TMUX1209: Can TMUX1209 be used as the multiplexer before ADS1118 to connect more thermocouplers?

Hi Bin,

I have a quick question to check: what are the expected voltages coming from the thermocouple with respect to ground and what is the power supply you are planning to use for the multiplexer / ADC ?  Based on the ADC I don't think there should be issues, but I want to make sure.

As for the application, you want a multiplexer that is going to affect the signal the least. Typically that means we need to focus on a few parameters:

On Resistance + Variance/Mismatch, On Capacitance, I/O Leakage Currents, and charge injection.

On resistance will degrade voltage as a current passes through it - For this application the input impedance of the ADC is very large so on resistance will basically be negligible for lower on resistances - The TMUX1209 has a typical  on resistance of 5 ohms at a VDD = 5V. It has a typical variance of 1.5 Ohms and a  typical channel mismatch of 0.15 Ohms. There is a max resistance of 9 ohms, at a 5V input there could be a typical max input current  of ~1uA which could would cause a 9uV drop across the TMUX1209.  This loss should be negligible  when compared to the ADC error. 

On Capacitance can distort the signal due to different delays for different frequency content. The TMUX1209 has a low on capacitance (42pF) and at the speed of a sigma delta ADC this should not be a major issue and can be neglected.

I/O Leakage currents: Leakage currents can cause errors due to more charge being injected into the ADC. When the TMUX1209 has chosen a channel the leakage from the on channel typically is  +/- 200nA - this is where we may have issues because this current can increase/decrease the voltage value read by the ADC, and this current is 1/5 of the appx. max current that the application should be using w/o leakages. However if the temperature goes near the max operating range of the mux we could see up to +/- 750nA. Depending on the accuracy the customer is trying to achieve this may be an issue, and with a sigma delta I'd assume this is what the customer is trying to get. 

Finally we get to Charge injection: When the channel switches a voltage transient will occur on the output. This is due to charge injection which follows the following equation Q_inj = (C_load + C_on) * delta V. If we divide our load capacitance from charge injection we get a approximated voltage transient. The TMUX1209 also has a higher charge injection at  +/-9pC   which means for every 10pF of capacitance on the output we will have a  +/-900mV transient occur. which will cause failed measurements at transition points.

The TMUX1209 is a good part, but due to some issues of leakage and charge injection there could be more error created by the device. However we do have a higher precision device that is better suited for these types of applications: The TMUX1109. Which is a higher precision version of the 1209. It has improvements on every category:

Ron - TMUX1109 Typical 2.5 Ohms - with a max of 4.9 Ohms < TMUX1209  Typical 5 Ohms - with a max of 9 Ohms

Con - TMUX 1109 Typical 38pF  < TMUX1209 Typical 42pF

On Leakage - TMUX1109 Typical 3pA - with a max of +/- 3nA < TMUX1209 Typical +/- 200nA - with a max of +/- 750nA

Charge Injection - TMUX1109 Typical -1pC < TMUX1209 Typical +/- 9pC.

Due to the ADC chosen I'd strongly suggest the TMUX1109 due to its lower leakage and charge injection as this will affect the thermocouple signal the least. 

Please verify the signals being muxed , mainly their voltage levels w.r.t. ground and the chosen power supplies.

Best,

Parker Dodson