Subject: Re: High Temp Pyrometer Elements
From: firstname.lastname@example.org (Phil Duclos)
Date: 26 Mar 93 19:36:54 GMT
K-type (Chromel/Alumel) and J-type (Iron/Constantan sp?)
thermocouples are quite easy to make. The pyrometers on the other hand
tend to be a little more expensive.
To make either a j or K type thermocouple, buy THERMOCOUPLE
wire, NOT thermocouple extension wire. Both types are available with
the extension wire being somewhat cheaper. Buy also some ceramic
insulators, I used the 1/4 round 6" long two hole type. Strip the wires
and insert through the insulators. Twist the two wires together about
1/2" in length after they pass through the insulator. Using an oxy/act
torch and a little borax flux, melt the ends of the two wires together
in as small a blob as is possible. What you want is an "intimate
contact" connection, not a jumbo weld. I also used thermocouple
connectors between my probe and the wire. The key factor is
thermocouples is an "intimate contact" and keeping the materials the
same throughout the sensor's entire length. Do not use copper wire.
A good pyrometer with controller can be had for ~$200. I bought
a cheap panel mount unit, 6 connectors, 30' of K thermocouple wire for
a little over that. I have found indicator units (no controller) for
One outfit I dealt with is called THERMX/Southwest. They are in
the San Diego area and were willing to sell to individuals. Or find a
K-type is the way to go for ->2500F and J type (cheaper) is
good for ->1500F. Don't waste time trying to find the individual
wires. Buy "K type thermocouple wire." A lot more people will know what
you are talking about than if you ask for Chromel or Alumel wire. There
are other types available too: S type, R type, Platinum, optical (IR),
RTDs and so forth. They are ALL more expensive or unsuitable for the
environment in a kiln or furnace. BTW, in anything other than an
electric kiln you will need a thermocouple well to protect the sensor.
Common types are inconel. ALL are expensive, stay away from them.
LIkewise jacketed thermocouples tend to be very expensive. The
typically have a stainless skin, etc and are really unsuitable for a
Hope this helps.
From: John De Armond
Subject: Re: Calibrating a thermocouple.
Date: Wed, 18 Apr 2001 16:53:05 -0400
Paul Stevenson wrote:
> What's the best way to calibrate the temperature display on my kiln?
> I set the controller to 500degC but I'm sure it reaches ~550degC. Which is
> no problem really because I just set 50deg lower than what I want to reach.
> But it would be useful to know what the real value is. Any ideas? Potter's
Cones won't work because they're a time-at-temperature device and
not just temperature.
Here's part of an article I wrote on this topic several years ago.
The most typical problem with inexpensive pyrometers is that they have no
cold junction compensation. A thermocouple generates a voltage
proportional to the temperature DIFFERENCE between the hot and cold
junctions. The hot junction is the one in the kiln and the cold junction
is the point where the TC metals transition to base metals - normally the
terminals of the instrument. If the instrument is not cold-junction
compensated, as the meter terminals warm (perhaps by being located above
the kiln), then the indication will drop. If the ambient reaches 270 deg
F (not unusual right above an operating kiln.), then the indication will
be 200 deg low because most pyrometers are referenced to 70 deg.
Cold junction compensation is done by several methods. One method, used
by that fancy instrument from Ford, is to simply measure the terminal
temperature using a thermistor or similar component, and then let the
microprocessor do the correction. Another method is a cold junction
compensator, a small electronic device that connects between the TC leads
and the indicator that measures the ambient and generates a millivolt
signal necessary to make the indicator read correctly. For a
non-electronic pyrometer, the most common compensation is a bi-metal
spring internal to the meter that mechanically shifts the zero of the
meter to make it indicate correctly. Frankly I've never seen this in the
cheap pyrometers sold to crafters. If your pyrometer is mechanically
compensated it will indicate so in the fine print at the bottom of the
scale, typically with the abbreviation "comp".
Another source of error with non-electronic pyrometers is incorrect loop
resistance. Unlike electronic instruments, the analog meter actually
draws current from the TC and so compensation for that current must be
provided for. Standard practice is to calibrate the meter for a loop
resistance of 100 ohms. A precision fixed resistance is added in series
with the TC to make the total resistance of the circuit at the meter
terminal equal to 100 ohms. Many times this resistor is internal to the
meter case. If the TC wire gauge or length is significantly different
from what the meter was calibrated for, then the indication will be in
The most reliable and precise method of high temperature calibration that
one can do at home or the small shop is to use the freezing/melting point
of metals. Indeed the International Temperature Scale (ITS-90) is
defined in terms of the freezing points of many materials. (check out
http://www.omega.com/techref/intltemp.html for a copy of ITS-90) This
procedure is very simple. Simply melt the pure metal of choice,
preferably in a refractory crucible to help retain heat. Place the TC in
the molten metal and stir vigorously. As soon as a layer of metal starts
to freeze around the sides, the freezing point is reached. The mix will
hold this temperature until all the metal has frozen. You will have
quite some time to work, depending on how well insulated the crucible is.
This is the laboratory calibration method of choice using chemically pure
metals. For our purposes, commercially pure metal is usable. here are
some handy melting/freezing points:
Deg C Deg F
Aluminum 660.45 1220.81
Copper 1084.87 1984.77
Gold 1064.43 1947.97
Lead 207.20 404.96
Mercury -38.84 -37.91 (Freeze with dry ice)
Potassium 63.71 146.68 (Requires inert atmosphere)
Plutonium 640.00 1184.00 (Just in case you have some
laying around :-)
Silver 961.93 1763.47
Tin 231.97 449.55
Zinc 419.58 787.24
Copper used in electrical wiring is >99.999% pure and is suitable for a
calibration metal. Aluminum foil is usually pure aluminum. 24kt gold
and silver are readily available. Potassium (or sodium) is handy for low
temperature calibrations. Either must be heated either in an inert
atmosphere (argon) or under oil. Pure tin solder is available. The zinc
in carbon-zinc ("classic") dry cells is pretty pure. Pure lead is
available for auto body work. Care must be used with lead, as even small
amounts of alloying metals such as zinc or tin radically alter the
If you want to be sure, chemically pure metals can be obtained in small
quantities from chemical supply houses. One I like dealing with is Strem
Chemicals (800.537.2278, http://www.strem.com). Fisher Scientific also
sells pure metals.
If your pyrometer does not have a temperature compensated reference
junction as described above, you only have a couple of choices.
Buying one that is compensated is one option. Good digital units
are now well under $100. Buying a cold junction compensator from
Omega engineering or Hades engineering is another option (about
$100). A third option is to put the pyrometer in an isothermal box
(a box filled with fiberglass insulation will do) and locate it in a
fairly stable thermal environment. The isothermal box will slow the
effects of ambient temperature change. To compensate for ambient
temperature, simply short the pyrometer terminals and adjust the
zero for the ambient temperature as measured with a thermometer.
From: John De Armond
Subject: Re: Calibrating a thermocouple.
Date: Thu, 19 Apr 2001 04:08:52 -0400
> Thanks John..... this clearly explains then the most likely reason I need
> different firing target temps for the same result in my different kilns.
> Other factors being fairly similar, I suspect they all transmit heat to the
> exterior at different rates...therefore the 'apparent' difference in
> internal temp readout, when in fact there is little difference inside when
> the same degree of firing is achieved.
> I understand that other factors will also affect the reading, but this seems
> like it could well be the factor of greater magnitude.
Yes. The lack of cold junction compensation adds a one for one error
to the reading for every degree of temperature deviation of the cold
junction from wherever the meter was zeroed. If the terminals of
the meter rise, say, 100 degrees because the pyrometer is above the
kiln and catches the hot draft, the the reading will be in error by
Another factor is the loop resistance. Analog meters are calibrated
for a certain loop resistance. The resistance of the thermocouple
and lead wire must be measured and the resistance subtracted from
the specified loop resistance. Then a precision resistor of that
value must be inserted in series with the thermocouple. Usually the
resistor is a hand wound wire resistor inside the meter case. This
scheme was arrived at so that a given pyrometer could be used with a
wide variety of TC wire gauge and length. Too little resistance
and the reading will be too high. This shouldn't be an issue if the
pyrometer and TC are purchased as a unit. It might become an issue
if you have to replace the TC or change the length of the lead wire.
A third issue is TC aging. It has been discovered that when a TC is
operated at high temperature, it slowly but continuously ages,
probably from oxidization and maybe metal diffusion. The effect is
a gradual decrease in emf for given temperature. NIST has
characterized this drift and has developed the equations for same.
There is a circular on this, though I can't seem to find the
number. We're pushing the limits of type K thermocouples for glass
furnace work so one must simply plan on replacing the TC every so
often - or to calibrate it using the freezing metal procedure.
> It is also yet another reason standardized 'firing schedules' won't export
> verbatim from one persons setup to another .
> Regards, Jacques Bordeleau