AD7767 Tsettle specification incorrect?

Hello everybody,

In the AD7767 datasheet (rev C) the value Tsettling in the table on page 5 of 24 is specified as:

(592 × n) + 2 in mclk cycles. (digital filter settling time plus 2 mclk cycles for power-up and reset I assume)

For the AD7767-2, n = 4 so we would have a settling time of 2370 mclk cycles. (2.893 ms)

If I measure the between the rising edge of the mclk after the sync goes high and the last rising edge before the drdy goes down, I measure 2.8767 ms.

If I export my oscilloscope data and analyze it in scilab, I count 2358 mclk periods in this time, so I would say there is a significant difference between what is specified in the datasheet for this time, and what is actually coming out of the hardware.

Am I missing something here or interpreting something not correctly or is the specification incorrect?

I need this settling time specification to get a proper phase alignement with other ADCs (different types) in my system.

Best regards,

Olaf