Help with heat treating drill rod

[EpicWill]

Interesting the display on the reader is limited to about 1/2 the range of the TC material. High temp K probes are limited by the jacket and the choice of insulation on the wires, not the wire itself. Of course the reader could be limited as well.

Pretty sure it's the reader. It's a Fluke reader but the plug-in probe is an Omega unit so both are high quality units.

 

[WobblyHand]

I have a cheap multimeter with a TC input. It doesn't give believable results. It comes with a TC to dual banana plug adapter which is not keyed. Neither orientation seems to be correct. Maybe the cold junction compensation is off. Or it has been damaged by the reversal.

Also have an old Tektronix dual channel thermometer DTM920. That unit seems to measure temperature correctly. It uses standard TC plug input and is keyed. Same probe that isn't believable on the multimeter, works fine on the Tek.

For what it is worth, you can buy 5 meters of high precision high temp type K wire rather inexpensively. Get some K connectors and you can make up some of your own custom thermocouples. You can weld the tips, but all that is really necessary is a tight twist of the wires together.

 

[RJSakowski]

Ideally calibration of any device should include a calibration point close to the intended use. Calibration should consist of a minimum of two points. Calibration using the freezing point and boiling point of water for use at temperatures around 1000ºF requires excessive extrapolation and potential errors.

Fortunately, thermocouples have very well known curves and a two point calibration is usuqlly sufficient. The voltage vs. temperature isn't linear but if you know the type of thermocouple being used, the temperature can be determined by reading the voltage across the cold end of the wires. This web site has calibration curves for various types of thermocouples. https://www.omega.com/en-us/resources/thermocouple-types. The problem that occurs is the accuracy of the algorithm that converts the millivolt reading to temperature, h3nce the need for calibration.

The ice point would be an acceptable calibration point. For the upper end, I would choose something like the melting point of copper. Copper electrical wire should be close enough to pure copper for a reasonable check point. The melting point of lead would be a good intermediate point.
.
The easiest way to calibrate is against known good thermometers. An infrared thermometer is a convenient way. However, my infrared thermometer won't go up to hardening temperatures and they are dependent on the correct emissivity for the object being scanned. My multimeter uses a type K thermocouple and reads to 1382ºF.

Another method, if you know the type of thermocouple, is to disconnect the thermocouple and use a variable voltage to insert a suitable voltage, as determined from the voltage vs. temperature curves for the thermocouple used. Use an accurate digital voltmeter to measure the input voltage.

 

[WobblyHand]

Ideally calibration of any device should include a calibration point close to the intended use. Calibration should consist of a minimum of two points. Calibration using the freezing point and boiling point of water for use at temperatures around 1000ºF requires excessive extrapolation and potential errors.

Fortunately, thermocouples have very well known curves and a two point calibration is usuqlly sufficient. The voltage vs. temperature isn't linear but if you know the type of thermocouple being used, the temperature can be determined by reading the voltage across the cold end of the wires. This web site has calibration curves for various types of thermocouples. https://www.omega.com/en-us/resources/thermocouple-types. The problem that occurs is the accuracy of the algorithm that converts the millivolt reading to temperature, h3nce the need for calibration.

The ice point would be an acceptable calibration point. For the upper end, I would choose something like the melting point of copper. Copper electrical wire should be close enough to pure copper for a reasonable check point. The melting point of lead would be a good intermediate point.
.
The easiest way to calibrate is against known good thermometers. An infrared thermometer is a convenient way. However, my infrared thermometer won't go up to hardening temperatures and they are dependent on the correct emissivity for the object being scanned. My multimeter uses a type K thermocouple and reads to 1382ºF.

Another method, if you know the type of thermocouple, is to disconnect the thermocouple and use a variable voltage to insert a suitable voltage, as determined from the voltage vs. temperature curves for the thermocouple used. Use an accurate digital voltmeter to measure the input voltage.

I'd agree that 0C and 100C calibration points are too close for real calibration especially if measuring things over 1000C. But they are convenient temps which are safely tested. These lower temperature cal points give us confidence that the TC is indeed type K, or whatever you are using. Beyond that we are trusting that the wire compositions and circuitry are ok. Nothing wrong with high temp cal, but it gets more and more difficult as the temperature rises.

 

[RJSakowski]

I'd agree that 0C and 100C calibration points are too close for real calibration especially if measuring things over 1000C. But they are convenient temps which are safely tested. These lower temperature cal points give us confidence that the TC is indeed type K, or whatever you are using. Beyond that we are trusting that the wire compositions and circuitry are ok. Nothing wrong with high temp cal, but it gets more and more difficult as the temperature rises.

Thermocouple temperature responses are stable and well defined. I would be worried about the circuitry converting the voltage to a displayed temperature. There was no mention as to whether this was a new or used oven. If it was used, there is a possibility that the thermocouple was replaced with the wrong type. While in theory, the thermometer could still be calibrated to read properly, provided there was that much adjustment range, using calibration points well outside the working range would lead to errors. Aside from that, there are always user induced errors which can creep in. Large extrapolations tend to amplify those errors which is why I prefer to use calibration points which stradle the range of use. If I can't do this, my next best option is to pick calibration points close to the range of use.

 

[EpicWill]

Thermocouple temperature responses are stable and well defined. I would be worried about the circuitry converting the voltage to a displayed temperature. There was no mention as to whether this was a new or used oven. If it was used, there is a possibility that the thermocouple was replaced with the wrong type. While in theory, the thermometer could still be calibrated to read properly, provided there was that much adjustment range, using calibration points well outside the working range would lead to errors. Aside from that, there are always user induced errors which can creep in. Large extrapolations tend to amplify those errors which is why I prefer to use calibration points which stradle the range of use. If I can't do this, my next best option is to pick calibration points close to the range of use.

The oven is brand new and had only been through it initial firing cycle (1600F for 2 hours). The TC meter is a brand new Fluke 561 however the KTC probe is a few years old but is an Omega unit so it's a good brand and is in good condition.

 

[RJSakowski]

The oven is brand new and had only been through it initial firing cycle (1600F for 2 hours). The TC meter is a brand new Fluke 561 however the KTC probe is a few years old but is an Omega unit so it's a good brand and is in good condition.

That leaves the material as the prime suspect. If the steel is a low or medium carbon steel, you will never achieve the hardness you want. I would suggest that you get a piece of known good O1 and go through the hardening process. If successful, you have found the culprit. If not, then either your process or the TC meter are the problem. You might try increasing the hardening temperature to the high end or slightly above the recommended range,

 

[homebrewed]

The curie temperature for pure iron could be used as a sanity check. According to Wikipedia it's 1418F. Nickel's curie temp is 669F, too low for what you want to do.

 

[EpicWill]

Ok I just tried doing the torch method and it seemed to work pretty well when heated to a bright orange color. I checked one of the other pieces from yesterday and it seems to be between 55-60 HRC according to the hardness testing file set we bought. I think my interpretation of the instructions regarding "marking" the material were too vague as the 45 HRC file did mark it but it definitely skates across the surface compared to the 60 that for sure bites in.

I'm going to fire it up to 1500 again and take a picture when I open the kiln door so you guys can tell me if the color looks correct for that temp. I also have a thermal probe that a K type thermocouple so it'll go to 2300F so I'll see what it says the kiln is at when it saying it's at 1500F.

Pretty sure it was all down to my interpretation of "marking" the material. I have the shafts I made in now and we'll see how they go

 

[EpicWill]

Alright something isn't right here. I fired these 9 shafts that I made. Fired to 1500F, held for 20 minutes to be sure. Pulled them, quenched them and not a single one is hard. The 40HRC file eats them up and I can bend them by hand.