LM60: About LM60 detection precision

Part Number: LM60

hi dear supporting team,

as for LM60,  the temp it reports  is the brd temp or the ambient temp?  how is the precision of +/-4degree get?(comparing to which position temp)

my customer is using it for detecting the temp in power module. and meet below issue ,  on the layout, LM60 is soldered above the output current trace which could have big to 20A current flow through:

1. when the power module is not loaded with high current,  LM60 could get almost same result with both the other probes on the brd and on its case. 

2. when the power module is fully loaded with 20A current,  LM60 will have bigger than 5degree error.  for an example, the brd probe will report 35degree temp, and the probe on LM60 case will report 29degree,  while LM60 will report 40degree. 

the temp data is gotten from the output voltage, and look up the table directly, so no ADC error involved/ 

and the issue shows up on multiple LM60s.  so is it normal ? tks a lot!

  • Vera,

    LM60 measures local temperature mainly through the leads.
    The board temperature will dominate the LM60 temperature measurement because the thermal conduction pathway is dominated by the metal leads.

    How accurate are the reference probes?
    You will have to add the reference probe's error to the LM60's error to get the max error limits. For example, if the LM60 has a max error of +/-4C and the reference probe has a max error of +/-1C, the worst-case error you can report is +/-5C.

    The 20A current will also produce a temperature gradient. So unless you have very good thermal contact with the LM60's case or board, the temperature reading will not be accurate. Thermal grease should help with the thermal conduction.

    Have the customer test the LM60 and the reference probe in a temperature oven at the temperature of interest as a sanity check?
    The oven will give them a uniform temperature test environment to verify the accuracy of the LM60.

    -Kelvin
  • In reply to Kelvin Le THS:

    hi Kelvin,

    thank you very much for the reply!

    the reference probe is very precise and only have 0.x level error. 

    and yes, they also did temp oven test as well, and reported at -40degree, the error is bigger than higher temp. while when I was on site I did not see the failed phenomenon.  I will follow up and turn to you once there is further issue. thanks a lot for the support!