K-type thermocouple with MAX6675

N

nickwest

Member
I am trying to build a thermometer for the exhaust on my car. It seems that a k-type thermocouple is the only device that would withstand the 1000+degC temperatures involved. I plan to use a MAX6675 to interpret the thermocouple and provide a digital output for a PICaxe. I've found a few threads suggesting this approach for various other applications, but not many people seem to have actually TRIED interfacing a MAX6675 and a PICaxe. Can anybody shed some light?

Thanks.
 
D

Dippy

Moderator
EH????? Well, you haven't tried very hard looking have you.

Check out the very recent thread about "FLUCTUATING temp readings" - most recent post was a few hours ago.

The first line of the first post says:-
"I'm using an 18X to read the temp of a K type thermoucouple via a MAX6675."

True, it is "not many people". But it's not something a lot of people want to do.

I'm sure greencardigan will go green with envy if you can get yours working well in an automotive application.
 
Last edited:
slimplynth

slimplynth

Senior Member
Type N or K?

Wouldn't choose to use a Type K because of the high temp and repeated temperature cycling from hot to cold - especially in air atmospheres.

1000C is within the operating range of type K, according to the maufacturers' guides

http://www.tcdirect.co.uk/deptprod.asp?deptid=190

Maybe you don't require the degree of accuaracy required for aerospace heat treatments but Type N thermocouples - maybe a leaf type. would be better for your operating range. My company stopped the use of Type K thermocouples a couple of years ago. When they were in place, use was restricted to applications lower than 350C.

Type N's (Nicrosil) withstand oxidation better than Ks, another TC killer. This might be a consideration if your thermocouple is difficult to replace.

It's something I've been pondering on my back burners because there are look up tables for each type of thermocouple to convert the mV output to a degree C value.

http://www.dataforth.com/catalog/pdf/an107.pdf

I've got some Type Ns at home, just not had chance to play with them and the picaxe yet. The mV/degC relation ship is nonlinear but looking at the PDF I'm guessing that applying a little math should compensate for this, depending on accuarcy requirements
 
Last edited:
N

nickwest

Member
Thanks for the info everyone, now that I am searching whole threads instead of just the titles I have found more results! Sorry about that.

slimplynth: I won't be using a bare bead-type thermocouple... you're right, it would die pretty quickly in its intended environment. I believe high-temperature versions are available, in a sealed probe with suitable insulation etc... As far as accuracy is concerned, 2% would be nice at a temp of ~900degreesC. The other reason for choosing a k-type thermocouple is it seems information, circuits and code to interface with these is more abundant. Alas my skills are not at the stage where I can start from scratch.
 
slimplynth

slimplynth

Senior Member
Yes you are right, bare bead is not a good idea in an oxidising environment, the ones we use are mineral insulated. they are bendy and very easy to mount - can be spot welded/brazed with some shim covering the thermocouple. I know what you mean though, I'm a complete newbie to picaxe myself.
 
B

barbababa

Member
Hello all!

I am trying to build a circuit that monitors the temperature of a filament in a high vacuum chamber with K thermocouple and simultaneously drive a high current source to that filament to heat it
according to the required temperature provided by the user. I thought just to continue this thread rather than making a new one. It would be greatly appreciated for any help or comment!
Here is my draft version of the project:

circuit.jpg

The PICAXE/MAX6675 interface was just copied from this forum. The currentsource is some old commercial high current power supply that we have at our institute so i don't have to meddle around with the mains.
The basic idea is to control the current into the filament with a pulse width control to drive the high power mosfet. I thought some optoisolation should be proper here, which is not yet implemented and the controlling system will just be some keypad/LCD system to control the required temperature and heating time (not yet implemented).
My question here is that, is this a proper way to do this? The actual temperature resolution is not that critical, 1 to 5 degree error is still ok. I would be satisfied with a maximum temperature of around 200 degrees,
but the better the higher it gets. Is there some issues with K thermocouples and high vacuum?

Thanks in advance,
Toni
 
G

Goeytex

Senior Member
Toni,

A colleague of mine and I developed a steam injector system for an automotive engine using a Picaxe to control everything.
Darren did the thermocouple interface using a MAX 6675 to measure exhaust gas temperature. For the prototype system,
the thermocouple was one of the very cheap K types found on the E auction site. It worked surprisingly well and has not
burned out in many hours of testing .

I do not see where vacuum will affect the thermocouple at all. Your approach seems fine.
 
B

barbababa

Member
Thanks for the reply!

I will start gathering the parts now and will update my project here if someone else is interested....
 
R

Reloadron

Senior Member
When choosing the exact Type K T/C configuration remember you will have the sensing in a vacuum chamber. That said, you may want to be looking at maybe a inconel or stainless steel sheath or protection tube that comes with a compression fitting allowing you use in a vacuum. While the newer Type N is taking over as everyone's darling a Type K should do fine for your application. The temeratures you mention are not that extreme nor is the environment.

Just My Take
Ron
 
You must log in or register to reply here.
Top