[R-390] Class D Amplifiers

[Charles Steinmetz]

Roger wrote:

>for communications purposes, it should be fine.  Usually they switch 
>between 80 and 100 kHz and the LPF takes all of the junk out of the 
>audio band.

Two problems:

(i) Yes, they switch in the neighborhood of 100 kHz, and this is 
above the audio band.  However, by pushing the switching garbage 
above the audio band, you've pushed it INTO the radio's receive 
bands.  Remember, the calibration oscillator in a 390 is simply a 100 
kHz square wave generator, coupled into the RF path by a tiny (1 pF) 
capacitor, and you can hear its harmonics loud and clear all the way 
to 30 MHz.  Same with the PWM output of a Class D amp, but because 
the PWM pulses are all different widths, the harmonics are not 
confined to multiples of the switching frequency -- it generates hash 
all over the HF bands and beyond.  The shielding and output filter 
would need to attenuate all of that by 120 dB or so for it not to be 
troublesome.

(ii) The amplifier module linked by the OP has NO output filter at 
all, and is open frame (no shielding at all).  The chip manufacturer 
says in the datasheet that it will not meet FCC standards without 
additional filtering, but the application circuit in the datasheet 
shows no additional filtering and the board manufacturers do not add any.

Best regards,

Charles