Post Go back to editing

LTC2983 one probe shadowing temps of another

I am testing an LTC2983 for engine CHT and EGT temperature measurement.

The LTC2983 is connected to an Arduino through ISP.

I have two type J thermocouples connected, one with differential measurement, the other with single wire.

The two probes are connected as shown in data sheet figure #4.

Both probes terminate on a 4 mm ring, and the TC joint makes contact with the ring, i.e., they are "grounded probes".

Test 1: probes rings grounded (connected to the PCB ground)

When probe 1 is heated, probe 2 follows similar behaviour, reporting ~ 50% of probe 1's temperature, even though probe 2 is not exposed to heat.

If probe 2 is heated instead, the mirror behaviour is observed

If the heated probe is disconnected, the other probe returns to the normal temperature.

Test 2: probes rings NOT grounded (disconnected from the PCB ground)

The probes behave normally, totally independent from one another.


I have an outboard diode (actually a transistor) connected to channel 20 for cold joint temperature, and that works well.

Other connections and components are per figure #31 in data sheet (except for sensors, only two TCs connected as described above).

Given that the probes will be screwed to an engine block, therefore grounded, how can the behaviour in test 1 be corrected??

  • Hi

        I am wondering if this might be due to a ground loop caused by having the tips connected to PCB ground as well as the ends of the TC.

        Along the same lines, I don't think you will get a correct result for the single ended probe. You are effectively cutting the TC in half, shorting out any potential developed on the negative wire. 

         My suggestion would be to try test1 again but with both probes connected differentially, and each lead left floating (i.e. tied to the LTC2983 input but neither tied to GND). The tip being grounded will keep the ends from floating outside the common mode range of the LTC2983 inputs. 

    -Logan