Heat Treat Oven Issues...help!

[nickandersonart]

Hey all, so I ran into a big issue with my heat treatment after building my own oven. Essentially, I think it's way off temperature. I know this is kind of a long post so please bear with me as I try to explain the gist of it :foot:

I realized that on some of my knives I was getting larger-than expected grain growth. With steels like 01 and 52100/Suj2 (The JIS equivalent of 52100) I didn't seem to have any problems with excessive grain growth, but other steels, when breaking off part of the tang and checking, the grain size would be larger than expected.

This time around, however, I did a batch of AEB-L steel, and one blade of CPM-154. It resulted in an epic fail, I lost 4 out of the 4 knives, and the CPM-154 blade literally fell apart in the oven. The AEB-L blades all warped big time, as well, even moreso after dry ice treatment, and after breaking off some of each blade, I found the grain size to be HUGE...so pretty much useless for these ones. I used foil for all of these, and proper HT protocols.

I decided that my oven must be way over temp, and decided I'd use my other controller with TC to cross-check the temps. I drilled another hole into the top of the oven and setup my other PID controller to measure temp simultaneously so that I could compare.

I built it with heavy duty resistance wire (13 gauge), components from Auberins, double SSR's and pretty much every "best practice" I could find (well, almost all...still room to upgrade But lately it's looking like my temp readings may be well under what the actual temp is inside the oven.

These are PID's and thermocouples from auberins:

The TC's are both the High temp model:
http://www.auberins.com/index.php?main_page=product_info&cPath=20_3&products_id=39

The PID's are (1) ramp/soak version and (1) simple set-point style that I'm using to cross-check temp:
http://www.auberins.com/index.php?main_page=product_info&cPath=1&products_id=4
http://www.auberins.com/index.php?main_page=product_info&cPath=1&products_id=3

So, upon heating and holding at 1500F, the results were DRASTICALLY off, around 70 degrees difference between the two controller units. I tried switching the thermocouples on the controllers, and again, there was a drastic temperature differential on the SAME controllers, which was really weird.

I also tried the salt melting calibration technique (salt should melt at 1474F) and it melted at around 1370F, meaning I could be off even by 100 degrees.

Essentially, I'm wondering what could possible going wrong here to where my temps could be so far off, and what I can do further to try to fix this. Could there be something I connected incorrectly that could cause everything to mostly work but with a temperature misreading? Is it possible that the TC wires could pick up some voltage across the wire that's affecting the readings?

This is one of the most important parts of my shop and of my whole process, and it's super important for me to have this equipment functioning as accurately as possible, so I'm willing to do whatever it takes to try to figure this out/remedy it. I've posted some pics below if that helps as well. Thank you everyone for any insight on what could be going on here and what I can do next to get this thing functioning as accurately as possible!

Stainless blades about to meet their fate

Massive grain growth

Showing the second TC installed in the oven

TC terminal block with brass connectors

TC plugs

Showing the massive temperature differential between TC measurements

View from the inside - looks like uneven heating, but it's the photo

 

[mete]

" fell apart "? If it really fell apart you were far above the temperature you should have been !!! Grain growth is much more sensitive to temperature than time !
Fell apart - you have been high enough temp to oxidize and perhaps even melt the grain boundaries .

 

[timgunn1962]

The first thing I'd look at is the thermocouple wiring.

The brass connectors are ringing alarm bells and it looks like you may need to read up on temperature sensing with thermocouples. Omega have some good information on their website and the Labfacility handbook covers things well too.

https://www.labfacility.com/media/pdf/temperature-handbook-1-to-4.pdf

If you have wired the thermocouple using the correct wire but have swapped the legs, you'll get an error in the reading that corresponds to twice the temperature difference between the controller and the connector block. I'm guessing this is your problem.

 

[Stacy E. Apelt - Bladesmith]

I'm with Tim. My first suggestion is to check that the wire is Type K wire and that the TC and the controller are all + to+ to + and - to - to -.

As mete said, massive grain growth isn't caused by being 70-100F too high for half an hour. Try 300-400F too high for an hour and you will get massive grain growth.

One reason I usually recommend purchasing a HT oven over making one is that the few bucks you save come with almost no guarantees or assurances of accuracy. You can be pretty sure the oven that you get from Paragon/Sugar Creek/Evenheat/etc. will be pretty close on temp an have the bugs worked out. If it isn't they will fix it or replace it free.

 

[nickandersonart]

Hey guys thanks a bunch for the suggestions! Yep Mete, I was really surprised as well - let's just say that the stainless foil had cracked open from the high temp and long soak (30 minutes) for the CPM154, and the surface looked just charred. I'll post a pic to show what that looked like, kinda crazy. And thanks Timgunn for the reference I'll read through that to see what I can find in regards to this oven.

Stacy, good advice and I totally would if they were available where I am but I've had to build most of my own equipment out here. This one is by far the most crucial of anything in my shop, so I'd really like to get this dialed as absolutely closely as possible if I can. Experiencing this makes me want to completely revamp the oven to make absolute sure I'm getting close to perfect readings from hereout.

My initial understanding with reverse wiring the thermocouple was simply that it would move in an inverse direction as the heat increased, and that you could just switch polarity if that was happening. I'll read more about this and check my wiring. I'm using stainless stand shielded TC wire from Auberins as well as connectors, though they did include some yellow insulated wire with the TC's that I should maybe use instead.

The brass fittings I actually just added in right before posting this, as I thought they'd help in case the TC wire wasn't making good enough contact with the TC terminal block. One thing, however - the TC terminal blocks also seem to have brass (or bronze) terminal connections as well with 4 holes - 2 for the TC legs and two for the wire connections, which I WAS using previously. Perhaps better to just have the wire touching the TC legs themselves.

I guess my questions would be these:

I'll start with checking polarity on all of the wiring. Would it be better to solder the TC wire directly to the TC? Or would the introduction of an extra "dissimilar" metal be grounds for another potential issue?

Also, what's the best way to calibrate one's oven for multiple different temperatures? I can use the salt method to calibrate at 1474F, and fine silver would get me 1762F, and then copper would do 1,984°F. I don't have tempilstix out here, so would that be a satisfactory method of calibrating?

Is there a way to use a multimeter/Ohm meter to check resistance for the TC to make sure it's what it should be?

I know it's a lot of questions, thanks for hearing me out. Once fixed, I'm going to keep two thermocouples in this oven from here out so that I can get the most accurate reading possible. Thanks again guys

 

[timgunn1962]

Reversing the thermocouple will indeed give readings that go down when the temperature goes up. It's usually a fairly easy error to spot.

The effect of getting the wires wrong is much more subtle and much harder to spot.

Oversimplifying only slightly, a thermocouple is just a pair of dissimilar metals in intimate contact. The junction of these metals will generate a Voltage that is dependent on the temperature of the junction. All pairs of dissimilar metals do this, but a small number of them do so in a way that is useful for measuring temperature. These have been standardized, with published compositions for the alloys used in each leg and tables of Voltage (milliVolts) against temperature. Most of them have also been assigned letters to make things easy.

The thermocouple Voltage is generated at the "Hot" junction and measured by the instrument which is at the "Cold" junction (though when measuring low temperatures, the "Hot" junction will be colder than the "Cold" junction. Is your head aching yet?).

Any wiring between the Hot and Cold junctions needs to be in the same alloy as the leg of the thermocouple it connects to. That way, we only have one Hot junction out in the circuit and the instrument reads the Voltage, looks it up in the table and displays the appropriate temperature. To make the circuit, the wires also need to be connected at the cold junction. The cold junction will also generate a Voltage that is dependent on the temperature of the junction. To compensate for this, an accurate temperature measuring system is incorporated into the controller (or pyrometer) and the Hot junction Voltage/Temperature becomes a known quantity.

If you use the correct wire, but connect the instrument to the thermocouple with the wires reversed, you will create 2 new junctions at the connection points. These will be the opposite way round to the Hot junction so will generate a voltage of the opposite polarity to the Hot junction. Back at the controller, the Voltage seen will be the Hot junction Voltage, minus the Voltage generated at the "extra" junction in leg A, minus the Voltage generated in the "extra" junction in leg B. The controller simply measures this Voltage, looks it up in the table and displays it.

The controller will then power the elements until the reading it sees is the one it has been told to control to.

In your case, it looks like the temperature of the connection block is around 35 degrees above the temperature of the controller, causing the oven to be 70 degrees hotter.

The specifications for thermocouples call out the allowable tolerances on composition for the alloys used in each leg. There will be variations in accuracy resulting from variations in composition, even within the allowable tolerances. For best repeatability, it is usual to specify "same melt" thermocouples in industry. If you are ordering a thermocouple, it may be a good idea to order several at the same time and from the same batch and keep them on the shelf. This will ensure that there is no change in composition when you next change the thermocouple.

Type K thermocouples are particularly prone to "drift" when used above about 1000 degC (1832 degF) and it may be a good idea to change them regularly, rather than wait for an obvious failure.

 

[nickandersonart]

TimGunn you are the man. Thanks for such a thorough description on how that works -- this stuff can really be quite a lot to take in but I do plan to build more ovens in the future so it really is great to know details like this. I'm actually going to take out all of the connection points (yellow male/female plugs etc) and just go for an uninterrupted connection between the PID and the thermocouple (will also have the bare TC wire making direct contact with the TC legs).

I did check, and according to the amazon seller that I bought from, my wire is indeed type K. Who knows how hot my TC actually got when the CPM-154 knife fell apart, this one hasn't been used more than 6 months, but I fear maybe it could have suffered a bit of damage. I may get another TC or two to make sure I'm starting with a blank slate.

For anyone else who has built their own ovens - I'd be really interested in what measures people would recommend for ensuring ongoing accuracy in their temp readings. Now that this happened, I kind of want to overbuild it so that there are fail-safes built in, or at the very least a cross-reference for reading temperature, in case one of the TC's has started to lose accuracy. I remember Kevin Cashen saying that he uses 2 TC's as well as a sort of "box" for blocking blades from radiant heat from the coils. I think I may choose to go for something like this, once I at least solve the immediate issue

 

[Drew Riley]

You might give Auberins a call. Used to be they had really good support for things like this. I'd also think about ordering a TC from them and direct connecting it without a plug. The plug shouldn't really make a difference, but it's just one more variable to eliminate.

Lastly, IIRC, there are a couple of parameters that you can change or calibrate on your PID. I recall having to do something with the "autotune" fuction to tighten up my setpoints. Initially I was getting a pretty wide temperature swing, but after autotuning, I hold within 1 degree.
Also, I believe there is a way to calibrate your PID using ice water. You'll have to read through your manual and see what the protocol is.

 

[Stacy E. Apelt - Bladesmith]

I wouldn't get rid of the TC plugs and connectors. They are polarized, and if the wires to them are right, it pretty much assures correct connection. Make sure all the other connections are correct. The TC is polarized, The PID is polarized, too. I highly recommend using a TC connector block on the oven where the TC comes out and the TC wires connect to it. Nothing will make a better joint than that.

 

[nickandersonart]

Yep i think that's good advice I'll give auberins a call. I know that these systems aren't exactly easy to setup initially but over time I've read quite a bit about setting up these temp controllers and built 4 of them now, and was at the point where they we're feeling fairly simple to setup--Perhaps I was a little over-optimistic. So far I haven't seen anything about people having drastic temperature misreadings like this, most of the issues I've seen are just readings that are a bit off.

I've used the auto tune and it really did help it to stay within a degree or so, but perhaps still at the wrong overall temp?

I'll go through the rest of the parameters too and make sure there isn't anything set weird, the only thing I think I messed with was the hysteresis but perhaps accidentally changed something else like Thermocouple type or something crucial. Probably worth just going for a reset actually.

Stacy I do really like having the plugs in there but I'm gonna get this figured out first without any additional connections and then can add them back in if needed once it's working as usual.

Thanks as always guys.

 

[Drew Riley]

Section 5.3.9 of the ramp/soak PID manuals mentions a "Pb" offset.

Have you tried sticking your TC in some ice water to see how close to 0C/32F it is?

 

[nickandersonart]

Hey Andrew I didn't try the ice method but just re-wired and tried with boiling water and it hung right between 211 and 212F. So that's at least a good sign for that! (Yay ) though, does that also mean it should be (fairly) accurate at high temperatures as well? I guess I should still do a salt-melt test to very this as well as the silver and copper melt test [emoji848]

Sent from my iPhone using Tapatalk

 

[nickandersonart]

I kinda want to go calibration crazy after losing that last batch of blades...that, and whatever other measures I can take to have this thing dialed in as much as possible


Sent from my iPhone using Tapatalk

 

[Drew Riley]

Hey Andrew I didn't try the ice method but just re-wired and tried with boiling water and it hung right between 211 and 212F. So that's at least a good sign for that! (Yay ) though, does that also mean it should be (fairly) accurate at high temperatures as well? I guess I should still do a salt-melt test to very this as well as the silver and copper melt test [emoji848]

Sent from my iPhone using Tapatalk

If I'm not mistaken, the millivolt to temperature conversion between TC and PID is non-linear, so any minor offset on a lower temperature would be amplified in the higher. For instance, if you put it in ice water and read only 30 degrees, that's a 2 degree offset at lower temperatures. However, that doesn't mean that at 1500 degrees you'll read 1498. The PID compensates for this non linear error with the previously mentioned Pb setting, BUT I think it's set up to work (or rather be calibrated) at 0C rather than 100C (32F vs 212F).

Then again, I may just be completely misunderstanding all of this.

The fact that you read somewhere close on the boiling water might be a good sign though...