Okay you are talking about passing RF to "ground".
from the schematic I suppose that the 100 pf capacitor goes to analog ground. The nearer to the high gain input stage (op amp, or first signal transistor base) the better.
Now try putting 33 pf or 47 pf or some disk cap from analog ground to case ground - the safety ground.
This 100 pf cap you used was a ceramic disc cap wasn't it? Wound plastic film caps have too much inductance to short out RF signals really well.
Your case is seriously steel isn't it? does this happen with the cover screws in? sometimes lanyards with barbed ring terminals have to be connected between the cover and the main case if the paint interferes with the grounding of the steel cover.
If the cover is aluminum, start over. Steel is better IMHO and I haven't had good shielding luck with stainless steel NEMA enclosures either. There is a drive by CBer who emits barking dogs yelping "Dixie" continuously from his P/U truck who is my RF interference test. To get him out of my disco mixer, I had to float the RCA jack panel from the case with rubber o-rings, separating analog and case ground. The circuit board already had 33 pf center terminal to analog ground, on it. I then grounded the case to the 3rd pin of the wall plug (safety ground), which was not done originally since it was a 2 prong line cord. I had a mag-phono shell connected from turntable to case of the mixer with a green wire already. I put the transformer out of the case for hum reasons, and filtered the 18 VDC coming in from a wall transformer with a toroid choke input hash filter salvaged from a PCAT switcher power supply. The Dc supply had the usual electrolytic filter caps, disk caps, and voltage regulator (pass 5W zener in this case).
from the schematic I suppose that the 100 pf capacitor goes to analog ground. The nearer to the high gain input stage (op amp, or first signal transistor base) the better.
Now try putting 33 pf or 47 pf or some disk cap from analog ground to case ground - the safety ground.
This 100 pf cap you used was a ceramic disc cap wasn't it? Wound plastic film caps have too much inductance to short out RF signals really well.
Your case is seriously steel isn't it? does this happen with the cover screws in? sometimes lanyards with barbed ring terminals have to be connected between the cover and the main case if the paint interferes with the grounding of the steel cover.
If the cover is aluminum, start over. Steel is better IMHO and I haven't had good shielding luck with stainless steel NEMA enclosures either. There is a drive by CBer who emits barking dogs yelping "Dixie" continuously from his P/U truck who is my RF interference test. To get him out of my disco mixer, I had to float the RCA jack panel from the case with rubber o-rings, separating analog and case ground. The circuit board already had 33 pf center terminal to analog ground, on it. I then grounded the case to the 3rd pin of the wall plug (safety ground), which was not done originally since it was a 2 prong line cord. I had a mag-phono shell connected from turntable to case of the mixer with a green wire already. I put the transformer out of the case for hum reasons, and filtered the 18 VDC coming in from a wall transformer with a toroid choke input hash filter salvaged from a PCAT switcher power supply. The Dc supply had the usual electrolytic filter caps, disk caps, and voltage regulator (pass 5W zener in this case).
Last edited:
Presumably that's how things got worse. Very hard to tell what happened without any specifics though (including amp model and schematic).
Sounds like a classic ground loop.But I also discovered that I don't need to connect the computer RCA terminals at all, I just need to touch the RCA shield coming from the computer, to any GND part of the amp, and the noise comes in, on both channels equally loud, independent of volume control and source. It's LOUD, you can hear it through the headphones without having them on. It basically just screams at you. EECHEEEECHCEEEECHCH ...
PE connection, you say? Man, that's a dead giveaway. Your amp seems to be an IEC Class I device (like your computer), and possibly star ground is tied to PE, hard. Class I devices and unbalanced connections just don't mix. (Which is why hi-fi gear tends to be Class II a.k.a "double insulated".) In any setup using unbalanced connections, there must be at most one, and only one, earth connection.
Here's what happens:
PC power supply return current couples to PE, creating some voltage.
Said voltage then drops over your audio connection ground when plugging into your amp on another, clean PE connection. Unfortunately, this voltage is indistinguishable from a real signal.
In order to minimize ground loop problems, Class II amplifiers tend to have mains transformers with very little capacitive coupling to mains, using dedicated shield windings if need be.
What you could try in your problem amp:
Identify central audio ground - PE connection.
Replace direct connection with these components in parallel:
* a 10-100R power resistor
* ~47-100 nF film cap
* a pair of antiparallel big-*** rectifier diodes (or bridge rectifier with AC terminals tied together)
Last edited:
Remember many 70s amps and receivers have high value resistors and/or death caps to chassis from active. Older unearthed units with US style 2 pin plugs can be rotated in the socket and the problem can disappear-reappear.
You are likely getting a modern version of radio breakthrough (the bane of 70s gear and high powered nearby AM stations), where the front end junctions or NFB loop (via the speaker cable 'antenna') is basically acting as a detector for the modulated HF hash/ carrier.
I run optical fibre across the room to a DAC to give complete galvanic isolation. There's nothing worse than mouse noises in an otherwise quiet setup. Also consider the differences in earth potential at different power outlets can cause currents to flow in loops, via rca sheaths etc.
You are likely getting a modern version of radio breakthrough (the bane of 70s gear and high powered nearby AM stations), where the front end junctions or NFB loop (via the speaker cable 'antenna') is basically acting as a detector for the modulated HF hash/ carrier.
I run optical fibre across the room to a DAC to give complete galvanic isolation. There's nothing worse than mouse noises in an otherwise quiet setup. Also consider the differences in earth potential at different power outlets can cause currents to flow in loops, via rca sheaths etc.
I always fit RF attenuation filters at the input to amplifiers.
I typicaly use 1k+680pF for a F-3dB=234kHz for single pole passive filtering.
You can use any F-3dB roll-off from 100 kHzto 400kHz. Some Builders use around 1MHz and some even higher.
I don't believe that high roll-offs give sufficient attenuation.
In addition I fit more filters at the entry of any cable into the amplifier enclosure.
47pF across the signal input and 47pF signal to chassis.
1nF speaker to chassis and 100nF+4r across the speaker terminals
IEC socket with integrated filter for the mains.
I typicaly use 1k+680pF for a F-3dB=234kHz for single pole passive filtering.
You can use any F-3dB roll-off from 100 kHzto 400kHz. Some Builders use around 1MHz and some even higher.
I don't believe that high roll-offs give sufficient attenuation.
In addition I fit more filters at the entry of any cable into the amplifier enclosure.
47pF across the signal input and 47pF signal to chassis.
1nF speaker to chassis and 100nF+4r across the speaker terminals
IEC socket with integrated filter for the mains.
So just to confirm that it's not signal that's at fault: when I connect a wire to the PC case, and touch the other side that wire on the amplifier, I get the noise. I guess putting something on the input of the power amp section is useless since the noise gets everywhere, somehow.
And, I get that exact same noise, and equally loud, on a tube amp that I built... Same deal here, that amp is grounded and so is the PC.
There is about 12 mV (DC) between the PC case and a grounded outlet. Normal? If I do the same measurement on a grounded amp, I get about 10 mV DC. Both readings are very stable and not fluctuating.
Yeah, I'm just gonna order a DAC with an optical input and hope that solves my problem.
And, I get that exact same noise, and equally loud, on a tube amp that I built... Same deal here, that amp is grounded and so is the PC.
There is about 12 mV (DC) between the PC case and a grounded outlet. Normal? If I do the same measurement on a grounded amp, I get about 10 mV DC. Both readings are very stable and not fluctuating.
Yeah, I'm just gonna order a DAC with an optical input and hope that solves my problem.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.