Testing balanced versus unbalanced headphone amplifier drive

[Hipper]

This was the original concept of balanced headphone amps:

https://www.headphones.com/pages/balanced-headphones-guide

I fell for it and bought the Blockhead, which was in fact two physically separate Headroom Max amps connected by a front and back piece with two short XLR cables connecting them at the rear for a 'crossfeed processor'. In effect, headphone monoblocks:

https://www.stereophile.com/headphones/614/index.html

This thread discusses 'why', actually 'if', it is necessary:

https://www.head-fi.org/threads/why-balanced-headphone-amps.539458/page-8

I like to think I've summed up this long thread in my post 112, plus later comments.

And here there's a discussion specifically about 'balanced headphone cables'::

https://www.audiosciencereview.com/.../are-headphone-balanced-cables-snakeoil.5589/

The summary seems to be that the only advantage is more amp power, which may be good, and crosstalk reduction, which is measurable but such a small improvement it has no practical effect on listening. I would also suggest that if you use an XLR terminal you may get a more secure connection.

Would another way of looking at it be that a balanced headphone amp is simply bringing headphones into the same arrangements as stereo speakers?

 

[solderdude]

which is measurable but such a small improvement it has no practical effect on listening.

Where are the measurements that show the effect of the common return wire of various headphones ?

I would also suggest that if you use an XLR terminal you may get a more secure connection.

A much better and sturdier connection but difficult to implement in portable equipment.

Would another way of looking at it be that a balanced headphone amp is simply bringing headphones into the same arrangements as stereo speakers?

Only electrically

 

[DonH56]

A perhaps more technical point:

The circuit design of the differential buffer, if present, also influences what happens when used single-ended. Some don't care, or care little, while others will exhibit significantly worse performance. Here's a very hand-waving analogy: consider a playground teeter-totter as a differential stage. In "balanced" mode, there is a person at each end, and it pivots in the middle. The middle is thus fixed in place and we can consider that the common-mode point of the circuit. In single-ended mode, some designs act like there is a person at one end and the pivot point is still in the middle. Swing (movement) is halved but the circuit is otherwise relatively unchanged.

In other designs, running in single-ended mode effectively moves the pivot to one end. The beam is fixed at that end, and the other end swings through the complete range of motion, instead of pivoting in the middle. The middle of the beam is now moving up and down. To a differential stage, that looks like a large common-mode signal, and some designs will not perform well with their common-mode operating point being moved up and down like that. They will have higher distortion, lower power-supply rejection, and so forth.

HTH - Don

 

[Hipper]

Where are the measurements that show the effect of the common return wire of various headphones ?

I don't know! Isn't that one of the possibilities behind this thread?

 

[solderdude]

Yes it is but no-one measured this yet.

 

[MC_RME]

Easy to do. The main limitation is both the cable resistance, the connector to socket resistance, and (sometimes) the PCB ground resistance. All these add up and turn a DAC's channel separation from >110 dB into a mere 70 dB up. Yes, in reality it can be that bad.

It is also true that a 4 pin XLR has about half the contact resistance than TRS, so the available channel separation is a bit higher. And of course even more high when the headphone uses separate ground wires left/right. Also true is that the audiophile's vinyl has mere 20 dB channel separation, an FM tuner about 30 dB. That seemed to have been enough.

Now the point is to put this into perspective. Is 70 dB crosstalk a problem? Or inaudible? If you have an ADI-2 Pro or DAC you can perform a simple experiment. It includes a control named Stereo Width which allows you to go from full stereo to mono in many small steps. Use a test sample that is extremely stereo, then step down to mono to see when you notice the difference. Use a one channel test sample to do the same until you notice it moves away from only one side a bit. It should be obvious that using headphones this test is much more precise/revealing than with speakers in a room.

Here is a list what dB value of crosstalk is reached when using the Stereo Width setting (measured at the rear outputs).

Width X dB
1.00 114,8
0,99 45,9
0,98 39,9
0,97 36,3
0,96 33,8
0,95 31,7
0,9 -25,6
0,8 19,1
0,7 15,1
0,6 12,1
0,5 9,5
0,4 7,4
0,3 5,6
0,2 3,6
0,1 1,7
0,0 0

So after a few steps you are already in Vinyl land. My hearing threshold was about 0.95 = 31.7 dB for stereo to mono and about 0.98 (39.9 dB) for the single channel test.

Using a 3 meter cable extension for your TRS headphones ground wil have around 0.5 Ohms, resulting in only 36 dB of channel separation. Still enough, it seems.

 

[solderdude]

Easy to do. The main limitation is both the cable resistance, the connector to socket resistance, and (sometimes) the PCB ground resistance. All these add up and turn a DAC's channel separation from >110 dB into a mere 70 dB up. Yes, in reality it can be that bad.

In reality it can be as bad at 24dB (measured this with AQ Nighthawk).
Not the the connector to socket resistance, and (sometimes) the PCB ground resistance part is the problem but the headphone cable.

The 'not so easy to measure this' remark was concerning the cable resistance of the common return wire combined with the impedance of the headphone. This differs from headphone to headphone and cable to cable.
The effect of this I have never seen measured nor is there a norm to test this too.

The amps usually are not the actual problem, unless really crappy in design or using virtual grounds.
The RME certainly will perform admirably on that point

Its the cables of the connected headphones that may not.
Balanced headphone cables have none of these issues by design.

 

[MC_RME]

Yep, fully correct.

 

[restorer-john]

Also true is that the audiophile's vinyl has mere 20 dB channel separation, an FM tuner about 30 dB.

Oh come on. Plenty of tuners exhibited separation of well over 50dB and many over 70dB.

 

[sergeauckland]

Oh come on. Plenty of tuners exhibited separation of well over 50dB and many over 70dB.

You may be right, but I've never seen more than 40dB. Bear in mind also that the stereo encoder at the transmitter end didn't do a lot more.

S

 

[Deleted member 65]

You may be right, but I've never seen more than 40dB. Bear in mind also that the stereo encoder at the transmitter end didn't do a lot more.

S

I've got the Revox B260 in our summerhouse, excellent piece of Swiss engineering, see specs below:

 

[restorer-john]

You may be right, but I've never seen more than 40dB. Bear in mind also that the stereo encoder at the transmitter end didn't do a lot more.

Here's a particularly good Sony ES tuner. One of the lowest distortion, highest performing tuners ever made.

Here's a Pioneer F-90

A Pioneer F-91

And a Yamaha TX-1000

These are just a few models I could think of off the top of my head that I have owned, or do own, all of which significantly exceed the ridiculously low bar of 30dB mentioned above by RME as a "true" number. Clearly, it is not true, is it?

I think there must be a whole generation or two who haven't experienced technically excellent tuners and the capability of FM broadcasting like we did.

Consider we had CD players in the late 1980s with >126dB channel separation (tested) at 1Khz and >108dB at 16Khz...

 

[solderdude]

The above discussion was not about vinyl but separation of the stereo decoder in FM tuners.
There are quite a few technical ways to make a decoder and in a lot of cases one could actually adjust these decoders.
Quite a lot of spread in separation between tuners/decoders as well.
The cheaper ones may well have had poor(ly adjusted) decoders resulting in poor channel separation.

 

[JohnYang1997]

The above discussion was not about vinyl but separation of the stereo decoder in FM tuners.
There are quite a few technical ways to make a decoder and in a lot of cases one could actually adjust these decoders.
Quite a lot of spread in separation between tuners/decoders as well.
The cheaper ones may well have had poor(ly adjusted) decoders resulting in poor channel separation.

Oh my bad.

 

[john61ct]

If you want a cheap balanced cable, check VE clan's website.

https://www.veclan.com/engappliance_sel_one?eng_ApplianceVo.eac_id=16

It gives many options to choose from for most headphones.

Beware of the VE Black Litz cable though, it's known to be microphonic while it looks cool.

Pricing ranges from 20$ to 100$+ depending on the options you choose.

What is the meaning of "microphonic" as used above? I infer that it is bad, so "picking up unwanted noise"?

 

[Nathan Raymond]

What is the meaning of "microphonic" as used above? I infer that it is bad, so "picking up unwanted noise"?

"Microphonic" when talking about headphone cables often means the degree to which a cable transmits the friction sound of the cable rubbing against clothes as you move into the earphone/headphone so you hear the cable rubbing in addition to sound from the drivers. Cables can transmit more or less of that sound due to the sheathing material and physical design of the cable.

 

[john61ct]

Aha, was in the ballpark then thanks!