jbruce

I'm the author of an open source kiln controller project. A recent feature request is Cone Firing Mode which most private companies offer in their controllers. I need some help in understanding cones and my mental model of them. I know very little about cones other than reading what manufacturers [like orton] make available for public consumption. Here's what I've done so far: I've used python to simulate a cone such that for a cone with the 3 rates and corresponding maturation temps as input, I can find the maturation temp for any heating rate [within reason 10C/hr through 300C/hr]. I use some math libraries to curve fit the three points. The fitted curve equation can be used to get the maturation temp for any rate. When I iterate over all the reasonable rates [like 15C/hr to 250C/hr] by 1 degree C for a given cone [like cone6]... for each rate, if I calculate the total heat work using a 2s cycle time [so that's 2 seconds X temp for that 2 secs] iterating every 2s from 1 hour in the past to maturation [to get the total heat work for those 1800 samples], I get a deviation of just a few percent across that for total heat work comparing different rates. This few percent equates to a minute or less actual firing time. As an example, here are two different heat rates for a cone6 cone... name,heating rate in C/hr, maturation temp, temp to start measuring, total heat work to bend cone (in degrees C X secs) cone6,30,1203.09,1173.09,4277147 cone6,89,1230.61,1141.61,4270090 So this data says when I heat at 30C/hr from 1173C to 1203C [cone fully bent], the cone6 cone absorbs the same energy as when I heat at 89C/hr from 1141C to 1230C [cone fully bent]. I can get these stats for any rate. The difference in total heat work (4277147 and 4270090) represents just a few seconds in kiln run time for these two examples. I think having total heat work numbers so close validates the model. Using this ability, I think I can examine a profile and estimate the spot in the profile when a specific cone would bend. A proper estimate requires only a constant heat rate during the last hour. Does all of this sound valid? or way off base?

@MikP if you ever suspect the pid controller is contributing to the issue [and I don't think so in this case], you can remove it and try to fire using manual controls to see if you can reach beyond 1080c.

@MikP there is also some software to graph a run https://github.com/jbruce12000/kiln-stats so you don't have to do it manually unless that's your jam. I use the graphs to help diagnose problems and help folks with tuning. You can see the sample is not tuned well.

Hi @MikP. I'm the author of this software. If you're still having trouble, lmk and I'll do what I can to help. It's best to open an issue on github at https://github.com/jbruce12000/kiln-controller/issues From what I've read, it looks like the elements need replacing. We can help with PID tuning after you get the heating problem sorted. For everyone else, this software acts like an on-off switch until it gets N degrees away from the set point, so in this case, since it is way under the set point, the controller is telling the elements to heat at 100% continuously. Once it is within the window, it uses a PID to control temps.

@Ju00Ls Hi. Glad you're following and glad you're no longer just a lurker. The metal box looks scary, but it is grounded. Aluminum boxes are common for electronics projects. In this case I chose it for its thermal properties and ability to quickly absorb lots of heat and transfer it from the SSR. The high voltage/high current cable is close to the thermocouple, but I'm not getting noise from that. The noise I'm seeing comes when a client connects to the server running on the PI. I believe that connections cause a spike in the processor which draws more current which impacts the 5v power supply connected to the PI. I think the problem is easily solved by using a larger PI power supply. This is a common PI problem. I have not proven this theory because the TC noise has zero impact on firing. The primary reason it has not had an impact is that the software reads the TC five times during it's duty cycle (2s by default) and uses the highest value. PID control is fascinating. Spend some time up front to tune it, and it stays amazingly close to the set point. All I keep thinking about is my wife's oven and the fact that it varies like 50F from the set point. To have such fine grained control and oversight of a kiln that a few months ago was just "Low, Medium, High" is awesome and has given me the ability to learn from mistakes faster by gathering lots of evidence with each firing. Having control over cooling of the kiln is nice too. I've never worked with crystalline glazes, but I'm already seeing small crystals and more character in the pieces I've glaze fired. Maybe I'll do crystalline glazes someday. It's also nice to be able to start a firing automatically on a schedule I choose using the api. I can also stop and restart at any point in a schedule which has been useful for skipping the last part of a soak you don't need [because the cone is at perfect 90]... or conversely going back to the beginning of a soak because it was about to start cooling and the cone wasn't quite perfect. I have a 19 lb Maine Coon named Snickers. What's yours?

Yeah @liambesaw, conservative in this regard. Mine is in my basement and I don't want to burn my house down. I guess the generally acceptable distances are: 3 ft from another kiln 12 inches from a concrete wall on a concrete floor

I think a cooled solid state relay will have long life. Wear and tear on SSRs is mainly from heating (expansion) / cooling (contraction) of the components in the package. The increasing amplitude of the waveform as your kiln gets hotter is due to cooling because of the 60s cycle time. A faster cycle time would be harder on your mechanical relays, but would yield more accurate ramps / soaks. WARNING !!! WARNING !!! WARNING !!! WARNING !!! WARNING !!! WARNING !!! WARNING !!! running a kiln in an enclosed space is extremely dangerous. some fumes are noxious. please dont. running a kiln within 8 feet of anything flammable is extremely dangerous. please dont. WARNING !!! WARNING !!! WARNING !!! WARNING !!! WARNING !!! WARNING !!! WARNING !!! This is the pot calling the kettle black (because I made some mistakes too), but the best thing you could do if you choose to run this in its current location & state is to automate a call to 911 when it starts. :-)

@High Bridge Pottery With a cycle time of 2s my controller switches on about 11,000 times in a 13 hour firing. I'm not sure what "a lot" means to you. I'm not worried about the number of switching cycles - as long as I keep my SSR cool - it will have a long life. If you're switching using a mechanical relay or contactor, I would set the cycle time to at least 10s. I'm not sure about 60s... I'd have to test that. By experience, my kiln drops a F degree per second when not being heated at 2000 degrees or more. Setting a cycle time of 60s means the temp swing could be 30-40 degrees or more. I want to be more precise than this. I think a 3 degree swing is acceptable.

@High Bridge Pottery Are you deciding if you should turn the relay off or on every 2 seconds? Yes. This is a configurable parameter. The code originally made this decision every .5s and that duty cycle was too high for kilns. I have heard some folks going as high as 10s. Since it is configurable, you decide the value that works best for your kiln. How did you come up with the PID values to use? I read many articles discussing PID tuning. Some were complex algorithms, some simple. I tried to find the simplest method that could be done in a few hours. The process I used is described here... https://github.com/jbruce12000/kiln-controller/blob/master/docs/pid_tuning.md The tuning it provided the first time around is good for my use. If you find a simpler, better tact, post a PR and I'll check it out. Part of our duties... @Min @neilestrick @Bill Kielb If you, Neil, or Bill feel like this has become too big a risk to remain on the site, I totally understand. The last thing I want is for anyone to violate local code or endanger themselves. I am committed to improving my implementation, but it will take time. I have a full time job and this is a side project / experiment for me. I cannot stress enough the value I have found in this forum. I thought I was pretty good at what I was doing. Turns out, I made lots of mistakes. If I made these mistakes, others will too. I think if folks had a resource where they could learn these things, it could prevent folks going down the same road I have. Not sure what else to say on this topic. Assumptions can be dangerous. Reflect carefully on risks others point out. Strive to become better.

@neilestrick The controller is a raspberry pi 3. It uses standard power supply which is 110V input, 5.1VDC 1A output. This connects to a micro usb port on pi. The power for the controller needs to be separate from kiln power. The controller is used to monitor the temperature even when the kiln is off. The power code for this currently dangles out of a gaping hole in the front of the bud box. This needs to be fixed. The load side of the relay is 48V-280V AC 50A. I believe I need another one of these to bring the controller up to snuff.

Please understand that this is just an interesting project / experiment. Please note that I have no intentions of creating a product. Please understand that this hardware is not finished and is in an alpha state. I do not want defend improper practices and I am here to learn from others and improve. I love that I have found a passionate, intelligent, experienced group of people in this forum. If I had a forum with information like this when I started the project, it would have changed the course of the project. So now, I'm left to make things right with folks so they don't follow the wrong path. Please help me do that. 1. Conductors on the input side need to be to code (could I use a replacement 50A dryer cable for this?) 2. Grommets for any conductor that goes into the bud box 3. connect the heatsink directly to the SSR (can I then mount the heatsink to the bud box?) 4. use an Ultra Fast Acting fuse at the rated amperage of the circuit to protect each of the SSRs - 50A 5. use an SSR on each leg, instead of just one 6. use mechanical safety relay to disable all SSRs in the event there is an SSR failure/short. I'm not at all sure how to implement this. Can someone post a link to a schematic? Did I leave anything out?

I want to make sure I answer all of the points. First question is about using extension cords. The conductors are 12 gauge and according to https://www.powerstream.com/Wire_Size.htm for short runs / chasis wiring, a single 12 gauge conductor can handle 41 amps. I am using three conductors for each connection. Experimentally with all elements running 100% of the time, the conductors are not getting warm. Even if it can only safely handle 60A, that is 20A above the max for the kiln and 10A above the breaker it is plugged into. So is using this wire dangerous, or just messy and something you'd never consider for a consumer product? Next point is about fuse protection for the SSR. What would you rate the fuse at? 50A? There is a 50A breaker box that the controller plugs into, so do I still need the fuse? Yes, if I was making a product for sale, I'd have one. Next point is about the heatsink being inverted and connected externally. Not sure about the inversion because when I received it, this is how it was oriented / mounted to the SSR. Yes, the heatsink is connected externally. There is thermal paste between the SSR, the bud box, and the bud box and the heatsink. My intent here was to have the bud box help in cooling. Experimentally when all elements are on 100% of the time for a long period of time, the heatsink is warm to the touch and the bud box is a little cooler than the heatsink. Next question, is there a constant 120V to the elements of the kiln? Yes there is. I have only a single SSR switching a single leg. I have a breaker box just behind my kiln and I shut that off before I open the kiln. This problem is easily solved by using 2 SSRs or a https://www.auberins.com/index.php?main_page=product_info&cPath=2_30&products_id=331 which is essentially 3 SSRs in one. When I first built this prototype I used two 30A SSRs. They were not zero crossing SSRs and so they did not switch at the same time. This caused an inrush current through one of the SSRs that exceeded it's rating and POOF. That is when I switched to using a single SSR with a 50A rating. At the time I could find no resources to definitively tell me how to best handle switching. I'm open to suggestions - especially when it comes to safety. I am open to modifying docs so folks who choose to build do so safely.

yes, ground.

@liambesaw Here is the pic I promised. Input is a nema male plug with three conductors (2 hot, 1 neutral). I used an old extension cord for this to make sure it could handle the current. Turns out it could handle about 4 times the 40A current max, so lots of safety margin there (and free wire). One hot leg is switched by the SSR, everything else runs straight to the output on the right hand side where the female nema plug is mounted. The RPI is in a plastic case and is covered by a breadboard with all the spaghetti wiring. You can see the max31855 in there. The thermocouple is the red/yellow wire. The rpi connects to the control side of the ssr. There is also power for the rpi (that black cable coming out the square hole in the front). That big square hole is from a previous incarnation of the controller that used a PID controller purchased from a company. Note the BIG heatsink with thermal paste and the BIG aluminum box. If you want your SSR to survive, keep it cool. There will be times where the SSR is on 100% of the time and that will generate significant heat.

Thanks Bill. I changed my schedule from 108F/hour to 120F/hour for that final 250F approach. We'll see if that changes the heat work a little. I expect that change will get me part of the way, but probably still a little high somewhere between cone 7 & 6.

I emptied the kiln this morning after the first glaze firing and everything looked good even though the kiln was over-fired by a cone. The witness cones showed that the kiln reached cone 7... just like the kiln sitter cone did. The maximum temperature according to the thermocouple was 2221 and cone 7 is 2262 at my firing rate of 108F/hour. So I guess the thermocouple is just off by 41F. At the same time, I'm reluctant to believe this because at cone 05, the cone bends right when it should. Maybe someone with more experience with thermocouples can provide advice. It's a K type thermocouple and I know I have the right polarity on connecting wires and the thermocouple itself. I'm using NIST linearization for the greatest accuracy. I cannot think of a reason it would be off.

Yep. It's a small kiln, so I just have one set of witness cones. In this case, 5,6,7 on the center shelf. In past firings, I've watched these to determine when to shut down (or move on to the next part of the schedule). This time I did not get the chance because the kiln sitter flipped before I thought it would. Next time.

The cone 6 glaze firing went pretty well. You can see in the image that my kiln sitter tripped just before reaching 2232. It had a cone 7 kiln sitter cone in it. I think I had a kiln shelf too close to it. No worries, I just by-passed the sitter with slow-flip and used the api to start the cooling phase. You can see I skipped the initial drying phase of this schedule. There is more error this time than previous runs because I included the initial warm up, the part of the schedule where my kiln could not keep up, and the kiln-sitter flip. It was usually within a degree F though. There were points in the schedule where the elements were on 100% of the time. schedule name cone-6-long-glaze schedule date 1/5/2019 average error in degrees F 3.25 solid state relay cycles 11934 schedule length in hours 11:25:12 elements on (s) 6:34:00 element percent on 57.50 element watts 9640 cost per kwh 0.126 schedule total cost $7.98

I'm excited! I'm in the middle of my first cone 6 firing using the rpi controller. I'll post stats when it's done. My kiln could not keep up with the aggressive schedule (345 F/hour) from 1880 to 1976, but it was pretty close. After 1976 it slows to 108F/hour and it caught up quickly. @liambesaw I have not forgotten I promised pics. I'll get 'em posted for ya.

ok, I'll open up the bud box I have and send you a pic. I have to go to dinner first, but sometime soon. basically it's a 220v male connector on one side, a female 220v connector on the other. an ssr is mounted to the box with some thermal paste, and the rpi is plopped in there with a plastic case and a breadboard on top. I have not etched a circuit board for this and might not do so. The bud box plugs into the wall, the kiln plugs into the bud box. simple.

@liambesaw I think that depends on how fast you want to cool. If you want to cool at a rate faster than the natural cooling of your kiln, that is not supported today. If you want to cool slowly , which I am guessing you want for crystal formation, then yes. Here is my cone6 schedule which cools at 436 F/hour until 1832 and then cools more slowly at 130F/hour to 1400 to maximize crystal formation. So what happens with the PID controller is that it just tries to maintain the current set point. It doesn't matter (too much) whether a ramp is going up or down or is static. If you REALLY want to cool super fast, the original software supported this, I just removed it from the code. Reflow ovens use a fan to cool faster than the original toaster oven could cool. This is probably not what you want.

Happy Holidays Potters! I did another bisque fire today and got very consistent results when compared to the last bisque firing. I started this firing at 2AM using the new API and it ended just before 3pm. I also used a new API feature that allows me to start a kiln run anywhere in a schedule. With this run, I skipped the first two hours because I knew my pots were bone dry. I added this feature for power outages, but it is handy for this too. schedule name cone-05-long-bisque schedule date 12/28/2018 average error in degrees F 0.78 solid state relay cycles 11086 schedule length in hours 15.17 elements on (s) 21955.36 element percent on 47.78 element watts 9640 cost per kwh 0.126 schedule total cost $7.41 The capacitors I added did not improve thermocouple noise or thermocouple errors. I discovered that thermocouple noise is caused by me when I connect a device via wifi. It is either causing noise from the wifi transmission being picked up by the thermocouple cable... or it could be power supply noise. I'm not sure, but it does not affect firing at all, so I'm going to ignore it until it becomes a problem. Next is a cone 6 glaze firing. I'll post results for that.

Today I added the ability to restart a schedule at any point in the schedule. I did this just in case I have a power outage, or something unforeseen happens. I also added an api. So far, only starting a schedule is supported, but soon, everything will be added. What this means to me is that I can do things like run my kiln on a schedule. I can pack it the night before, automatically start the schedule [ware drying cycle] at dark thirty and the whole process will finish during the day instead of late at night. In other news, I twisted my thermocouple leads - this did not stop the noise I'm seeing. I have capacitors on order to see if they can help in reducing noise

@Bill Kielb I have the SSR attached to an aluminum bud box with thermal paste in between. With the elements 100% on, the temperature of the box is just warm, not hot. I have not measured this temperature. I don't have shielded SSR leads. I used the leads I got from auber instruments with my thermocouple. I've read that I can: use shielded leads and ground the cold side twist the leads once every 2 to 6 cm to reduce noise I don't have fuse protection on the SSR. There is fuse protection on the circuit feeding the kiln rated at 50A. It's likely that my SSR would survive a kiln element short. The SSR UL testing shows 100k cycles at 50A resistive load.

I'm getting two errors, both thermal related and increasing in frequency with temperature. Thermocouple short to ground (starting at 1657F) Thermocouple short to Vcc (starting at 1832F) I'm sure I don't have either a short to ground or vcc. It's something else. I need to go back and read the spec sheets for the Max31855K. I've read they recommend using capacitors on the thermocouple. Maybe that's the problem?