Help with Thermocouple K-Type Circuit

[thetrueman]

I've searched for subjected circuits but could not find proper professional circuit. Please give some working circuit hint to use Thermocouple K-Type to measure temperature. My limitation is to use LM358, LM324 or OP07.

Also is there necessary to make arrangement of Cold Junction Compensation? Please advise... Thanks.

[solutions]

Look for app notes, not circuits - I think there are some from Linear Tech and National Semi that talk about the design in great detail.

cds.linear.com/docs/Datasheet/LTK001fa.pdf

Drop the limitation if it does not do the job.  For saving $1.25 you will wind up spending weeks trying to make something else work.

[darck_khronos]

ahhh.... instrumentation Opam

[oldvan]

Here's what I've used successfully:

Maxim MAX665:  Cold-Junction-Compensated K-Thermocouple to-Digital Converter (0°C to +1024°C)

The MAX6675 performs cold-junction compensation
and digitizes the signal from a type-K thermocouple.
The data is output in a 12-bit resolution, SPI™-compatible,
read-only format.

This converter resolves temperatures to 0.25°C, allows
readings as high as +1024°C, and exhibits thermocouple
accuracy of 8LSBs for temperatures ranging from
0°C to +700°C.

[TomJackson69]

Go to example from Proteus, there is one in the “D:\Program Files\Labcenter Electronics\Proteus 7 Professional\SAMPLES\VSM for PICMICRO\VSM for PIC18\MAX6675 Thermometer” folder. This is working example. I have done this before.

Tom

[solutions]

Here's what I've used successfully:

Maxim MAX665:  Cold-Junction-Compensated K-Thermocouple to-Digital Converter (0°C to +1024°C)
....exhibits thermocouple accuracy of 8LSBs for temperatures ranging from 0°C to +700°C.

Nice for wide range, but an accuracy of about 3°C is kinda sucky. I guess that means you have to have a cal table, since they have 0.25°C resolution.

[thetrueman]

Frisky there are many professional temperature meters which are designed with LM358 or OP07. So why we can't think in the box when everybody is using the same ICs?

[LabVIEWguru]

go to analog devices - my favorite is the AD620. I used an LTC1100 for excellent results.

Linear Tech also has about a million examples.

Both companies will give a couple samples out to students working on a project.

If you are looking for a particular temperature range you can have a (fairly) small look up table

You'll need at least 16 bits of resolution. If you want accuracy over the entire range you can do a least-squares fit
I got about a million hits on that.

You can do an easy (REAL FAST) linearization of data by summing samples/dividing by # of samples, take the square root of that number (floating point) and multiply by full scale range. 16 to 24 bit sigma/delta from Analog Devices worked great.

Maxim-IC makes ASICs that will do almost everything for you except the programming, and they give you examples for that.
Analog Devices has mixed-signal microcontrollers that do all kinds of cool things.

At 16 to 24 bits, noise is a problem. Also Pyro-(I forget the term) (pyrophoria??) induces noise in the circuit. Pay particular attention to grounding. Analog Devices has a couple real good app notes on grounding in a mixed-signal system.

So, figure out what temp band you need out of the temp range of the thermocouple, figure out what resolution you need, how fast you want it to update (not very fast needed) what processor you are good using (or maybe just read from a voltmeter?) and where you will get the parts. Even if you don't use their parts, you can use the manufacturer's app notes to see how the circuit works and them use the part you have on hand. Everyone does that.

I just gave you six months of my life from about 1997. As a couple of my colleagues here have pointed out, you need to do the homework and work out the circuit. Then ask for specific help on specific problems.

[solutions]

...and if you truly have better things to do with your time, just buy one of these (has RS232/USB interface) and work on the other stuff: http://www.sears.com/shc/s/p_10153_12605_SPM3011557602P?prdNo=15&blockNo=15&blockType=G15

[thetrueman]

Actually the application is not so sensitive and +/- 1 Degree can be accepted. So I want to keep circuit very cheap.

Yesterday I checked a professional temperature controller and now I'm guessing what is style of amplifier. Anyhow thanks for your replies and clues. Hopefully soon I'll get command on thermocouples.

[optikon]

Actually the application is not so sensitive and +/- 1 Degree can be accepted. So I want to keep circuit very cheap.

Yesterday I checked a professional temperature controller and now I'm guessing what is style of amplifier. Anyhow thanks for your replies and clues. Hopefully soon I'll get command on thermocouples.

+/- 1 degree is not trivial. Your temperature error budget includes the accuracy of your temp measurement of cold junction and also your math for converting the small voltage back into a temperature.. the voltage measurement that error too.  I agree with Frisky Ferret, you can forget about doing this with a cheap opamp.

[thetrueman]

The manufacturing tolerance of an ordinary Class 2, Type K thermocouple is +/- 2.5 °C from −40 °C to 333 °C.

Thanks Emotional Frisky for adding value to my knowledge. In fact I take things as challenge. With costly precisely designed ICs anybody even a student can make a circuit but it is hard to make things out of the way. In my life I did many tasks out of the way getting same results.

Yesterday I got a temp. controller which has LM358 and 9013 for cold junction compensation. The problem was cold junction (CJ) not amplification. Now I've an idea for ever how to implement CJ. Now even I can copy the circuit or design my own, no problem.

Thanks a lot for everybody who gave worthy comments to reach some useful point.

-----------------------------------------------------------------------------------------------------------
Hardwork is like Stairs and Luck is like Lift... Sometimes Lift may fail BUT Stairs will always take you at the TOP!

[solutions]

http://www.eetimes.com/design/industrial-control/4217170/Modern-thermocouples-and-a-high-resolution-delta-sigma-ADC-enable-high-precision-temperature-measurement?Ecosystem=industrial-control

[metal]

I was reading this topic, how can I find out the shortest length of thermocouple to use, for example if I want to design my own soldering station, and I only want to use thermocouple wire inside the soldering iron? Does it depend on the gauge of the wire?

[kukumar]

Thermocouple wire must go :from place, where temperature is measure, to IC (close as possible)which have compensation circuits.
It is not allowed to use other wire for....(use 1cm of thermocouple wire , and other Cu wire....No!!)
Price of thermocouple wire is not so high!

[xor]

I found few good's from an old project but isn't complete.
Missing few files due to HDD failure on my old PC...I try to find the missing files next days

[TomJackson69]

I was reading this topic, how can I find out the shortest length of thermocouple to use, for example if I want to design my own soldering station, and I only want to use thermocouple wire inside the soldering iron? Does it depend on the gauge of the wire?

Found this on the internet. It may help:

"
What is the maximum length of thermocouple wire?
There are many factors that can impact the usable length of a thermocouple. As a guideline, under 100 feet with 20 AWG or thicker wire in an area free of electromagnetic interference usually is fine. Two of the main factors in determining useable thermocouple length are total loop resistance and preventing electrical noise getting into the signal. Because different thermocouple wires are made of different materials, the resistance will vary based on the type as well as the wire diameter and length. The allowable loop resistance is affected by the input resistance of the amplifier circuit to which it is attached. But as a guideline, typically the objective is to keep the total loop resistance under 100 ohms. Loop resistance is determined by multiplying the length in feet by the resistance per double feet (remember 1 foot length of run includes 1 foot from each of the two t/c wires) as shown in "Resistance Vs Wire Diameter" table for thermocouple type and gauge. Remember in your calculations to include the probe (when used/applicable) in addition to the wire length. The second major factor in running a thermocouple wire is to keep it away from any electromagnetic fields. Thermocouple wire creates a low voltage signal and should not be run near power wires, motors, etc. To help minimize noise pickup, a metal over braid or twisted shielded wire is commonly used.
"

You should have the specification of the TC you are using. Check for "Resistance Vs Wire Diameter" spec.

Regards,

Tom

[titi]

Hi,

Microchip has an application documentation about interface thermocouple and Pic Micro controlers.
Here some documentations to interface thermocouple.

SiliconChip published on december 1998 an article "Thermocouple adaptor for DMMs" using OP07, LM335 and ZR423.

May be this helps.

Best regards.