How can i connect a Thermocouple to ADC

hello Everyone!

i need a little help, i want to connect a thermocouple to A\D converter
so that i could show a temperature in a 7 segment.
i wrote the code for that and i'm using FPGA component.
if someone can show me a sketch of the connection it will be very helpful.
thank's

In theory it is simple. In practice it is a nightmare. The DC voltages are very small, microvolts, so you have to use some precision amplifiers to feed the ADC. Omega is a company that deals in these kind of circuitry, you could probably find everything off the shelf, but they also offer technical information, such as voltage to temperature coversion tables.

Your best bet is to use an chip designed for the purpose.

Analog Devices have thermocouple amplifiers AD594/AD595 with cold junction compensation that interface the thermocouple to an ADC.

Maxim have MAX6675 a thermocouple amplifier with on-board ADC and SPI output that will easily interface to a microcontroller.

Other chip manufactures have similar products.

I've had good results with the Analog Devices chip Alex mentions. I've not run the o/p to an ADC for my applications (just used a voltmeter & read temp directly). But it should be a simple matter to divide the signal from the amp and feed it to the ADC.

Thinking about it more... and specifically about the AD595...

If the temperature range of interest is no higher than 300 to 700 C, one could run the amp o/p (no more than 3 to 7 volts) directly to an appropriate ADC.

If one wanted 1200 C, one would need the voltage divider. Accuracy would be lost, unless one could find very precise resistors.

If one one wanted the full range of -300 C to +1200 C, one would need to throw in an offset (possibly a summing amp) prior to the divider. I suspect this would result in even more loss of accuracy.

Anyone know a good source for 0.1% resistors?

Digi carries a line of Susumu chip resistors in .1% tolerance. $.58 per 805 chip. That covers the standard 1% values.

There is no escape.

Those Susumo resistors would allow the AD595 + ADC to be accurate within 4 or 5 C over full scale.

Looks like the MAX6675 is a better choice if one does not need the wider range of measurement. It has better accuracy and costs less. I might just have to find an excuse to get a couple... Of course I would have to learn SPI protocol.