# AD5933 upper frequency limit vs MCLK

Question asked by ignasi on Oct 10, 2012
Latest reply on Jan 7, 2013 by rms1981

Hello,

I'm want to use the AD5933 eval. board for electrochemical impedance measurement in a broad range of frequencies (1Hz- 100kHz).

I read in UG-364 and EVAL-CN0217-EB1Z docs that I have to scale the system clock to analyze frequencies below 1kHz.

The table "Experimental Lower Frequency Limits vs. MCLK" (table 2 in UG364) shows  the experimental lower frequency limit in function of the system clock but I want also to calculate the upper frequency limit but I'm not sure how to do it.

In documentation (UG-364 and EVAL-CN0217-EB1Z) people of AD say that DDS reference clock is MCLK / 4, and the max DDS frecuency output is DDS reference clock / 8. I understand that Fout,max = Fdds * M / 2^27 = Fdds * 2^24 / 2^27 = Fdds / 8. So, for a 4MHZ system clock Fout,max = 125kHz but the maximum frequency range of transmit stage is 100kHz (from AD5933 datasheet).

Moreover, the example in doc says that for a 4MHz system clock the DDS reference clock is 1MHz and the maximum output frequency is 1/32*1MHz = 31.25kHz. So, there is a factor of 4 (125 / 4 = 31.5) that I don't know  where it comes from. It seems that system clock is divided by 4 twice ( MCLK/(4*8*4) ), why?. Does anybody know if it it's the rigth way to calculate the upper frequency limit? And where does this factor of 4 come from?