[RFI] Noisy PC Speakers

[Hare, Ed W1RFI]

The ARRL Lab received a report of interference from a Class D audio amplifier that had been installed about 500 feet from an amateur.  We purchased a set of them, set them up in our C63.4 conducted-emissions test room, connected two resistors as loads and measured radiated and conducted emissions. (Our conducted-emissions measurements are pretty accurate. For radiated emissions, we are making measurements on HF and VHF in the presence of a lot of other conductors, so we are looking for gross problems only.  The amplifier passed conducted emissions testing, did not create a lot of noise on HF and seemed to be below the limits on VHF radiated emissions.  

This didn't jive well with the report we received. I had been to the site where the noise was occurring and it was real and it was heard over that distance, so something was clearly going on.

We then hooked up 16 feet of wire and real speakers.  Bingo!  The noise level went up dramatically. It was my guess that the speaker leads were radiating, so we tried a common-mode choke and bypass capacitor. It made a bit of a difference, but it still was not nearly as quiet as it was when we terminated the speaker connections with resistors.  We tried a better filter on the speaker leads and the noise actually went up slightly. 

Then, we remeasured the conducted emissions, just as a thorough test and, to my surprise, replacing the resistive terminations on the amplifier output with speakers connected to wires caused the *conducted emissions* to rise well above the FCC limits.  VERY interesting.  It would be typical of a lab to measure conducted emissions on one test fixture where, for simplicity, it terminated the speakers with resistors. So, a unit that "passed" when tested under laboratory conditions failed emissions testing badly when installed under real-world conditions. I just would not have expected emissions conducted onto the ac mains to vary with the length of speaker wires connected to the output, but that is exactly what we saw.

We have also seen the same effect with lighting devices.  I was testing an LED bulb for conducted emissions and saw that the conducted emissions increased dramatically when I unscrewed the bulb, placing my hand near it. Okay, that's not a really big deal, but the bulb was a spotlight bulb, so instead of testing it in a open fixture, I bought a recessed can that would be very typical of the way the bulb is used in real life.  The conducted emissions were about 8 dB higher, putting the bulb over the limits when used in a real-world environment.

I also saw the same effect on a power supply that passed emissions limits when terminated in a resistive load at its output, but having conducted emissions increase dramatically on 7 MHz when terminated with the same resistive load attached to the supply with a 33-foot twisted pair.  

I made a proposal to the IEEE EMC Society to do an "Experiments and Demos" presentation of these phenomena, because this is information that really should be made part of standards for the ways that EMC is measured in laboratories.  Unfortunately, the in-person event was cancelled due to COVID-19 and I didn't have access to the ARRL Lab until only several days before the deadline for submitting a video for the virtual event that took its place.  Now that I am back in the Lab pretty much every day, I intend to make that video in case the May EMC Symposium in Raleigh, NC is not held in person. I will also document this in an article that I can submit to one of the EMC Society peer-reviewed journals.  That is probably the best say to introduce this into the standards arena. I want to see if we can get C63.4 modified in some way so that these resonance effects can be included in the standard to ensure that testing done will match the ways devices perform in the real world. 

Ed Hare, W1RFI
ARRL Lab


-----Original Message-----
From: RFI <rfi-bounces+w1rfi=arrl.org at contesting.com> On Behalf Of Tony
Sent: Saturday, September 19, 2020 12:10 AM
To: Rfi List <rfi at contesting.com>
Subject: [RFI] Noisy PC Speakers

All:

Purchased a set of USB-powered PC speakers which turned out to be very noisy on HF. The speakers are made by Mica model PB20i.

See: 
https://www.amazon.com/gp/product/B07VTS62L6/ref=ppx_yo_dt_b_asin_title_o02_s00?ie=UTF8&psc=1

Tony - K2MO


_______________________________________________
RFI mailing list
RFI at contesting.com
http://lists.contesting.com/mailman/listinfo/rfi