Potato Cakes
Well-known member
Hello, Everyone,
I had created several threads over the last couple of years regarding my struggle with SMPS supplies and what was needed to get them to work with class A microphone preamps and other circuits that used a BA283 style amplifier. I have now finally found a solution to this problem of mine and I would like to share everything here in one spot instead of replying to the other threads.
First, the audio circuit is the preamp section of a 1073 channel strip. I have been building this circuit using kits and boards from some of the exceptional members here. I then began working on a solution for IEM mixing for singers and found myself using BA amp circuit blocks to achieve the results that I wanted and in prototypes solved the complaints one would get from singers when using wireless IEM transmitters. The goal was a self contained circuit with PSU in a 1U chassis. There were also going to be components that needed a bipolar voltages, so that would need to be taken into consideration. I knew from previous builds of this circuit that this particular preamp is very susceptible to noise, especially from a linear PSU using a toroidal transformer. I started to experiment with switch mode power supplies and found that the individual amp sections could be powered from a 500 series power supply using the +/-VDC connections and it was very quiet. I moved onto using an enclosed 24V single source SMPS with no filtering for the entire circuit and after some rerouting of the ground connections, I had a very quiet unit. I also found that when using modern equivalent components that I the circuit could be powered using a 48V single source power supply (convenient for +48V) but it would need a bigger heatsink as this made the output transistor become much hotter than before. However using a +/-24V PSU gave me the same headroom but did not generate nearly the same amount of heat.
After having success with the non-filtered 24V PSU I started building the next prototype and used more compact SMPS units as I was going to need two of them to make a bipolar power supply. I did not use any filtering as on the first build but now there was all of this oscillating and other unwanted noises. I did not know why this was happening as these were medical grade SMPSs which are supposed to have much lower noise than the low cost versions. This is when I started posting inquiries regarding using SMPSs and proceeded buy more units from different manufacturers but never actually finding a solution. I went back to the original PSU and the noise went away. It seemed that I randomly found an off the shelf switching power supply that was low noise in this circuit.
More inquiries and internet searching regarding my issue and I was lead to CLC filtering, which did not seem to solve the issue. I was told to look at capacitance multipliers by Bo Deadly (thanks!) but at the time when I was reading up on them the information didn't make sense. It was not until I was directed to the same articles as I previously been reading that I found a schematic with values I could try. I made of these circuits and all of my noise issues are gone. In fact they are better than any result previously achieved. Attached is an image of the noise of the whole preamp circuit set to unity gain after being burned in and another one when 0dBu of pink noise is applied to the input using the same gain setting. The noise floor does not seem to seem to increase with changing the gain. These results were achieved with SMPS units that were previously causing excessive noise.
The schematic for the capacitance multiplier that was implemented with the noise figures shown. I used 2200uF, 220uF, and TIP122 for 4700uF, 470uF, and TIP35 respectively. These values can be tuned to best suit the capacitance rating of the type of SMPS used as well as voltage needed.
I am going to do more testing with this circuit and other makes and models of SMPSs. For now this is a big sign of relief having solved an issue that has been plaguing me for nearly three years.
Thanks!
Paul
EDIT: The first image was with the preamp cranked to +40 and no input signal after the burn in process. The second image is with the gain set to unity with 0dBu pink noise injected to the input.
I had created several threads over the last couple of years regarding my struggle with SMPS supplies and what was needed to get them to work with class A microphone preamps and other circuits that used a BA283 style amplifier. I have now finally found a solution to this problem of mine and I would like to share everything here in one spot instead of replying to the other threads.
First, the audio circuit is the preamp section of a 1073 channel strip. I have been building this circuit using kits and boards from some of the exceptional members here. I then began working on a solution for IEM mixing for singers and found myself using BA amp circuit blocks to achieve the results that I wanted and in prototypes solved the complaints one would get from singers when using wireless IEM transmitters. The goal was a self contained circuit with PSU in a 1U chassis. There were also going to be components that needed a bipolar voltages, so that would need to be taken into consideration. I knew from previous builds of this circuit that this particular preamp is very susceptible to noise, especially from a linear PSU using a toroidal transformer. I started to experiment with switch mode power supplies and found that the individual amp sections could be powered from a 500 series power supply using the +/-VDC connections and it was very quiet. I moved onto using an enclosed 24V single source SMPS with no filtering for the entire circuit and after some rerouting of the ground connections, I had a very quiet unit. I also found that when using modern equivalent components that I the circuit could be powered using a 48V single source power supply (convenient for +48V) but it would need a bigger heatsink as this made the output transistor become much hotter than before. However using a +/-24V PSU gave me the same headroom but did not generate nearly the same amount of heat.
After having success with the non-filtered 24V PSU I started building the next prototype and used more compact SMPS units as I was going to need two of them to make a bipolar power supply. I did not use any filtering as on the first build but now there was all of this oscillating and other unwanted noises. I did not know why this was happening as these were medical grade SMPSs which are supposed to have much lower noise than the low cost versions. This is when I started posting inquiries regarding using SMPSs and proceeded buy more units from different manufacturers but never actually finding a solution. I went back to the original PSU and the noise went away. It seemed that I randomly found an off the shelf switching power supply that was low noise in this circuit.
More inquiries and internet searching regarding my issue and I was lead to CLC filtering, which did not seem to solve the issue. I was told to look at capacitance multipliers by Bo Deadly (thanks!) but at the time when I was reading up on them the information didn't make sense. It was not until I was directed to the same articles as I previously been reading that I found a schematic with values I could try. I made of these circuits and all of my noise issues are gone. In fact they are better than any result previously achieved. Attached is an image of the noise of the whole preamp circuit set to unity gain after being burned in and another one when 0dBu of pink noise is applied to the input using the same gain setting. The noise floor does not seem to seem to increase with changing the gain. These results were achieved with SMPS units that were previously causing excessive noise.
The schematic for the capacitance multiplier that was implemented with the noise figures shown. I used 2200uF, 220uF, and TIP122 for 4700uF, 470uF, and TIP35 respectively. These values can be tuned to best suit the capacitance rating of the type of SMPS used as well as voltage needed.
I am going to do more testing with this circuit and other makes and models of SMPSs. For now this is a big sign of relief having solved an issue that has been plaguing me for nearly three years.
Thanks!
Paul
EDIT: The first image was with the preamp cranked to +40 and no input signal after the burn in process. The second image is with the gain set to unity with 0dBu pink noise injected to the input.
Attachments
Last edited: