I've been getting crazy with the Ez-Kit 21489 due a crosstalk problems between the analog outputs.
I was programming some filtering routines, and testing them with a 4 way cross-over example, I thought that
my filter implementations were wrong... See the blue graph below that is the filter frequency response I measured
with a 4th order Butterorth band-pass filter at 5kHz and 15kHz. The response at low frequencies is really bad,
and I believed it was a rounding and/or filter structure.
After hours and hours of coding and testing, I discovered that when I switched off the signal to the other outputs
(with the rest of band-pass filters) the filter response was ok (the red curve)....
I tested again with the C-Block Based examples that make a pure by-pass, and it happens the same.
This is the spectrum with the by-pass example. The blue is the output spectrum without any input connected
to the analog inputs. The red is the output with pink noise at one input that feeds the signal to the other outputs.
It seems that the low-frequency cross-talk between the outputs of the AD1939 is so high.
Is it a common problem? or ist it a problem of my Ez-Kit 21489?
Looking at the design of the Ez.Kit, probably the capacitors C254 and C257 are too small and not the best type of cap.