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.

AFE58JD32: TX_TRIG and odd/even internal mux behavior

Part Number: AFE58JD32

We are using an AFE58JD32ZBV in our design with the JESD interface. I am noticing something that I cannot explain and am looking for help.

The following figure is Figure 81 in the datasheet. First, can anyone explain the dotted vs solid line of TX_TRIGD? That may explain what I am seeing. I am not sure. 

 

What I am seeing is that the even/odd conversion seems to be 1 system clock cycle delayed from what this diagram is showing. We have an external 4-to-1 mux (yielding a total of 128 ADC channels). For my experiment, I am connecting the output of the 4-to-1 mux to AIN16 of the AFE58JD32. I have 1 of the 4 mux inputs connected to 200mV and the other 3 connected to GND. If I switch the mux control lines at the neg clock edge of system clock where the conversion starts, the 200mV signal shows up split across 2 of the 4 mux channels. That mux switch timing should allow for the mux to switch and the AIN16 input sampling cap to be fully charged on the next clock cycle before the conversion on the following clock cycle. See the diagram below, also from the datasheet.

 

 

Here is a diagram that illustrates how we are switching the external mux control lines relative to the expected sample and conversion window.

What I have discovered is that when I delay the mux control line switching by 1 system clock cycle, then the 200mV only shows up on just 1 of the 4 muxed channels as expected. However, I cannot explain this and would like to understand why this is the case.

 

Some other things:

1) I have tried using a TX_TRIG that is a half system clock cycle (falling to rising) in duration like the diagram above shows, and I have tried a full cycle TX_TRIG (falling to falling). Both yield the same behavior.
2) 200mV erroneously split across two of the 4 muxed channels:

3) 200mV showing up on 1 of the 4 channels as expected, but only after 1-cycle delaying the mux control lines.

4) Scope picture showing TX_TRIG (ch3) and the mux select lines (B2 is odd, B1 is even) relative to ADC_CLK. You can also see the analog signal changing with even mux select lines. Please disregard the noise on the analog signal.

Thank you!

  • Hi Kyle,

    I am sorry to hear about your issue.

    Please give me some time to understand and think about it. I will get back to you early next week.

    Thanks & regards,

    Abhishek

  • Hi Kyle,

    Apologies for delayed response.

    From what I can understand so far is that you are providing kind of a step input to the device channel. When you switch the mux, essentially the device sees it as a step signal coming in. As you know there is a LNA and all kind of filters at the input. 

    So, if you pass this kind of input through the signal chain the LPF will act and makes it slow. And when sampling happens you see some residual voltage. On the other hand when you move the switching of mux by half a cycle, it might be that it is giving more time to settle and hence you are seeing correct result. This is just a theory that we could think of.

    You can verify this by increasing the LPF corner frequency and see if the voltage you measure decreases.

    If that is the case, than most likely this is the reason.

    Let me know if this works for you.

    Thanks & regards,

    Abhishek 

  • Hi Kyle,

    I am following up on your issue. Was my suggestion helpful?

    Has your issue been resolved?

    Regards,

    Abhishek

  • Hi Abhishek,

    I apologize for the delay but I have been too busy to look at this. I have a workaround in place so it has been lower priority. I will get back to it as soon as I can.

    Kyle

  • Hi Kyle,

    I am closing this thread. Please feel free to re-open when you have any concerns.

    Thanks & regards,

    Abhishek