Thermocouple reading using PROP up to 2300F

eagletalontim

I am working on building a programmable heat treat oven for knives and need a way to accurately measure the temperature inside the oven +/- a few degrees Fahrenheit. I am leaning towards the K or N type thermocouple for the price, but the R or B type looks to be better suited, but more expensive. Been reading about different single IC's that can connect a thermocouple directly to the IC and give a reading, but most I have seen go up to 700F or 1250F which is just not high enough. The MAX31855 looks promising which can do up to 700C on a K type but would it looks like it would require I go with a more expensive thermocouple which can do the higher temperatures, but struggle at the lower temps...or maybe I read that wrong.

The temperature range I am looking to accurately read is between 200 and 2300F. Most common will be between 400 and 2000F

Is this the best IC for what I am looking to do? Any suggestions on which thermocouple I should go with? Thanks for any help in advance!

avsa242

@eagletalontim
I've used a K-type with a MAX31856 (driver), though not nearly as high temps as you're doing ...only at kitchen oven temps.
The datasheet has a table that lists K-type range as -200 to 1372C, which looks to cover your expected range.

JRoark

Type R & S thermocouples tend to be fussy. I've used them, but would suggest that you avoid going there if you can. For what I was doing (in the same temp ranges you are contemplating), they were just added expense and drama so I went back to K's and they worked fine. If I was doing this over again, I'd be inclined to use Type N just to play with it.

The K/N range is specified as -200 to 1300C, but there are caveats. Nickle undergoes a voltalic discontinuity around its Curie point. If you're "just passing thru" to another temperature, this may not be an issue, but getting accurate readings right around that Curie zone gets tricky. Likely not an issue for you.

If you happen to have access to a TIG welder, you can make your own thermocouples using just Type K extension wire. I used a jig to position the wires and then zapped them gently with the TIG under argon. You can make a half dozen of these for the price of one store-bought bead-type K, and the accuracy is comparable. This TIG trick also works a treat to eliminate splices using M/F connectors. Every one of those darned things introduces noise/bias whereas a welded same-to-same wire gives you a clean signal without mechanical/electrical artifacts.

K2

Agree with JRoark. Type K extension wire works fine - the type with woven fiberglass insulation. At 2300 F, thermocouple life will depend on the gauge of the wires used. But even 24 gauge wire lasts a while. From places like Sheffield you can find 8 gauge!

While shielding gas is the best situation, you can absolutely make thermocouple junctions with nothing more than any old arc welder that you crank way down, and the carbon electrode out of a carbon/zinc AA cell from the dollar store. Keep the metal cap on...it will facilitate electrical connectivity. Hold twisted end up, aim carbon rod down, and strike arc for a fraction of a second. (University spoiled me - I had to get over not having argon purge and automatic pulse timing.)

The MAX31855 has worked great for me. There are also 4-20mA current loop thermocouple transmitters that provide cold junction compensation and are very nice if distance or immunity to interference is needed.

DigitalBob

The max31855 works pretty good, but If your going to have long extension wire runs use shielded to prevent faults. Proper wire gauge is important on long runs.

eagletalontim

I do have a TIG in my shop and didn't know I could do that. Why is it so much more expensive buying the ready made vs the extension wire? I plan on ordering a few of those MAX31855 and I only need about 6ft of thermocouple wire as I plan to have 2 to 3 sensors for the front, middle, and back that will control different elements to have a good even heat treatment. They would be sensing about 3ft from the IC.

Another issue I have had is flyback noise from switching a relay and plan on using solid state relays that will be kicking the elements on and off.

JRoark

eagletalontim wrote: »

I do have a TIG in my shop and didn't know I could do that. Why is it so much more expensive buying the ready made vs the extension cable?

I think its mostly because: 1). Not many folks know this trick, and 2). Its a goodly bit of change for a good TIG.

I dearly love my old Miller Synchrowave 250. Totally decked-out with every factory option it was $4k ~30 years ago. You can weld a razor blade to a boat anchor with a bit of practice.When you mash that pedal, the lights dim and the UV flies! Its huge, clunky, has all copper windings and makes a darned Skippy good jumper box for dead cars with the high freq turned off.

eagletalontim

Been looking at Digikey for some parts, but I can't seem to locate the right stuff. I found this, but the temp rating it significantly lower than others of the same "type"... why is that? I don't mind spending ~$20 per sensor, but not $100+ per sensor...

https://www.digikey.com/product-detail/en/digilent-inc/240-080/1286-1099-ND/5418219

eagletalontim

I am a bit confused now... Been researching more into the K Type extension wire and there is apparently a Positive and Negative lead which would have to be bought separate? Omega offers bare wire Here, but it is rated for only 1100ºF which is significantly lower than what I am looking for and I must buy 50ft where I only need about 10ft. Any suggestions on where I can buy this short of length? Is the 1100ºF rating due to the gauge or am I missing something?

K2

eagletalontim wrote: »

The next step after woven fiberglass insulation is 1) a metal sheathed thermocouple from someone like Omega or 2) two-hole porcelain sleeves that you slide over the bare thermocouple wires.

I've used both types. But 90% of the time I use neither. The wire with fiberglass insulation works fine. The fiberglass survives on the portion of the wire inside the furnace wall. Past the furnace wall, the wire is bare. The fiberglass definitely provides enough support and separation for the wires as long as they don't extend way into the furnace.

JRoark

@eagletalontim Here ya go. Cheap fun:

https://www.amazon.com/gp/aw/d/B0768H132M/ref=sspa_mw_detail_1?ie=UTF8&psc=1

The insulation will ablate at the high temps, but the wire will mostly survive. 30 feet gives you lots of practice before you have to have your welding technique down. If the wire gets soggy from the extreme heat, stuff it inside a ceramic tube (also on Amazon). You only need the very tip in the heat. A furnace should be a very turbulent environment so sticking the measuring point way out likely isnt needed for a good reading. At the temps you are running, radiative transfer across the furnace will be huge, but olay with it and see.

DigitalBob

I'm thinking 3-6 ft you should be Ok. You could use a ready made probe assembly. I use TC direct for my thermocouple supplies. I bought some Amazon thermocouples assemblies, but they where cheap Chinese Smile and broke real easy. The K type are yellow (all thermocouples are color coded). Extension wire for K type will have one yellow wire and one red wire. The red wire is actually the negative terminal and the yellow is the positive. The same holds true with all types of thermocouples. With J-type one wire is red and the other is white. The K-type are made from alumel and chromel wire.