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.

ADC10D1000: require some user guidance

Part Number: ADC10D1000
Other Parts Discussed in Thread: ADC12D1000

Hi TI experts,

I have some technical questions on ADC10D1000, as listed below.

1.  I see Vcmo is internally buffered. Does TI recommend connecting this pin directly to the common-mode input pin of an fully differential amplifier (FDA), or externally buffering it before connecting it to an FDA?

2.  If the Tdiode+- pins are not used, is it OK to leave them floating (not connected)?

3.  In typical performance plots shown in datasheet, how many samples are used to perform FFT?

4.  The equivalent circuit of CalDly shows that it is a CMOS input pin. Why the recommended pull up/down resistor is 1k Ohm? What are the considerations?

5.  If the ADC is operating at DESI mode and 1:4 Demux mode, can I latch DQ and DQd with DCLKI and leave DCLKQ floated (since DCLKI and DCLKQ are always in phase), or terminate DCLKQ with a 100-ohm resistor but not connect DCLKQ to FPGA? If ORI is high indicating VinI is out of range, how do I determine which sample is out of range, since both DI and Did are clocked out at the moment? Are ORQs active in 1:4 Demux mode?

6.  Does the part require a specific power supply ramp rate?

7.  What is the recommended termination method for the unused AC-coupled input in DES mode? Table 18 says one should tie unused+ to unused-, while the datasheet (AC-coupled Input Signal section, page 60) also mentions that the unused AC-coupled inputs should be connected to ground through capacitors.

8.  In both DES and Non-DES modes, the unused DC-coupled inputs should be tied to Vbg, right? Is there any requirement on how should we connect them? For example, should we connect them through traces with a specific characteristic impedance, or through low-impedance planes?

9.  How to terminate the unused SPI pins in Non-ECM mode?

10.  What is timing requirement of calibration as shown below?

Looking forward to your reply. Thanks.

  • Hi Yubing,

    I will look into all your questions and get back with you.

    Regards,

    Rob

  • Hi Yubing,

    1.  I see Vcmo is internally buffered. Does TI recommend connecting this pin directly to the common-mode input pin of an fully differential amplifier (FDA), or externally buffering it before connecting it to an FDA? RR: Adding a buffered follower amplifier is always recommended. If only one connection to an amplifier is needed, then this is probably okay as well.

    2.  If the Tdiode+- pins are not used, is it OK to leave them floating (not connected)? RR: yes, leave them floating.

    3.  In typical performance plots shown in datasheet, how many samples are used to perform FFT? RR: either 16k or 32k FFT points were used.

    4.  The equivalent circuit of CalDly shows that it is a CMOS input pin. Why the recommended pull up/down resistor is 1k Ohm? What are the considerations? RR: this ensures a logic low when powering up the part by default

    5.  If the ADC is operating at DESI mode and 1:4 Demux mode, can I latch DQ and DQd with DCLKI and leave DCLKQ floated (since DCLKI and DCLKQ are always in phase), or terminate DCLKQ with a 100-ohm resistor but not connect DCLKQ to FPGA? RR: I would follow the timing diagram in the datasheet and connect both DclkI and DclkQ. If ORI is high indicating VinI is out of range, how do I determine which sample is out of range, since both DI and Did are clocked out at the moment? RR: There is probably no exact way to locate the exact sample, as there is pipeline delay within the ADC that would need to be factored in and all bits will start to look like a stream of 111111’s Are ORQs active in 1:4 Demux mode? RR: yes, ORQ’s in 1:4 demux mode

    6.  Does the part require a specific power supply ramp rate? RR: no, it does not.

    7.  What is the recommended termination method for the unused AC-coupled input in DES mode? Table 18 says one should tie unused+ to unused-, while the datasheet (AC-coupled Input Signal section, page 60) also mentions that the unused AC-coupled inputs should be connected to ground through capacitors. RR: correct, 0.1uF capacitor then 50ohm in series to ground. This is specified also in the EVM user guide.

    8.  In both DES and Non-DES modes, the unused DC-coupled inputs should be tied to Vbg, right? RR: correct, same as above, 0.1uF capacitor then 50ohm in series to ground. This is specified also in the EVM user guide. Is there any requirement on how should we connect them? For example, should we connect them through traces with a specific characteristic impedance, or through low-impedance planes? RR: low impedance planes is fine.

    9.  How to terminate the unused SPI pins in Non-ECM mode? RR: all terminations are included as shown in the datasheet pin descriptions.

    10.  What is timing requirement of calibration as shown below? RR: the picture below, is difficult to read. Not sure what is being asked here. Please follow the datasheet, calibration timing on page 35, figure 10.

    RR: It might also be helpful to read the datasheet entirely and also see the following datasheets (ADC12D1000/1600/1800) that of the same family. This information is given in the first page of the datasheet on the ADC10D1000.

    Regards,

    Rob

  • Hi Rob,

    Thanks for your quick response.

    Q4: What I mean is that in practice 1k-10k ohm resistors are used as pull-down resistors. Why TI recommend specificly 1k ohm? What are the considerations?

    Q8: I am confused. Unused DC-coupled inputs should be connected to ground through 0.1uF capacitor in series with 50ohm? This is contradictory to Fig. 18 on page 60.

    Q9: I could not find any description on how to terminate unused SPI pins. But I see SCSb/SCLK/SDI are internally pulled-up and SDO is tri-stated if SCSb is high, so I believe it's fine just leave them floating.

    Q10: The picture shows the timing relationship of CAL register (0h bit15). I thought Fig.10 is for pin-controlled mode. Does it apply in ECM mode? For example, the user maintains this bit=0 for minimum of tCAL_L and then set this bit=1 for minimum of tCAL_H to initiate on-command calibration and then manually reset this bit to 0.