated

and you have a nice day too since you seem more intent on going at my posts that at the minimum offer up an opinion than in adding anything constructive to the conversation..........

rick out........................

 

[tomb]

Happy New Year, Rick.