ozshadow
500+ Head-Fier
- Joined
- Apr 29, 2006
- Posts
- 620
- Likes
- 10
The Shuffle uses direct drive for the headphones, as opposed to AC coupled on the other iPods; this is actually a cost/space saving measure, as the 100-220uF output decoupling capacitors are not required to remove the DC bias of the headphone amp. The chip used in the Shuffle, the Sigmatel 35xx, supports both types of headphone drive - Apple probably chose direct drive in order to save space and make the product more compact.
So, what IS the difference between direct drive and AC coupling? There now follows a long post which many people will find supremely boring
Simply put: to make the "cone" in the headphone's mini-speaker move both in & out with regard to the rest state, you need to put both a positive and negative voltage on it with respect to the headphone's ground pin. Most portable electronics are single-rail, meaning that there are only positive voltage supplies inside the product.
On these AC coupled products, the headphone amp chip outputs a wholly positive signal which idles at a mid voltage, which we'll call Vmid. As capacitors only pass changes in voltage, and not static (or "DC") volages, this offset is not seen at the headphone output, which uses system ground as the audio ground. When the voltage swings up beyond Vmid, this positive change is transferred acrosss the capacitor and provides power to drive the cone out, with the current returning to system ground. Conversely, as the voltage swings below Vmid, the negative change is transmitted to the headphones. Technically, the output capacitor and the resistance of the headphone coil forms a high-pass filter - the low resistance of "decent" headphones, combined with Apple using lower-than-recommended size output caps (saves space & money) mean that the lower the frequency the worse it will be transmitted across the capacitor. Hence, you get bass attenuation when headphones are plugged in.
The actual formula is f=1/(2*PI*R*C) - this is the 3dB down point, ie where the signal is halved in amplitude. For an iPod 3G with 100uF output caps and 16 ohm headphones, this is 1/(2*3.14159265*16*0.0001) = ~100Hz. If you're using 32 ohm earbuds, you get 50Hz, which is lower than the earbuds can deal with anyway). Also note that had Apple used Wolfson's recommended output capacitor size of 220uF, the situation is much better - 45Hz rolloff with 16 ohm headphones.
In general, earbuds are 32 or 64 ohms, and big headphones are 16 ohms (though some, eg the Etymotic ER4P's, are something like 300 ohms - this then brings its own problems due to not being able to transfer much power to the load due to the limited voltage swing available).
Direct drive systems work by making the headphone's ground pin Vmid and just driving the left & right channels around this point. When silent, all 3 wires will be at Vmid (typically ~0.9v-1.7v). There's no output capacitor, so low frequency/DC signals are faithfully reproduced to the limit of the amplifier's drive capability. Sounds ideal, but there *are* problems with this approach. You may have noticed that the iPod shuffle is totally plastic; this is good, because if it was metal and you sat it on, say, a metal-cased hifi amp and plugged the headphone lead into the amp's line in, it'd get very upset. The poor little player would be trying to drive the headphone "ground" pin to Vmid, but it wouldn't be able to as this line is connected to the amp's input ground, which (through the case to case contact) is also its ground. Result - the amp goes into protective shutdown to stop itself being damaged and your audio goes away or crackles furiously as the amp attempts to maintain normality.
People who used certain headphone plugs with the original iPod minis or the original Rio Carbon are witnessing this effect. Note that for "metal amp" you can substitute "USB cable / PC line in" too. See why people, especially those who make dockable players, like AC coupling?
fogd00d's graphs slightly miss the point in that AC coupling, when done properly, is just fine. People don't listen to 20Hz square waves, they listen to music which is rarely square wave based. The same tests with a sine wave - which is what is used in real frequency response testing (and rather closer to what a bass drum or guitar looks like than a 20Hz square wave) - will look a lot cleaner on all the players, AC coupled or not.
There's also a third way, which involves generating a negative voltage rail on the player to deal with the negative output swings. This way, you get to DC-couple your output but still use system ground - the best of both worlds - and often you'll get much more output voltage swing - great for people with problem headphones. Maxim make some headphone amp chips that do this onboard, but the downside is, as always, cost.
So, what IS the difference between direct drive and AC coupling? There now follows a long post which many people will find supremely boring
Simply put: to make the "cone" in the headphone's mini-speaker move both in & out with regard to the rest state, you need to put both a positive and negative voltage on it with respect to the headphone's ground pin. Most portable electronics are single-rail, meaning that there are only positive voltage supplies inside the product.
On these AC coupled products, the headphone amp chip outputs a wholly positive signal which idles at a mid voltage, which we'll call Vmid. As capacitors only pass changes in voltage, and not static (or "DC") volages, this offset is not seen at the headphone output, which uses system ground as the audio ground. When the voltage swings up beyond Vmid, this positive change is transferred acrosss the capacitor and provides power to drive the cone out, with the current returning to system ground. Conversely, as the voltage swings below Vmid, the negative change is transmitted to the headphones. Technically, the output capacitor and the resistance of the headphone coil forms a high-pass filter - the low resistance of "decent" headphones, combined with Apple using lower-than-recommended size output caps (saves space & money) mean that the lower the frequency the worse it will be transmitted across the capacitor. Hence, you get bass attenuation when headphones are plugged in.
The actual formula is f=1/(2*PI*R*C) - this is the 3dB down point, ie where the signal is halved in amplitude. For an iPod 3G with 100uF output caps and 16 ohm headphones, this is 1/(2*3.14159265*16*0.0001) = ~100Hz. If you're using 32 ohm earbuds, you get 50Hz, which is lower than the earbuds can deal with anyway). Also note that had Apple used Wolfson's recommended output capacitor size of 220uF, the situation is much better - 45Hz rolloff with 16 ohm headphones.
In general, earbuds are 32 or 64 ohms, and big headphones are 16 ohms (though some, eg the Etymotic ER4P's, are something like 300 ohms - this then brings its own problems due to not being able to transfer much power to the load due to the limited voltage swing available).
Direct drive systems work by making the headphone's ground pin Vmid and just driving the left & right channels around this point. When silent, all 3 wires will be at Vmid (typically ~0.9v-1.7v). There's no output capacitor, so low frequency/DC signals are faithfully reproduced to the limit of the amplifier's drive capability. Sounds ideal, but there *are* problems with this approach. You may have noticed that the iPod shuffle is totally plastic; this is good, because if it was metal and you sat it on, say, a metal-cased hifi amp and plugged the headphone lead into the amp's line in, it'd get very upset. The poor little player would be trying to drive the headphone "ground" pin to Vmid, but it wouldn't be able to as this line is connected to the amp's input ground, which (through the case to case contact) is also its ground. Result - the amp goes into protective shutdown to stop itself being damaged and your audio goes away or crackles furiously as the amp attempts to maintain normality.
People who used certain headphone plugs with the original iPod minis or the original Rio Carbon are witnessing this effect. Note that for "metal amp" you can substitute "USB cable / PC line in" too. See why people, especially those who make dockable players, like AC coupling?
fogd00d's graphs slightly miss the point in that AC coupling, when done properly, is just fine. People don't listen to 20Hz square waves, they listen to music which is rarely square wave based. The same tests with a sine wave - which is what is used in real frequency response testing (and rather closer to what a bass drum or guitar looks like than a 20Hz square wave) - will look a lot cleaner on all the players, AC coupled or not.
There's also a third way, which involves generating a negative voltage rail on the player to deal with the negative output swings. This way, you get to DC-couple your output but still use system ground - the best of both worlds - and often you'll get much more output voltage swing - great for people with problem headphones. Maxim make some headphone amp chips that do this onboard, but the downside is, as always, cost.
