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.

PCM6260Q1EVM-PDK: Open input diagnostic

Part Number: PCM6260Q1EVM-PDK

DC102 A, PPC v3.1.1, take channel 1 for example, Micbias is setup to 9V,

1)Open inputs diagnostic

Hardware setup: J3,J4,J7,J8 installed, J11, J12 uninstalled,  Remove J36, J37, VIN1M=0.035V, VIN1P=8.63V,

In order to detect no input issue, we need to set short to Micbias detect threshold to 420mV.

 

2)INM short to Micbias

Hardware setup:J3,J4,J7,J8,J36,J37 installed, J11, J12 uninstalled,  J33(Micbias is installed), the SW2 switched to C1, press SW1, 

IN1P from 8.21V to 8.84V, CH1_LTCH[3] and CH1_LTCH[2] are set to '1'.

When I adjust short to Micbias detect threshold from 420mV to 180mV or lower, only CH1_LTCH[2] is set to '1'.

 

There is a contradiction about short to Micbias detect threshold setup between 1) and 2).

Do you have other suggestions? 

 

 

  • Hello,

    Welcome to the forum! I don't think I understand the contradiction right now. If I understand correctly, in situation 1 you were externally forcing the input pins to 8.63V and 0.035V correct? Since IN1P is 370mV from MICBIAS it makes sense that you wouldn't see this trigger until you set the threshold to 420mV. Note that since the diagnostic ADC is only trimmed to 8-bit accuracy, the LSB size is ~45mV so there can be a bit of error in the actual voltage that trips the threshold (which is why you may not see it trigger at the 360mV setting because this is within the measurement error).

    In the second test, you are biasing the inputs with MICBIAS and the input levels are based on the MICBIAS voltage, the bias resistors, and the impedance of the onboard microphone. Then you short IN1M to MICBIAS using SW1? When I do this, I see IN1P sitting at approximately 8.78V, which is pretty close to the same as what you have measured. With your short threshold set to 420mV shorting IN1M to MICBIAS should report that both IN1P and IN1M are shorted to MICBIAS since IN1P is also pulled within this threshold. When you set the threshold to 120mV though, you should notice that only IN1M is flagged as being shorted to MICBIAS, which makes sense because IN1P is sitting 160mV away from MICBIAS when IN1M is shorted and it no longer falls within the threshold. 

    Please let me know if I have misunderstood anything!

    Best,

    Zak

  • Yes, you have caught my understandings. But the confusing problem is the setup about DIAG_SHT_MICBIAS.

    If we need open input diagnostic and INM short to MICBIAS diagnostic meantime,should we adjust DIAG_SHT_MICBIAS from 120mV to 420mV at open input diagnostic or from 420mV to 120mV at INM short to MICBIAS diagnostic at one application?

    In my opinion, there is no need to adjust the threshold after the threshold, scan rate and other parameters are setup completely and the diagnostic for input channels are enabled. 

    So, it seems not reasonable about the setup of short to MICBIAS detect threshold at open input diagnostic (420mV)and INM short to MICBIAS diagnostic(180mV).

    And I have another two doubts,

    1) I tested the IN1P(8.63V) and IN1M(0.035V). The pull-up and pull-down resistance is 1.1k, the input impedance is 50k(differential input).

    VIN1P=9*(50+1.1)/(50+1.1+1.1)=8.81V not equal 8.63V,  VIN1M=9*1.1/(50+1.1+1.1)=0.189V not equal 0.035V

    Could you point my mistake or give me an equal circuit?

    2) As we have concern about power dissipation, could your tell me about the structure of VBAT_IN pin?

  • Hi,

    I think perhaps I may have misunderstood your initial statement. In situation 1 above if your MICBIAS threshold is set to 360mV or below then you should not see any faults other than INM1 short to GND (unless you have also set this threshold below 35mV. You may also see VBAT faults if that pin is floating). An open input fault is a combination of a short to MICBIAS on one pin and a short to GND on the other, so once you raise the short to MICBIAS threshold to 390mV or greater you should see the short to MICBIAS as well as open input faults. Is this not what you are seeing?

    Below is an equivalent input model you can use to better understand the biasing.

    The VBAT pin is just a high impedance input with ~1Mohm input impedance to monitor the battery voltage. The power dissipation through this pin is negligible.

    Best,

    Zak

  • Hi Zak, 

    You are right. In the practical application, if I need open input diagnostic and INxM short to Micbias diagnostic at the same time.

    I setup the threshold of short to Micbias to 390mV and higher.

    When CH1_LTCH[3] and CH1_LTCH[2] are set to '1', I must consider this event is INxM short to Micbias issue.

    It seems no problem technically , but comparing to CH1_LTCH[2] is set to '1' individually, there is a little confused about CH1_LTCH[3] and CH1_LTCH[2] are set to '1'.

    If you have other good methods, please help to put forward, thanks.

    Based the equal circuit, in the AC coupled rext calc, I found the R1=16k, R2=R3=8k, this value is different the circuit you given. Which one is right?

    And at the input of op, is it more reasonable at R3 right side needn't connect together based on virtual open?

  • Hi,

    I think I understand now. When you short IN1M to MICBIAS, this will also pull up the other input through your bias and MIC impedance. The exact value this gets pulled up to depends on the impedances. If your short to MICBIAS threshold is really large (like 390mV or higher in this case) then this can cause the device to also report an IN1P short to MICBIAS because it is within the threshold to MICBIAS you specified even though it technically has not been directly shorted to MICBIAS. When you don't have anything connected between the inputs, there is no path for IN1P to also be pulled up, so only an IN1M short is reported.

    I think that the values in the model you attached were chosen as a worst case approximation with possible process variation. It is not likely you would actually see a device that behaves this way and you will notice the values I gave provide a more accurate estimation.

    Also remember the inputs to an op amp act like a virtual short, not a virtual open.

    Best,

    Zak