Analog Multiplexer for Thermocouple

I need to interface 3 channels of thermocouple(K- type) with MAX31850 thermocouple signal conditioning IC chip. The output of thermocouple will be in the range of few µV/ °C. Can I use analog multiplexer for my application? Required temperature accuracy is 2°C.

There are many analog switching ICs available, but they all have an on state impedance of a few ohms. It will depend on the output impedance of your circuit (i.e. where is the signal going) whether you can use them or not because low output impedances will cause excessive voltage drops over these switches. You may also consider mechanical relays, if that is undesirable.

udhay_cit said:

I need to interface 3 channels of thermocouple(K- type) with MAX31850 thermocouple signal conditioning IC chip. The output of thermocouple will be in the range of few µV/ °C. Can I use analog multiplexer for my application? Required temperature accuracy is 2°C.

Click to expand...

Due to the low voltage of a thermocouple, I would not recommend multiplexing at the input, but rather at the outputs of preamplifiers.
Analog multiplexers are many and good but they introduce voltage errors that are difficult to remove.
After a preamplifier (opamp, gain >10) such error could only contribute 1...5 mV to a "good" signal of >100 mV.
Mechanical (relay) multiplexing by reed relays may be preferred but all thermocouple circuits and relays must be screened (located in a metal enclosure plus using screened wiring).

I can understand that its better to pass amplified thermocouple signal through analog multiplexer. In my case I'm using MAX31850 for amplification & digital conversion where it will not allow amplified input.

Relay switching is good choice, but I'm afraid about channel switching speed, transient switching noise & noise injection from relay coil.

If I'm get out from MAX31850, hmm..., amplifier with analog multiplexer will be the best choice...

Any other suggestions?:shock:

udhay_cit said:

Any other suggestions?:shock:

Click to expand...

In his breakdown demonstration of the DS1054Z oscilloscope, Dave Jones demonstrated last month that its analog front end uses the Fujitsu FTR-B3 Relay for analog switching, so unless you require high switching speeds this might also be a good choice for your application.

Hi,

Required temperature accuracy is 2°C.

Click to expand...

you need a cold junction compensation. At a small place there should be the absolute temperature sensor and the connection of the themocouple wires to the PCB.
The problem is the low voltage with thermocouple. Every other voltage error causes error in temperature reading.

The problem with relays is, that inside there are different metals used causing additional thermocouple voltage. And additionally the relays get warm when activated. this amplifies the error.
There are special "low thermocouple effect" relays. Use relays with two poles in one package for the two thermocouple wires. This should further help to minimize thermocouple errors.
(for relays, maybe reed relays: look at COTO)


Klaus

Using OP-AMP buffers as shown in ther original post is a bad idea, because they add large offsets to the thermocouple output voltage.

I don't see how CMOS analog multiplexers would add large temperature errors. MAX31850 has specified input currents < 100 nA, so multiplexers with < 100 ohm on-resistance achieve < 10 µV voltage drops.

The cold junction compensation is inside the MAX31850, the design must assure that the input terminals have the same temperature as the chip.

Take a look at LMP90100 from TI. It will serve the need you have.
Remember to allow sufficient settling time after you switch the channels.

I hope that this helps.

FvM said:

Using OP-AMP buffers as shown in ther original post is a bad idea, because they add large offsets to the thermocouple output voltage.

I don't see how CMOS analog multiplexers would add large temperature errors. MAX31850 has specified input currents < 100 nA, so multiplexers with < 100 ohm on-resistance achieve < 10 µV voltage drops.

The cold junction compensation is inside the MAX31850, the design must assure that the input terminals have the same temperature as the chip.

Click to expand...

Thanks for for your advice. The available space to accommodate my circuit is too small. I can't place 3 reed relays in my circuit area.The cost of the relay & driving circuit also high. Very low input offset OPAMP's are available but too expensive. I don't have any target price in my mind but trying to minimize the cost as much as possible without affect the performance.
So I decided to place three individual MAX31850 chip for each channel. I feel its better to spend money on MAX31850 in place of analog multiplexer or relay or OPAMP.

Commands on MAX31850 designs are welcome

- - - Updated - - -

bimbla said:

Take a look at LMP90100 from TI. It will serve the need you have.
Remember to allow sufficient settling time after you switch the channels.

I hope that this helps.

Click to expand...

This is an excellent chip & trying to get more information about this chip. Do you have any experience with this IC Chip? How to program this chip after assemble it in my board?