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.

PCA9306: PCA9306 not working

Part Number: PCA9306

Hi,

 I am using PCA9306 for level translation from 3.3 V to 1.8V on i2c Bus.

1. First doubt I have is on the orientation of the IC as the one supplied to me does not have a reference dot mentioned in the datasheet, it rather has a laser marked line on one side. Not sure this is for indicating pin1?

2. When used in the orientation shown in the schematic (laser marked line as reference for pin 1), we find LDO losses regulation. With changed orientation LDO starts working but the I2C master is unable to pull the input lines (pin 7, 8 on P1 CONNECTOR) to low, they always remain at high state.

  • Hi Mohit,

    I believe your initial orientation is correct (with the mark at the top of the device). If rotated and driven in the way you describe, i believe the results would look consistent with what you were seeing.

    Does the controller device your using (P1) have internal pullup resistors? SCL2 and SDA2 need to be pulled up to Vcc. Without these, the lines would only reach Vdd which may not be readable by the controller. Are these lines where you're seeing the loss of regulation?

    Could you supply a waveform of the Vdd supply when the device is placed in the initial orientation? The value that it reaches or any oscillations would help us debug.

    Regards,
    Eric
  • Hi Eric,

    Yes the controller device driving pins 7 and 8 (at connector P1) has strong internal Pullup resistors.
    I am finding loss of regulation on the LDO output with regular orientation of PCA9306. I have done some measurements through the multi meter at LDO (TLV70018) output pin 5 with normal orientation below are the observations:

    Sample No.  Vin Actual              Vout                  Expected Vout Voltage
    Board1          2.993V                   2.055V             1.8V
                         4.990V                    3.917V             1.8V

    Board2         2.993V                   2.046V              1.8V
                         4.990V                   3.901V              1.8V

    Thanks for the response, I will try to share the waveforms too.
    Mohit.

  • Hi Mohit,

    Could you also measure the voltages at some of the pins of the SCA9306:

    EN, VREF2, SCL1, and SCL2.

    Also, are the 0.1uF and 1uF capacitors for the LDO and the 200k at EN populated according to the schematic?

    Regards,
    Eric
  • Hi Eric,

    Below are the traces required : 

    attachment 1:   EN or Vref2: Enable in blue trace, VCC in yellow.

     

    attachment 2 :  LDO input  in Blue and output (Vref1)  in yellow.

    attachment 3 :  SCL1 (PCA9306 Output) in blue and SCL2 (PCA9306 input)  in yellow.

    Above traces are recorded with normal orientation of PCA9306 (attachment 4):

    0.1uF and 1uF capacitors for the LDO and the 200k at EN are as per the schematic.

  • Above traces were taken with normal orientation of PCA9306 (U3 in the Board Image) and I am getting that LDO is unable to regulate the output to 1.8V but PCA9306 works properly. When we selectively remove the components and check output of LDO. The LDO regains the regulation only when PCA9306 is removed. Due to these observations and finding no laser marking mentioned on datasheet (for marking orientation ) we suspected orientation.
  • Hi Mohit,

    It's good to see that data is being successfully sent through the device. This means that this orientation is correct!

    On the scope shots, it appears that the EN/Vref2 pins are higher than Vcc and both of these values may be slightly greater than 3.3-V. Could you confirm your supply is 3.3-V with a cursor in the scope shot to get a more precise reading of the voltages?

    The fact that the EN/Vref2 pins appear to be very close to the Vcc voltage indicates a problem somewhere either in the layout or the device itself. These pins should stay around 2.4-V when Vref1 = 1.8-V and Vcc = 3.3-V.

    Could you ensure that the EN and Vref2 pins are in fact connected on the board? It appears that the trace may be under the device but doing a continuity test with a meter would make sure this is not the problem.
    The 200k resistor appears to be populated correctly (code 30D mapping to 200k-Ohm), but could you measure with an ohmmeter to be sure? Low values for this resistor are the most common scenario we see with EN/Vref2 being so high.
    You may want to try replacing the device with a new one. It's possible that soldering and removing the device has damaged it.

    Regards,
    Eric
  • Hi Eric,

    I can confirm that Vcc was 3.3 Volts. As I am using a Digital multi-meter too at the supply output, it showed 3.3V at output.

    Checked En and Vref2 continuity through multimeter and found that they are connected.

    Checked 200K resistance, it is 199K by multimeter. 

    Also checked capacitance at Vout pin of TLV70018 found it 3uF.

    I removed C5 to measure it's exact value but i think 100pF is too low for Multimeter i have.

    Voltage at EN is 3.0V with 3.3V Vcc (after removing C5)

    One more interesting thing i missed to tell, when we un-populate the PCA9306, LDO TLV70018 regains the regulation and Vout is set to 1.8V. Un-populating any other component does not make any impact.

    Thanks,

    Mohit.

  • Hi Mohit,

    Thank you for taking these measurements. It's helping us narrow down the possible reasons for this behavior.

    Have you tried replacing the device? It's possible that a damaged device could have a short and is allowing leakage voltage to the LDO output.

    Would it be possible to measure the current coming out of the LDO's Vout? It's possible that current from the PCA9306's Vref1 (negative voltage to the LDO) is causing it to lose regulation. I'll look closer at this possibility.

    It seems difficult given the way the PCB is laid out, but could you disable the device (EN and Vref to GND) to see if the loss of regulation persists? (Note: ground these pins on the far side of the 200k resistor)

    Regards,
    Eric
  • Hi Eric,
    I tried with other similar part but the problem remains same, LDO not regulating the output.
    Removed the 200K resistor and checked if the LDO regains the regulation, LDO does not regains the regulation.
    Then I removed C5 and put a short there to ground to Enable pin of PCA9306, LDO is able to regulate in this case. LDO Output is 1.8Volts with PCA 9306 disabled.

    I have tried measuring the LDO output current before, what i got is, there is no current flow observed when it is out of regulation.
  • Hi Mohit,

    I believe the loss of regulation could be from the LDO not being able to sink the current coming from the PCA9306 device's Vref1 pin. When idling, this pin leaks about 5uA of current. With not other path, this current may be sinking through high impedance paths creating the voltage that deregulates the LDO.

    To sink this current on the board you have, you could replace one of the LDO capacitors (C1) with a 300k-Ohm or smaller resistor. This would allow the current to dissipate and the LDO to maintain its set voltage.

    Let me know the results of this modification.

    Regards,
    Eric

  • Hi Eric,

    I put 200k in place of C1. It works!! LDO ouput is 1.8Volts. Not checked the communication yet, I will update you on that too. Thanks for the support. 

    One more question I have, what are those parameters i should look in the level translator datasheet or LDO datasheet if i have to avoid this kind of situation in future?

  • Hi Mohit,

    Glad to hear that solved the LDO problem!

    Unfortunately it is a pretty specific set of properties that caused this interaction. The leakage current from the PCA9306 (5uA) is specified on the datasheet but only as an input. The TLV70018 also doesn't clearly state it's behavior when trying to sink current because this is an unusual use case.

    In another similar system with more loading on the 1.8V line or that used another means of level translation (such as a buffer) may not experience this problem at all.

    I look forward to hearing how the rest of the tests go.

    Regards,
    Eric
  • Hi Eric,

    I checked the I2C communications, they are good now ! . Thanks allot for all your support.

    Mohit.

  • Hi Mohit,

    Very good to hear. Good luck on the rest of the project!

    Best,
    Eric
  • Issue Resolved. Query can be closed now.