We have an application where we need to measure temperatures with unshielded, welded K-type thermocouples. The thermocouples are spot-welded onto a conductive surface, so the thremocoules are grounded. In a different application the thermocouples are spotwelded onto conductive surfaces that have a couple of volts potential between them (no more than 1V).
Currently we are measuring temperatures with multiple thermocouple sensor ICs connected to isolated power-supplies and sending the measured data through opto-isolators.
Would the LTC2986 be able to measure temperatures on multiple channels like described above (with grounded probes or probes with potential between them) so we dont have to have a convoluted isolation system? Also we pick up a lot of noise with the current system, the LTC2986s noise rejection will surely help.
Thank you in advance for your answers.
The inputs can go 50mV below ground with no change in linearity. This allows grounded tip thermocouples to be digitized directly without biasing or isolation. In this case since the tip is grounded the TC+/- should connect directly to the LTC2986 without connecting TC- to GND as is shown in the datasheet - that is done in the examples just so the TC doesn't float outside the common mode of the part.
For the case where there is a potential between thermocouples, you'll need to be careful not to exceed the part's ABSMAX (VDD - 0.3V to GND -0.05V). If the potentials are something like 0V, 1V, 2V, 3V, and you're on a 5V supply wire them up the same as the grounded tip example above. Otherwise if they are like 0V, -1V, 2V, -1.5V I would suggest looking into powering the part with a bipolar supply of +2.5V as Vdd and -2.5V as Vss (labeled GND). That way mid-supply for the LTC2986 is at system GND.
Let me know if that makes sense and if you have any questions.
Thank you for your detailed answer. Ill purchase a device for testing from an AD supplier.
Have a nice day,