I spent my quarantine the past year designing a particular preamp and its now my obsession to tweak it the best quality possible.
My concern now is the input grounding. Ive been using a star ground method to summate the input ground, output ground and the feedback resistor ground to a single point as close to the regulator output as possible and thats been hum free so far and sounding good.
But thinking about it wouldnt it be a better practice to summate the input ground and the output ground to the feedback resistor node (of the amplifying circuit) then trace that with one wire to the psu ground? Because after all the feedback ground is where the correction occurs.
Please enlighten me
Thanks
My concern now is the input grounding. Ive been using a star ground method to summate the input ground, output ground and the feedback resistor ground to a single point as close to the regulator output as possible and thats been hum free so far and sounding good.
But thinking about it wouldnt it be a better practice to summate the input ground and the output ground to the feedback resistor node (of the amplifying circuit) then trace that with one wire to the psu ground? Because after all the feedback ground is where the correction occurs.
Please enlighten me
Thanks
Any conductor with current flowing in it cannot be 'ground' at both ends, because current flow requires a voltage drop. If there is no voltage drop there is nothing to make current flow.
Regarding your question, if the one wire to the psu ground has current flowing in it to power the circuit, then it can only be at ground potential at one end. The other end will have a 'noise' voltage on it that is a combination of both the circuit bias current (including ripple) and current that is correlated with the signal being amplified.
If these voltages appear in series with the wanted signal, they will be amplified by the next gain stage which cannot discriminate between wanted signal and noise signal, but simply amplifies the difference between signal + and signal ground+noise.
Regarding your question, if the one wire to the psu ground has current flowing in it to power the circuit, then it can only be at ground potential at one end. The other end will have a 'noise' voltage on it that is a combination of both the circuit bias current (including ripple) and current that is correlated with the signal being amplified.
If these voltages appear in series with the wanted signal, they will be amplified by the next gain stage which cannot discriminate between wanted signal and noise signal, but simply amplifies the difference between signal + and signal ground+noise.
Depends in part on the type of preamp. A line level preamp can be made that works quite well using a ground plane. A moving coil phono preamp can be a different matter. Way to learn is to study grounding techniques and ground problem troubleshooting, and by building and experimenting with circuits.
https://www.jensen-transformers.com/wp-content/uploads/2014/08/generic-seminar.pdf
https://www.jensen-transformers.com/wp-content/uploads/2014/08/generic-seminar.pdf
Last edited:
I think a better question to ask might be: what are the kinds of noise that can be introduced into the signal when multiple sources are connected to the same circuit?
You have used star grounding in your preamplifier, however star grounding generally can not be maintained throughout the audio system, and this is the problem. Star grounding depends on having signals having only one return path for signal currents and so signals do not share return paths. This avoids crosstalk being introduced between the signals because of the common voltage drops over the shared path. However, star grounding does not prevent other kinds of ways that noise can be introduced.
To understand how noise can be introduced, we must keep in mind that every closed loop of conductor is a loop antenna for magnetic fields, and every high impedance isolated conductor acts like the plate of a capacitor and picks up electric fields. So to minimize magnetic pickup, we must either minimize loop area or break loops. To minimize electric pickup, we have to place a grounded electrostatic shield around a conductor, or space the conductor away from high voltage noisy conductors.
Furthermore, we have to distinguish between two types of noise: differential mode noise and common mode noise. Because the audio signal is transmitted as the difference in voltage between two conductors, if there is a differential noise source that introduces a difference in voltage between them, that contributes to the noise at the output. Common-mode noise adds a common voltage to both conductors, and so in principle does not contribute noise. However, common-mode noise is often converted to differential noise and therefore is audible.
Magnetic fields can introduce differential mode noise as it introduces a noise voltage inside the loop formed by the two conductors of a signal cable. Electric fields will generally introduce common-mode noise, as both conductors will be charged nearly the same if they are in the presence of an electric field. However, common-mode noise can be converted to differential mode if, for example, the two conductors are charged unequally due to having an unequal capacitance. The coaxial nature of audio cables tends to help both types, with the surrounding shield both minimizing the loop area with the inner conductor and serving as an electrostatic shield.
The ground loops for several sources attached to a preamplifier comes about because each source is generally connected to the amplifier both through the conductors of the audio cable, and through the safety ground to which many sources are also grounded. The preamplifier is also generally connected to safety ground as well. The safety ground and external signal ground of the audio cable form a large loop which can pick up magnetic fields. Because the highest currents in the environment tend to be mains power cables, this produces hum. Similarly, mains voltages tend to be higher and therefore can electrostatically introduce charge into nearby conductors.
So one could disconnect the safety grounds to remove the loop (not recommended) or use a "ground lifter" which inserts heavy duty diodes into the safety ground paths to provide a path for dangerous currents. However, this does not help a preamp much, as the loops are formed between the equipment connected to the preamp as well as the preamp itself. Furthermore, an improperly implemented ground lifter can be a serious safety hazard, and I don't think regulatory agencies would approve of it, and therefore I wouldn't use it. Furthermore for a preamp, a ground lifter would need to be used on all the components (or all but one) to break all of the ground loops.
Another approach is to use a hum breaker resistor. This breaks the ground loops between the components by introducing a small resistance into the shield connection of the audio signal (typically 5 to 100 ohms). This reduces the current in the loop so that the voltage of the noise on the shield of the audio cable is reduced.
If you are going to reference the op amp amplifier stages all to the input ground, and pass the input ground through, then you are going to have to isolate the input ground from the preamp supply voltages, but provide enough of a path for a return current. Something like a hum breaker resistor may work for this. I used a similar approach in my phono preamp. But this assumes that the source of noise that you are trying to address is ground loops.
You have used star grounding in your preamplifier, however star grounding generally can not be maintained throughout the audio system, and this is the problem. Star grounding depends on having signals having only one return path for signal currents and so signals do not share return paths. This avoids crosstalk being introduced between the signals because of the common voltage drops over the shared path. However, star grounding does not prevent other kinds of ways that noise can be introduced.
To understand how noise can be introduced, we must keep in mind that every closed loop of conductor is a loop antenna for magnetic fields, and every high impedance isolated conductor acts like the plate of a capacitor and picks up electric fields. So to minimize magnetic pickup, we must either minimize loop area or break loops. To minimize electric pickup, we have to place a grounded electrostatic shield around a conductor, or space the conductor away from high voltage noisy conductors.
Furthermore, we have to distinguish between two types of noise: differential mode noise and common mode noise. Because the audio signal is transmitted as the difference in voltage between two conductors, if there is a differential noise source that introduces a difference in voltage between them, that contributes to the noise at the output. Common-mode noise adds a common voltage to both conductors, and so in principle does not contribute noise. However, common-mode noise is often converted to differential noise and therefore is audible.
Magnetic fields can introduce differential mode noise as it introduces a noise voltage inside the loop formed by the two conductors of a signal cable. Electric fields will generally introduce common-mode noise, as both conductors will be charged nearly the same if they are in the presence of an electric field. However, common-mode noise can be converted to differential mode if, for example, the two conductors are charged unequally due to having an unequal capacitance. The coaxial nature of audio cables tends to help both types, with the surrounding shield both minimizing the loop area with the inner conductor and serving as an electrostatic shield.
The ground loops for several sources attached to a preamplifier comes about because each source is generally connected to the amplifier both through the conductors of the audio cable, and through the safety ground to which many sources are also grounded. The preamplifier is also generally connected to safety ground as well. The safety ground and external signal ground of the audio cable form a large loop which can pick up magnetic fields. Because the highest currents in the environment tend to be mains power cables, this produces hum. Similarly, mains voltages tend to be higher and therefore can electrostatically introduce charge into nearby conductors.
So one could disconnect the safety grounds to remove the loop (not recommended) or use a "ground lifter" which inserts heavy duty diodes into the safety ground paths to provide a path for dangerous currents. However, this does not help a preamp much, as the loops are formed between the equipment connected to the preamp as well as the preamp itself. Furthermore, an improperly implemented ground lifter can be a serious safety hazard, and I don't think regulatory agencies would approve of it, and therefore I wouldn't use it. Furthermore for a preamp, a ground lifter would need to be used on all the components (or all but one) to break all of the ground loops.
Another approach is to use a hum breaker resistor. This breaks the ground loops between the components by introducing a small resistance into the shield connection of the audio signal (typically 5 to 100 ohms). This reduces the current in the loop so that the voltage of the noise on the shield of the audio cable is reduced.
If you are going to reference the op amp amplifier stages all to the input ground, and pass the input ground through, then you are going to have to isolate the input ground from the preamp supply voltages, but provide enough of a path for a return current. Something like a hum breaker resistor may work for this. I used a similar approach in my phono preamp. But this assumes that the source of noise that you are trying to address is ground loops.
Here's my bible regarding grounding: https://www.hypex.nl/img/upload/doc/an_wp/WP_The_G_word.pdf
I highly recommend reading it more than once... I designed my power amplifier layout around these principles. There's only one ground reference on the whole board and that is the output speakon's ground connection. From there an extra trace goes to the feeback resistors and the servo's noninverting input. The rest of the pcb is just a single, solid ground plane for power supply decoupling and running the power supply current to the output speakon. If anybody wants to have a look: Jurgen Herrmann / AMP-DR * GitLab
I highly recommend reading it more than once... I designed my power amplifier layout around these principles. There's only one ground reference on the whole board and that is the output speakon's ground connection. From there an extra trace goes to the feeback resistors and the servo's noninverting input. The rest of the pcb is just a single, solid ground plane for power supply decoupling and running the power supply current to the output speakon. If anybody wants to have a look: Jurgen Herrmann / AMP-DR * GitLab
For me best GND management is only possible in cases, where exist independent power supplies include separate transformer winding for each gain stage.
At least an independent power supply for each channel of line stage and each channel of phono stage, if unbalanced mode is chosen.
In cases of only one power supply for the whole pre-amp device you don't get best acoustical performance - amongst other things a star grounding is necessary and no possibility for not isolated mounting of the cinch/RCA plugs at the enclosure.
Basically an disadvantage but unavoidable with a single power supply.