Connect with us

making thermocouple temp probes?

Discussion in 'Electronic Basics' started by JazzMan, Apr 24, 2005.

Scroll to continue with content
  1. JazzMan

    JazzMan Guest

    I've got some 1/8" OD 316 stainless tubing on order from
    McMaster-Carr, already have some thermocouple wire rated
    for 2K°F, and will be buying some fittings for mounting
    the probes in the exhaust manifold of my engine. My thought
    was to TIG one end shut on a 2" length of the tube and use
    a stainless compression fitting to lock the probe in place.
    I have several questions. One is, do I need to pot the
    thermmocouple wire inside the tube? If so, with what? Would
    it matter if the potting agent was conductive? Do I need to
    somehow weld the thermocouple junction to the end of the
    tube being used as the probe body? Or would simple contact
    be suffient?

    For reference, the engine's been modified and I've been
    editing the fuel maps in the ECM. Knowing the EGT at
    individual cylinders allows me to address flow balance
    issues in the intake manifold, for instance.

    Please reply to jsavage"at"
    Curse those darned bulk e-mailers!
    "Rats and roaches live by competition under the laws of
    supply and demand. It is the privilege of human beings to
    live under the laws of justice and mercy." - Wendell Berry
  2. Chris

    Chris Guest

    Like your motto, Jazz.

    Exposed thermocouple beads can be grounded at the point they're
    measuring (sometimes there's no other way), as long as your instrument
    measuring the thermocouple is floating (like a battery-powered
    thermocouple meter). As long as you've got a floating meter, a
    grounded bead and a SS compression fitting is a great idea. You might
    want to put a sliver of mica between the compression fitting and the
    T/C bead to ensure you're not shorting out the T/C by grounding it at
    more than one point. Other than that, you're in good shape.

    Sounds like you've got a good plan.

    Good luck
  3. JazzMan

    JazzMan Guest

    Thanks! I stole it from someone else a while back. I have no
    idea who the wendel guy is, but the saying just seemed to embody
    an ideal I thought was worth knowing.
    I appreciate the advice. The t-couple wire pair is insulated in
    a single woven material, and each individual wire is also
    insulated. I'll have to remove the main insulation to get
    the two wires to fit inside the tube. The potting agent would
    be used to seal the two wires going into the tube to keep
    water and dirt out. The patting agent I looked at in the catalog
    has aluminum in it, so it would presumably be conductive. I
    don't know if the agent would soak through the woven insulation
    around the individual wires.

    Please reply to jsavage"at"
    Curse those darned bulk e-mailers!
    "Rats and roaches live by competition under the laws of
    supply and demand. It is the privilege of human beings to
    live under the laws of justice and mercy." - Wendell Berry
  4. message ....

    I would not worry about the potting agent, unless
    it was sold as a moldable conductor. The aluminum
    is probably an oxide, or particles that are oxidized
    on the surface, and not particularly conductive. The
    thermocouple acts as a low voltage, low impedance
    source, so a little stray current would not change
    the sensed voltage enough to matter anyway.
  5. Dick

    Dick Guest

    Don't know if this info is of any help but I made up a lower
    temperature probe, based on a Pt100 sensor & thin walled 3/16" brass
    tubing, by brazing in a very short length of brass rod to block off one
    end of the tube.
    Slid the sensor into the tube as far as it would go and back
    filled the rest of the tube with a very fine, dry sand, then sealed open
    end of tube with high temperature epoxy adhesive.
    The sensor leads & joints are protected by some lengths of
    narrow high temp sleeving and the sand was simply garden sand, washed
    thoroughly, dried in an oven, sieved through a fine grade kitchen sieve
    & finally ground very fine in the pestle & mortar usually used for herbs
    & spices (best to do this when herself is out of the house !). To get
    the sand past the sensor leads & into the tube fix a small paper cone
    around the open end of the tube, fill it with the fine sand and then
    gently tap the tube .........
    I've been using this arrangement for some time as the sensor for
    a PID temperature controller & it works very well, strong mechanically &
    with a fast thermal response time.
  6. Bill

    Bill Guest

    No, you do not need to pot the wire to the tube but it would be a good
    idea to keep out water, grease, etc.

    If the potting agent is conductive and if it causes a low impedence
    between the 2 wires then you could get a secondary junction there. I
    doubt that will be a problem with what you're doing.

    You do not need to weld the TC junction to the tube cap although that
    would probably give you a faster and more accurate measurement assuming
    electrical noise is not an issue. A grounded junction would have less
    susceptability to thermal gradient measurement errors in this

    Make sure you are getting sufficient immersion depth into the exhaust
    stream otherwise you could get an error in your measurement due to
    thermal shunting. You might try locating your sensor into the stream
    and then move it slightly to get more or less immersion. If slight
    changes in immersion depth change the indicated temperature by too much
    it could mean you have insufficient immersion into the exhaust stream.

    If you are getting an indication that you have insufficient immersion
    you could try exposed junctions where you do not cap weld the tube but
    rather let the wire junction be exposed to the exhaust stream and just
    use the tube for mechanical support. You would need to seal the tube
    at the cold end to prevent exhaust leaks or use a ceramic potting at
    the hot end.

    Bill Schuh
  7. Jack Hayes

    Jack Hayes Guest

    The exact TC you need are available commercially, normally they are sold
    with a plug connector.

  8. JazzMan

    JazzMan Guest

    Got a link? I've looked through the McMaster-Carr catalog
    and typically their sensors are far too long (clearance issues
    in the engine bay around the manifolds) and too large in diameter.
    I want to run as small a diameter as possible so as not to
    affect exhaust flow. My perfect sensor would be 1" long and
    less than 1/8" diameter, with a 1/16" NPT base so that I
    could drill and tap the edges of the manifold flanges at
    the heads and install the sensor directly into the exhaust
    stream. I've never seen anything like that except in photos
    of complex test equipment setups.

    Please reply to jsavage"at"
    Curse those darned bulk e-mailers!
    "Rats and roaches live by competition under the laws of
    supply and demand. It is the privilege of human beings to
    live under the laws of justice and mercy." - Wendell Berry
  9. Bill

    Bill Guest

    I hate to send you to a competitor but for a one off order try ($25).

    I'd suggest you buy a 0.063 or 0.125" diameter mineral insulated probe
    whatever length you need with some lead wire. These probes can be bent
    to fit into the engine spaces as needed.


    Buy the NPT fitting you want in a plug form and drill it out to accept
    the probe then braze the fitting onto your probe sheath at the desired
    immersion depth. I'd use a plain steel fitting instead of a SS fitting
    since if you don't design for the thermal expansion of your metals you
    could run into some problems. Use a lot of antiseize on the threads.

    Bill Schuh
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day