Switching thermocouple signals using an analogic Mux???

Thread Starter

esm.

Joined Jan 20, 2012
30
Hello all.

The voltage signal of a thermocouple type T is aproximately 40uV \ °C for a temperature range of 0 to 100 °C.

I want to read 15 type T thermocouples, all of them have same reference (ground reference, negative wire, minus wire).

I want to use an analog muliplexer (16:1 mux, model 74hc4067, it will be powered from a +3.3V rail) to switch the positive wire (positive voltage signal) of these 15 thermocouples into an OPAMP (I am going to use the LTC2050 from Linear Tecnology, it has a maximum input offset voltage of +- 3uV and very low offset drift generated by temperature variations).

I know the analog switches of an analog multipler have an internal resistance when conducing, but I think there is no problem because the non-inverting input of OPAMP has very high impedance.

I will use a non-inverting amplifier with a gain of around 200x, and its output will be connected to a 18 bit precision ADC with internal precision voltage reference and I2C interface to communicate with a microcontroller.

The main doubt now is:

Is it possible to conduce very a small voltage signal through an analog switch????

Because the type T thermocouples generate a voltage of around 40uV / °C.
Somebody know if a there is a minimum acceptable voltage level in order the analog switch can conduce the signals???????
 

MrChips

Joined Oct 2, 2009
19,154
There is no minimum voltage. The analog switch will look like a resistor and is linear with voltage. The problem is the ON resistance is not guaranteed to be constant and stable. You may experience problems at thermocouple current and voltage levels. Someone else can correct me if I am wrong.
 

Thread Starter

esm.

Joined Jan 20, 2012
30
There is no minimum voltage. The analog switch will look like a resistor and is linear with voltage. The problem is the ON resistance is not guaranteed to be constant and stable. You may experience problems at thermocouple current and voltage levels. Someone else can correct me if I am wrong.
Hello friend.
Why do you think that???

I did an analog ground track carefully in the PCB (it is not ready yet), there will be a very very small current flowing through analog ground tracks (and they will have 20mils width). I also predicted an LC filter between the thermocuple output and OPAMP input and LC filter for power supplies of OPAMP and the ADC converter.

My doubt about the analog switches is because they are like some mosfet in anti-paralel, conducting te signal in both ways, thats what I understood after check the datasheet (74HC4067)

I think there is no problem with the ON resistance value, because the noninvert input of the OPAMP has a much higher impedance.
 
Last edited:

crutschow

Joined Mar 14, 2008
23,140
The mux should be fine to switch the signals. But make sure that the return line of the thermocouples goes directly to the common for the gain resistor of the op amp. Best would be a separate trace directly from the pad where the resistor connects to common. You don't want any other current running through any part of the thermocouple ground connection.
 
Top