Thermocouple connector question

[throbscottle]

I'm turning my ordinary soldering iron into a temperature controlled one using a cheap PID temperature controller from eBay **broken link removed**
and a thermocouple such as this **broken link removed**.
I found a female k-type thermocouple connector on there: **broken link removed**
but I have no idea what these connectors are made of.

Question is, do I need to use thermocouple wire to connect the female connector to the PID controller, or do I use copper wire?

Thanks in advance

 

[alec_t]

I can't answer that, but I'm curious as to how you are going to keep the thermocouple in good thermal contact with the iron tip?

 

[throbscottle]

I was planning to hide it in the space behind the tip, in the body of the iron - or possibly make a little hole in the side of the collet (collet? I think that's the right word!) - I know the temperature won't be accurate but I'll know how far off it is so can adjust accordingly. Or I can make a little ring to hold it to the side of the tip for better accuracy - though I'm not interested in absolute accuracy, just in being able to make the iron run cooler.

I was originally planning to run the iron using a timer with an adjustable on/off interval - inspired by no less than the iron's power cable developing a break in it and running intermittently, but actually gave a very useful lower temperature (until the fizzing and smoking started at the break...). It's a very cheap and old 30W iron so it doesn't matter if it gets vandalised. But these temperature control units are pretty cheap on eBay, and then there is always the chance to use it for something else...

But I /think/ I've found the answer to my own question - it appears these connectors *should* be made of the thermocouple metals, so it will be thermocouple wire all the way.

 

[jpanhalt]

Question is, do I need to use thermocouple wire to connect the female connector to the PID controller, or do I use copper wire?

Thanks in advance

Here is a nice discussion of how to connect thermocouple wires: **broken link removed**

In brief, the Law of Intermediate Metals (see above) allows you to use whatever metal you want to use. The key is that at each junction between different metals, you create a new junction voltage. Now, if you have only one junction that is hot, and the other junctions are all at the same temperature (i.e., room temperature -- your reference temperature), the different junction voltages cancel out and you get a good reading. Nevertheless, it is common to run the two thermocouple leads all the may to the terminals of the measuring device.

Another version of the question asks whether the measuring junction must be welded or can it be soldered or brazed. Again, you can use solder or brazing, so long as the metals in the measuring tip are all at the same temperature. Of course, solder may not be a good choice for the thermocouple in a soldering iron.

John

 

[throbscottle]

That's what I love about this forum. Ask a simple question and I learn stuff I never even knew to ask about! Thanks!

 

[KeepItSimpleStupid]

There is thermocouple wire and thermocouple extension wire. In the TC world RED is negative. The extension wire has the same behavior and is more rugged and is of a much high gauge. Extension wire is cheaper.

Thermocouples can be purchased pre-made or can be spot-welded. I've done both. One spot-welding technique for small thermocoules is to use a carbon block and a high current power supply. You can twist the wires and drag across the block. This works for J and K types for sure and probably T. R and S TC's can be fused together with a torch. Those were the TC's I was most familiar with. We did use a Type C type TC in one process though.

You don't want multiple twists at the end. This creates multiple junctions and the effect is to average. Yep, thermocouples in parallel inherently average. A quick test for thermometer is to short the terminals of the thermometer together with a piece of wire. The thermometer should read room temperature if the readout will read room temperature.

The major importance is that each 1/2 junction is in thermal equilibrium of the other 1/2 junction. A LT of the thermocouple measuring stuff I have seen really does work well. There was this one where if you blew a fan over the USB box, the temperature would change when the TC was located 10-15 feet away.

Getting a temperature from a voltage is more difficult than you think. You have to KNOW the temperature of the terminals at the measuring instrument absolutely using a silicon sensor usually. Then find out what the voltage would be at that temperature using a cold junction of 32 deg. So it's an inverse function. Then the voltage at the instrument is measured and either added or subtracted (forget which) to find the voltage that corresponds to the temp using tables or polynomials.

 

[throbscottle]

This is why I've just paid £8.20 for a PID temperature controller with built in cold junction compensation. I can stick it in a box and wire it up, and it's way cheaper, easier and more accurate than buying parts and trying to roll my own

 

[KeepItSimpleStupid]

Yep, but what I'm suggesting is that your controller could have a sensor mounted NEAR the terminals and thus senses the temp around the terminals and not the terminals themselves. Also, I wanted to give you the general "heads up" with thermocouples.

 

[Rbeckett]

I gotta go with throbscottle on this one. I have been experimenting with and building an automated brewery controller and PID's are really the easiest way to go. Then you can switch them on and off according to need with a relay or transistor as a switch. I was experimenting with DS18B20's,and K-type sensors and thermistors. Less parts=less problems and more robust design options and more ease of interface. You can do everything you want with a couple of Pid's linked to an Arduino and just about any 16F series chips and build a nice system/platform for pretty decent money. Cuts way down on interference and ballancing discreet components for noise and correction. I am using single output and dual output PID's linked to a 16f877a to set up the timing and do the basic logic decision making. You might also try an FPGA to programs all of your logic funcions on and save lots of time and building too. If you want to change the decisions you can reprogram an FPGA quick with minimal rewiring, Just my humble opinion from my experience with this brewtroller build. But I am also definately doing it Betty Crocker Cookbook Style with bits and pieces from everywhere on the net and in text books too. Hope this helps get you going.
Bob

 

[KeepItSimpleStupid]

I would too. But then I've built some really crazy control loops before. The easy one was to use the proportionl band of a controller to control the temperature of another RTD whereas the normal RTD was located in the chamber. Kept inching up the setpoint, but never higher than the proportional band to control a heat/cool environmental chamber from -80 to +200 C. Kinda cool.

I know what the difference is between a $1200 thermometer and one for $250 and I've seen some stupid ones.

I'm partial to Eurotherm temperature controllers.

 

[throbscottle]

Yep, but what I'm suggesting is that your controller could have a sensor mounted NEAR the terminals and thus senses the temp around the terminals and not the terminals themselves. Also, I wanted to give you the general "heads up" with thermocouples.

I'm generally out of my depth with thermocouples anyway (even even though I've now read the excellent ni.com article), but the PID controller does give the option to poke a sensor out of the casing next to where the thermocouple is connected, if a three terminal RTD is not being used - you poke the sensor out the hole for the third terminal. Is this the kind of arrangement you mean?

This is the datasheet for the unit that's in the post: https://www.electro-tech-online.com/custompdfs/2012/07/c100inst.pdf - it's way higher specced than I need but it was only £8.20 & post free after all.

But thanks for the heads up - interesting posts - thanks

 

[MrAl]

Hi,


One of the biggest problems in thermocouple literature and understanding thermocouples is the use of the word "junction". As soon as an electrical guy hears "junction" it brings up thoughts of two wires connected together and the ends that form a junction. That would be correct...correct had we been talking about an electrical circuit. However, when we talk about thermocouples we are not always talking about an electrical circuit usually we are referring to a thermal circuit or just a thermal conductor. This means when we talk about the HOT junction and the COLD junction we are not necessarily talking about an electrical junction.

So the first item on the list then is to figure out what exactly a JUNCTION really is.

There are two types of junctions used with the thermocouple. The first type is an ELECTRICAL JUNCTION, and the second type is a THERMAL JUNCTION. Let me list them here:
1. Electrical
2. Thermal

Now an electrical junction is usually the first thing we think of. That's where two wires are brought into close proximity to each other. In fact, they touch so that means an ideal junction has no voltage across it.

A thermal junction is the thermal counterpart of the electrical junction, but it is totally, totally different. A thermal junction is a place where two thermal conductors come into close proximity to each other. They dont have to touch though, just come very close to each other so that the temperature difference between them is zero.

So the electrical junction has no voltage drop, and the thermal junction has no temperature drop.

The big difference here is that the thermal junction can be formed from two metals yet never touch, while the electrical junction has to have conductors that actually touch. But because the wires are so close the electrical junction is also a thermal junction. So the electrical junction is both an electrical junction AND a thermal junction, yet the thermal junction is only a thermal junction.

Now we can talk about the HOT and COLD "junctions". Note the word "junctions" appears here and in all the literature.

The HOT junction is an electrical junction, so it is also a thermal junction.
The COLD junction is ONLY a thermal junction, so the wires dont touch yet they are at the same temperature.

The HOT junction has two wires that touch, the COLD junction has two wires that do not touch. So this means the HOT junction is an electrical connection, but the COLD junction is just a thermal connection with no ability to conduct current within it.

This brings us to the subject of using other metal conductors with thermocouples. We need to know the temperature of the cold junction in order to calculate what the temperature of the hot junction is. This is so because the two dissimilar metals produce a known voltage difference between the hot junction and the cold junction, and the difference between the ends at the cold junction allow us to measure this voltage difference.
However, because we MUST know the temperature of the COLD junction, we can not use ordinary wire between the COLD junction and the measuring device if that measuring device is also measuring the cold junction temperature at the ends of the ordinary wire. This is going to be the case with temperature meters that use a thermocouple wire with a connector at the end. That's because the meter depends on knowing the temperature of the actual wires at the connector. If we insert wire between the connector and the meter, we dont know that anymore. We need to know BOTH the electrical voltage difference AND the temperature at the cold junction, not just the electrical difference. Using ordinary wire, we would only know the electrical difference and loose the temperature measurement.
If we could somehow still measure the temperature of the ends of the dissimilar metal wires we could still do this, but usually that's not the case. That would mean we would need two more wires to send the cold junction temperature back to the meter.

There is a saving grace here however, and that is if the two extra wires can be kept at the same temperature at both ends. That would mean we could still know the true cold junction temperature. Usually this isnt that easy to do however because what is preventing that wire from getting one end heated more than the other.

 

[KeepItSimpleStupid]

if a three terminal RTD is not being used - you poke the sensor out the hole for the third terminal. Is this the kind of arrangement you mean?

In a nutshell, yes. It makes the measured temperature very susceptible to air currents around the back of the instrument.
Take a fan to the back and watch the temperature change when the TC is far away.

The USB one that I had was really annoying. In fact, it was this one: https://www.mccdaq.com/usb-data-acquisition/USB-TEMP-Series.aspx It would not do Thermocouples right at all with a fan blowing nearby. Fortunately, we were using RTD's which it did well.

In a very well designed thermometer the terminals would be massive and in very close contact with each other. e.g. a shoulder washer and a Sil-pad (thermal washer) with the junction measuring thingy in intimate contact with the metal. This is usually way to much work.

 

[throbscottle]

MrAl - I have been trying to get my head around this for quite a while now and read a few different sources and still been scratching my head - yours is by far the clearest explanation I've seen - thanks

 

[MrAl]

MrAl - I have been trying to get my head around this for quite a while now and read a few different sources and still been scratching my head - yours is by far the clearest explanation I've seen - thanks

Hi throbscottle,

Thank you much. I tried to be brief but as usual i find myself torn between brevity and a clear explanation of the important facts. I strive to get the important stuff out there but it sometimes takes more words than i'd like. So if you take anything from my little 'discussion' take the difference between the two types of junctions (electrical and thermal) and the fact that we must be able to measure the temperature of the cold (THERMAL) junction, and that wires between the cold junction and the measuring device wont work unless they have little temperature gradient or the measuring device has some other means of measuring the cold junction temperature, or of course they are made of the same material as the original thermocouple wires.
And of course good luck with your project

 

[KeepItSimpleStupid]

One misconception: Extension wire for thermocouples are NOT made of the EXACT same material as the thermocouple itself. Here is some extension wire stuff: **broken link removed** but they will have the same thermal characteristics. They won't necessarily have the same corrosion or temperature range as the original materials.

 

[KeepItSimpleStupid]

Click here: https://en.wikipedia.org/wiki/Seeback_effect#Seebeck_effect and on one of the "thermocouple" highlighted link which will also take you back to Wikipedia.

 

[MrAl]

Hi again,


Yes sometimes other types of metals are used for the extension but also sometimes at the cost of reduced temperature range or reduced accuracy over the temperature range.

 

[throbscottle]

I'll post up the results when all is up and running

 

[strantor]

I'm partial to Eurotherm temperature controllers.

They are the best. That's all I deal with.

In answer to OP, all theory aside, in practice, the machines I deal with always use the red/white extension wire, that (I believe) are the same metals used in the thermocouple. Never copper wire, and never going through terminal strips; straight run all the way. This is because as mentioned, junctions of dissimilar metals mucks things up.

EDIT: oops, I didn't realize there was a second page to this thread.

EDIT2: having read the second page now, I find that my own understanding is a little off.