We can do a little more debugging to figure out that noise/interference source. First, I suggest shorting the amp input right at the 2-pin input connector on the board. Use a short wire staple (piece of component lead for example) for this. Shorter is better as a long loop of wire can pick up interference.
In this configuration, the amp should be dead quiet except for maybe a very, very faint hiss (white noise) that you can only hear when you press your ear up against the speaker.
If you don't get a quiet amp under these conditions, I suggest lifting the ground. You can do this by simply disconnecting the mains ground wire at the IEC inlet. It's not a safe way to operate the amp in the long term, but for a simple experiment it's fine to do so.
If the noise goes away under this condition, you're dealing with a ground loop.
If your source is a PC driving a USB DAC, I suggest using an optical link between the PC and the DAC. I've seen several examples of noisy PCs injecting noise through the ground. An optical link always fixes this.
A final option is that the amp could be "talking to itself". You want to make sure that the wires going between the amp boards and the power supply board are tightly bundled. Use zip ties for this.
Tom
In this configuration, the amp should be dead quiet except for maybe a very, very faint hiss (white noise) that you can only hear when you press your ear up against the speaker.
If you don't get a quiet amp under these conditions, I suggest lifting the ground. You can do this by simply disconnecting the mains ground wire at the IEC inlet. It's not a safe way to operate the amp in the long term, but for a simple experiment it's fine to do so.
If the noise goes away under this condition, you're dealing with a ground loop.
If your source is a PC driving a USB DAC, I suggest using an optical link between the PC and the DAC. I've seen several examples of noisy PCs injecting noise through the ground. An optical link always fixes this.
A final option is that the amp could be "talking to itself". You want to make sure that the wires going between the amp boards and the power supply board are tightly bundled. Use zip ties for this.
Tom
I generally aim to keep the signal wiring away from the transformer and power supply. However, the supply wiring is low impedance, so it can run over the transformer if need be. Same for the output connection. Definitely keep the input wiring shielded and away from the transformer and mains wiring.
Tom
Tom
All the panels and heat sink should connect together. Sometimes anodization or powder coating gets in the way, so you might have to sand that off in places. Once you have all of the chassis connecting together, you can start filling the chassis.
The best place to ground the mains is right by the mains input ground. Use as short a wire as possible.
Tom
The best place to ground the mains is right by the mains input ground. Use as short a wire as possible.
Tom
ok, i can take it to the back panel right next / above the mains inlet. thanks of the advice!
i'm using a galaxy 2u chassis with additional base plate:
Galaxy 2U
Base Plate
i'm using a galaxy 2u chassis with additional base plate:
Galaxy 2U
Base Plate
If you can hear it, it is probably not rf interference ;-)
Another thing to realise is that with just a few mV input your S/N ratio is relatively low giving the impression of highish noise, which will disappear with 'normal' listening levels.
Jan
Unless the RF is demodulated by a diode junction somewhere and turned into bouncing DC. Happens all the time with GSM cell phones. The RF bursts of the GSM signal can cause the DC offset of an input stage to bounce which results in the BRRZT-BRRZT that you hear in the speaker as a result. That's one of the many reasons I put the RF filter on the input of the LM3886DR.
I recall the design review for the LMV851 EMI hardened opamp to be rather interesting. The applications team had fun trying to "break" the chip. You can see the results, including the cell phone call I mentioned above, on page 16-17 of the LMV85x data sheet.
True. You can't move electrons without causing noise. That said, with a medium-efficiency speaker (say 85-87 dB) an LM3886 amp should not generate enough noise to be bothersome at the listening position. I measure 43 uV RMS 20Hz-20kHz (unweighted) on the output of the LM3886DR.
Tom
i finally finished the enclosure for my 3886dr project last night. to my great satisfaction the noise problem i had earlier disappeared almost entirely once the enclosure was finished and the ground was securely connected. what remains is tiny amount of white noise which is only audible if i hold my ear against the tweeter and listen very hard.
i'll post more later - for now i am very happy the way it turned out.
i'll post more later - for now i am very happy the way it turned out.
Awesome! Yep. That noise you're hearing is the 40 uV RMS (unweighted) of the LM3886 along with whatever acoustical noise is in your head.
It's normal.uhm .... ok, i'll shut up about noise now
Sounds like a plan.
I'm glad the amp worked out well for you. How about some pictures, we like pictures too
Yes, please. I'm curious what you ended up with.
Tom