• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 30 subscribers
  • Views 428 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.

2-wire LMT86

Iosif Szilagyi
Prodigy 20 points
Other Parts Discussed in Thread: LMT86

Hi,
I report here a 2-wire connection to LMT86, using only the GND and Vs lines.
The Vs line connects a commutable port of a MCU with the output of the LMT86,
thru an 1k current limiting resistor.
The dominant state of the port is high level output state(+5V).
The supply is provided locally by means of a 10uF high temperature tantalum capacitor,
connected btw. Vcc and GND. This capacitor is charged from the Vs line thru a Schottky diode
to~4.5V.
Right before the AD conversion, my program commutes to the ADC input and waits S&H time.
The output voltage vs. supply voltage diagramm is remarkable flat, so here is room for
different conversion speeds.
I made a lot of tests with a 2m long line(2x,twisted,28AWG).I fund loss of accuracy
less then 1/LSB at 12bit resolution.
An 1nF decoupling capacitor reduces the input noise, however some dither is welcom.

My questions are :may I think about series production? Is here some hidden error in this
schema?It would be possible a future version of LMTxx which incorporates the limiting resistor
and the diode?

  • 0
    TI__Expert 4905 points

    Hi Iosif,
    What is your temperature range?
    I am surprised this works. How did you measure the accuracy?
    I would Would mainly be concerned with loading of the temp sensor and self heating. 
    I would need the temp range to do some calculations. But basically the Sensor would be sinking the current from the 1k resistor so let say the output wants to be at 1V and the supply is 4V the the current that would flow into the device is 3 mA and 3 mW would be dissipated by the sensor therefor the self heating would approach 1 C or so. I highly recommend you increase the resistor to 100k.
    Since the current would drop dramatically you may find that you could also drop the size of the capacitor.
    Take care
  • 0 in reply to Emmy Denton
    Prodigy 20 points

    Dear Emmy,
    Thank you for your kind answer.
    I made a comparative measurement schema for relative accuracy determination, using the same MCU and DUT,
    with an auxiliary port(directly to Vcc), and switching btw. 2-wire and 3-wire by soft.
    Any offset and slope differences can be resolved by software.   
    Actually, I have a periodic short sequence of ..HiZ/+5V(10msec)/HiZ,wait/ADC1/wait/ADC2/wait/ADC3/HiZ..with
    a period of 100msec.I am using 3 consecutive conversions for self-heating correction by
    extrapolation (something like Keithley's Delta Method)
    I am planning to play with this period,maybe I can obtain some 50/60Hz rejection.
    The main problem with this 2-wire solution is the commutation noise,it exist an optimum
    waiting time for de-ringing before the first ADC ; unfortunately this is line length dependent.
    This is why I'm not very happy to increase the 1k resistor.
    Better, I will make efforts to improve the thermal bridge btw.LMT86 and case.
    The size of the capacitor is determined only by LMT86 supply current.In ADC state the diode is
    reverse polarized,blocking the way from the capacitor to the resistor.

    Iosif