EEVblog Electronics Community Forum

General => General Technical Chat => Topic started by: blueskull on August 05, 2017, 03:05:30 am

Title: Big company has no idea what it puts in its datasheet
Post by: blueskull on August 05, 2017, 03:05:30 am

I was lurking around for a low jitter, small package clock buffer, and I came across this:

Quote

The above plot captured using Agilent E5052A shows Additive Phase Noise of the NB3N551 device measured with an input source generated by Agilent E8663B. The RMS phase jitter contributed by the device (integrated between 12 kHz to 20 MHz; as shown in the shaded region of the plot) is 16 fs (RMS Jitter of the input source is 104.08 fs and Output (DUT + Source) is 119.77 fs).

This is a section excerpted from here: http://www.onsemi.com/pub/Collateral/NB3N551-D.PDF (http://www.onsemi.com/pub/Collateral/NB3N551-D.PDF)

Why would they think RMS_total=RMS_a+RMS_b? Isn't it supposed to be RMS_total^2=RMS_a^2+RMS_B^2?

Both IDT, Silabs and TI did their calculation correctly, and I wonder why it's so hard for OnSemi to get the numbers right?