Heat Treat Oven Build + Kanthal Wire Calculator

[Dick Sexstone]

In the world of salt baths a capped off stainless tube is used to protect with thermocouple wire….would that work in a Kiln?

 

I have a commercial kiln that uses the bare wire so maybe a stainless sheath to protect the thermo couple wire will not work in traditional kiln? 

[Brandon Bearden]

Whew. I can't believe how long it has been. I had some life circumstances that got in the way, job change, family, personal...

 

Anyway, I finally got back to it today. 

 

I made the metal base, I cut out the CFB base and then cut the walls into two pieces and took out 2" to put in support for the shelf. I also assembled the first leg of the walls. I still have some more sealing to do, but I ran out of time. Here are some pics of the progress.

 

Next step is to finish sealing it, run the wire on the first half and then "glue" in the shelf supports and the top section of the walls and finish out the wire lay.

 

 

 

 

 

 

 

 

 

 

 

 

 

[Brandon Bearden]

Well I fought the wire into the base today and cut the pieces to hold the shelf. I just set the top section on for now. I will glue later this week. Here are some updated pics. I don't think the wire will need any staples. There is a LOT of tension in the wire and it pushes out on its own. It was rough to put in. It is very stiff and unwieldy, especially alone. (550 some odd inches of coil) I still have some futzing to do to get it to sit right, thought I am not that worried that it goes in every crevice. It can go a couple of inches unsupported. Though I will certainly try! I am actually a little worried to take the strap off. I will have to think about some removable temporary bracing before the side panels go on. Anyway, here are some pics. 

 

 

 

 

 

[Dan Hertzson]

Over time your coils will sag from heating.  All that "spring" will go out of them and you will need something to hold them on the shelf you have formed.  I usually cut my grooves at an angle, so gravity holds the coils in place.  I would pin each coil in at least 3 places on yours.  Needless to say, make sure your pins don't go thru to the kiln metal jacket.

[Brandon Bearden]

12 hours ago, Dan Hertzson said:

Over time your coils will sag from heating.  All that "spring" will go out of them and you will need something to hold them on the shelf you have formed.  I usually cut my grooves at an angle, so gravity holds the coils in place.  I would pin each coil in at least 3 places on yours.  Needless to say, make sure your pins don't go thru to the kiln metal jacket.

 

Experience gained is directly proportional to equipment destroyed. Thanks for the tip. I should have thought the wire would relax through heat cycles. I will pin it. This entire thing gets wrapped in 2" of ceramic blanket. In your experience, how much heat will escape though the walls at that point? I was going to add another layer of blanket between the wall and the control box but would like to skip that if it is not necessary. I was also thinking of not gluing the baffle shelf and top section together so I could more easily reach the wire if needed in the future, not sure how leaky it would become then though.

[Dan Hertzson]

It all depends on the relative temperatures of the interior and exterior of the kiln, the exterior surface area, and the thermal resistance of the assembled insulation (to a smaller degree the relative airflow around the equipment which carries the heat away convectively).  Fortunately the resistance of the two layers is additive, so the formula is pretty straightforward.  It appears that you will have essentially two layers of insulation: 2" of "light" insulating brick" and 2" of ceramic blanket.  In it's simplest form the equation is Q=U x A x (Ti-To)  where Q= heat loss over time, A= average surface area, U= overall rate of conduction in appropriate units, and (Ti-To)= differential temperature.  There are likely calculators online to help you with this.  My gut says that for typical heat treatment ranges you won't exceed a outer skin temperature of 350 deg. F.

[Alan Longmire]

I think that's enough insulation.  The top may get hot, but the sides probably won't.  My reference point here is my Evenheat HT kiln.  When I do AEB-L folder springs it will do the austenitizing cycle of go to 1650 and hold ten minutes, then ramp up to 1925 and hold ten minutes, quench, then temper at 1135 for two hours without the sides getting warm at all.  The top does, of course.  It has no kaowool, it's just two inches of insulating brick.  After the temper cycle as the kiln slowly cools, the upper sides may get warm enough that I don't want to leave my hand on it for more than a few seconds, but the lower sides and bottom remain cool to the touch.  I imagine if I did a full process anneal that takes 24 - 36 hours it would get warmer, but I don't think I'll be doing that.

 

Makes me feel silly that I built an aluminum flashing heat shield around its spot on the bench, but better safe than sorry.

[Will Robertson]

On 10/8/2022 at 12:42 AM, Brandon Bearden said:

am going to use 2" thick ceramic fiber board

Ceramic fiber board is available in two variations - with organic binder and without organic binder - the sort will organic binder will produce a lot of evil smelling smoke the first time you heat it up.

Ceramic fiber board can dissolve like candy floss in contact with fluxes at high temperature - I don't think this should be a problem at the temperatures in a heat treatment oven but bear it in mind if you think about taking the oven to a higher temperature - I think the refractory mortar may reduce the risk of this happening significantly though (depending on the chemical interaction between the two - best to confirm compatibility with the suppliers).

There's also the option of using light insulating refractory bricks - they can be significantly easier to work with than ceramic fiber board and cheaper.

For Kanthal (FeCrAl) heating elements I bought the heating elements as ready-made coils for the power output and supply voltage that I wanted and connected as many of these as needed (star, delta or parallel depending on the power supply and coils). The coils need to be stretched to the length you want then heated to anneal them in the new length.

It's often good to have an independent second temperature sensor connected to a system that can shut off the power or raise an alarm if the temperature goes too high - if anything goes wrong (for example nearby lightening strike messing up the electronics, semiconductor failure in the case of SSRs or welded shut contacts in the case of electromechanical relays) and a furnace like this overheats the consequences can be horrible.

You're right about the advantages of the fast switching speeds of SSRs but the semiconductors in SSRs do sometimes - rarely nowadays - fail - sometimes through no fault of the SSR designer - for example because of an unexpected high voltage surge on the incoming mains or a serious fault somewhere else injecting a DC bias into the AC mains. AC SSRs are usually and SCR / TRIAC and optoisolator inside.

The interior "hot face" of some furnaces is intentionally built from refractory materials with a higher thermal conductivity than the outer layers to help give a more even temperature inside the furnace.
 

On 10/8/2022 at 12:51 AM, Joël Mercier said:

12000W means around 55A at 220V. You might want to rethink that...

 

On 10/8/2022 at 1:42 AM, Alan Longmire said:

True.  Evenheat kilns max out at 20A 250V, something may be off.  

 

On 10/8/2022 at 1:48 AM, Brian Dougherty said:

Buy yeah, 12kW is off the charts.

I agree with Joël, Alan and Brain - 12kW seems like a light industrial power supply rather than a domestic power supply. If the supply can provide enough current the snag with pushing this much power into a furnace is that it will heat up fast but it may not reach thermal equilibrium - heating it up at this speed it will reach target temperature very fast but there may be significant temperature differences inside it when it reaches target temperature that would take time to even out.

Using an SCR / TRIAC to reduce the power it would be necessary to run through phase angle calculations -
if an SCR / TRIAC controller reduces the power output to 85 % of the original the current consumed will be significantly more than 85 % of the original (would need to go into AC theory to fully explain this).

A final thing about FeCrAl heating elements is that a reducing atmosphere will damage the oxide layer on their surface that protects them and significantly shorten their life so that tends to be best avoided for long periods.

The best digital PID controllers have a self-tuning feature that may be a significant help.

Not sure if that's any help or not...

[Will Robertson]

On 10/25/2022 at 9:54 PM, Alan Longmire said:

but because it acts as thermal mass to even out fluctuations once it's up to heat.

Yes - often the inner refractory has a significantly higher thermal conductivity than the secondary and outer refractories (3 layer refractory design) to help even out internal temperature differences in the furnace faster and to conduct heat away from the FeCrAl heating elements and into the rest of the furnace to help extend the lifetime of the heating elements.

[Will Robertson]

On 11/25/2022 at 3:54 PM, Dick Sexstone said:

In the world of salt baths a capped off stainless tube is used to protect with thermocouple wire….would that work in a Kiln?

 

I have a commercial kiln that uses the bare wire so maybe a stainless sheath to protect the thermo couple wire will not work in traditional kiln? 

Typically, thermocouples are protected with a stainless steel sheath, and alumina sheath or a zirconium dioxide "zirconia" sheath depending on the temperature range intended.

Exposed thermocouples can be vulnerable to mechanical damage. The stainless steel sheath protects the thermocouple and the connecting wires from electrical noise but can't withstand as high temperatures as alumina. Zirconium dioxide "zirconia" is only used for specialist high temperature applications.

[Will Robertson]

17 hours ago, Dan Hertzson said:

Over time your coils will sag from heating.  All that "spring" will go out of them and you will need something to hold them on the shelf you have formed.  I usually cut my grooves at an angle, so gravity holds the coils in place.  I would pin each coil in at least 3 places on yours.  Needless to say, make sure your pins don't go thru to the kiln metal jacket.

Another trick that I've used to mitigate this is to avoid running the FeCrAl elements at full rated power - running them at lower than rated power typically reduces sagging and extends element life.

[Will Robertson]

On 10/14/2022 at 10:00 PM, Brandon Bearden said:

Knife Engineering

That's an interesting sounding book that I didn't know about!

What are folks' thoughts on it?

[Will Robertson]

On 10/14/2022 at 10:33 PM, Alan Longmire said:

My Evenheat will do 2200, but they warn against it.  Apparently it significantly shortens the life of the elements to go much over 2100, but I also don't think they used the fancy resistance wire you have.

 

@Alan Longmire Your furnace will use FeCrAl (trademark "Kanthal") heating elements as well - operating them at over 2100 °F is possible but - like you say - will significantly shorten their lives. Some borosilicate glass melters use FeCrAl heating elements in their furnaces - the FeCrAl elements can reach those temperatures but have a short lifetime.

(FeCrAl elements can withstand much higher temperatures than the NiCr heating elements in domestic electrical appliances and low temperature ovens.)

For long-term use above about 950 °C or 1740 °F SiC (Silicon Carbide) or "Molly" molybdenum disilicide heating elements are usually used.
 

On 10/14/2022 at 10:33 PM, Alan Longmire said:

Speaking of which, have you thought about adding an argon purge option?

Yup - I reckon an argon purge sounds better than burning gas to remove oxygen - FeCrAl elements at these kind of temperatures rely on forming a protective oxide layer - if burning gas to consume oxygen accidentally creates a reducing environment that will reduce and destroy the protective oxide layer on the elements and significantly reduces the lifetime of the elements. This can cause problems for potters if they repeatedly do reducing firings in their FeCrAl kilns for example.

In research we had stuff called "reducing gas" which was mostly nitrogen with a little hydrogen - the concentration of hydrogen was low enough that it could never form an inflammable or explosive mixture with air - no idea where to get it though - some specialist gas companies may do it.

 

[Brandon Bearden]

Whew, lots of good stuff. 

 

Again, experience gained is directly proportional to equipment destroyed!  Hopefully, it doesn't come to that, but I am aware this may fail to work as designed. If so, I will try to salvage what I can and go back to the drawing board.

 

I am not as patient as @Will Robertson with all the awesome quotes, so here is a vomit of response.

1. I am using A1 wire, not D. This should extend the life considerably for low use application. 
2. I didn't think about a thermal sink in the kiln with different materials. Would placing a hunk of steel in the bottom serve a similar function for temp stabilization? 
3. The CFB I purchased was from Cera Materials. It is the high density version.
4. The controller is an Omron E5CC-T programable model, it has an auto-tuning feature and is very robust from what I can tell. 
5. I have a wiring diagram somewhere in this thread, but there is no dedicated power supply. I am hooking into 240 mains off a 60 amp breaker. The control box will have two "rails" of power. Not literal like a breaker box but hooked into terminations and breakers before feeding out to the SSR. Basically, I am building my own power supply. It will have a sticker on it that says high voltage don't be dumb, frame grounding etc. Hopefully I did not misunderstand the EE principles involved here to make this work. 
6. I like the idea of a secondary safety sensor/alarm. I will see what I can do.
7. At this point, I will just be stoked if it all works and then I can think about inert atmosphere - probably not even then heh.
8. Knife Engineering is an awesome book 10/10 must read, imho.
9. I don't understand the comment about flux as this will only be for heat treatment. What am I missing?
10. Running the elements at less than full load is an interesting idea. 

At this point, I am thinking of building the control box before sealing up with an enclosure to make sure everything works. Maybe I will even just grab a board and build the entire control box breadboard style just so I can hook everything up and was the sparks fly. Not bringing it to temp or anything, just watching the wire get hot. *Shrug*

 

I always seem to step off the deep end with my projects. Go big or go home or something like that.

Edited May 1 by Brandon Bearden

[Brandon Bearden]

Well, it works... I mean at least it powers on and everything works as expected. Obviously in this setup I didn't let things heat up beyond about 10 seconds, which even then, it was hot... fast. It is going to take a minute to figure out the software and the controller completely, but so far, what I see I like. Here are some pics.

 

 

 

 

 

[Brandon Bearden]

What is strange to me is that the relay has leakage voltage of about 60 volts AC. I don't understand why. When I measure the voltage while nothing is applied to turn it on, it measures about 60 volts. If I measure the other leg, with nothing, it reads about 35M ohms of resistance. I don't understand why this. Basically, it concerns me because the system is hot enough to shock all the time, or I am guessing it is, maybe the current is so low it doesn't matter - I haven't touched to test. Obviously, this won't do. This relay is specifically designed for resistance wire control heating. So I don't know what gives. I emailed the company to see what they say.

Edited May 6 by Brandon Bearden

[Brandon Bearden]

Well, shoot. The event input operations of the controller do not have a way to temporarily disable and re-enable the program control. Meaning, I don't have a built in way to shut off power to the element with a simple switch. Now, I have to think of the simplest way to interrupt the relay. Luckily, I can just wire this right through the switch I have already and it will work because it is rated for 12v/5a - stoked on that! I was worried I would have to source another part which is such a pita. It sucks though because I could have found a controller way easier if I didn't need the event inputs. There are some cool other uses I can think of though in the future, such as run, stop, or advance that I could hook buttons to later. 

 

Side note, learning temp PID terminology is a pain. I am so glad I have the Thermo CX software to use instead of having to cycle through 5 button menu. It is really nice to be able to see all of the options, and dang there are a lot. Anyway, this is starting to become exciting. 

[Joël Mercier]

Simply add a switch in series with the PID's power supply. That's what I did with mine.

[Brandon Bearden]

Yeah, I am going to interrupt the line to the relay. If I cut power to the controller then it resets the cycle. Better to just cut power to the element. Lower power to deal with switching too.

[Brandon Bearden]

So, I have realized that I have to wrap the CF blanket and build the enclosure to install the remaining top pieces. This thing is not easy to put together and I hate working with the CFB - nasty, messy stuff. I need the outside pressure to hold it together.

 

I bought a 4x8 sheet of 16 gauge steel yesterday. Now I need to get my plasma cutter setup and get to work. Not sure when I will be able to get to it. So close, yet so far. Hopefully by mid June I can be done with the build and start programming and tuning. I figure if I have a problem with hysteresis, I can always limit the current to say 85% of max load to help. Hopefully, I will not have to do that. 

[Brandon Bearden]

I was able to start cutting the sides today. I also got the first section of the blanket cut and put some ITC100 on the interior. Getting setup for this and layouts was a job. That little plasma cutter works great and was only $200 - chinisium brand. The walls are a chicken and egg problem. I am going to weld up three, put the blanket in and then will have to weld the other three. I will have to be careful with the heat then. Finally I can finish the assembly of the base and start on the lid and control box. I see light and the end of the tunnel.
 

Edited May 22 by Brandon Bearden

[Brandon Bearden]

I was able to get three sides on today. The next two are going to be rough. It’s going to be a neat trick to squeeze it all together.

 

 

[Brandon Bearden]

Well, my first hiccup, and a complete temporary hack for now. The coil wasn't long enough. The thing is just such a bear to work with, I am going to create a chase to feed it the rest of the way. I am about 4 hours away from having the entire base, including the control box done. Which means I am 10-20 hours away from having the entire base, including the control box done ;). Then, I can start on the lid. Here are some pics. 

 

[timgunn]

Have you gone out through the wall with a single strand of FeCrAl, or is it doubled and twisted? The top photo from May 6th "looks" like a doubled-and-twisted tail into the UBAL 50 terminal block, but between screen resolution, aging eyes and technological ineptitude, I can't be sure what I'm seeing in the bottom pic of the last post.

 

When I started using a local (UK) supplier for elements, the supplier "very strongly suggested" using nuts/bolts/washers for the Tails-to-Copper connections. I maintain stuff for a living and recognized the signs of someone who spends a lot of their time fixing stuff that only needs fixing because someone in the design office thinks they know better than the guys who have been doing it for decades. I've only ever used bolted connections with large-diameter stainless steel washers either side of the wires (bolt head, washer, Kanthal, washer, Copper, washer, nut). If you are using terminal blocks, please let us know how they hold up.

[Brandon Bearden]

@timgunn Yes, it is double twisted at the end. I will for sure let you know about the terminal blocks. It is an experiment that I hope works out. If not, ah well, I will bolt them. I was actually completely unsure how to go from FeCrAl to copper. I knew some sort of mechanical connection was in order, but not entirely sure. I have a suspicion the plastic will not hold up. I am going to keep a close eye during testing.