• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 7 answers
  • Subscribers 30 subscribers
  • Views 367 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.

PGA305: Issues with OWI activation

Matias Ljøen
Prodigy 50 points
Part Number: PGA305
Other Parts Discussed in Thread: USB2ANY

Tool/software:

Hi,

We are currently unable to activate OWI at around 30-50% of our sensors using PGA305 with 4-20mA mode.
This has not always been a problem and we have had 100's of these sensors working perfectly fine, but now there is a big issue with activating OWI that was introduced just recently.
We have tried all kinds of combination of Rloop and additional voltage, nothing's working. And we have also measured the voltage and timing of OWI activation signal which seems to be correct according to datasheet, but there is no reply signal from PGA305.

This activation issue is only an issue when we have the electronics with PGA305 installed in a sensor. If we take the electronics out of the sensor, we are able to activate OWI, but not when it is installed in a sensor.
When the electronics is installed in the sensor it is connected to a wheatstone bridge of around 2.7k impedance and a temperature sensor.
The PGA305 is also programmed with calibration coefficients and scaling factors that do not correspond to the wheatstone bridge it is connected to, and the current we measure is around 3.1 mA to 3.4 mA.
We have found that if we delete the calibration coefficients and scaling factors, change Vbrg to a lower voltage or increase iTemp we are able to activate OWI.
But we have not found the root cause of why we are not able to communicate.

We have made some changes in EEPROM from the sensors that was working fine to the sensors with this issue.

See below picture for EEPROM data in sensor with issues to left and to the right is EEPROM data in sensor with no issues activating OWI:

  • 0
    TI__Intellectual 1980 points

    Hi Matias, 

    Let me ask a couple questions so I can better understand your issue:

    1. Just to confirm, you can operate PGA305 by itself 100% of the time, but when it has a sensor connected, only 30-50% of your sensors work?
    2. You mention that previously, you were able to get 100s of sensors to work. Has anything changed in your system? Are you using a new sensor?
    3. What voltage are you measuring at the PWR pin (pin 5)? What voltage do you measure between PWR and FBP (pin 5 and pin 12)?
    4. Are you communicating with the GUI over OWI or I2C? Are you using a TI EVM?

    Best,

    Maggie

  • 0 in reply to Maggie Lee
    Prodigy 50 points

    Hi, thank you for your quick reply, I will try my best to reply to your questions.

    1. Yes, that is correct.

    2. There has been some changes to SW and EEPROM variables, but the sensor has not been changed.

    3. We have tried adjusting additional voltage in 2 different ways to make sure. One of the way is adjusting additional voltage such that the voltage at PWR pin (pin 5) is normally 5V, activation high is well above 5.95V and OWI data low signal is between 3.9V and 4.2V. The other way we tried is adjusting additional voltage such that the voltage between PWR and FBP (pin 5 and pin 12) is at the levels just mentioned.

    4. We are communicating with the GUI over OWI, using TI EVM.

    Best regards,

    Matias

  • 0 in reply to Matias Ljøen
    TI__Intellectual 1980 points

    Hi Matias, 

    Thank you for answering my questions. Can you please share a picture of your EVM or tell me which jumpers are installed? 

    Are you able to send me a scope shot probing the following:

    TP4: OWI activation signal

    TP5: OWI_TX

    TP20: VPWR/OWI

    I'd like to confirm that the activation signals are as expected. 

    Additionally, could you send a scope shot probing the following:

    TP7: Input to OWI_RX Comparator

    TP8: Output to OWI_RX Comparator

    U6, pin 3: Threshold voltage for OWI_RX comparator

    For the RX scope shot, if you are able to take one shot of a device where OWI is activating and one scope shot where OWI is not activating, that would be very helpful.

    If the activation signals are as expected, then there might be an issue with the RX side of the OWI communication.

    Best,

    Maggie

  • 0 in reply to Maggie Lee
    Prodigy 50 points

    Hi,

    The jumpers I have installed is J12, J5, J4, J9, J19 (pin 2 and 3).

    I have used TP20 and TP45 to connect to an external PCB with PGA305 mounted.

    See below pictures for probe shots probing the requested TP's.

    TP4:

    TP5:

    TP20:

    TP7 (Not activating):

    TP8 (Not activating):

    TP7 (Activating):

    TP8 (Activating):

    U6, pin 3:

  • 0 in reply to Matias Ljøen
    TI__Intellectual 1980 points

    Hi Matias, 

    Thank you for taking the scope shots and confirming your jumpers. For the TX signals, TP4 and TP5 look fine, but TP20 is a little off. For reference, here is what my logic analyzer looks like: 

     

    It looks like your Vdd is getting pulled down during the activation pulse. Are you supplying an additional voltage to Vdd on your external PCB?

    I'm curious about your grounding scheme on your external PCB. On our EVM, the loop current on the output is referenced to IRETURN (TP45, TP46), which is different than ASIC_GND (TP39, TP43, TP44), the ground PGA305 is referenced to. There is a voltage difference between those grounds, so if they are shorted, then there will be an additional voltage offset referenced to IRETURN. This can lead to measurement issues (outputting an incorrect current) and issues with OWI. The EVM has pretty tight limits for OWI for transmission and reception. 

    Looking at your RX signals (TP7 and TP8), it looks like the comparator is not consistently triggering when there is a pulse on TP7. This may be due to an unexpected offset in the signal chain. 

    I would recommend checking your external PCB to ensure IRETURN and ASIC_GND (or the equivalents) are not shorted. If that does not resolve your issue, I would recommend lowering the threshold voltage on U6. There is a powerpoint attached to this E2E thread which describes a couple of modifications you can make to the EVM which will change the thresholds required to activate OWI communication. 

    Best,

    Maggie

  • 0 in reply to Maggie Lee
    Prodigy 50 points

    Hi,

    We have not shorted IRETURN and ASIC_GND on external PCB. So the current on the external PCB will flow from the ASIC_GND through the internal 40 Ohm resistor in PGA305 and out FB+ pin through a 10 Ohm resistor on PCB which is then connected to TP45 (IRETURN). 

    I will further try as you recommend, lowering the threshold voltage on U6. 

    I did try using a different TI EVM, which resulted in a different activation pulse on TP20. See below picture for the result. This time the VDD is not getting pulled down as much during the activation pulse, and the OWI is still not activated.

  • 0 in reply to Matias Ljøen
    TI__Intellectual 1980 points

    Hi Matias, 

    Thank you for describing your external PCB. I don't see any immediate issues based on your description. Please let me know how changing the threshold voltage goes. 

    Best,

    Maggie

  • 0 in reply to Maggie Lee
    Prodigy 50 points

    Hi,

    I have now tried changing the threshold voltage and I am still not able to activate OWI communication. 
    I've tried adjusting additional voltage, but this has not helped.

    Is it possible we have written something wrong to the EEPROM that makes it more difficult to activate OWI?
    We have made some SW changes to what we write to the EEPROM, but there's not been any HW changes to our external PCB.

  • 0 in reply to Matias Ljøen
    TI__Intellectual 1980 points

    Hi Matias, 

    Can you please share specifics on what software changes were made?

    I will load the two sets of EEPROM you shared in your original message to see if I can observe the same behavior. Please note, the US is on holiday on Monday, so I will not be able to test it until Tuesday.

    Best, 

    Maggie

  • 0 in reply to Maggie Lee
    TI__Intellectual 1980 points

    Hi Matias, 

    Thank you for your patience. I took at fresh PGA305 and performed the following procedure with the two EEPROM sets you provided:

    1. Set an EVM in current mode. (J19 (2-3), J9, J4, J5, J10, J14 (1-2), J12 enabled). 22V input to the EVM. Power supply connected to VinPP and VinPN (TP34, TP35) ranging from 0mV to ~30mV. No input to VinTP/VinTN.
    2. Put a new PGA305 device in the socket. Power on the board and plug the USB2ANY into my computer. Launch PGA305 GUI and select digital interface mode over OWI.
    3. Write the "good" EEPROM you provided. Read it back to verify successful programming.
    4. Switch device to compensation mode. Use a multimeter while the device to verify the output is changing with the input. (I didn't calculate the Vin to output you're expecting, I was just doing a quick test to make sure device is running).
    5. Put the device back into digital interface mode.
    6. Write the "bad" EEPROM you provided. Read it back to verify successful programming.
    7. Switch the device to compensation mode. Use a multimeter to verify output is changing with the input.
    8. Switch the device back to digital interface mode.

    I did switch the device between compensation mode and digital interface mode several times with the "bad" EEPROM. I did that with various voltages across TP34 and TP35, ranging from 0mV to 30mV. I could reliably activate OWI (by switching digital interface mode on) when my input to TP34/35 was 0mV and 10mV. I started having issues when the voltage across TP34/35 was 20mV. I could still active OWI, but not consistently. I did not observe this behavior with the "good" EEPROM. I was able to active OWI all the from 0mV on the input to 40mV on the input. 

    Between the two EEPROMs, most of the parameters are the same, except for the calibration coefficients. With the "good" EEPROM set, I notice that the device output clamps at around 16.4mA,  while with the "bad" EEPROM, the device output clamps at around 29.6mA, and the PGAIN is set to 133. 133*20mV = 2.66V, which is outside of the PADC's input range. So it looks like the issue is occurring when you have an input close to the full-scale input range of the PADC. 

    Based on this testing, I think there is an additional voltage drop caused by your sensor when it is close to the PADC's full-scale range. This is preventing OWI from activating reliably. 

    When you try to communicate with the device in your system, is it possible to either disconnect the sensor or ensure it is not outputting a full-scale signal?

    If there are any issues with my procedure, please let me know.

    Best,

    Maggie

  • 0 in reply to Maggie Lee
    Prodigy 50 points

    Hi,

    Thank you for testing this, this is very interesting for us seeing how OWI activation depends on the input signal. For our case I believe it is the same, only that our input signal is on the opposite side and below 4mA.

    Our sensors signal is very low, and below the calibrated PADC range. The output signal varies a bit, but tends to be around 3.1 - 3.4mA. This is because the device has been calibrated using a different sensor and then swapped to a new one, which makes the new sensor signal below the calibrated input range, which then we are often not able to activate OWI.
    What we have done for those we are not able to activate OWI on, we are disconnecting the sensor and then we are able to activate OWI every time. Then we are able to delete calibration coefficients, connect to sensor again and make a new calibration with successful OWI activation after this procedure. We have had no issues with this so far. With the new calibration coefficients in the device, the output signal is 4mA.

    It seems to us that when the input signal is way below the calibrated PADC range, causing the output signal to be way below 4mA, there is issues with OWI activation. And from your testing it seems the issue is related to troubles activating OWI when the output signal is far outside the expected range of 4-20mA.

    Best regards,
    Matias