Why do some amps provide more output power when using a balanced connection?

I have an Audio Research LS 25 mk I which has a higher gain when using the balanced inputs and outputs. I just acquired a Schiit Jotunheim headphone amp that has a higher output with balanced as well.

 The increases are not at all insignificant.  The ARC preamp goes from 12db of gain with single ended up to 18db when using balanced.  The Schiit Headphone amplifier goes from 800mw to 3000mw ( with a 50 ohm load) when moving from single ended to balanced.

 Can someone explain why this takes place? 

I'll leave the answer to your question to those better suited to answer it than I, however, I can tell you that 6 db difference between SE and XLR is *not* insignificant. 
gdhal, did you misread Mark's post? He described the difference between 12dB and 18dB as "not at all insignificant", same as you.
This is an Almarg question.Or, Atmasphere.

circuit design as to how an SE output is derived

can I convince you to run Balanced and get all you can from the nice pre-amp?

I have a Mk II BTW
Post removed 
gdhal, did you misread Mark's post? He described the difference between 12dB and 18dB as "not at all insignificant", same as you.

@bdp24  (and OP)

Yes, I did misread the OPs post. Please accept my sincere apology. 
Apology accepted :-)

So the reason I wrote this post, besides sheer curiosity, is that it effects my configuration with my new headphone amp.  

A bit of background: I am using a Schiit Jotenheim to power a pair of Audeze LCD-2 headphones and connect to my main system with fancy DAC, turntable, and ARC preamp, etc. When the baby is asleep and I can’t listen to the speakers in the main system, I will be enjoying music with the headphones. I am connecting the Jotenheim with XLR cables to the tape out on my tube preamp. The entire system is balanced.

Since the Audeze LCD-2 is a bit on the power hungry side, and the Jotenheim is perfect for that, I will also be using the Jotenheim at my computer desk during the work day. That means moving the amp to my home office and connecting to my Chord Mojo running in DAC only mode. The output on the mojo is 3.5mm stereo jack. The amp section will be disabled and it will simply feed an analog line level signal to the Jotenheim.

Finally, I get to my question: Will the Jotenheim deliver the higher output that the specs suggest with the headphones connected to the balanced output, and the input on the Jotenheim using single ended? Or does the balanced input as well as the balanced output need to be used to get the higher watt output? I would like to be sure I am driving the LCD-2’s with the full power available.

I have yet to purchase a cable to connect the Chord Mojo to the Jotenheim. I can get a 3.5mm to XLR cable, if that is required to get the proper output. If not, I can get a simpler 3.5mm to RCA.

I posed this very question to Schiit, and they replied:

"The Jotunheim is balanced differential. Meaning it will not turn a single-ended connection to a "Fake" balanced connection. You may notice more power from using the balanced headphone output with a single-ended input but I cannot speculate on that."

One of the reasons I chose the Schiit Jotenheim was its output power using a balanced config, in my main system.  All good.  The computer desk use is secondary, though I will likely use it more frequently there.  If I can get the 3000mw output rather than 800mw at my computer desk, my headphones will be happier.  Apparently this is not going to work, according to the gent I heard from at Schiit.

Is there anything I am missing?

I use balanced headphones .....balanced from the preamp through the balanced amp and then balanced to headphones.  The sound is quite wonderful...and recommended.  Your question is of less noise and increased clarity with balanced configuration. 
Hi Mark,

The preceding responses are generally correct, but I’ll elaborate on them a little further and then address the Jotunheim question specifically:

When an unbalanced signal is received, usually via an RCA connector, what is sensed by the receiving circuit is the voltage difference at any instant of time between the signal (on the RCA center pin) and ground (on the RCA ground shell). When a balanced signal is received, usually via an XLR connector, what is sensed in most cases is the instantaneous voltage **difference** between the two signals in the balanced signal pair (which are usually provided on XLR pins 2 and 3, with XLR pin 1 being ground). Those signals are usually made equal in magnitude but opposite in polarity, with the difference between their voltages at any instant of time therefore corresponding to twice the voltage of each signal at that instant.

When a component provides both RCA and XLR inputs or outputs, it is often the case that rather than those inputs or outputs being handled by separate and independent circuit stages, the center pin of the RCA connector is simply wired directly to one of the two signal pins on the XLR connector (usually pin 2) and the ground shell of the RCA connector is wired directly to the XLR ground pin (pin 1). In fact you can see in the schematic for your LS25 that exactly that is done, for both inputs and outputs. You’ll find the schematic for both the original and MkII versions of the LS25 near the bottom of the following web page. Click on the left-most of the three schematic pages shown for each version, to expand the relevant figure:


So in designs that are implemented in that manner, the relevant voltage on the XLR connector (i.e., between pins 2 and 3) and the relevant voltage on the RCA connector (i.e., between the center pin and the ground shell) will differ by a factor of 2. A factor of 2 voltage difference corresponds to a difference of 6 db, based on the relation db = 20 x log(V1/V2), where "log" is the base 10 logarithm.

Now, regarding the Jotunheim:

First, keep in mind that gain and maximum power capability are two different things, which are not directly related. What gain affects are the volume control settings that will tend to be used, and the volume control setting at which maximum power is achieved, assuming that the gains, sensitivities, and output levels of the components that are involved are such that there is never a need to turn the volume control higher than its max setting.

Most balanced power amplifiers will provide the same maximum power capability when driven single-ended as when driven balanced. But not all. Notable exceptions are many ARC power amps, which are fully balanced and ONLY provide XLR inputs. In those cases, if the amp is provided with an unbalanced signal via an RCA-to-XLR adapter its maximum power capability may be reduced as much as 75%, and sonics will suffer as well. But those designs are unusual in that respect. And since the Jotunheim provides both RCA and XLR inputs I would suspect that it is not similar to those unusual cases.

So while in the absence of specific knowledge of the design I can’t be totally certain, my guess is that you’ll be able to get to 3000 mw into 50 ohms via the RCA inputs of the Jotunheim. (Although that will correspond to about 3000 x 50/70 = 2143 mw into the 70 ohm impedance of the LCD-2). However, I doubt you’ll ever want to do that, given that 1 mw (0.001 watts) into the LCD-2 is spec’d as resulting in an SPL of 101 db!

I’m also basing that conclusion on the Mojo’s spec’d maximum output in "line level mode" of 3 volts, and the specified 14 db max gain of the Jotunheim (although it isn’t made clear which combinations of balanced and unbalanced inputs and outputs that 14 db spec applies to). Those numbers should assure that you won’t run out of range on the headphone amp’s volume control.

Finally, be aware that a 3.5 mm to XLR cable will **not** provide balanced signals to the XLR connectors, since a balanced pair of signals for each channel does exist on the 3.5 mm connector. A 3.5 mm stereo connector provides an unbalanced left channel signal, an unbalanced right channel signal, and a ground connection. The cable would simply route those signals to separate XLR connectors, probably on pin 2 of each XLR connector, and would probably connect the ground of the 3.5 mm connector to pins 1 and 3 of each XLR connector. The result being no benefit relative to a 3.5 mm to RCA cable.

Whew! Best regards,
-- Al

Al, you rock.  I've said it before and I'll say it again.  I think I will have to re-read your post a few times to properly digest it and understand it fully!  You have some serious knowledge.  Thank you!

- Mark
+1 @marktomaras.
Al amazes me every time I read and try to digest some information he provides. Amazing, isn't it? And the best part is, there is ZERO bias in his responses.
Maybe we should have a "Almargopedia" section on audiogon :-)
.... and not only does Al possess a vast subject matter knowledge, but his ability to articulate his knowledge, resourcefulness in finding other relevant information (including schematics) on the Internet and overall kind demeanor is exemplary.

I think if there were a "most valuable Audiogoner" award, it should be presented to @almarg

Hello Al,

so, I have read, re-read, researched, and digested. I found some other threads discussing balanced and unbalanced that you also posted on. I now feel that I have a reasonable working understanding of the concept.

I do have some more questions of course!

1. Is the design that Audio Research used in my LS-25 a reasonable way to deal with a balanced signal?  Or is it a “fake” balanced connection? I see they use the same method in their Ref 2 and Ref 3 line stages, Though my ability to accurately read a schematic  is questionable :-)

2. If balanced connections add more gain, why do you suppose Schiit is showing a difference in wattage output based on the balanced versus unbalanced operation in the Jotunheim?  

3.  Is there a way to test the difference in the driving Power into the headphone?  I mean, should I be able to detect an audible difference in wattage delivery If I compare fully balanced operation to fully single ended up operation?  

If I use RCA cables to connect my audio research preamplifier to the Jotunheim, and the standard single ended headphone cable, I should be getting 571mw of power. If I switch back to XLR from the ARC preamp to the Jotunheim, and use my balanced headphone cable, I should be back up to the 2143mw at 70 ohms that Schiit suggests in the specs (after applying your formula to adjust for the different load) But now after your explanation,  I am confused why the Jotunheim is outputting more wattage with the balanced connection at all.

In fact, I am less interested in the amount of gain that the Jotunheim delivers to my LCD-2 headphones, both settings, low gain or high gain deliver enough volume.  I am concerned about the wattage delivery.  I want the power and the control of the wattage to properly drive the headphones.  Audeze recommends 1-4 watts to drive them, and that is one of the reasons I chose the Jotunheim.

Thanks very much for the nice words, gentlemen!

@marktomaras, regarding your questions:

1)The LS25 has a fully balanced/truly balanced/genuinely balanced internal signal path, and if XLR inputs and outputs are used the entire signal path from input to output will be balanced. The approach of having RCA input and output signals common with one of the two signals in the corresponding balanced signal pair, rather than having them handled by a separate circuit stage, is not at all uncommon, as it significantly reduces cost, complexity, and requirements for internal physical space. The downside of that approach which most commonly arises is that if both outputs are used simultaneously, for example to connect the XLR output to a power amp while connecting the RCA output to a powered sub, impedance compatibility issues and issues involving sensitivity to cable capacitance can arise which can affect the signals to both destinations.

Note, for example, that ARC recommends a minimum load impedance of 20K for the LS25, as they do for most of their other line stages and preamps. Powered subs commonly have line level input impedances in the area of 5K to 20K, which in combination with the input impedance of a power amp would result in the preamp "seeing" a load impedance even lower than that.

2)It follows from some of my earlier comments that the 3000 mw vs. 800 mw difference is most likely just a function of which of the two **output** connectors are used, not which input connector is used. Providing the unit with unbalanced inputs, whether via the RCA input connectors or the XLR input connectors using an adapter cable, will probably affect gain but most likely not maximum output power capability. As I indicated previously, though, I can’t say that with total certainty without specific knowledge of the design.

And if I understand correctly there is no reason you can’t use the headphone’s balanced cable for your Mojo/computer desk application as well as for your main system application.

3)No, I can’t think of a convenient and practical way of determining maximum output power capability, without test equipment that most of us don’t have. But even in what I believe to be the unlikely event that using the Jotunheim with unbalanced inputs/balanced outputs provides you with a maximum output capability of only 571 mw, I don’t see that as being an issue. Calculating from the efficiency spec, 571 mw into the headphones would produce an SPL of 129 db! And regarding your reference to "control," that is a function of output impedance, not maximum power capability. Output impedance is specified as "less than 0.1 ohms, balanced or SE, at both gain settings," which is negligible in relation to the 70 ohm impedance of the phones.

Best regards,
-- Al
"The Jotunheim is balanced differential. Meaning it will not turn a single-ended connection to a "Fake" balanced connection. You may notice more power from using the balanced headphone output with a single-ended input but I cannot speculate on that."
Just a FWIW: this response really doesn't make sense! So I can understand there being some confusion.

If the input to the "Jotunheim is balanced differential", then it will not matter if you are driving it single-ended or balanced. This is because a differential amplifier does not care what its inputs are - it just amplifies what is different between its inverting and noninverting inputs. So if one of  those inputs as at ground as is the case with a single-ended input, it will amplify that. So its gain will not change nor will its distortion. That is, if it is balanced differential.

You won't hurt anything by trying the hookup! Our products are also balanced differential and what this means is that it usually does not matter to the gain of the amp if the XLR or RCA connection is used.
Just a FWIW: this response really doesn’t make sense! So I can understand there being some confusion.

And, unfortunately, that is a common theme where Schiit customer service and technical support are concerned.
Wow, I knew this was an Almarg and Atmasphere question.