Digital Amplifier Questions

I Don't know why but I thought any RF leakage from class D amps would be closer to the FM band range.

Thanks,
Tim
Noble100

Noble100, don't know why you were thinking this???
FM operates in the 88MHz - 108MHz range i.e. 88 million cycles per second --> 108 million cycles per second.
the class-D switching amp is probably running at 1MHz switching freq (if that high at all. more likely in the 500KHz region). I.E. 500,000 cycles per second --> 1 million cycles per second.
As you can see, 500,000 --> 1 million cyc/sec is far away from 88 million --> 108 cyc/sec.
Let's say that the class-D amp was switching at 1MHz, then the 89th harmonic of this 1MHz switching freq would be 89MHz but it's power would very, very low to the point that even if it interfered with the FM radio station tuned to 89MHz it would produce a very low amplitude distortion that would most likely be un-hearable.
(Just wanted to point out that the switching waveform of the class-D amp is most likely a square wave & square waves have only odd harmonics).

Like Almarg pointed out the AM freq would be right in the vicinity of the class-D switching frequency & much more likely to create interference.

My class D amp switches at 500kHz. This square wave is filtered by the Zobel network at about 60kHz. Harmonics are most likely filtered out and main problem is about 1% residue of 500kHz switching frequency present on the speaker cable. This 500kHz frequency represents wavelength of 600m. Antenna below 1/10 of the wavelength becomes very ineffective - and that is 60m. Typical 3m cable is practically non-radiating. It is still a good idea to keep cables apart to avoid capacitive coupling.

03-29-15: Kijanki
My class D amp switches at 500kHz. This square wave is filtered by the Zobel network at about 60kHz. Harmonics are most likely filtered out and main problem is about 1% residue of 500kHz switching frequency present on the speaker cable.

makes sense Kijanki - 500KHz is 3.05 octaves away from 60KHz. Zobel network/filter is most likely a 2nd order filter which is 12dB/octave implying 36.6dB of attenuation of the 500KHz. That works out to 1.48% of residual 500KHz on the speaker binding posts.

Thanks for the info Bombaywalla and Kijanki.

I tried the test again using the AM radio. I couldn't even find an AM station that was tunable. There was too much buzzing on the whole range. I'm thinking this is caused by RF leakage from my 3 class D amps, all located within a 5ftx 2ft area.

I plugged the radio into an outlet in my adjacent dining room (less than 20ft away) and the buzzing (hash?) went away and I was able to tune in a station in the low AM band.

I then moved the radio as close to my amp rack as the cord would allow (about 10ft away) and there was no interference; the station continued to be clearly audible without hash/buzzing. I was thinking the interference would increase as I moved the radio closer to my amp rack (from about 20 feet to about 10 feet way) but it did not.

These somewhat mixed results have me a bit confounded. I think further testing is called for, perhaps with a better AM radio, before conclusions are drawn. My current thought is that at least one of my amps is producing RF interference but I don't know for certain at this point.

I'm going to get a better AM radio, plug it in in my dining room using a 20ft extension chord, tune in a station on the lower AM band and retest by walking toward my amp rack in the adjoining living room. I'll post again with results.

If anyone has a better idea, please let me know.

Thanks,
Tim

Hi Tim,

If you are going to get another AM radio, consider purchasing a portable one that is battery powered. Also, the test will be most sensitive if you tune to a frequency at which no station is present. If a station is present the radio's AGC (automatic gain control) circuit will reduce its sensitivity, especially if the station's signal is strong.

Best regards,
-- Al