ADL5535 as ADC Driver

Hello,

I am trying to drive an ADC with a high impedance differential input by using two (single ended) ADL5535 amplifiers to serve as the differential input. A few questions:

1. Is it wise to use two amplifiers in this manner or should I stick to using either one differential amplifier or by using a 1:1 transformer to convert from single ended to differential?

2. Is it better to implement anti aliasing filtering before or after transforming from single ended to differential?

3.I am trying to drive a 16bit ADC from TI. the ADC16DX370EVM. I believe I may have an issue with a common mode voltage between the two amplifiers. I have read that it is wise to place the output of your amp as close as possible (with filtering) to the input of the adc....Currently for testing purposes I am using a 1 inch long jumper cable soldered to the board to make the connection....could this also be giving bad performance or will the degradation be minimal for testing purposes. I am using a 30 gauge wire twisted pair jumper cable and have a third wire that ties the ground of the adc board to the ground of my amp test board.

Any considerations, feedback or notes that help push me along would be great.

  • 0
    •  Analog Employees 
    on Aug 26, 2014 2:45 AM over 6 years ago

    Hello Thomas,

    1. Use one fully differential amplifier. You will get better performance (lower noise and distortion). Transformer is not necessary. Fully differential amplifiers will convert SE to diff.

    2. Antialiasing filter goes between the diff-amp and the ADC.

    3. Place diff-amp, filter and ADC as close to each other as possible. Need to minimize PCB trace lengths.  Your current setup (with wires) will not get you good enough resolution. Common mode is not a problem with diff-amps.

    Looking at the TI ADC resolution specs, you can expect an effective number of bits, ENOB, of 11.3 bits from the ADC. You need to at least match this but preferably better this resolution with the diff-amp driver. You can find a good selection of fully differential amplifiers on the ADI website to drive this ADC. You will find examples of SE-diff conversion example without using transformers in the datasheets.  

    Take care,

    Zoltan

  • Hey Zoltan,

    Thank you so much for your reply.

    In regards to the question about the anti aliasing filter. The ADC evaluation board I'm using has two transformers in the signal path that are used to balance the signal. I'm thinking there may be an issue because my anti aliasing filter is on a test board after the amp. In order to get my amp, filter and adc as close as possible it seems i should bypass these transformers. If I decide to use the connector on the eval board and follow this path, will I run into an issue of my anti aliasing filter being before the transformers on adc eval board? this will greaten the length between the adc and amp/filter.

    Thomas

  • 0
    •  Analog Employees 
    on Aug 28, 2014 5:02 PM over 6 years ago

    Hi Thomas,

    The TI ADC eval board is set for single ended input, uses the transformers to convert SE to diff and has single pole RC filters located right at the ADC inputs. If you still intend to use an amplifier to drive this ADC you connect the amplifier output to the SMA connector on the ADC eval board. For evaluation purposes this arrangement is sufficient.

    If you intend to drive the ADC with a diff-amp, you can modify the TI eval board to suit. You can add the additional SMA connector and leave the rest of the eval board as is for ac coupled configuration or you can remove the transformers and short the paths with zero ohm resistors for dc coupled applications.

    I don't see a schematic of the TI eval board so I cannot tell if TI put additional footprints at the ADC inputs for higher order filters, so you might not be able to change the filter configuration without hacking the board (not recommended). But you can adjust the filter cutoff frequency by changing the capacitor values already on the board. This single pole filter configuration should be sufficient for your evaluation.

    The board is optimized for performance so the physical length between the transformer and the ADC would be "tuned" by the PCB design to form a transmission line. This eliminates the length problem, provided you don't cut into the traces and you don't attempt to modify them.

    These eval boards are meant to demonstrate the performance and do not attempt to optimize PCB area. In real life you would close the distance between the transformers and the ADC, you might replace the transformers with a diff-amp, you would not use SMA connectors and you would compact the PCB design to minimize PCB area.

    So, for your initial evaluation, you can leave the TI ADC PCB as is. If you intend to use a diff-amp to dive it, add an extra SMA to the TI board and connect the diff-amp outputs using short coax cables. If you intend to use the diff-amp in a c coupled configuration, you can remove the transformers from the TI board and short the paths with zero ohm resistors.

    Take care,

    Zoltan