[Jaywalk3r]

What are the practical advantages of a push-pull output compared to a single-ended output, aside from superior sustained bass output?

I have a 1st generation iPod Shuffle, which came equipped with the SigmaTel ST-MP3-5xx. As I understand it (and I could be mistaken), this SOC provides the Shuffle with push-pull output. Testing resulted in superior bass performance by the Shuffle 1G compared to the iPod 3G (Wolfson) and the iPod Mini (Cirrus). What are the other advantages and disadvantages of a push-pull output?

 

[Salm0n]

I'm no amp expert, but i do understand that push pull amps have less harmonic distortion due to cancelation of even order harmonics. i believe that they are also more energy efficient, but im not sure on that one.

 

[stv014]

Quote:

i believe that they are also more energy efficient, but im not sure on that one.

Yes, it is more energy efficient, especially if implemented as AB or B class, and most modern solid state amplifiers are push-pull in AB class.
 

 

[jcx]

I think you are reading badly worded, confusing web posts, an engineer doesn't use single-ended/push-pull to distinguish DC coupled output from AC coupled with DC blocking Cap
 
the amp chip in the 1st Shuffle uses a "active splitter" to avoid the DC blocking output Cap in series with each channel
 
I emailed to try to get Machrone to fix his wording, also pointed out that he misinterpreted some obviously clipping waveform in his square wave test - he doesn't care enough to make the needed corrections, everyone seems to have just kept on citing bad info

 

[Jaywalk3r]

I think you are reading badly worded, confusing web posts, an engineer doesn't use single-ended/push-pull to distinguish DC coupled output from AC coupled with DC blocking Cap

the amp chip in the 1st Shuffle uses a "active splitter" to avoid the DC blocking output Cap in series with each channel

Thank you for the clarification. I'm confident we are referring to the same test. (http://home.comcast.net./~machrone/playertest/playertest.htm)

Could you please clarify the benefits and drawbacks of using an "'active splitter' to avoid the DC blocking output Cap in series with each channel"? (I assume there is a benefit, even if it is only cost or physical size.) Is that different from a push-pull output?

Please forgive my ignorance. I'm a math guy, not a sound engineer. Still, I like to learn new stuff.

 

[BlackbeardBen]

Quote:

I'm no amp expert, but i do understand that push pull amps have less harmonic distortion due to cancelation of even order harmonics.

Just to clarify, the cancellation of even order harmonics is with "balanced" amplifiers which is an entirely different distinction from amplifier class (i.e. A, A/B, B, D).  The gist of it is it's two amplifiers running with opposite polarity - the reverse polarity signal at the "negative" speaker/headphone terminal.
 
Remember that the transducer only "sees" the voltage potential (difference) between the two terminals - headphones and speakers are fundamentally electrically balanced devices.  Thus, there is nothing "more balanced" about how the transducer is driven - the main advantages are of course in the near-elimination of even order distortion and the potential benefit of using two less powerful amplifiers instead of one more powerful one.  It also means that each amplifier sees half the impedance of the load, so with many amplifiers the level of distortion is increased and the stability decreased.  That's why bridged speaker amps can't handle low impedance loads well.
 
Oh, and as an aside, the doubled slew rate of a balanced amplifier design is a non-issue in audio applications.

 

[fubar3]

Quote:

 

What are the other advantages and disadvantages of a push-pull output?

I believe you need a center tapped transformer for a push-pull design.  These were often found in transistor radios of old.  These circuits were simple and provided some even harmonic cancellation.  But transformers add cost and weight, plus color the sound.  They are easily eliminated in modern direct-coupled circuits.
 

 

[jcx]

http://datasheets.maxim-ic.com/en/ds/MAX4411-MAX4411B.pdf is an example of the principle - I can't find a tear down analysis for the 1st gen shuffle with part #
 
the maxim op amp chip has extra circuitry to invert the (single polarity battery) supply V with a high frequency switched C converter using a very small cap
 
there is only one amp per channel, the Input has to be AC coupled but the input impedance is so high that only 2 small caps are needed
 
the output swings symetrically about the negative supply of the battery = gnd of the switching supply inverter, using its own switched C inverted power for the swings below "gnd"

 

[Jaywalk3r]

I can't find a tear down analysis for the 1st gen shuffle with part #

The first generation Shuffle uses the SigmaTel ST-MP3-5xx. The data sheet for the chip is here.

 

[jcx]

thanx, the Headphone amplifier section shows a "common mode amplifier" so this is more like a tl2426 buffered active gnd "supply splitter" - so there are  3 amps, one buffering 1/2 Vbatt and is used as "gnd" for both headphone R,L drivers