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

[whatsthegeek]

Hi, I'm trying to design my own controller for a JCB series 245 iron (I know there's a project on dangerous prototypes that does it but I's quit complex and I won't learn anything by just using it ^^). I want to use an all-in-one chip but I need to know what type of thermocouple is used :/

Some people says it's a type K and others say it's not, so it's quit confusing...

Maybe if one of you guys have one and have a bit of free time, you could heat up the tip to 200° and measure the voltage drop between the metal body and the last contact,
It could tell me exactly what type of thermocouple is used.

Thanks in advance ! 

[wraper]

If you have a cartridge, you can easily measure it yourself. Type K is a most common thermocouple type and it is well suitable for such task. Say put it into boiling water compare measured voltage with a chart. Or if you have  multimeter with thermocouple function, you could connect it and see if measured value is matching.

[whatsthegeek]

I don't have the cartridge yet :/ I want to be sure I can easily make a circuit to control it before buying it, It's super expensive...

That's why I'm asking here, otherwise, I would be already done 

[AndyC_772]

OK, I've tried a very simple and not very scientific experiment:

- fit a cartridge into the handpiece and set to 200C
- apply a blob of solder, which does just about melt, to aid thermal conductivity
- place the tip of a second cartridge (C245-907, for what it's worth) against the tip of the first cartridge
- connect DVM probes between the body of the second cartridge and the contact right at the end
- wait a couple of minutes and watch the voltage increase

Final result was about 4.14mV. It takes much longer to reach anything like a steady state than I would have expected.

[whatsthegeek]

Ok, thank you sooo much ! Comparing it to the chart, the closest thermocouple type is a type N

On a type N, 4.14mV corresponds to 144°C. Since the conctack between the cartridges wasn't perfect, it looks like it matches pretty well !

I don't know why they would use a type N instead of a type K though ? Maybe because it has a longer life span ?

[AndyC_772]

Another data point for you: at 350C, I get 7.12mV. Does that correlate?

[wraper]

This is a completely lousy measurement, you cannot just touch the tips and expect another side to heat up to the right temperature when there is basically a huge heatsink on the other end.

[pigrew]

I don't know why they would use a type N instead of a type K though ? Maybe because it has a longer life span ?

Seems that type N has better stability. The two have very similar temperature curves, so it'd be hard to differentiate them. Note that without the "cold-junction" temperature, you can't tell which thermocouple you have (from a single data-point). You either need two known temperature-voltage pairs, or a (T_thermocouple,T_coldjunction,voltage) tuple.

[wraper]

Ok, thank you sooo much ! Comparing it to the chart, the closest thermocouple type is a type N

On a type N, 4.14mV corresponds to 144°C. Since the conctack between the cartridges wasn't perfect, it looks like it matches pretty well !

I don't know why they would use a type N instead of a type K though ? Maybe because it has a longer life span ?

You made assumption based on incorrectly done measurement.

[wraper]

I don't know why they would use a type N instead of a type K though ? Maybe because it has a longer life span ?

Seems that type N has better stability. The two have very similar temperature curves, so it'd be hard to differentiate them. Note that without the "cold-junction" temperature, you can't tell which thermocouple you have (from a single data-point). You either need two known temperature-voltage pairs, or a (T_thermocouple,T_coldjunction,voltage) tuple.

You can estimate temperature on the cold end and get a precise enough result. There is quiet high difference between K and N to not be able distinguishing them.

[whatsthegeek]

Another data point for you: at 350C, I get 7.12mV. Does that correlate?

I get 236.03, very far, but still the closes of all.

This is a completely lousy measurement, you cannot just touch the tips and expect another side to heat up to the right temperature when there is basically a huge heatsink on the other end.

The only other way would be to power the connect a wire inside the handler and measure the temp while on which is hard to do and dangerous.

The two have very similar temperature curves, so it'd be hard to differentiate them. Note that without the "cold-junction" temperature, you can't tell which thermocouple you have (from a single data-point). You either need two known temperature-voltage pairs, or a (T_thermocouple,T_coldjunction,voltage) tuple.

How could we get the could junction temperature ? (I don't know a lot about thermocouple)

You made assumption based on incorrectly done measurement.

Still better than nothing ^^

Thanks again 

[wraper]

Still better than nothing ^^

Thanks again 

No it's worse.
EDIT: and I can say why. There are a lot of people who screwed up calibration of their soldering iron thinking they are adjusting it to display correct temperature. Put regular thermocouple on the tip and think they make correct measurement while actually it may be like 50^oC off best case. Error is much worse in this case and you already made an assumption which, most likely, is wrong.

The only other way would be to power the connect a wire inside the handler and measure the temp while on which is hard to do and dangerous.

As I said, put the end of cartridge into boiling water.

[whatsthegeek]

As I said, put the end of cartridge into boiling water.

Maybe AndyC_772 could try this ? But, isn't this bad for the tip ?
And also, at 100°C I don't think the difference will be easily visible on a common multimeter :/

[wraper]

And also, at 100°C I don't think the difference will be easily visible on a common multimeter :/

If we assume cold end is at around 25^oC temperature, there will be around 3mV vs 2mV. Not so tiny difference I can say.

[Ysjoelfir]

I have made many measurements on jbc cartridges in the past for my diy jbc station. righ now I am not near my lab, but I can take a look after christmas. I for myself came to the conclusion that it is pretty much an N type thermocouple, not an K type.

[kubeek]

I guess you could also do it the other way, and put a know voltage into the soldering base and see what temperature it thinkgs the tip is at. Should be a bit more precise and provide a definitive result.

[wraper]

I guess you could also do it the other way, and put a know voltage into the soldering base and see what temperature it thinkgs the tip is at. Should be a bit more precise and provide a definitive result.

I don't think it will be that easy to do. Most likely there is also a current sensing, so it will just show an error if no load present. And if there is a load it probably could show an error because it does not respond to heating.

[AndyC_772]

you cannot just touch the tips and expect another side to heat up to the right temperature when there is basically a huge heatsink on the other end.

The thermocouple junction has to be right at the tip, by design, and the two tips are thermally connected by liquid solder which is an excellent conductor of heat. The two junctions will be much closer in temperature than you're assuming.

What's not known is the temperature of the junction between the cartridge body and the meter probe; it's different between the two leads, and does change over time as the cartridge warms up, which means the measurement shown on the meter isn't a true representation of the thermal EMF produced by the thermocouple.

I wonder how JBC deals with that issue.

[alanb]

Do you need to know whether it's N or K type at this stage? You can get an IC that caters for both types of thermocouple (MAX31855 for example). I assume the different type of thermocouple is selected in the firmware.

[RoGeorge]

The thermocouple can be considered as a linear device for soldering temperatures, so all you need to determine is how many microvolts per degree it has.
Dip the unpowered soldering tip into water with ice (0°C), then into boiling water (100°C), and measure the generated voltage.
Any fine tuning corrections for the uV/°C constant of the thermocouple will be made later, at calibration time.

When calculating the temperature at soldering time, don't forget to add the cold junction compensation offset (the room temperature).

[wraper]

Do you need to know whether it's N or K type at this stage? You can get an IC that caters for both types of thermocouple (MAX31855 for example). I assume the different type of thermocouple is selected in the firmware.

MAX31855 has separate version for each thermocouple type.

[SeanB]

Cold junction compensation can be one of a few methods, using either a diode or thermistor to read ambient at the controller and adding it to the measured voltage, or the more common thing done in these and simply assuming ambient is 25C and having a fixed offset. The fixed offset is quite common, as it does not worry too much if the ambient varies 10C either way, the variation in tip temperature in any case varies more just from orientation, or the amount of solder on the tip, along with it being in contact with the objects being soldered.

If you want true cold junction compensation you need either a simulated cold junction using a proper thermocouple amplifier and conditioner, or just add the scaled offset. Otherwise you need a cold junction with the melting ice bath and all to have the cold junction referenced to.

[wraper]

you cannot just touch the tips and expect another side to heat up to the right temperature when there is basically a huge heatsink on the other end.

The thermocouple junction has to be right at the tip, by design, and the two tips are thermally connected by liquid solder which is an excellent conductor of heat. The two junctions will be much closer in temperature than you're assuming.

First of all there is some distance. Don't expect it being closer than 5-10mm in each of cartridges. Secondly, on the other side of it, there is much bulkier part which acts as a heatsink. And your measurements do not even remotely match nor N, nor K type. You apparently did not take into account what I wrote about big measurement error even when using usual thermocouple with exposed junction which does not have such thermal mass and a heatsink on the other end.

EDIT: I just made some tests with thermocouple as on the pic below. When just inserting junction into a solder blob (like most people do), measurement is around 30^oC lower than actual tip temperature.  Only if inserting it into solder blob together with around 4-5 mm of wire, measurement is correct, because those wires no longer can act as a heatsink. In your case, situation is an order of magnitude worse.

[whatsthegeek]

Well, I think I can safely assume it's a type N. it's the only curve that is close to K, goes higher than 400°C and gives lower voltage for the same tamperature. Since I can't buy any MAX ic that handles type N on LCSC, I'll just convert the temperature result from a MAX6675 all-in-one type K ic with this code:


if (t < 412){
   return in * 1.256;
}
else {
   return in + 126.65;
}


it's precise to +-2°C which, I think, is enoguh for a soldering iron 

[wraper]

Well, I think I can safely assume it's a type N. it's the only curve that is close to K, goes higher than 400°C and gives lower voltage for the same tamperature.

The only thing safe to assume is that measurement was done incorrectly. Yes, reading is way lower than soldering station set temperature, however how much lower, you can only guess. 200^oC (170+30) at thermocouple junction, assuming given conditions, 350^oC set temperature and around 30^oC at cold end seems completely possible to me.

I'll just convert the temperature result from a MAX6675 all-in-one type K ic with this code:

And get cold junction compensation working incorrectly. Therefore reading dependent on ambient temperature.