Power supply
As I mentioned already, I want to use a very simple passive PSU for this amplifier - mainly to keep it simple and discrete, just like the rest of the project. Also, I don't want any power semiconductors in the PSU-section that would need cooling. For this reason, I want to go with a simple RLC-lowpass filter that smooths out any noise that may come from the external 24 V power-supply, and will provide each channel with its own filter to avoid crosstalk.
Schematic
Starting off, the core of the PSU is made up of two parallel resistors isolating the external power supply from the passive load of the RLC low-pass. While an inductive load should not be a big problem for the internal control-loops of most AC/DC converters, I don't want any risk for incompatibilities and oscillations. After that, the power inductor follows. This is a Murata 1400-series power inductor of 4.7 mH, capable of a DC of 600 mA. This should be just about perfect for this application without pushing the limits of the inductor too much. In parallel to the inductor, a flyback-diode is placed to avoid voltage spikes at switch-off. Another two parallel resistors follow between the inductor and the bulk capacitance, to avoid oscillations forming in this LC-network. The bulk capacitance is the last part of the circuit and provides a low-impedance voltage-source for the amplifier. To avoid large inrush-currents, a NTC is placed at the input of the filter network. The exact value of this has to be figured out, but this should be enough to avoid any damage to the amplifier or the AC/DC converter.
Simulation
The network should provide roughly 65 dB of attenuation at 1 kHz and over 100 dB at 10 kHz. With the high switching frequencies of well over 10 kHz of most power-supplies, as well as the PSRR of over 50 dB or so of the amplifier, this should avoid any audible coupling of PSU-noise onto the output of the amplifier.
PCB
Just like the other PCB, this one has two layers with an (almost) solid ground plane on the bottom and large copper pours for the connections between the components. In the screenshot, the input is on the left. It directly connects to the NTC and then into the first bank of parallel power resistors. Each of these resistors can dissipate 2 W, so both of them result in a 2.35 Ohm resistor capable of 4 W dissipation. With a nominal dissipation of R*I^2 = around 0.59 W, this should be plenty of headroom to also cover current bursts due to inrush-current or fault conditions. At this point, the diode parallel to the inductor is also placed. The power-inductor follows, with its output feeding the second resistor bank. lastly, the bulk capacitance of 10 mF made up of three 3300 uF capacitors is connected.
The design of the PCB is again kept as symmetric as possible. To keep the amplifier reasonably thin, this PCB will lie on its side. Each channel will have its own PSU section, to avoid any crosstalk between the channels due to the passive PSU design. The symmetric PCB allows me to lie it on one side for the one channel and on the other side for the other channel, keeping the amp symmetric as a whole.
All in all, this is a really simple approach, but I think it will work well in this amp. I just want to see how simple an amp can really get while keeping good sonic performance, and am excited to see how all of this will work together.
As always, thank you for all the previous and future replies!
As I mentioned already, I want to use a very simple passive PSU for this amplifier - mainly to keep it simple and discrete, just like the rest of the project. Also, I don't want any power semiconductors in the PSU-section that would need cooling. For this reason, I want to go with a simple RLC-lowpass filter that smooths out any noise that may come from the external 24 V power-supply, and will provide each channel with its own filter to avoid crosstalk.
Schematic
Starting off, the core of the PSU is made up of two parallel resistors isolating the external power supply from the passive load of the RLC low-pass. While an inductive load should not be a big problem for the internal control-loops of most AC/DC converters, I don't want any risk for incompatibilities and oscillations. After that, the power inductor follows. This is a Murata 1400-series power inductor of 4.7 mH, capable of a DC of 600 mA. This should be just about perfect for this application without pushing the limits of the inductor too much. In parallel to the inductor, a flyback-diode is placed to avoid voltage spikes at switch-off. Another two parallel resistors follow between the inductor and the bulk capacitance, to avoid oscillations forming in this LC-network. The bulk capacitance is the last part of the circuit and provides a low-impedance voltage-source for the amplifier. To avoid large inrush-currents, a NTC is placed at the input of the filter network. The exact value of this has to be figured out, but this should be enough to avoid any damage to the amplifier or the AC/DC converter.
Simulation
The network should provide roughly 65 dB of attenuation at 1 kHz and over 100 dB at 10 kHz. With the high switching frequencies of well over 10 kHz of most power-supplies, as well as the PSRR of over 50 dB or so of the amplifier, this should avoid any audible coupling of PSU-noise onto the output of the amplifier.
PCB
Just like the other PCB, this one has two layers with an (almost) solid ground plane on the bottom and large copper pours for the connections between the components. In the screenshot, the input is on the left. It directly connects to the NTC and then into the first bank of parallel power resistors. Each of these resistors can dissipate 2 W, so both of them result in a 2.35 Ohm resistor capable of 4 W dissipation. With a nominal dissipation of R*I^2 = around 0.59 W, this should be plenty of headroom to also cover current bursts due to inrush-current or fault conditions. At this point, the diode parallel to the inductor is also placed. The power-inductor follows, with its output feeding the second resistor bank. lastly, the bulk capacitance of 10 mF made up of three 3300 uF capacitors is connected.
The design of the PCB is again kept as symmetric as possible. To keep the amplifier reasonably thin, this PCB will lie on its side. Each channel will have its own PSU section, to avoid any crosstalk between the channels due to the passive PSU design. The symmetric PCB allows me to lie it on one side for the one channel and on the other side for the other channel, keeping the amp symmetric as a whole.
All in all, this is a really simple approach, but I think it will work well in this amp. I just want to see how simple an amp can really get while keeping good sonic performance, and am excited to see how all of this will work together.
As always, thank you for all the previous and future replies!
Attachments
