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.

TMP422-Q1: Polarity of DX+/DX- signals

Part Number: TMP422-Q1
Other Parts Discussed in Thread: TMP422

We have several boards with remote temperature value problems.  For some, the value takes several minutes to change, for instance, at power on it reads 0C, after some minutes it finally drifts up to 22C (correct value).  Others read extreme negative values - for example, -50C for a room temp measurement.  The local sensor is consistently correct.

We are using 2N2907A transistors, in the transistor configuration, on cables about 24 inches from the PCB with the TMP-422.  We are in extended temperature mode (and accounting for the 64C offset) and sampling at 1 second.  Our schematic is exactly Figure 21 in the datasheet, with Rs = 50 ohms and Cdiff = 100pF.

One thing I've noticed is that, on good channels, the DX+ and DX- signals are at approximately 0V, then jump to something like 1.5V and 1V, respectively, during the measurement interval.  On the bad channels, the signals are at about 2 or 3V, and drop down to 1.5V and 1V during the measurement interval.  This seems wrong to me as it puts a voltage across the transistors all the time.

Any ideas?

  • Hi Bailey,

    Can you please send waveforms of the DX+ and DX- pins?  Also, what type of cables are you using? If the temperature is drifting up after start up it sounds like this could be a charging effect and could be caused by parasitic capacitance or noise coupling on the cables. 

    You are correct there should not be a voltage across the transistor at all times.  Please see the below links for common issues regarding remote temperature sensors. 

    https://e2e.ti.com/support/sensors/f/1023/t/892915

    Thanks,

    David

  • This is part of a sensor system for the Orion space capsule, so cables are military style - circular connectors, twisted shielded pair.  I've attached four scope pics - 

    64 - DX+ (good channel)

    65 - DX- (good channel)

    70 - DX+ (bad channel)

    71 - DX- (bad channel)

    Are these two different polarities indicative of the different transistor connections (transistor configuration  vs diode configuration)?  It's possible we have some of our cables wired incorrectly.

    DX+ (good channel):

    DX- (good channel):

    DX+ (bad channel)

    DX- (bad channel)

  • HI Bailey,

    This is very strange to see the DX pins continuously have a voltage across it whether in diode or transistor configuration.  Typically when the DX pin goes close to V+ this is an indicator of a short. 

    What is your supply voltage to the TMP422-Q1? 

    Can you put the device into shutdown mode (Register pointer 0x09 bit:6 = 0)?  There should be no voltage on any of the DX pins. 

    Can you also try replacing the transistor in case it has been damaged?  

    Thanks,

    David

  • Supply voltage is 3.3V

    It would be difficult to modify our s/w to try shutdown mode.

    We've tried different sensor cables with the same results.  I've ordered some 2N2906 transistors that should be in today that we can test with, independent of our cables.

  • Bailey,

    Please keep me updated on the results using the 2N2906 transistor.  

    Thanks,

    David

  • We replaced the TMP-422 on 3 units and saw no change.  We tried 2N2606A transistors and saw no change.  I'm inclined towards believing this is possibly a leakage issue.  The voltage outside the conversion interval varies from unit to unit.  It's always at zero for the good channels but higher for the bad channels.  Am I correct in assuming that the DX+ and DX- pins are turned off (floating) outside of the interval?  Or are they switched to ground?  If they are switched off then these voltages could indicate leakages on the signals.  I can test that by removing the series resistors and looking at the voltages on the transistors.  We're currently having the Cdiff caps removed to see what that does.  

  • Hi Bailey,

    The DX+ and DX- pins on the TMP422 are floating outside of the conversion cycle.  Please let me know if removing the capacitors and series resistors fixes your issue.

    Thanks,

    David

  • Removing the caps had no effect.  When I removed the resistors, and the transistor cables, the DX+ and DX- signal traces were effectively isolated - but the voltage levels I observed outside of the conversion interval were still there.  So I believe I have a leakage issue with my PCBs. I'm investigating further.

  • We've positively identified the issue as leakage resistance apparently caused by moisture in our PC boards.  Thanks for your help.