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.

LMT84-Q1: Misbehavior @ 108deg C

Part Number: LMT84-Q1
Other Parts Discussed in Thread: LMT84,

Hi Team,

One our customer is using LMT84-Q1 and observing misbehavior around 108 deg C, the output reference from the device starts increasing at 108 degC indicating that the temperature is getting reduced after 108deg. The device is used to measure the temperature of 3phase inverter power stage and near to the switching mosfets, 2 LMT84 are used at different locations to measure the power stage temperature and given to controller ADC for average temp measurement. Misbehaving issue is observed at one location were the output track is routed between power tracks,we cut this track and connect the signal directly to ADC using wire  but the same issue repeated.

Attached is the schematic, layout and temperature plots for your reference. Can you please help with  feedback and observations to rectify this issue.

Temperature sensor layout.pdf

Plot-1.pdf

Plot-3.pdf

Thanks in advance for your help and support. 

  • Hi,

    I have a few questions for you:

    The curves you supplied do not have x y axis labels. Can you describe to me what they are showing (exactly what nodes in your schematic are you probing, lable the xy axis etc..)?

    When you bypassed the track did you bypass both sections you show in the layout snip? (did you used shielded cable, shield grounded at one end to quite ground)

    Does the behavior always happen at the same temperature?

    How many boards have you seen this behavior on?

    What ADC are you driving?

    What exactly are you doing to change the temperature?

    Take care,

  • Hi Emmy,
    Please find the below reply.

    The curves you supplied do not have x y axis labels. Can you describe to me what they are showing (exactly what nodes in your schematic are you probing, lable the xy axis etc..)?
    X-axis is Time (change in temperature) and Y axis is temp sensor output voltage, in the 1st plot error reading starts after 19mins and in 3rd plot error reading starts after 42mins ( two plots were taken at 2 different output load condition). Probing at TP156.

    When you bypassed the track did you bypass both sections you show in the layout snip? (did you used shielded cable, shield grounded at one end to quite ground)
    Only the error giving section is bypassed. The track is cut at TP152 and connected to TP156. No shielded cables are used with shield ground are used.

    Does the behavior always happen at the same temperature?
    Multiple times observed at same temperature.

    How many boards have you seen this behavior on?
    Almost all the boards out of 15-20 boards.

    What ADC are you driving?
    Driving ADC inside the controller, inbuilt SAR ADC.

    What exactly are you doing to change the temperature?
    The power stage is 300W inverter stage driving the motor/charging the battery, the temperature change is due to the change in output load condition.

    Thanks in advance for the help and support.

    With Regards,
    Shinu Mathew.
  • Hi Mathew,

    Thanks for you reply. I am concerned that something is happening in your system since initially the voltage is fine for quite some time. Could your load be shutting down causing the temperature to decrease? (perhaps there's a thermal shutdown taking effect in the regulator or some other device?)

    Thanks for your patience!

    Take care,
  • Hi Emmy,

    Thanks for reply.
    The load is not shutting down, if that is the case the same variation should happen over the second LMT84 sensor also. We hope this is sensor misbehavior when exposed to high frequency switching circuit. We have one more LMT84 on the same board placed over the other side and the same is giving perfect reading.

    Please help with your thought to rectify this issue.

    With Regards,
    Shinu Mathew.
  • Hi Shinu,

    You can look for inductive coupling as well (try to find out what is physically around the LMT8x that is misbehaving).

    Send a scope photo of the VDD for the LMT8x that is miss behaving.

    Make sure they cut both sides of the track not just one - if one side is connected then you have a big antenna hanging on and the impedances are fairly high. They could try to parallel the caps they have with smaller (higher frequency) cap (100pf of so).

    find out what happens in the system when the LMT8x output rises (is the motor spinning faster? the charger is turned on?). Identifying this will give us clues where to hunt further.

    Thanks to you and the customer for your patience!

    Take care,