Why use the inductor in the output of the LM3886?
When should I use?
What it does? (I read the note in the datasheet, but I'm kinda dumb.
)
Thanks!
When should I use?
What it does? (I read the note in the datasheet, but I'm kinda dumb.
)Thanks!
Its called a zobel network. Normally it is a cap and resistor in series with each other, some of the fancier designs use a coil too, to turn it into a thiele network. All it does is eliminate impedance rise with frequency. It also helps alittle with oscillation since the network makes it so the amp cant output above 150khz frequencies.
Cue in Andrew.
Cue in Andrew.
Last edited:
the output inductor can be seen as "disconnecting" the amp from the cable/load at very high frequencies ( > 200 KHz, to have little effect on audio)
this is most needed when there is lots of hi Q capacitance most commonly from long cable runs, there are a few capacitve tweeter technologies out there too
if there is too much C seen directly by the chip amp output at frequencies near the loop gain intercept it can cause oscillations
this more important when using the minimum gain of 10, higher gains actually mean less feedback, are typically more stable
if the chip amp is being used in a powered monitor with a dynamic tweeter only inches of wire away you can probably skip the inductor
this is most needed when there is lots of hi Q capacitance most commonly from long cable runs, there are a few capacitve tweeter technologies out there too
if there is too much C seen directly by the chip amp output at frequencies near the loop gain intercept it can cause oscillations
this more important when using the minimum gain of 10, higher gains actually mean less feedback, are typically more stable
if the chip amp is being used in a powered monitor with a dynamic tweeter only inches of wire away you can probably skip the inductor
Last edited:
Besides the primary function that has already been explained well, above, the inductor and resistor ALSO help to keep RF from infiltrating into the amplifier circuit, through the output. Some speaker cables, and some types of planar speakers, can make great antennas. RF getting around the feedback loop and into the inputs can degrade the sound quality.
I would want to ALWAYS use the inductor and the parallel resistor.
This seems to be one of those cases where the datasheet doesn't tell you anything about what types of components should be used. So here you go:
At audio frequencies, basically everything will be going through the inductor (and not the resistor). So the inductor's type and quality are very important to your sound quality. I would use an air core inductor, there, to avoid the distortion from hysteresis that other core types would have, and to avoid the possibility of saturation that other core types would have. The inductor needs to be able to carry a lot of current (i.e. ALL of the output current), so it should be made with large-enough-diameter solid magnet wire.
You could VERY easily wind your own coil, in this case. There is a coil calculator at Pronine Electronics Design - Multilayer Air Core Inductor Calculator .
I entered 0.7 uH and a coil inner diameter and coil length of 0.66 inches (for both), and specified 14 AWG wire, and it calculated that you would need to use exactly 9 turns of 14 AWG wire on an inner diameter of 0.66 inches (i.e. all turns are of the same diameter, in a single layer).
It also reported that it would require 1.56 feet of wire (but you would need to leave extra length for the leads). You should use "magnet wire", which has a thin coating of insulating lacquer or epoxy or something.
The 10-Ohm parallel resistor basically only conducts at very high frequencies. So there "shouldn't" be too much power dissipated in it and it shouldn't affect the sound quality. But if your amp somehow oscillated at a high frequency, then in that case there could be a lot of power dissipated by the resistor. I don't know what is a typical power dissipation rating that is used, for that resistor. But the rectangular cement-encased power resistors are cheap, even for high wattages, and they are fairly rugged.
To be very safe, with overkill, I guess you "could" use a resistor power rating of twice the maximum output power that your amp could produce, and then optionally replace it with a lower-wattage resistor after you see whether or not the amp oscillates. But I would probably just start with a 5-Watt cement type, probably a vertical-mount one, like the Xicon model that should be part # 280-PRM5-10-RC at mouser.com. They should still be under $0.50 each for quantity 1.
And as a free bonus, if that resistor gets warm to the touch (or gets so hot that it instantly vaporizes your skin, or catches fire or explodes), then you won't need an oscilloscope to know if your amp is oscillating.
Have fun!
Tom
I would want to ALWAYS use the inductor and the parallel resistor.
This seems to be one of those cases where the datasheet doesn't tell you anything about what types of components should be used. So here you go:
At audio frequencies, basically everything will be going through the inductor (and not the resistor). So the inductor's type and quality are very important to your sound quality. I would use an air core inductor, there, to avoid the distortion from hysteresis that other core types would have, and to avoid the possibility of saturation that other core types would have. The inductor needs to be able to carry a lot of current (i.e. ALL of the output current), so it should be made with large-enough-diameter solid magnet wire.
You could VERY easily wind your own coil, in this case. There is a coil calculator at Pronine Electronics Design - Multilayer Air Core Inductor Calculator .
I entered 0.7 uH and a coil inner diameter and coil length of 0.66 inches (for both), and specified 14 AWG wire, and it calculated that you would need to use exactly 9 turns of 14 AWG wire on an inner diameter of 0.66 inches (i.e. all turns are of the same diameter, in a single layer).
It also reported that it would require 1.56 feet of wire (but you would need to leave extra length for the leads). You should use "magnet wire", which has a thin coating of insulating lacquer or epoxy or something.
The 10-Ohm parallel resistor basically only conducts at very high frequencies. So there "shouldn't" be too much power dissipated in it and it shouldn't affect the sound quality. But if your amp somehow oscillated at a high frequency, then in that case there could be a lot of power dissipated by the resistor. I don't know what is a typical power dissipation rating that is used, for that resistor. But the rectangular cement-encased power resistors are cheap, even for high wattages, and they are fairly rugged.
To be very safe, with overkill, I guess you "could" use a resistor power rating of twice the maximum output power that your amp could produce, and then optionally replace it with a lower-wattage resistor after you see whether or not the amp oscillates. But I would probably just start with a 5-Watt cement type, probably a vertical-mount one, like the Xicon model that should be part # 280-PRM5-10-RC at mouser.com. They should still be under $0.50 each for quantity 1.
And as a free bonus, if that resistor gets warm to the touch (or gets so hot that it instantly vaporizes your skin, or catches fire or explodes), then you won't need an oscilloscope to know if your amp is oscillating.
Have fun!
Tom
Last edited:
Gootee,
I saw some design got a capacitor in series with parallel resistor. What is the function of it?
Thanks.
I saw some design got a capacitor in series with parallel resistor. What is the function of it?
Thanks.
It depends where that is used. But in general, at low frequencies where the capacitor looks like an open circuit, only the resistor matters. At higher frequencies, where the capacitor can conduct, the impedance would be the parallel impedances of the resistor and the capacitor, with the capacitor's impedance being inversely proportional to frequency times the capacitance. At high-enough frequencies, the capacitor would act like a short circuit and the resistor would have very little effect. So a capacitor in parallel with a resistor can function as a high-pass filter.
Tom
Last edited:
- Status
- Not open for further replies.