kiln PWM parameter for PID

[coppercone2]

So the kiln I have has a electromechanical controller that is basically a 20 second duty cycle thing.

I.e. 25% means 5 seconds on 15 seconds off.

With my reflow oven, the PID basically applies continuous power to the elements until they get near the set point, then it starts pulsing them to reduce over shoot.

If I put a PID on it, do I need to maintain this duty cycle? The elements are embedded in some kind of ceramic.

Naturally the PID will enable the SSR continuously until it gets near the setpoint (say its 900C for brazing), then it will start reducing it. However, this means the elements will get continual power for quite a while.

The turning on and off means that they suffer more temperature changes, but on the other hand it means that there is more time for the heat to equalize or whatever inside of the oven.

I was planning on equipping it with a CN9000 series OMEGA PID controller with a zero crossing SSR (derated 50%)

[coppercone2]

basically I think I need to find a manaul for a manual kiln.


I wonder if the instructions tell you how to run the god damn thing. Like "limit power to 20% for 30 minutes before each use" or something. I feel like it probobly has some crap like that


anyone use muffle furnaces and know this from experience?

I don't have a sense of the time constants in there. Can the wire overheat and burn out before the thermocouple responds if its getting floored?

Or are they designed with the expectation that people will set it to 100% to get their lab work done faster? Then turn it down.

How was this done in the 70's?

[rogerggbr]

How was this done in the 70's?

Typical on/off controllers, whatever the method of control (P, PI, PID) will assume that continuous output is ok if far enough away from the set point and measured value moving slow enough. The ones we have on our ovens also have a high level controller, sometimes called a supervisor or similar, in case the main contacts get welded. I assume a modern solid state system can also fail short circuit so a high level trip/lockout may still be a good idea.

[coppercone2]

hmm so basically their designed to be OK at 100% so long its not near the limit?

Is this still true for embedded element heaters, where its a tile with wires?

Safety issue too, I made sure to get the smallest one, is only 9 amps. Not sure what kind of safety I can put on a SSR. Putting a contactor in there to act as safety is maybe possible but it might be getting cramped.

The old one was busted so it looks like when you over heat this type, it just melts the element and you get a drop of molten metal inside of the furnace.


I think with that controller I can wire output #2 to be an alarm if its connected to a normally closed latch open relay at a over temperature event. Then it should disconnect the heater lol

[Doctorandus_P]

There are (legal!) limits to on/off cycling times of big current consumers. It probably does not matter for a DIY smt oven, but for commercial designs you have to make sure you don't switch too often. The PWM frequency does not matter much as long as it is about 4x to 10x faster then the time constant of the oven itself. But of course it depends also on how accurate you need to track the temperature.

For the rest. If the oven is cold and you want it to get hot, it takes a while to get "close" and in that time the integrator will wind up. The whole PID algorithm does not work properly in that case. And instead of generating a slow temperature rise that can be tracked by the oven and it's PID regulator, it's faster and simpler to just turn the oven on on full power until the temperature get close, and only then start the PID controller.

You can also:
1. Turn on the oven until it's hot.
2. Turn off the oven completely, and track the overshoot.
3. Subtract the overshoot temperature from the overshoot.
4. In a second test run: Turn off the oven when this setpoint is reached.
5. Start the PID algorithm at maximum temperature of the overshoot peak.

When done correctly, the peak of this "overshoot", now coincides with the final temperature setpoint. You can even preload the integrator with a value that is close to steady state operation. A bit of guesswork is OK, as (when adjusted correctly) the PID algorithm will adjust for the rest.

[coppercone2]

i was worried that the elements will get melted in the while its running at 100% power. The burried elements are very funny to me. the walls just start glowing red

the method you give looks like the method for manual PID tuning in the manual for the CN9100A

the PID has a thing where you can limit it maximum power to say 90% if you want. But it disables auto tune. They say its when the heater is too big. In this case its just unlike everything because the heater is encased totally in the wall, no element groove


because my measurement shows me, at 80% power, I am at 1000 degrees of the 1100C limit. I assume if I ran it continuously at 100% power it would burn out.



If there was say a big block of copper in there, the furnace would be OK trying to get to setpoint at 100%? Say I want to braze a fitting on a ... copper heat exchanger of some sort. And the setpoint was maybe 875. Or to braze some feet on a sizable copper heat sink. The main idea is to use less acetylene for dan gelbert style prototyping

would I ever wanna limit maximum power to elements under any load condition? I know to keep the load from the walls etc by some recommended ratio

[coppercone2]

Well so much for trying to run it without a PID lol. it burned out at 80% power. For some reason the temperature sky rocketed to 1300 + within the hour. Maybe the coating, because I used the same settings before and it would not go past 1100! it melted the thermocouple. It got totally glassed. :'(

I think its because I was curing the coating on my little tray and it was too close to the walls.

well that was a painful expense. I think I will just use a thin alumina sheet on the bottom to keep it clean

I had already bought the PID, otherwise I would have thrown it into the trash.



I see why people use open element kilns, because they are cheap to fix. this one kicks you in the nuts on the 4th of july


I am gonna limit power on this thing too in the PID algorithm



tldr : don't try to wing it with unrecommended oven loading parameters

[coppercone2]

alright I got a 25A Panasonic SSR from digikey to use with this and a cheaper heatsink on ebay (the skeleton ones but with a fairly thick back bone). The nice hs have 100mm foot print, I only have 100mm to spare total

The thing is rated at 9A so I think it should be good enough

I made the front panel out of 12 gauge galvanized steel that I plasma cut. I am going to paint this with the flame proof paint like the rest of the kiln.

I am going to make a heat shield that goes between the SSR and the kiln bottom out of this stainless steel high nickel content sheet I have (bearly bubbles in contact with H2SO4 and HCl mixture after a long time).  its shiny, fairly easy to bend and corrosion resistant. I am going to make a L shape that get sandwiched between the front panel and PID about 1/2 an inch from the top of the kiln and 1/2 inch from top of the PID module.



The only thing  I don't like is that the PID module has fast-on tabs for the thermocouple wires. For screw terminals I usually get the proper lugs made out of the correct metal. I can't do that in this case and I need to use regular fast on connectors.

It looks like the PID module has a tube socket on the back with a adapter into the fast-on blades, so I guess it won't be super accurate anyway even if I did get type K fast-on connectors.

[coppercone2]

for a heat shield, I can either make a sheet metal cut out that basically floats above all the electronics or basically a roof that covers everything with a gap.

Does it work better for convection if its just a line of sight thing between the components and the hot bottom of the kiln or if its a entire layer?

Or should it be a layer with preforations?


They have a partial heat shield in there already over the gauge, it only covers a bit of it and extends 2 inches, but its only a few mm away from the kiln floor. I wanted to attach my heat shield to their partial heat shield so I don't have to manufacture another bracket.

It seems that since the channel is narrow, perhaps just doing a line of sight thing would be cooler for the components then making essentially a whole layer that has ventilation on the edges?

I.e. I was going to extend 1 plate over the PID and another plate over the SSR heat sink a few mm away from the top by extending the original shield with some aluminum strips, or I can cut a big square of aluminum and basically make a whole floor. I am worried that the hot air might bunch up in this tight space and it won't cool the PID and heat sink as much as just a vertical direct line of sight only thing from the floor.


But using a whole sheet is appealing because there is two mounting bolts that got replaced with a ceramic terminal block. I could bolt a sheet to this and have it float under the kiln box floor by the distance of 1 or 2 nuts and some washers. I would only need to cut wires through for the element leads and the mounting holes. Then I can lose the stupid segment of heat shield the kiln came with.

[Poroit]

The gauge of thermocouple wire used and its position in the kiln will affect the response time.

Overshooting to 1300°C with the controller set to 1100°C sounds like a parameter programming issue with the Omega Controller.

I have successfully upgraded numerous kilns by fitting an SSR and a good quality controller with SSR output.

What are the current and voltage ratings of your kiln?

[coppercone2]

oh no it over shot and destroyed itself set to 85% on the mechanical timer regulator (85% duty cycle of 10 seconds).

Because I swore before I left it at 80% and it was maybe 1000 after 2 hours. When I set it to like 86% the next day, 2 hours later it was 1300. Maybe there was still moisture that got baked out and it suddenly got more efficient. . I just wanted to cure a base plate I made. I guessed that it might get to 1050-1100. BUt I came back and the TC was reading 1300.

its like 9A 120VAC


The good part is I know how it fails if you let it go or it malfunctions, basically becomes totally inert. I was worried that the exterior walls might start glowing and it could be a molten sheet metal dripping fire hazard. It appears so long its put in a safe spot, it won't do anything crazy even if it totally malfunctions. For me, it seems like a really scary device. I feel alot less scared of it now that I saw it fail

[Poroit]

Sorry...I misread that as I thought you had installed the Omega Controller.

I suggest you push on and install a PID Controller and SSR as it will greatly improve the performance of your furnace.

I have been involved with the instrumentation in foundries who also use the brazing technique and they normally use a Ramp/ Soak controller for this application.
Brands I have used are RKC, Shimaden, Fuji, Brainchild, Eurotherm etc.

You need one that has the following features;
Type K Thermocouple Input
3-32V DC SSR Output
Auto Tuning

Your 25A Solid State Relay will be fine but must have the compatible 3-32 VDC Control Input.

A couple of questions re your Muffle Furnace.
Does it have a thermal fuse fitted?
How hot does the existing controller enclosure get? Can you touch it with your hand?

[coppercone2]

nah it would burn you it got to like 100C. But I built everything so it floats off the perforated plate walls in the under body on brackets and it looks like every component will have a heat shield on it.

I have no idea how to install a thermal fuse, it looks like it does not get hot enough to ignite things even if its on the verge of burning out after operating for a long time. I am paranoid about it so its only probobly going to be used on the steel welding table and that area is kept clean from scraps and trash because sparks cause fires

I thought about making a  secondary mesh shell around it eventually so its touch safe


WHat is the point of  a thermal fuse if the thing just burns out without anything happening as I have seen? It will probobly be unstable to try to fuse the chassis with heat because of air temperature changes. I could glue something to the bottom... but it looks like its safe even if you let it burn up. The elements did not even drip nichrome into the furnace body, when I took it out, it was all contained.. the elements fuse literary like a fuse (look exactly the same from the outside)


I think I just need to live with the fact that if the element wants to burn out, it burns out (assuming there is some kind of malfunction with the SSR or controller).


I could see if the elements were exposed coils some kind of dangerous condition could happen if they start drooping in the thing to short out and explode, but in this case its embedded in the manufacturers fire brick material, so they can't move at all... it just melts in side of the refractory and opens circuit. And its only 1 circuit, that is a single monolithic element, so it either works or does not work., if you have multiple elements the behavior might get funny if one burns out and the other one makes it drip or something


And a thermocouple just measuring air in the enclosure where the PID would be is ~30C in a fairly hot weather (maybe 28C)

I just thought the ceiling might irradiate the enclosure since its black, so it will get a aluminum umbrella

[Whales]

Sorry to hear you melted your element, that sounds expensive.

The performance difference between your two runs might have been related to the gaps in the kiln and how well sealed they were, including the peephole.

[coppercone2]

I did have some coatings in there, ITC296. Maybe its highly non linear where it really kicks into gear at 1000C reflecting alot of heat. Or it finally cured totally

again, slightly sad, but now im fully confident its pretty damn safe, I was border line scared of it before I knew how exactly it fails. I thought the whole thing might start glowing red dripping molten metal and spewing sparks and smoke lol, my imagination was off by maybe 100kW of power. And its power cord was fine too, and the outlet was not really hot or anything like I thought it might be. Its a small price to pay for peace of mind IMO

[coppercone2]

this one does not have a door seal. It has a flange with a brick frame that the door goes inside of, like a labarnyth path.


I see I can buy ceramic fiber rope material to make a gasket seal between the door brick and the bricks making the exterior heath frame


How much efficiency does it gain if I did that? These mods are getting a bit pricey. I did try to jam kao-wool and put ICT on all the gaps to seal it up as much as possible because things were getting a little loose in there, but the seal needs to be woven to maintain integrity, felt will make a mess there

Think it would be more then 20 watts saving?


also, what glue would you recommend for the door seal adhering ceramic fiber material to fire brick material if its worth doing?

I do see a red glow through it if its hot and closed around the ring like a light in closed room in a dark hallway deal. I assume it might be substantial if I can see it glowing through the gap


ALso I was thinking lol, if the front and back are coated with IR reflector, is the kiln not sort of a laser, with a aperture instead of a silvered mirror? A low q thermal laser? It makes me think of lasers. Its just too big

[coppercone2]

I think I found a case for using the thermal fuse. If your using this for chemistry I can see possibly a crucible full of molten metal rupturing and flooding the bottom of the kiln and turning it into a short circuit. Not sure what exactly would happen but maybe it would allow it to get hotter then normal. Then even if the monolithic heater fuses some where, the pool on the bottom might conduct electricity. I would expect it would blow the fuse unless you get very unlucky. if molten metal shunted the nichrome in some bad spot maybe it could cause trouble that could result in the chassis going bonkers. I would think it would trip the GFCI.

Maybe I will glue that on. Not really something I can test, it seems like it would be one of those things you either got a contingency or it goes nuclear

I wonder if this case of flooding the circuitry with molten metal is a concern for liquid metal reactors or a leaking ladle in a foundry. thats some mega bullshit. The circuit box is full of molten steel again

I think what I can do is make a leaf spring that presses a thermal fuse into the bottom of the chassis using the old mounting studs for the nichrome wire that I replaced with a ceramic terminal block. It should fit between them and get pressed in by a M shaped spring thing for good thermal contact

Maybe put that at 175C? It should never activate unless things have gone seriously foul.

[Poroit]

The furnace may have reached 1300°C because the thermocouple was already open circuit.
If the thermocouple open circuits, the furnace keeps heating trying to reach to setpoint.

A Thermal fuse looked like another thermocouple and was normally housed in a Ceramic sheath protruding into the workspace.
They were made from a precious metal combination. It would burnt out at a higher temperature than the workspace rating, turn off the power and protect the furnace from thermal runaway.

Modern Furnaces/Kilns don't use this method and use the thermocouple break feature of the PID Controller and a second contact output to turn off the power at higher temperatures.
Some also have a second controller and thermocouple for this belt and braces approach.

[coppercone2]

no when i burned it it was just running at 85% of its duty cycle in 10 or 20 second period with a relay oscillator with no feedback. I thought I could get away with it. I really wanted to make a ceramic PCB on the 4th of july, and I needed to cure a little kiln tray in a hurry

i was very enthusiastic about the project, because basically it was totally fine before running for a few hours at 80% the other day. its just non linear.

you gotta totally watch it if you wanna go hot with out feed back control. not worth doing

i previously had alot of failure with a microwave kiln attempt and I thought I could finally knock a trace out lol

[coppercone2]

well I managed to crack the bottom protective plate, meaning either I spend 3 hours with a die grinder and saws trying to recut it, so i just fixed it with furnace cement. i hope that holds.

I like the bait and switch tactics of laboratory materials suppliers too. its supposed to be america but it works like alibaba

no wonder R&D is being cut everywhere, when they keep getting robbed on their equipment budget for repairs

[johansen]

can't really help you when you've not included any information on who made this kiln, or where you got the heating elements, what modifications you made, etc.

[coppercone2]

i got plenty of help already

[coppercone2]

glad i started to dig into that old cable. I thought I could reuse it but it suffered from brittle insulator, which turns into maccaroni fragments. effectively, it was just separated by wool/fiberglass in the cord. Always check the cord out. IMO those rubber cords offer a false sense of security because its packed with soft stuff and rubber, so it feels good on the outside ,but if you look in there its often totally compromised

[coppercone2]

Does anyone know what the error means in the CN9000A

EE5 - Outside Time Limit


Its when I try to run a autotune to 850C.
https://assets.omega.com/manuals/M1191.pdf

Not sure what its talking about exactly

Something with the integral? I think someone said something to me before about the I being problematic with kilns?

[Poroit]

EE5 means Auto Tune outside of time limits.

What were your PID values after you ran the Auto Tune?