I’ve been trying to track down some pesky rectifier noise in one of my amplifiers. I hooked an oscilloscope up to the speaker connection on one channel and noticed some periodic spikes in the signal with no music playing. The spikes are roughly every 5ms, so 200hz. I sort of thought this might be rectifier noise since I live in Europe (50Hz) and I have two bridge rectifiers in the PSU. Despite all my efforts of moving the grounding around, I couldn’t get it to go away.
Then I decided to take a break, and shut the amplifier off by turning off the smart plug it’s connected to (which kills the power but leaves earth connected), and surprisingly the spikes didn’t go away. I then started unplugging devices one by one from the power bar that powers that area of my entertainment center. In general the more devices I unplugged, the smaller those peaks got until everything was unplugged, and all that was left was the background noise.
I’m trying to figure out what the cause is. Is it really just a bunch of rectifier noise that’s been passed onto the earth connection by random devices on that outlet? If so, any way to counter that inside the amplifier (ie use a ground lift or something?)
Then I decided to take a break, and shut the amplifier off by turning off the smart plug it’s connected to (which kills the power but leaves earth connected), and surprisingly the spikes didn’t go away. I then started unplugging devices one by one from the power bar that powers that area of my entertainment center. In general the more devices I unplugged, the smaller those peaks got until everything was unplugged, and all that was left was the background noise.
I’m trying to figure out what the cause is. Is it really just a bunch of rectifier noise that’s been passed onto the earth connection by random devices on that outlet? If so, any way to counter that inside the amplifier (ie use a ground lift or something?)
Attachments
The line itself has lots of harmonics because of several reasons: saturation in transformers, leackage in isolators, industrial appliances, etc.
But if you want to circumscribe to your home, several electric and electronic devices includes a line filter which has low value capacitors from line and neutral to earth or ground wire. Usualy are called Y capacitors and its value oscillates between 1nF and 10nF. Particularly those with SMPS inside, not only conducts noise from L and N to E, some RF also escapes. For avoiding oscillations in SMPS and PFC'c, a similar cap is placed between negative side of line rectification to earth and from secondary to primary negative. It is so done in order to give a path for internal capacitance of the transformer or inductors, in which a RF current is established from primary to seconaries, but also make a capacitance divider. So small amounts of RF also are conducted via input cables backwards. Worsen designs grounds the main power MOSFET heat sink(s) HS to ground, a definitive error. Those hs must return to any of the line rectified DC poles, generally the negative one. If not, in the drain to hs capacitance also flows some RF currents to ground.
Also take in mind, some unavoidable capacitances from L and N to ground, for example, in the motocompressor of a refrigerator, air conditioners, and a large number of et ceteras.
But if you want to circumscribe to your home, several electric and electronic devices includes a line filter which has low value capacitors from line and neutral to earth or ground wire. Usualy are called Y capacitors and its value oscillates between 1nF and 10nF. Particularly those with SMPS inside, not only conducts noise from L and N to E, some RF also escapes. For avoiding oscillations in SMPS and PFC'c, a similar cap is placed between negative side of line rectification to earth and from secondary to primary negative. It is so done in order to give a path for internal capacitance of the transformer or inductors, in which a RF current is established from primary to seconaries, but also make a capacitance divider. So small amounts of RF also are conducted via input cables backwards. Worsen designs grounds the main power MOSFET heat sink(s) HS to ground, a definitive error. Those hs must return to any of the line rectified DC poles, generally the negative one. If not, in the drain to hs capacitance also flows some RF currents to ground.
Also take in mind, some unavoidable capacitances from L and N to ground, for example, in the motocompressor of a refrigerator, air conditioners, and a large number of et ceteras.
Just to confirm - your amp has a protective earth connection to the mains supply? Does your speaker have a low ohm resistance connection back to the amp's protective earth pin? Similarly, does the Rigol have a protective earth connection, and does the BNC shield have a low ohm resistance connection to that protective earth pin?
Yes, internally in the amp I connect the earth to the main ground at the star ground. And yes, the IEC socket in the amp connects to the mains supply, which is earthed.
I'm not sure about the oscilloscope. It's a Rigol DS1054. I clip the ground clip on the probe to a wire connected to my star ground and the signal part of the probe on the positive speaker terminal. the Rigol is connected to the mains via an IEC cable. That's it for the hookup.
So a few more observations today.
1) If I take the ground reference of the oscilloscope at the RCA shield, the noise goes down a lot with the amp powered. I'm guessing the currents at the star ground are a lot noisier as all the power amplifiers connect there. The RCA shield is grounded too internally, but via the PCB. So it goes RCA shield -> LM3886 PCB -> PCB Signal Ground -> PCB Power Ground -> Amplifier Star Ground. So probably it makes more sense to take the reference at the RCA shield
2) My Mac Mini is the main source of audio now.. If I use my iPhone instead (via the headphone jack), the oscilloscope garbage is dramatically reduced. So I'm guessing some of what's on the signal is junk from the Mac Mini's SMPS.. Interestingly, the Mac Mini isn't grounded internally as it just uses a 2 wire power cable.
1) If I take the ground reference of the oscilloscope at the RCA shield, the noise goes down a lot with the amp powered. I'm guessing the currents at the star ground are a lot noisier as all the power amplifiers connect there. The RCA shield is grounded too internally, but via the PCB. So it goes RCA shield -> LM3886 PCB -> PCB Signal Ground -> PCB Power Ground -> Amplifier Star Ground. So probably it makes more sense to take the reference at the RCA shield
2) My Mac Mini is the main source of audio now.. If I use my iPhone instead (via the headphone jack), the oscilloscope garbage is dramatically reduced. So I'm guessing some of what's on the signal is junk from the Mac Mini's SMPS.. Interestingly, the Mac Mini isn't grounded internally as it just uses a 2 wire power cable.