Connect with us

cheap temperature measurement 0-200C

Discussion in 'Electronic Design' started by Naveed, Feb 26, 2004.

Scroll to continue with content
  1. Naveed

    Naveed Guest

    Hello,

    How would you guys measure the temperature off an IC surface that cam
    go upto 200C. I need such a high temperature for thermal runaway.

    I have been seeing a very high prices for Thermistor and RTD above
    150C range.

    I have a 12-bit ADC, a PGA (1,10,20 gain), and a microprocessor
    already available on the board.
     
  2. If cost was not a limitation, I'd use a metal film RTD. Depending on
    quantity, should be < $5-10, and good for way more than 200°C (though
    there are some low grade types only good for 150°C).

    I might also try to cheat by using some kind of on-chip diode as a
    sensor, if that could be sufficient.
    Best regards,
    Spehro Pefhany
     
  3. Robert Baer

    Robert Baer Guest

    As seperate, inexpensive temperature measuring device, look at
    National Semiconductor; the LM35DT which is a TO-220 package and reads
    digrees Centigrade (100mV/C).
    They have other packages and versions that read degrees Farenheit.

    If you want to "directly" read the chip temperature of some IC, then a
    "trick" must be done.
    A fair number of ICs have a built-in protection diode at each input
    pin, and can be forward biased with a current to produce negative diode
    voltage.
    The second half of the trick is to measure that voltage at two
    different currents (say 1mA and 2mA), to get a self-calibrated
    temerature representation.
    These two currents could be relatively short pulses, so that the
    operation of the device is not unduly interrupted, and that the heating
    of the chip is not unduly altered by the measurement.
     
  4. Naveed

    Naveed Guest

    As seperate, inexpensive temperature measuring device, look at

    The problem with all these methods are they won't measure temperature
    above 125C or 150C max. Semiconductor diodes, or LM35 are not rated
    for above 150C. I know that I have to use Thermistor or RTD for upto
    200C, but the question is how to do it without using those expensive
    assemblies made by temperature sensor companies
     
  5. Here it is in eight easy steps:-

    1) Take the sensor element
    2) Stick it onto the surface with adhesive or whatever.
    3) Connect a precision resistor in series with it.
    4) Ground one wire of the sensor
    5) Connect the resistor to a reference voltage, say 2.5V
    The resistor should be something like 2K49 in this case.
    6) Connect the common point to the input of your PGA, set
    the gain to around 8.
    7) Read the ADC voltage and use a LUT and interpolation to
    spit out the temperature.
    8) Enjoy.


    Best regards,
    Spehro Pefhany
     
  6. Robert Baer

    Robert Baer Guest

    Bullshit on the "won't measure temperature" statement.
    I have reliably used the LM35DT to 200C with no problems or loss of
    accuracy.
    Just because a diode or other part is not rated above 85C (a more
    commonly seen limit these days) or above 125C does not mean that it is
    not useable or reliable at higher temperatures.
    The major issue for most semiconductors above 165C, is due to P-N
    junction leakage, which doubles every 10C.
    Most semiconductor devices have bothersome leakage problems above
    175C, but i stated 165C because in almost all band-gap references, the
    internal leakages become dominate and compete with the current
    set-points for normal operation. So, (almost) all low power band-gap
    references are unuseable at 165C and higher.
    The trick for reliably using a diode for temperature measurements to
    200C or so, is to use currents that are at least ten times the leakage
    current at that temperature.
    And 1mA fits that bill (and then some) for most silicon low power
    diodes.
     
  7. Robert Baer

    Robert Baer Guest

    ....err, what does "LUT" stand for?
     
  8. On Sat, 28 Feb 2004 08:53:06 GMT, the renowned Robert Baer

    Look Up Table. You could use the equations too, but it would be more
    computationally intensive on a microcontroller with no FPU (Floating
    Point Unit)- it's not far off linear to begin with (maybe a few % of
    span), and the equations involve transcendental functions.

    Best regards,
    Spehro Pefhany
     
  9. Naveed

    Naveed Guest

    The trick for reliably using a diode for temperature measurements to
    That is very interesting...

    I will try that in the lab, and see if it really works at that
    elevated temperature.

    This is what I will do..

    Use Analog Devices, ADT7461 (Dual remote diode temperature sensor,
    with built in current source and ADC. can measure upto 158C)
    A cheap 2N3904 transistor in TO-92 package.

    Heat the TO-92 package to around 150C and see if the temperature
    drifts within 24 hours. Also check, if 2N3904 can be thermally
    shocked (25C to 150C) multiple times and still works OK.

    Naveed
     
  10. Metal case transistors such as the (real) 2N2222A are actually rated
    for Tj of 200°C, but getting them to accurately read a surface
    temperature would be rather difficult.

    Best regards,
    Spehro Pefhany
     
  11. Conrad.nl do a si diode temp sensor that is specified to work from
    -60 to +300 C and costs 1Euro59.
    Model KTY84-130 (Google?), and order nr 18 34 58-02
    they may still be intoxicated from the carnaval, like last week...,
    so make sure you get what you ordered ;-) for the price you
    ordered it.
    JP
     
  12. IIRC, a Euro manufacturer has a cheap Si resistance sensor (in a diode
    package) good to much higher temperatures (for self-cleaning ovens).
    But a decent RTD would be my choice if price is not overwhelmingly
    important.

    Best regards,
    Spehro Pefhany
     
  13. Yes could be this one, it is a diode package, and it says in the catalog:
    1000 Ohm +/- 3%, not a simple si diode.
    It all depends on what you want to do I guess.
    I did temp- sensing with a simple .05 cent si diode in some big industrial
    equipment, where a comparator was used to get a max value.
    Made the max value 37.5 C and calibrated it by warming it up with my hand,
    then setting the comparator to just activate.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-