Questions bout the intput and output of the AD8302?

Q:

1. the formal input signal is shaken and distorted by the attenuation network, how can I eliminate the distort?

      A: Which attenuation network are you referring to? It looks there a two: one before the ADA4807, and one after (one per channel).  Also, why are you driving the AD8302 from the middle of the feedback path of the ADA4807?  Instead of driving the AD8302 from a low impedance or a 50 Ohm match, you are driving it with a 3k impedance.  This is probably not preferred.

2. I have change the value of R167 and R178 to 1KΩ, what is  the influence of this changes?（If I set the two resistors to 52Ω , the input signal will disappear ）   

      A: It affects the voltage division in the feedback path of the op-amp.  With a 1k resistor installed, the voltage is reduced by a 0.2 factor, whereas with a 50 Ohm installed it will reduce by 0.016 factor.

3. no matter I change the input signal, the average of VMAG stays at about 1V and VPHS stays at about 1.04V, where is the problem?

      A:  I would look at the input signals with an oscilloscope before worrying about VMAG and VPHS. Also, what do you expect the difference in magnitude and phase to be between your two signals?

Thank you for your reply!

I tried to introduce a buffer circuit ,just like ADA4807, but I forgot the impedance matching. Now I changed my schematic, the circuit as follow.

(ps: the function of ADG836 is choose two signal among three)    

Q:

1. I have moved the ADA4807 and changed the value of R13 and R15 to 51Ω, but  the signal is still shaked and distorted. Now ,where could be the problem ?

2.  A:  I would look at the input signals with an oscilloscope before worrying about VMAG and VPHS. Also, what do you expect the difference in magnitude and phase to be between your two signals?

     A: you are right, I also  would look at the input signals with an oscilloscope before gaining  the VMAG and VPHS. I mean that no matter how the input signals change, the output of the  AD8302 will not change.  Therefore,  is there any other reason besides the chip is broken?

• C16 and C17 need to be AC grounded
• The signal being distorted and shaky could be a lot of things.  I would remove C15 and C18 and probe R13 and R15.  Is the signal still distorted?  I would work my way towards the ADG836, probing at various points, checking the signal to see if its distorted. 
• Is it distorted at the inputs of the ADG836?
• How are you changing the signal amplitudes?  For best behavior and the greatest measurement range, one of the two signals needs to be set at -30 dBm (7.07mVrms), while the other can range from -60 dBm (223 uVrms) to 0 dBm  (223 mVrms); equivalent voltages are assuming 50 Ohm terminations.

Thank you.

1. I just forgot to make GND out.

2. The signal is formal when R18 and R19 are large resistance, for example,3K,4K....;Therefore, I think the signal distortion is caused by the noise introduced into GND by small resistance(for example 62Ω).

Q: Do you have any better advice for the attenuation circuit?

3.Actually, one of the signal is about 100mV, 100khz; another signal is about 200mV. 

Q：Should the chip output an even incorrect value when the amplitude of the input signal changes? or my chip have damaged？

looking  forward your reply!    thank

• If you have that much noise/"ground bounce" then the signal should be distorted everywhere on your board.  Is this the case?  Have you probed right where you generate the signal?
• How do you know that the signals applied to the AD8302 are 100mV and 200mV?  Did probe the INPA and INPB pins?  
  • It is not clear what 100mV and 200mV actually means.  Is this peak or peak-to-peak or RMS or DC or etc. 
  • You could be past the limit of the AD8302 measurement range.  I would reduce the signals by 10x and remeasure.  You should get some sort of response on VMAG if one signal is 10mV (something) while the other signal is 20mV (something)

Thank you for reply!

1.If you have that much noise "ground bounce" then the signal should be distorted everywhere on your board.  Is this the case?  Have you probed right where you generate the signal?

A:Only using the small resistance to connect GND in the attenuation is informal.  Of course, it is informal behind the attenuation too.

2.How do you know that the signals applied to the AD8302 are 100mV and 200mV?  Did probe the INPA and INPB pins?  

3.It is not clear what 100mV and 200mV actually means.  Is this peak or peak-to-peak or RMS or DC or etc. 

A: The 100mV and 200mV refer to peak-to-peak. The INPA and INPB are 4.71V when use the circuit as the schematic, and the  voltage will not change when I change the voltage amplitude of input signal.

4.You could be past the limit of the AD8302 measurement range.  I would reduce the signals by 10x and remeasure.  You should get some sort of response on VMAG if one signal is 10mV (something) while the other signal is 20mV (something)

A: I have attenuated the signal to a smaller signal. The output voltage of Vmag and Vphs have a small range data fluctuation, when I change input. the For example, the Vphs  fluctuates from about 990mV to 1.14V, the Vmag fluctuates from about 936mV to 1.05V. But the output voltage still is not obvious.  

      I have try to remove C15 and C18, now, the voltage as same as attenuation when I probe the INPA and INPB pins. But the Vmag is about 950mV, the Vphs  is about 1.99V, and remain unchanged.

( I think the most of the data fluctuations are caused by the shake and noise of the signal, so I can't acquire enough information. But now I still can't  eliminate the noise in the attenuation)

Perhaps the part is damaged.  I would try another part and re-run your tests.