AnsweredAssumed Answered

Driving the AD9643 250MSPS

Question asked by Zartagal on Aug 31, 2011
Latest reply on Sep 13, 2011 by Qui

Hello,

I'm having some questions about the AD9643 ADC driver. I'm designing an ADC board with wideband input in baseband (0-100MHz, althougt it's really 2-100MHz). The previous stage must to be an DVGA, in particular I was thinking to use the AD8376 becouse it complies with the requirements of the system. It has a differential output, and the ADC has differential input too.

 

In this driver I have to include a filter, and an impedance adaptation to transform the high track & hold impedance in a more benign load. In the datasheet there are 4 possible adaptation circuits. First: with a chip driver; second, with a differential transformer-coupled; third, the interface with the DVGA (AD8376) I'd like to use; and four, with a differential double balun input.

 

My driver should be something between the seconone and the thirdone, I think. In the thirdone (figure 49) there is a resonant aproach for narrowband bandpass, (tank resonant) and it's not proper for me. The filter is bandpass too, becouse it's suposed to be used in a different Nyquist Zone, so it doesn't serve neither.

 

The second circuit, (figure 48) can be adjust for 0-100MHz. The problem is that in this design I'm trying to reduce components and I would like to remove the transformator. In the other way, it's a conversion from single-ended to differential and It's not the case I'm considering, althought it could be change to differential  input and add the filter before.

 

I think that the first circuit (figure47) could be a solution too. althougt it implies space too.

 

I'm thinking to use the impedance matching in the figure 48 (just the right part of the transformator), removing the transformator and follow (from right to left) with 49R9 resistor from P and N lines to the CMV, a fourth order filter, and then the pull-up inductors with a dc-block capacitor of 10nF (to get a 300KHz dc-blocking filter) to make compatible the CMV between the two components.

 

1.- ¿Do you think it's an appropiate driver?

2.- If I use the transformator, ¿Why it says that in applications where SNR is a key paramenter, differential transformer coupling is the best recommendation? (In this application the SNR is a key parameter).

3.- Understanding the right part of the figure 48 circuit, I think that the R1/C1 couple is probably a RC filter, but ¿How does it work the R3/C2 serial network?

 

Thank you very much

 

Gonzalo

Attachments

Outcomes