Soldering Constantan Wire (Nickel/Copper Alloy) for Thermocouples
Q. Subject: Solder Type "T" thermocouple extension wire
(copper/constantan). Solder constantan to constantan with minimal molecular contamination to the base metal from flux, cleaner or solder elements.
The thermocouples (T/C) are used to monitor the temperature of spent nuclear fuel in dry cask storage. It is incomprehensible to me why this topic isn't adequately covered in the instrumentation or nuclear fields. While my search of finishing.com yielded some related information, I didn't find information specific to soldering constantan wire.
Type "T" thermocouple wire typically consists of a 100% copper wire and a constantan wire composed of 45% nickel and 55% copper. Wire manufacturers may add trace impurities to manipulate the physical characteristics of the wire/cable.
I need to extend the thermocouple (T/C) cables by joining the copper lead to the copper wire and the constantan lead to the constantan wire. Welding would be ideal but difficult. The presence of a third metal (solder) in between the wire and lead will not affect the accuracy so long as the constantan junction is at the same temperature as the copper junction.
The technique that I am trying to use is to mechanically abrade the AWG #20 constantan wire with 150 grit cloth, clean the wire with 70% Isopropyl Alcohol [linked by editor to product info at Amazon] and solder using lead-free solder and 140 watt soldering gun with no flux. I am tinning the lead to obtain a consistent surface over the exposed wires. The two wires are then gently twisted together and then soldered for mechanical soundness.
Issues of concern are:
1. An even and consistent layer of the third metal between the lead and wire, which must be a homogeneous alloy free of contaminants.
2. Molecular contamination of the base metal by the flux, (this can change the thermo-electric characteristics of the junction thereby affecting the calibration of the T/C system).
3. Overheating of the insulation, (same concern as above).
4. Overheating the base metal, (same concern as above).
5. Excessive cold working of wire by bending, (same concern as above).
1. How should the constantan be prepared\cleaned for tinning?
2. Can the nickel/copper alloy wire be soldered (tinned) without flux?
3. If flux is required, what type of flux could be used and how best can the solder connection be kept free of flux contamination?
4. What should be used to remove excess flux?
5. Which solder will yield the best compromise of ease of application and minimal potential for contamination, (60/40, lead-free, silver)?
6. What wattage should be sufficient to tin and join two AWG #20 constantan wire without degrading the insulation or base metal?
7. What products (solder, flux & cleaners) are recommended and what suppliers provide these products?
8. How can I verify that I have obtained an evenly wetted tinning and a consistent layer of solder over the constantan wire?
- Rowe, Massachusetts
A. I just used 400 grit carbide paper until both wires were bright, twisted them together and soldered the tip only. I had very little problems and since the solder was beyond the twisted junction, there was zero cold junction problem.
I started out with soft silver solder and later changed to electronic grade 60-40 solder, which was cheaper and easier.
- Navarre, Florida
Q. Thank you James, I will change the specification on the abrasive cloth to 400 grit carbide. When you say that you had "very little problems", what was the nature of the problems that you did experience? I also want to reiterate that I will be joining constantan to constantan and copper to copper.Timothy Snyder [returning]
- Rowe, Massachusetts
A. Why not just buy the thermocouples made to the length that you need, or the length of the longest one and cut the others. It is basically what you are trying to do now.
Another thought might be to twist them together (after cleaning )and then bent the end back half way and slightly twist. Cover with a vinyl dip material or use a heat shrink teflon sleeve to isolate the wire from oxidizing atmosphere. No solder as a possible cold junction. You will not break the wires if you are very careful in the stripping operation and do not use too much brute force.
My thermocouples were in corrosive plating solutions. The company I worked for was cheap.
- Navarre, Florida
Q. The intent is not to fabricate a T/C but to extend the extension wire. Our system was designed as two separate cable pulls. The type T cables were pulled from the instrument enclosure to a stanchion terminating at an Omega Engineering multi-pin T/C connector. Another cable was pulled from the industrial grade 12" T/C to the mating end of the multi-pin connector. The multi-pin connectors were not properly installed (by design) and have subsequently degraded due to environmental conditions. Therefore, I am proposing that the multi-pin connectors be replaced with soldered splices. This means that I would be joining the copper wire to the copper wire (easy) and joining the constantan wire to the constantan wire. Since the wires are #20 solid conductors, a mechanically crimped connection is not advised. As stated previously, a third metal interface within the splice will not de-calibrate the T/C system so long as the same metal interface exists in the complementary lead. The difficulty is in joining the constantan wire to the constantan extension wire. When tinning the constantan the solder doesn't wet well or flow evenly over the wire. There are numerous aspects of the tinning, splicing and soldering that I need to develop the techniques for producing a sound splice with a minimum of de-calibrating characteristics. 1)Wire preparation, 2)flux selection and application, 3)mechanical joining, 4)solder selection, 5)soldering heat source, 6)soldering temperature monitoring/control, 7)flux removal, 8)electrical insulation and sealing.Timothy Snyder [returning]
- Rowe, Massachusetts
A. When soldering constantan without any treatment other than applying flux, you most likely will not get a good bond. Rather, you must apply an acid to the constantan (we use an acid here, however, I don't recall its name at the moment...if anyone knows what acid would work, please reply) so that the solder will bond molecularly with the constantan and so that the copper wire can then be soldered to the constantan. You could use the same acid to treat the copper also, instead of standard flux. Of course, once you have finished soldering a junction, you should wash the junction in water to remove any remaining acid so it doesn't continue to eat away at the junction.
Of course, one would ask, why solder the junction anyway? If you have a welder available, it would be much better to weld the junction.
- Quakertown, Pennsylvania
October 13, 2008
A. For the purpose of joining wires in the above situation the most reliable joint is going to be a crimped tubular link with heatshrink over it. It is also cheap and quick.John Leinster
- Gladstone, Australia
August 5, 2009
A. A neat solution would be to use Din mount Thermocouple Terminal Blocks. Omega can supply them for about $8 each and have types K, J, E, T and R available.Joubert van der Merwe
- GEELONG, VICTORIA, AUSTRALIA
April 9, 2010
Q. Why is thermocouple special extension cable required. If I used any cable between thermocouple and controller what what happen?
If I cut the extension cable in between and put copper cable, so thermocouple show the reading or not? If yes, what is the temperature difference compared to actual value?
employee - Bharuch
October 3, 2012
Q. In an earlier post someone mentioned using an insulated barrel splice. It's my understanding that you don't want to junction the TC wire with any metals other than those similar to the wire. If you are going to join a constantine wire shouldn't you use a connector made of constantine? RFQ: If such splice connectors are available please let me know as I have been searching and can not find anything. I am splicing J thermocouple wire.gary Nelson
- New Brighton, Minnesota USA
This public forum has 60,000 threads. If you have a question in mind which seems off topic to this thread, you might prefer to Search the Site