Simulating AD8333

Hello, My questions are actually basic. I am using AD8333 in my project for bioimpedance measurement. The main aim is to study the food quality using 4 electrode method (auto balancing bridge method). So I wanted to use AD8333 to extract real and imaginary parts of impedance. But I am not getting any proper waveform, its completely distorted. And I wanted to know what would be the relation between the amplitude of output voltage after AD8021 and the impedance value, I am facing some difficulties in understanding that from datasheet. I hope I can use this IC for my particular application. Moreover, I am simulating in Multisim, where I cannot find any LVDS driver so I have used another LNA for LO inputs. Will it be okay? And mainly, I need your help to know if I have made some mistakes with connections or if there are any errors you can rectify for me. Kindly guide me. In my simulation for now, I have used a simple RC parallel circuit to represent bioimpedance in food samples. I am also attaching the picture of the waveform I am getting. Sorry for the circuit being a little messy.

attachments.zip
  • 0
    •  Analog Employees 
    on Nov 5, 2015 11:34 PM


    Hello,

    Thank-you for the email. We like to receive AD8333 inquiries as it is such a specialized part that it's nice to know that people find it of interest.

    Unfortunately your schematic is very small and I think not enough digital resolution to enlarge for better visibility. Since your question involves a simulation, (I don't have access to Multisim) I cannot build and test the circuit to determine if I can improve the quality of your results. At first glance my sense is that you may have a version with limited node capacity, or you may be running a low resolution transient analysis which would be problematic. If you want to see this through you might see if the circuit solution has converged by looking for the correct dc values on the pins.

    The main question though is why use the AD8333 for a single channel application when it was designed for diversity receiver applications where parallel a plurality of channels are made coherent for signal clarity? If you just need  a single channel I/Q-demodulation there are many to be found on our product selection website. I'm not conversant with these products but you can take a look at the website and get in touch with our central applications group through this forum.

    I hope this helps,

    js

  • Hello,

    Thank you so much for your reply, it definitely helps. And yes with Multisim, sometimes I face convergence issues for some values. So I will definitely follow your suggestion and try to check for transient analysis details. I am using AD8333 as the other I/Q demodulators have RF input range starting from 50MHz or 400MHz and for my application, I needed an IC that starts from lower values. Plus, when I started simulation, initially the Spice model of AD8333 was not available. Should I send you or upload my Multisim file so that you can get better overview? So that I can atleast be sure if I have correctly connected AD8333.

    Regards,

    Hetashri

  • 0
    •  Analog Employees 
    on Nov 6, 2015 2:36 AM

    Hi Hetashri,

    OK, understand, and I've heard this before. I actually don't have Multisim, ADI has changed to ADIsimPE. However we do have some SPICE experts to support such questions as yours. BTW  there is an LVDS driver shown in the AD8333 data sheet in the eval board section. I'm pretty sure this was tested and works in SPICE but the topic hasn't been raised in several years. From what I can tell from your attached schematic the LNA + ac coupling may be OK but  we have always used dc coupling to the AD8333, and the input circuit shown in the schematic on pg 29 of the attached. Also based in the scope response in your schematic, the output looks a lot like a sine wave with some LO in the output, especially if you are using it at a lower frequency without changing the LP filter in the AD8021 feedback. What is your input frequency? If you're using the circuit in the data sheet bear in mind it was designed for an input frequency of 5MHz. For any other frequency you'll have to scale the filter appropriately.

    Now that I think of it, there is a pretty good (maybe even better) design guide in the AD8339 data sheet.

    I hope some of this helps,

    js

    AD8333_Final_Approval.pdf
  • Hello,

    Okay so now the doubt is pretty much clear. I will go thoroughly through AD8339 datasheet to check my values of LPF. I was actually using 2MHz input frequency and 8MHz for LO! Thank you so much for your guidance. Hoping it will work now. Once again, thank you so much.

    Best regards,

    Hetashri