• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 46 subscribers
  • Views 528 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Clock generation for multiple AD10D1500

Sergio Fernandez
Prodigy 230 points
Other Parts Discussed in Thread: LMX2531, LMK01000

I want to use a LMK01000 clock buffer in the clock generation because I want to place more than one ADC10D1500 in this board and I would like to clarify some questions about the LMX2531-LMK01000-ADC10D1500 connection.

In am using as reference the design in ADC10D1x00RB board but I need to place a LMK01000 between the LMX2531 and the ADCs. I want to keep the possiblity to choose between the LMX2531 output and an external clock. I could use the LMK01000 for this since it has two clock inputs but I rather use a control signal as it is in the ADC10D1x00RB design than programming the registers in LMK01000 to choose the clock. So that I have placed a relay to choose the input clock. In ADC10D1x00RB board the output of the relay is connected to a transformer in order to generate a differencial clock and then this differencial clock goes to the ADC.

Since the LMK01000 allows the use of either single ended of differencial clock inputs I can skip the use of the transformer and connect the single ended output of the relay directly (through a decoupling capacitor) to one of the inputs of the LMK01000 and connect the complementary input of that clock to GND (AC ground as it is specified in the datasheet). Is this OK? Is there any advantage in using a differential clock as input for the LMK01000?

After that, the differential outputs of the LMK01000 can be directly connected to the ADC clock inputs by using just decoupling capacitors in series (AC coupling is a must for ADC10D1500). I mean, I do not need any 100 Ohm resistor since the CLK+/- pins in the ADC10D1500 have them internally. Am I right?

For the uWire communication with LMX2531 and LMK1000, I will use the scheme depicted Figure 2 in the LMK01000 datasheet. Since all the pins are inputs and are 1.8V operation  compatible, I will use a microcontroller supplied at 1.8V. I guess there is no risk for the IO pins of the microcontroller (no 3-3.3V signals expected at these pins).

And finally, a simple question about the power supply. If I want to generate a signal to monitorize the voltage supply of the ADC, could you recommend me a chip (a kind of power on reset) to be used to monitorize this 1.9V supply?

  • 0
    TI__Genius 13335 points

    Hello Sergio,

     We have experimented with using a relay and ultimately found out that the insertion loss affects the slew rate too much and causes performance (specifically SNR) to be degraded. For this reason we’ve revised the EVM (schematic attached  2570.ADC12D1X00RF_RB-SCH_B.pdf) to have an external resistor solder option to switch between internal or external clocking. If you want to do a control signal clock that does not pass through either the LMX2531 or LMK01000 then this is the option we would recommend as well. However, since you have more than one ADC one your reference design you would need a way to split the clock to all of the different ADCs. You may be better off connecting both the LMX2531 and the external clock source directly to the LMK01000 and power down the LMX2531 when it’s not being used as shown below.

    Note that when using this technique, the unused clock source should be powered down to prevent any crosstalk.

    To monitor the supply voltage, we are typically concerned with adding noise to the supply lines. As long as you use an analog power supply monitor like the TPS3126E18 that should be ok. 

    You can choose one with the appropriate threshold as needed.

  • 0 in reply to Luke LaPointe
    Prodigy 230 points

    Thank you very much for your answer, Luke.

    I guessed the relay was not the best option. It was the easiest solution in order to select the clock source with a control signal generated by the FPGA or even a jumper in the first prototype. The 0 Ohm resistor is not a option for the final design where the clock source must be selectable via software, I will check your suggestion with the LMX2531 and the external clock connected to the inputs of the LMK01000.