Heat treat oven accuracy and thermocouple response rate

[JTknives]

I have been wondering about this topic for a long time and thought it would be good to invite some other minds into the circle. The big question on everyone's mind when thy build or buy a oven is "can I trust it", well can you? After building mine I noticed that there is a flaw in the general excepted concepts. Now I could be off base and probably am but I thought I would put this out there. I have noticed that the thermocouple sticking into my oven never reaches the same color as the inside of my oven. Well that's not exzacty true, it will get close after sitting at temp for an abnormally long time. This got me thinking, I know different materials have different IR emissivity index (0 to 1.0) but does that affect what we see or just IR laser heat guns. What I'm wondering is is kinda a two part question. A: will the junction of a TC be lower in temp because the leads are going outside the oven and acting as a heat sink. B: what is the response rate of the large TC we use in ovens.

I started researching response rate and what I found was quite shocking, large TCs have very long response times. In the amount of and exceeding a min. This is a long time when you consider your pumping in X amount of heat energy into an insulated confined space. Most oven TC have leads that are up to 1/8" diamater. The response time on these are crazy. This could mean the TC is lagging by hundreds of degrees. In something that sits at temp for a long time like a ceramic kiln it would not matter but on something like an oven where we often bring the blade up to temp with the oven it could mean serious over heating.

Next I was thinking about the leads acting like a heat sink and drawing heat out of the junction. I don't know how much this would affect the actual accuracy but it can't help considering most oven TCs are already large and lagging behind the actual temp anyway. My thought is to try some good quality butt or bead wire TCs which have response times of 1-5sec plus thy have very thin leads which should draw less heat away from the junction. I have tried a cheep bead TC in my oven and it worked good as you could really see the jumps the oven was making. It was not a smooth increase like the larger TC would show. It caused the PID to shorten the pulse to the elaments which I think created a more even temp. I stopped using it because of its quality and I was questioning its accuracy.

So what do you think, am I off base or on the right track?

Here is a response time chart from omega for there thin wire TCs. If you notice in the discretion it says time it takes to reach 63.2% of instantaneous tempsturestemps change. So even thin TCs are not instant makes you wonder what the large TCs are doing.

 

[timgunn1962]

I looked into this a few years back when I built my first HT oven using mainly bits that I'd scrounged over several years working on landfill gas burners.

I had a 6mm diameter Mineral Insulated type N TC with insulated junction as my control thermocouple because it was free. I had a bunch of other thermocouples as well, of different diameters.

I was particularly interested in how a blade would be affected by radiated heat from the elements and tried to simulate the thinning profile of a blade by using progressively thinner thermocouples: 1.5mm MI, 0.5mm MI and a fine-guage thermocouple with an exposed junction.

I set up a datalogger and set the control to a (tempering) temperature of 250 degC (482 degF). I was using a high-end controller borrowed from work.



I let things cool, then ran a ramped warmup to the same temperature



Overshoot was massively reduced with the ramp and started me looking for affordable ramp/soak controllers.

I did a little more testing, though I've since lost the data in a hard-drive failure, before I settled on using a Mineral-Insulated type N thermocouple with a grounded junction: my reasoning being the MI construction and use of type N instead of type K "should" provide long-term stability. The grounded junction "should" provide reasonably fast response. I do not actually ground the thermocouple sheath as a rule, because I have seen an increase in signal noise when I do so.

I always give the oven a run up to near-maximum temperature before running the autotune: it gets an Oxide layer on there which changes the response to radiated heat from that of a new, shiny thermocouple.

The 1.5mm thermocouple in my original tests was already Oxidized. The 1.5mm and Exp were not.

Insertion depth is important. It's not easy to find a simple rule of thumb for our application, but at least ten diameters seems to be the recommendation for insertion in flowing gas lines. I tend to aim for at least 20 diameters for my ovens: quite easy with an MI thermocouple.

 

[Stacy E. Apelt - Bladesmith]

The problem with some charts is you have to compare apples to oranges. We aren't talking about an instantaneous temperature change as in JT's chart. That is for things like figuring how long you need to allow for a good reading when you stick a probe up a tail pipe or in a hot oil bath. In a HT oven, the probe changes temperature more or less at the same rate as the oven. The leads are no "sucking heat" from the tip. The tip may radiate heat back down the leads, but the reading is at the tip. Something that most folks can't grasp is that there is no such thing as cold ... only heat. Heat moves from the hottest place to a less hot place by convection. It doesn't matter if the less hot place is -100F, the heat moves at the rate of the materials coefficient of heat. "Cold" does not suck heat from hot.

Tim's charts show the apples to apples info, with the conclusion that once preheated and the chamber refractory is soaked, the regulation is pretty tight. During start-up, there can be overshooting and excessive radiation heating of the blade.

Moral of the story - preheat your oven before adding the blades, or ramp at a 50% rate (5000 instead of 9999).