Building Quiet Audio Power Amplifiers Requires Best Wiring and Layout Practices

There is a misconception that noise pickup is reduced by placing the input RCA connectors (and similarly XLR input connectors in a fully balanced system) on opposite sides of the amplifier rear panel and then running short wires to the amplifier modules. This arrangement is the worst of all worlds. The internal loop areas available for the power transformer’s magnetic field with which to interact with are huge, even in transformers fitted with a GOSS band. Secondly, to reduce common mode RFI pickup, good practice dictates that the incoming interconnect screen be tied right at the input connectors directly to the chassis. At RF, this makes the source, the interconnect shield and the receiving power amplifier a single enclosure and offers good protection from external radio frequency interference. However, if there are any HF noise sources inside the amplifier – like an SMPSU or, more commonly, diode turn-off ringing, the internal HF loop area between the source and separated input connectors is very large, leading to noise pickup. By keeping the input connectors co-located with the signal returns bonded together and using a single RFI cap (2-5 nF disc ceramic or if SMD, an 0805 NPO/COG type – do not use a film capacitor) to the chassis right at the input, the loop area for HF pickup is greatly reduced.

These two slides show how to considerably reduce the internal loop area wiring in an audio power amplifier to better achieve low noise.

Comments

7 responses to “Building Quiet Audio Power Amplifiers Requires Best Wiring and Layout Practices”

  1. Philippe says:

    Hello Andrew

    Nice slides as usual.
    One question:
    It seems to me that it is easier to keep a small grounding loop surface area inside the amp by putting the main transformer in the front side of the amplifier (we see this on many commercial design, Pass for exemple).
    And with this, the main transformer is also outside the grounding area. Of course main AC wire are crossing the whole amplifier. But is it a big problem ?

    Regards
    Philippe

    • Bonsai says:

      Thanks for your comment Philippe. You are correct, there is no fixed way to do this and placing the transformer towards the front can offer benefits. With respect to the mains supply crossing the other cabling, just make sure the mains wires are twisted or closely coupled (as in using sheathed mains cable for example) and that there is a good gap to the other cabling.

      Regards

      Andrew

  2. Mathew says:

    Hi, I have read your suggestions on grounding, wiring, and loops and have implemented your suggestions on one of my repairs, a Hafler DH200. I was both suprised and impressed at how quiet the amp became. Thank you. I do have one problem though and I may have misread your directions. The amp works perfectly, but when I connect it to an Earthed source it burns the HBR a 2.2 ohm on the one channel. The amp has Class 2 wiring. Should I change this to Class 1 or connect the input to the chassis through a 2.2 ohm resistor as suggested on some websites and as it was originally wired? My thought is that the source ground return is now through the HBR to the chassis when it should go directly to the chassis. Any help or suggestions would be much appreciated.

    • Bonsai says:

      Hi Mathew, glad to hear you made some progress. Could you sketch out how your input receptacles are connected? You should not have any voltage between your input receptacles signal return and the signal return on your source gear. what you couple do is apply power to your amp and the source but not connect them, and then measure the voltages between the two on each channel. You should read 0V DC and 0V AC.

  3. Mathew says:

    Hi, The amp is taken apart again to replace the HBR. The HBR was 10ohms and I did not connect the two inputs together as they are far apart on the DH200. I also did not install the second bridge rectifier nor connect the earth wire as the amp only had live and neutral ac input. I did measure the source to amp inputs (amp plugged in but off) and there is a 0.5 volt ac difference. This was supprising….
    I am going to change the HBR’s to 2.7ohm, connect the input grounds together, then connect the input grounds to 0v central ground, reassemble the amp, and do the measurements you suggested. Thank you thus far for your help.
    Can I email you the diagram?

  4. Mathew says:

    Hi. I measured the voltage from my source, which is earthed with a 3 prong 240ac plug, to my amp 0volts (chassis/input) and measured 70vac and 6mv dc. When I turned the 2 prong ac plug of the amp around in the ac wall socket I got 200vac between earth and amp chassis. There is 0 amps current though. The amp works when connected and only 3 millivolts and 0 amps between house earth and amp chassis. Maybe I should just change the amp to class 1 wiring? Will this be legal though….. adding an earth to a class 2 device.

  5. mathew says:

    Hi, I found my problem…I connected my signal generator and oscilloscope to the amplifier to do some measurements. When conected the fuse from one channel blew. I was using different ac sockets to power the equipment and assume the house has a wiring issue. The signal generator and oscilloscope are earthed, class 1, and the amp is class 2. I grounded the amp inputs to the chassis with 2.7 ohm resistors on both channels and this solved the issue. The HBR’s are also 2.7 ohms. The amp is still very quiet…no hum or noise from the speakers. I am happy now, thank you. The house wiring will need further inpection though….

Leave a Reply to Bonsai Cancel reply

Your email address will not be published. Required fields are marked *