What type of thermocouple is used in JBC irons ? (Maybe type N)

[tboy32]

Just did my own experiment.
Tip: C245-903
Ice water: -0.600mV
Boiling water: +0.970mV

Notes: The boiling water experiment was tough. I think bubbles hitting the tip caused fluctuations. I tried keeping the tip+heating section under water but not touching the bottom of the pan. Steam condensed on the top and the alligator clips holding onto the contacts. This was tap water so I wouldn't be surprised if impurities caused increased conduction between terminals and thus a smaller output voltage than it should have been, but I have no idea how to estimate how much of a difference it was. I tried shutting off the heat source (eliminating bubbles) and immediately taking a measurement but it looks like (according to https://us.flukecal.com/Thermocouple-Temperature-Calculator) +0.970mV on an N type thermocouple with a 25C reference temp (guess) is only 60C. I did see measurements above 1mV (like +1.025mV IIRC) but was unable to reproduce it. Also: shorting clips together lead to a measurement of about -0.003mV.

[RoGeorge]

Just did my own experiment.
Tip: C245-903
Ice water: -0.600mV
Boiling water: +0.970mV

That would be about 15.7uV/°C, which is not a sensitivity for standard types of thermocouple in this table: http://www.analog.com/en/analog-dialogue/articles/measuring-temp-using-thermocouples.html

Maybe the soldering tip does not have a standard type of thermocouple.
Since almost any 2 different metals can be used as a thermocouple, they might probably used as a thermocouple whatever 2 different metals were there anyway for the mechanical build of the tip.

[whatsthegeek]

Just did my own experiment.
Tip: C245-903
Ice water: -0.600mV
Boiling water: +0.970mV

That would be about 15.7uV/°C, which is not a sensitivity for standard types of thermocouple in this table: http://www.analog.com/en/analog-dialogue/articles/measuring-temp-using-thermocouples.html

Maybe the soldering tip does not have a standard type of thermocouple.
Since almost any 2 different metals can be used as a thermocouple, they might probably used as a thermocouple whatever 2 different metals were there anyway for the mechanical build of the tip.

That would be pretty annoying :/

Just did my own experiment.
Tip: C245-903
Ice water: -0.600mV
Boiling water: +0.970mV

Notes: The boiling water experiment was tough. I think bubbles hitting the tip caused fluctuations. I tried keeping the tip+heating section under water but not touching the bottom of the pan. Steam condensed on the top and the alligator clips holding onto the contacts. This was tap water so I wouldn't be surprised if impurities caused increased conduction between terminals and thus a smaller output voltage than it should have been, but I have no idea how to estimate how much of a difference it was. I tried shutting off the heat source (eliminating bubbles) and immediately taking a measurement but it looks like (according to https://us.flukecal.com/Thermocouple-Temperature-Calculator) +0.970mV on an N type thermocouple with a 25C reference temp (guess) is only 60C. I did see measurements above 1mV (like +1.025mV IIRC) but was unable to reproduce it. Also: shorting clips together lead to a measurement of about -0.003mV.

These are pretty weird results, I would have thought the voltage drop would have been higher at 100°C :/

[Bgdn]

Hello,

i did today  some measurements in a electric oven.

Placed a type K termocouple connected to a multimeter and osciloscope channel 1

Placed a jbc tip C245-931 connected to osciloscope channel 2.

I tried to place the thermocouple as close as posible to the proximity of the jbc tip.

I used the averaging function of the scope, because the raw signal is quite noisy from both thermocouples.

Bellow are the measured values and the temperatures wish small aproximation maybe.

type K        JBC
5.24 mV - 0.61 mV - 132 degrees
5     mV - 0.54 mV -  128 degrees
4.84 mV - 0.5 mV -   125 degrees
4.6  mV  - 0.42 mV - 120 degrees
4 mV     -  0.36 mV - 105 degrees
3 mV     -  0.26 -       80 degree
2.04      -  0.2  -        50 degree

What do you think?

[wraper]

What do you think?

I think you have no idea how thermocouples work and placed a whole cartridge connected by usual copper wires into the oven. The problem is that thermocouple measures temperature difference across its length, not absolute temperature.

[Bgdn]

The type K thermocouple wires which comes with my multimeter its not from copper, i think its something "silverish".

I dont have a reference junction .

Maybe just flood both into temperature controled oil? without exposing the wires to temperatures?

[wraper]

The type K thermocouple wires which comes with my multimeter its not from copper, i think its something "silverish".

That's why you had a correct reading with thermocouple. In most cases their wires are just a continuation of thermocouple.

I dont have a reference junction .

Well, you have them, just don't understand it. Reference junction is where thermocouple is connected to normal wires. It's just that temperature where thermocouple ends must be known to calculate a correct result. On the picture it's in water at 0^oC. Multimeters and thermometers usually have a resistive temperature sensor in them.

Maybe just flood both into temperature controled oil? without exposing the wires to temperatures?

You can do that. Just avoid cartridge heating up as a whole. Terminals must remain cold or special thermocouple extension wire must be used.

[Bgdn]

I googled a bit about cold junction compensation...

Now i am wondering, since i connect the osciloscope probe directly to the C245 cartrige.

 Based on the picture above,  the point of conection between the probe and the cartrige shall be at 0 degrees, right?

But if its at room temperature? Shall i consider my reading to be the heat temperature + rom temperature?

Its a bit puzzling, i see alot of discussions on the internet about these jbc tips.

[RoGeorge]

A termocouple is made out of two different metals, in this case two wires of different metals.
It does not measure the absolute temperature, like a thermistor or a mercury thermometer.

The two different wires are tied together at one end.  At the other end you connect a voltmeter.  The points where you connect the voltmeter is often called the "cold" junction, but in fact the cold junction can be at any temperature.  Can be even at a higher temperature than the tip of the thermocouple.

The voltmeter will indicate a voltage proportional with the temperature difference between the so called "cold" junction and the tip where the two wires are twisted together.

For a multimeter to measure the temperature with a termocouple, it needs to know the temperature of the "cold" junction (usually the cold junction is at room temperature).  The very cheap DMMs will asume the room temperature is always 25*C or so, but most of the DMMs will have a thermistor inside, so the DMM can measure the room temperature.  You can check this on a DMM with temperature by unplugging the thermocouple.  When you unplug it, the DMM should indicate the cold junction temperature (in this case - probe unplugged - will indicate the room temperature).

When you put the tip of the thermocouple inside the oven and the other end at the room temperature, the voltage on the oscilloscope will tell (T_oven - T_room).

The cold junction is considered the place where the two type of metals (the two wires of the thermocouple) are connected to a 3rd type of wires of the same type (here the oscilloscope probe is the 3rd type of metal).

For example, if you have the oscilloscope probe connected to two copper wires (to extend the oscilloscope probe's lenght), then those two copper wires are connected to the thermocouple wires inside the oven, the "cold" junction will be also inside the oven, so you will measure very close to zero volts.

Even more, if you let the cold junction inside the oven and put the tip of the thermocouple outside, you'll measure a negative voltage. 

For your K thermocouple, the cold junction is where you connected the oscilloscope (or a voltmeter).  For the C245 cartridge, the cold junction is somewhere inside the cartridge, on the opposite side of the soldering tip.  To measure correctly, you will need to keep inside the oven only the soldering tip, and the other end (where the cold junction is) must be kept outside of the oven.  If you do so, the voltage will be proportional with the difference in temperature between the two ends of the thermocouple (T_oven - T_room).

To recap, if you have 2 wires of metals type A and B, the cold junction is where you connect them to a 3rd type of metal, C.  Any two metals put in contact will form a thermocouple, weather you want it or not, just that usually the voltage is very small and hard to measure.  Thermocouples like K type or J type have the wires made out of metals that gives higher voltage with temperature, so to be easier to measure.

Even more, the effect is is reversible! 
If you put a difference of temperature you'll read a voltage.  If you put a voltage you'll force a temperature difference in the thermocouple.  That's how the Peltier elements are working.

[Miti]

You can check this on a DMM with temperature by unplugging the thermocouple.  When you unplug it, the DMM should indicate the cold junction temperature (in this case - probe unplugged - will indicate the room temperature).

Nope, you forgot to short the TC inputs.

[RoGeorge]

Have you checked?  Mine still sees 0V (and thus will display the room temperature) with the thermocouple unplugged, so no need to short the inputs. 

[Bgdn]

Thanks George for the usefull explanation.
So i wil try to do the following experiment, hopefully i will get usefull results.

On a copper plate 1-0.5mm thick, i will try to solder on  the tip of the C245 together with a k thermocouple, and try to externally heat only the copper plate and see what results i will get.

What do you think?

[RoGeorge]

Don't solder the tip on a copper plate.  Any force between the copper plate and the soldering tip might peel the tip and ruin your C245.  If you want to keep the soldering tip and the thermocouple tip at the same temperature, cover them together in a small blob of Aluminium wrapping foil, or wrap them together with some copper wires and heat the wrapping with something else.

Keep the other end at room temperature (for both the thermocouple and the C245), where you connect the oscilloscope.  Make a table with all the measurements and post the results.  Then, you can check the measurements against the expected voltage for K type and for C245.

[Miti]

Have you checked?  Mine still sees 0V (and thus will display the room temperature) with the thermocouple unplugged, so no need to short the inputs. 

Check this out.

[Bgdn]

Hello,

i gaved up on the copper plate experiment.

But fixed the termocouple tip very tight to the c245 tip, and submerged both of them in aprox 5mm of melted solder ( in a small metalic copper box heated with hot air station from underneath).

The reading was stable enough to do some readings on the scope connected to the c245 cartrige.

What do you think?

[RoGeorge]

Very well! 
I think you should be precise with the questions, ask what you want to find out.

Now, you can put those numbers in a spreadsheet and plot the voltage versus temperature.  It is expected to show a tilted line, ideally a straight tilted line.  Inherent measuring errors should still show the tilting trend, though the line will look crested instead of being straight.

Then, you can compare the results with others, for example the results shown in this other page:
https://www.eevblog.com/forum/beginners/jbc-t245c245-thermocouple-thermal-coefficient/

[RoGeorge]

you forgot to short the TC inputs.

Check this out.

Good advice to short the inputs before reading the room temperature with a DMM, thank you.

For the cheap clone of Mastech MY64 the reading shows the same temperature, shorted or open.
With an expensive Metrix MTX3283, the room temperature read is incorrect with open inputs.

First pic is with open inputs instead of a thermocouple, second pic is with a short circuit (a piece of a paperclip for MY64, and the red wire in the back for mtx3283) instead of the thermocouple.  The room temperature was about 26*C.

[Bgdn]

Very well! 
I think you should be precise with the questions, ask what you want to find out.

Now, you can put those numbers in a spreadsheet and plot the voltage versus temperature.  It is expected to show a tilted line, ideally a straight tilted line.  Inherent measuring errors should still show the tilting trend, though the line will look crested instead of being straight.

Then, you can compare the results with others, for example the results shown in this other page:
https://www.eevblog.com/forum/beginners/jbc-t245c245-thermocouple-thermal-coefficient/

i wonder why i missed that thread, there is a very usefull characterization of the c245 tip.

After seeing this, it makes sense to consider this to be a very likely type k inside the c245 tip.

But at least i have now all the info needed, i will proceed to design the analog front end for this jbc tip.

Thanks everyone!

[RoGeorge]

If you just want a controller for that C245, it would be simpler and cheaper to order a clone soldering station, or to use an existing design.

If you want to design your own controller, that would be a nice project and a lot to learn from it, go ahead.  The main characteristic of a thermocouple is how many uV/*C it produces.  From your last measurements, it doesn't look to be a type K, it looks more like type C.

Image source:  https://www.thermocoupleinfo.com/thermocouple-types.htm

[Bgdn]

Hello,

today i connected a thermocouple to the circuit shown, and trying to read the values over the 24bit adc, running at 8192 samples.

In paralel with the tip i have connected a keysight u1272a which can go down under 10uv measure scale.


I have around 19-20 degree room temperature, and i measure across this C245 tip:
-  24uV from ADC
-  27uV measured by the multimeter.

Tomorrow i will do another hot solder pot test, to see if the values match the last time i did this...

Not sure if i need larger biasing resistors 

Latter eddit... Since ad8221 needs under 1nA input bias current,  i tryied biasing  the tip with 1.5Mega resistors, the same result... surprisingly after letting the tip to sit for a while in the setup at room temperature... the voltage drops quite low.. arround 3uV on both instruments.

I have also tried to remove the 1M5 resistor from GND, the value this time stays arround 50uV. 

Not sure if this behaves liniar at such low temperatures, because if touch the tip with my finger, immediately jumps to 100uv and stays there some time, then it drops back to ~60uV quite slow

[Bgdn]

Hello,
sorry for the short delay
i did the hot solder pot test, where both thermocouple & tip share the same melted solder "lake", heated externally.

These are the values read on the c245 tip

Also i tried to apply power to the tip, and read the voltage across internal thermocouple ( 150hz pwm wave provided with 10-20% dutycycle), and the signal across the thermocouple is read during the "pwm off" period.
I tryied not to heat to much the tip, because both type k tk & heater tip share a solidified solder ball 

I left some time to reach a constant temperature, then recorded the voltage acros jbc tk.

These are the values read aswell.

Somehow the thermal inertia seems to cause diferences between hot pot test & direct voltage aplied to the tip. But 17uV/degree is not so far away from 21uv/degree


What do you think?

[wraper]

What do you think?

That your circuit is not suited for thermocouples. And has several issues beyond that. Not to say why do you use in-amp to begin with? PWM to tip temperature is even more meaningless exercise.

[Bgdn]

What do you think?

That your circuit is not suited for thermocouples. And has several issues beyond that. Not to say why do you use in-amp to begin with?

why not? The GND from the measurement circuitry is isolated from the heater power circuit

[wraper]

Because you pass quite a bit of current through the thermocouple (R1, R6) which you certainly shouldn't do as it creates significant voltage drop across it. If thermocouple has say 10 Ohm resistance typical to thin longer ones, it's already  1.6mV of voltage on it. JBC thermocouple should have lower resistance but it will still have a significant impact. Also this circuit has horrible CMRR, it would be better with no C1 and C2 at all than like this, proper input filter circuit attached. There is no need for in-amp and you only overcomplicate the thing this way. No cold junction compensation, but let's assume it's okaish if you know ambient temperature.

[wraper]

R2, C12 also is not how you should do it. I won't look into in-amp specs but it certainly can cause problems with some in-amps or op-amps. Not to say it's a really piss-poor way how to do a low pass filter.