Hi all,
I’d like to have your help about a 50Hz problem I’m facing with my diy preamp.
Few days ago, I already posted this question in Salas hotrodded DCB1 thread but considering it’s not a DCB1 problem I guess it’s better to have a separate thread.
I completed the preamp you can see in the attached image:
1. LDR attenuator managed by an Arduino to have full fine control of the balance and the attenuation through digital pot;
2. DCB1 buffer to interface the LDR to the final amp. The schema and resistance values are the ones you can find in DCB1 thread, no change in that;
3. power supply based on Salas shunt reg (I just added a +5V and +12V lines using a standard 78xx+TIP32 configuration and a relais to activate the shunt reg itself;
4. all the logic managed by the Arduino (mute/unmute, input selector, soft latch power on/off, management of the attenuation level and selected input when turned off…).
Now, everything works just fine, the only issue is the 50Hz noise when I put the preamp in mute. A part from the volume level, when the selected input is floating I can read from 0 to 2mV AC as output (mainly 50Hz). As soon as I put the preamp in mute, the AC voltage goes to 120-150mV and I can hear it from the speakers.
It seems to me that the problem is in the 220V stage because:
1. the noise decreased a lot when I removed all the faston I used to connect 220V plug, fuse and trafo. Consider that before removing all the faston (directly soldering the wires), the 50Hz noise was so strong that I could hear it through my final amp even if the pre amp was switched off 😱
2. if I measure the Vac at the output of the power supply I have less than 1mV (including +5V and +12V lines)
Now, few other clues:
1. consider that the case is not connected to 220V earth or to the ground of the preamp;
2. as a test, I moved all the 220V component outside the case and closed it but the noise was still there;
3. when in mute, the Vac decrease from 150mV to 10mV when I shortcut the ground of the RCA output plug to the aluminum case with my hand
I also tried to directly connect the audio ground to the 220V earth with a very little impact on the noise. Considering the last scenario, do I have to connect the 220V earth to the case and to the GND?
Any test I could do or any good advice?
I’d like to have your help about a 50Hz problem I’m facing with my diy preamp.
Few days ago, I already posted this question in Salas hotrodded DCB1 thread but considering it’s not a DCB1 problem I guess it’s better to have a separate thread.
I completed the preamp you can see in the attached image:
1. LDR attenuator managed by an Arduino to have full fine control of the balance and the attenuation through digital pot;
2. DCB1 buffer to interface the LDR to the final amp. The schema and resistance values are the ones you can find in DCB1 thread, no change in that;
3. power supply based on Salas shunt reg (I just added a +5V and +12V lines using a standard 78xx+TIP32 configuration and a relais to activate the shunt reg itself;
4. all the logic managed by the Arduino (mute/unmute, input selector, soft latch power on/off, management of the attenuation level and selected input when turned off…).
Now, everything works just fine, the only issue is the 50Hz noise when I put the preamp in mute. A part from the volume level, when the selected input is floating I can read from 0 to 2mV AC as output (mainly 50Hz). As soon as I put the preamp in mute, the AC voltage goes to 120-150mV and I can hear it from the speakers.
It seems to me that the problem is in the 220V stage because:
1. the noise decreased a lot when I removed all the faston I used to connect 220V plug, fuse and trafo. Consider that before removing all the faston (directly soldering the wires), the 50Hz noise was so strong that I could hear it through my final amp even if the pre amp was switched off 😱
2. if I measure the Vac at the output of the power supply I have less than 1mV (including +5V and +12V lines)
Now, few other clues:
1. consider that the case is not connected to 220V earth or to the ground of the preamp;
2. as a test, I moved all the 220V component outside the case and closed it but the noise was still there;
3. when in mute, the Vac decrease from 150mV to 10mV when I shortcut the ground of the RCA output plug to the aluminum case with my hand
I also tried to directly connect the audio ground to the 220V earth with a very little impact on the noise. Considering the last scenario, do I have to connect the 220V earth to the case and to the GND?
Any test I could do or any good advice?
Attachments
From these two clues, it seems to me a grounding problem. A circuit diagram would be helpful to say more.
"Earth" wire must be connected to the case. Not to circuit GND.
The third wire in the mains cable is Protective Earth (PE). It is there to help you stop killing you or any operator of your electrical equipment.
The PE must be mechanically and permanently attached to the Chassis. No exceptions.
If you don't know that then STOP !
Don't try to build any mains powered equipment.
There is a second rule.
All exposed conductive parts should be connected to the protected Chassis.
These two Safety rules must be adhered to.
Note in all of the above there is no audio.
Audio circuits and how they are wired has nothing to do with those Safety rules.
The PE must be mechanically and permanently attached to the Chassis. No exceptions.
If you don't know that then STOP !
Don't try to build any mains powered equipment.
There is a second rule.
All exposed conductive parts should be connected to the protected Chassis.
These two Safety rules must be adhered to.
Note in all of the above there is no audio.
Audio circuits and how they are wired has nothing to do with those Safety rules.
You are 100% right and I well understand your strong advice: I've already bought faston to connect any single side of the aluminum case to the PE.
The reason why the case has not already been connected to the 'earth' is related to the fact I need to add two 3.5mm jack connectors to manage the 12V trigger output for the final amp and the phono preamp.
I've not been able to find an electrical insulated jack and the GND of the jack (that is the GND of the power supply and of the rest of the circuit) will be connected to the case too and consequently to the 'earth'.
Anyway tonight I'll try to connect the case to the earth and I'll let you know, if the noise will disappear I'll find another solution for the 12V trigger jack.
Unfortunately no major improvement regarding the noise.
Now I can hear the 50Hz noise even when not in mute and it increases when I increase the volume. It seems now the noise is injected into the input.
Anyway the noise cannot be heard when I play a song because it's not so strong.
I'd like to measure the noise without any source, what's the best practice? shortcut the input, use a high/low resistance? This way I could have something to compare to when I try to change the scenario.
Now I can hear the 50Hz noise even when not in mute and it increases when I increase the volume. It seems now the noise is injected into the input.
Anyway the noise cannot be heard when I play a song because it's not so strong.
I'd like to measure the noise without any source, what's the best practice? shortcut the input, use a high/low resistance? This way I could have something to compare to when I try to change the scenario.
short both inputs of the stereo amp at the input sockets.
Measure the hum plus noise (H+N) at the outputs.
Attach an interconnect cable to one input and short it's far end and measure the H+N on the output.
Swap the interconnect and measure the H+N at the outputs.
Attach a second interconnect to the spare input.
Short it's far end and measure the H+N at the outputs.
Short the two barrels of those two interconnects together at the far end. Measure the H+N at the outputs.
You now have a record of all the different input configurations and should be able to identify the problem one/s that increase the output noise above 0.1mVac.
Measure the hum plus noise (H+N) at the outputs.
Attach an interconnect cable to one input and short it's far end and measure the H+N on the output.
Swap the interconnect and measure the H+N at the outputs.
Attach a second interconnect to the spare input.
Short it's far end and measure the H+N at the outputs.
Short the two barrels of those two interconnects together at the far end. Measure the H+N at the outputs.
You now have a record of all the different input configurations and should be able to identify the problem one/s that increase the output noise above 0.1mVac.
Last edited:
- Status
- Not open for further replies.