[SolarWPA]

WHY NOT MAKE YOUR OWN THERMOCOUPLES?
by Bob Hill of Solar Atmospheres

Thermocouple manufacturers often charge you two times what it costs for you to make your own most common type K TC’s. If you have only ONE person trained to make all of your TC wires required, I guarantee that your wires will perform BETTER than those that you buy.

At Solar we use both refrasil and stainless steel braided type K. We only use refrasil sheathed wire when performing surveys because they are used once and they are cut up and recycled. The SS sheathed TC’s are more robust and are typically used as daily work TC’s. Our SS TC’s are used for up to five runs maximum and then cut up and recycled. (Any aerospace jobs we limit the use to one heat treat cycle).

Here is a step by step primer on the CORRECT way to make type K thermocouples.

1. Cut wire to your desired length.

2. Peel the SS braided material back a minimum of 6” exposing the refrasil.



3. Pull the refrasil back at least 6” as tightly as possible.



4. Make sure the wires are parallel and not crossed while sliding the ceramic insulator on. Note- it is critical to make sure the wires are not crossed at the back of the ceramic.



5. Keeping the insulator tight against the bunched up refrasil, tightly twist the wire at least 1” long. Cut off extra loose wire.



6. On the plug end repeat steps 2 and 3.

7. Find the magnetic wire (typically the dull wire) and orient it to one side. Cut both wires about 1” long.



8. Insert both wires into the plug keeping the magnetic wire on the left and the non magnetic wire on the right or + terminal.

9. Tighten terminal blocks well. Push back the refrasil into the plug to be sure the bare wire is well insulated. Give a tug on the wire to make sure it is tight and replace cover.



10. Test every wire by plugging into an Omega Model HH-25KF Thermometer. You will know right away if your TC is made properly. To double check, expose the twist to heat from a hot air dryer. Do NOT use an open flame or a match.



This is a picture of 350 feet of refrasil wire used for a survey in a 24 foot furnace. All wires were made and tested by one person (this eliminates finger-pointing if they do not work). The total wire material cost for these 21 wires (various lengths) is $472.50. The high temperature plugs and ceramics are recycled. It took our employee 1 ½ hours to make and test these wires. To buy this wire already made from a manufacturer it would cost approximately $850.00!



So buy yourself a 1000 foot roll of certified thermocouple wire from a reputable manufacturer, save yourself money, and have better more reliable thermocouple wire at the end of the day!

For more information, feel free to visit Solar Atmospheres of Western PA at www.solaratm.com

[admin]

Appropriate hot junctions can only be made by arc welding or the use of special tips type K which you can apply with a crimping tools for making hot junctions.
The twisted wires as decribed will oxidate due to heattreatments and the oxidated wire will act as insulators.

The large twisted wires shown on the foto have a big mass. Thus the reaction time of the hot junction will not respond properly to temperature deviations in the furnace and important information as overshooting etct will be missed.

Don't forget a thermocouple is a precision instrument which has to transport uV's.
Loss of uV's will give deviations in your readings.

I'll suggest to read some literature as PRI information, STP-470, ASTM E220.
Posted by Frank Schellevis

[admin]

Hi,
I'm an Applicaions Engineer for Omega Engineerng. First if all thank you for choosing Omega products to do your testing. I also worrked for Inductotherm Corp and Consarc/PVT where I was Incharge of all inhouse testing including the AM2750D. Gerard is correct in that NIST tracibility is required.

When using type K thermal couple wire you have to remember that the Tollerence for standard limits of error is +- 4 deg F or .75% which ever is greater from 32 degF to 2282 deg F (temps start to follow the .75% curve as the temp gets higher). The could mean as much as 17.115 deg error @2282 deg f. Special Limits of error Type K wire error is +- 2 deg F or .4% which ever is greater from 32deg F-2282 degF..or 9.128 degF error at 2282 deg F. You also must add in the error of the meter, which for the hh25 is up to 0.4%rdg +0.6 degC.\

If you get a NIST certified probe you know how that probe is performing at a specific calibration points. At Omega we offer a Cal-3 (three point we chose) or a Cal-4(you pick the points). A Cal-3 is about $75 and for a Cal 4 its normally a $50 setup fee plus $25 per point.as memory serves me at this late hour.

I feel your best bet is to get a sheathed probe (inconel over inconel like the omega XL series long lasting and stable thru the life span) with a high temp ceramic connector, ungrounded(they will last much longer then the exposed) and to connect up to them with themocouple grade wire inside the vacuum(don't use extention grade). Get your probes with NIST calibration certificates. one 12" long xl probe is about 1/4" dia is about $46.00.

By doing this you will have 3 great benifits...1) lower cost big time because you will be able to use them over and over, 2) NIST tracibility. 3) You will now know the differences from furnace to furnace chamber...accurately.

Note that Omega also sells the calibrators as well.
This will gove you 2 huge advantages
If you need more help please feel free to call me in the office at 1800-872-9436 ext2307

Posted by Ben Box

[admin]

I have consistently used sheathed probes with ungrounded junctions, and remain convinced that it's money well spent.
Posted by Joseph Greene

[admin]

That sounds like an interesting and educational project. However, if any of the heat treating you intend to do requires NITS traceable documentation you'd be in a difficult situation to prove the acuracy of the thermocouple. I highly recomend looking into the ISO standard for Heat Treating known as AMS 2750D and others. My current position requires me to perform Instrument Calibrations, Temperature Uniformity Surveys, and System Acuracy Tests. All of these require tracable documentation for the Correction Factors and the accuracy of calibration as outlined in the AMS 2750D specifications.

If your need is not critical to acurate temperature conversions, I'd be interested to learn how you'd go about the project and see what your progress is.

By Gerard Mondriaan, Technician at Temperature Processing CO, Inc.

[admin]

Making your own thermocouples is like most other manufacturing activities i.e. if you see an economic benefit are prepared to dedicate the resource then sure ... for most buyers ( OEM and End-user ) the numbers are upside down.
Posted by Dave Chlystek

[admin]

I reviewed the description on the Heat Treating Forum, but I don't have a login there so I'm commenting here.
Someone else already mentioned a few problems with unshielded thermocouples and your manufacturing method so I won't bother discussing any of that.
The biggest problem with making your own thermocouples is that you can't show accuracy or, as mentioned by others, NIST traceability. If you're doing aerospace heat treatments like suggested in your article then you've got a big problem as there is no way this will meet those requirements and the only way you'll get this past an auditor is if they don't understand the difference between purchased 'calibrated thermocouple WIRE' and a 'calibrated thermocouple SYSTEM' (sorry about the caps, but there doesn't seem to be any way to add emphasis to text in this new group discussion system).
Anyway, you can meet NIST & AMS 2750 requirements and still make your own thermocouples, but you will need a NIST traceable and calibrated temperature source to be able to calibrate your system (reader, all connectors, and thermocouple itself) to the specific temperature points you need. It's best to do the whole temperature system as opposed to just the thermocouple as that way you have the exact total variation as opposed to figuring out each individual component's error and then adding them up to give you a maximum possible error which might actually exceed your certification limits. Of course that temperature source will have to be calibrated and checked annually (or more often depending on your certification requirements, range & frequency of use, etc.) by a NIST traceable instrument qualified to Field Inspection level or better.
Of course you also need to do furnace profile and stability checks, and usually the easiest way to do that is by bringing someone in from outside with the appropriate traceability and instruments calibrated to FTI level or better, so if you have a more permanent setup you can record the test run with your own instruments at the same time and compare the profiles to determine your variation and show that your system is reading within the allowed variation. That's a way of killing three birds with one stone so-to-speak. Of course in that case your thermocouple placement will have to be within the distance limits to the Test thermocouples, and it won't help if you're replacing your thermocouples every time, so you'd need to be able to shield them and use a material combination that will allow multiple uses.
Posted by Carl Adams

[engine76]

Frank, you are correct the twisted wires will oxidize over time but arc welded junctions will oxidize right away. The large wires have nothing to do with temperature reaction time and are compliant to AMS 2750D. In commercial heat treating this is the norm, and in a lot of cases most people don't even use work thermocouples. Captive shops especially runnning the same loads this is very evident.

[SolarWPA]

There have been a number of cautionary responses to this thread, some with merit and some without. I will not entertain addressing each one individually, rather hit the main concerns.

First, I should have clarified the use of the sensors outlined in this thread. These are “expendable” thermocouples for use as load sensors and TUS/SAT sensors.

Secondly, there was a comment about the twisting of the hot junction pertaining to oxidation of the wires and errors associated with this. Our use is always a vacuum furnace, i.e. non-oxidizing atmosphere, and any other use should include a welded hot junction. In our application, a twisted end will not mask overshoot and will be imbedded in something of greater mass (than the tip) anyway, mass from parts or TUS heat sinks conforming with AMS 2750D section 3.5.10.1.

Concerning calibration, making thermocouples from spool wire is a perfectly acceptable practice per AMS2750D and Nadcap. When procuring it, one must also ask for NIST traceable calibration. There was a comment about AMS 2750D being an ISO standard, it is not. It is an SAE document, however it is recognized internationally. Two members of our organization sit on the committee that wrote this specification and we are as well versed in it as anybody. This specification is designed for aerospace use, and is the most stringent standard for Pyrometry in existence. Nothing in this thread violates any of the AMS2750D requirements.

When procuring the wire referenced in this thread, one must specify “special tolerance” limits outlined in ASTM E230 to ensure AMS2750D compliance with regard to accuracy. When procured this way, a calibrated spool will have correction factors listed from a calibration taken from each end of the spool. In practice, the “front” and “back” (each end) correction factors are averaged for application to TUS or SAT results. Connectors are to have the same thermoelectric properties per AMS 2750D paragraph 3.1.1.5. These are easily procured from numerous suppliers. The entire probe need not be calibrated per AMS 2750D paragraph 3.1.1.6.

It could be argued an individually calibrated thermocouple will have a more accurate correction factor than a calibrated spool, I will not dispute that, however I can tell you from over 25 years experience with the method described in this thread, it has proven to be very cost effective, useful, practically effective, and accurate. As for the rest of the “system”, this is always calibrated during standard pyrometric calibration practice. This means not disconnecting the wires from the backs of the instruments and calibrating only the instrument, but calibrating through all extension wire and jack panels.

In some instances pre-made Inconel-sheathed thermocouples are definitely more useful and appropriate, such as titanium heat treatments where water vapor from the cloth insulation would be detrimental, however the cost is substantially more (by a magnitude of 4 times) and Inconel-sheathed thermocouples have their downfalls as well. We are well versed in the use of this type of sensor as well.

Finally, Pyrometry is certainly a subject of great concern to the heat treating population, and there are as many opinions as there are heat-treaters. The fact is, when making sensors as in the original thread, with wire procured in accordance with AMS2750D requirements and with spool calibration traceable to NIST, one can save a lot of money and still conform with the most stringent Pyrometry standard in the world, while having accurate, useful data.

Posted by Bob Hill of Solar Atmospheres