i am using an SI8920 to observe a current over a shunt (what else..). The result is differentially amplified and is fed in a differential ADC. Now the ADC readings give some head ache since i am seeing some tones generated from the SI8920.
I have attached 2 images: one shows the base-line spectral noise of my system (via FFT on a 933kSPS signal stream) and one with the SI8920BC with shorted inputs.
Those tones are device-specific (85kHz in this example). That means, every SI8920 on my PCB (i am using 4 in parallel on 4 ADC-Channels) has a different noise-frequency peak and when desoldered and swapped around the specific tones follow the chip.
Those tones vary widely within the spectrum. I have a PCB where, by mere luck, all tones are so high, they are swallowed by the aliasing filter. Then i have a PCB where all are right in my spectrum of interest.
If anyone knows where this tone comes from and how to influence the frequency of this disturbance, i am all ears.