Geeks only - Raspberry PI controlled kiln

[neilestrick]

14 minutes ago, MarkTilles said:

I really want to use an S-type thermocouple instead of K.

Some things to consider- type S TC's are only really necessary if you're firing above cone 6, doing long high temp holds, or require great accuracy. For typical pottery firings you won't gain anything, and you can buy 15 type K for the price of one type S. Plus if you crack into the type S with a shelf and break it, it's an expensive replacement.

[MarkTilles]

Understood, but I already have two extras of them and my two ovens also came with them. So for consistency ...

[newps]

On 2/28/2021 at 1:29 AM, MarkTilles said:

@jbruceIs there any chance you could offer assistance to configure your kiln controller with the max 31856 breakout board instead of the 31855? I really want to use an S-type thermocouple instead of K. Thanks in advance,

Mark (in Sweden)

Hi Mark,

I did get the J Bruce's code up and running  tweaked to use the 31856, I am using an  Adafruit Universal Thermocouple Amplifier MAX31856 Breakout, unfortunately as we are moving house, the kiln, pi and SD card with the latest working code are all in storage.

I did add some other tweaks to the code to email at various stages of the boot up and also the firing. First stage would be to look at Stephen P Smith's MAX31856 driver (https://github.com/steve71/MAX31856) 

When I get the Pi back from storage I will try and dig out the rest of the code

[MarkTilles]

Thank you Newps!

I know nothing about python coding, as I haven’t touched a computer in 10 years, so I will need to wait until you’ve moved and unpacked. I’m sure Bruce will look forward to looking through your code as well!

My Adafruit card should arrive any day now from Holland, but I’m still waiting for my solid-state relays and heat sinks, so I’m not ready to roll just yet anyway.

Looking forward to hearing back from you!

Best regards from the country of Sweden,

Mark

[newps]

I was new to python coding at the time, and there are parts of the tweaks that possibly could be better coded else where in the different parts of code. As I am not using solid state relays but controlling using the original contactor I probably need to do a document explaining my setup and part of the reason for the revisions to different parts of the code.

I must say it was great to be able to sit in the living room and monitor the firing in comfort all by having the pi connected to the home network, so a big thanks to J Bruce and the other people who helped developed the code

[MarkTilles]

Yes, @jbruce (and clearly others) did put together a fantastic system, and the possibility for any heating curve desired - well, that’s what I am looking for. My current controller on my big oven, well I have to go in at least twice an hour to adjust the effect just to make sure it’s not going to fast or too slow.

I back up my raspberry pi boot partition regularly to my laptop, as I don’t want to take a chance of losing my small tweaks in the pi’s OS if the card should get fried or corrupted. I used to work in the computer service industry, and watched customer after customer lose critical data because they didn’t “back up” in a reliable regular fashion.

Once you again gain access to your memory card I would love to give it a try - just as it is. What’s important to me is that it works. Anything will work better than what I have right now :-)

Edited March 6, 2021 by MarkTilles

[newps]

Hi Mark,

@jbrucehas just been updating the code and added support for the Max31856, I haven't studied it fully but I sent him parts of my code so that he could see how I did it. I tried to send a full dump of my system but google wouldn't allow the attachments and there were some executable bits in it.

It may be worth relooking at https://github.com/jbruce12000/kiln-controller and giving it a try. @jbrucehas done a lot of reworking of the code so that configuring is better than my haphazard way.

For the record I was using a type R themocouple, rather than an S. 

 

 

[MarkTilles]

48 minutes ago, newps said:

Hi Mark,

@jbrucehas just been updating the code and added support for the Max31856, I haven't studied it fully but I sent him parts of my code so that he could see how I did it. I tried to send a full dump of my system but google wouldn't allow the attachments and there were some executable bits in it.

It may be worth relooking at https://github.com/jbruce12000/kiln-controller and giving it a try. @jbrucehas done a lot of reworking of the code so that configuring is better than my haphazard way.

For the record I was using a type R themocouple, rather than an S. 

 

 

(Don't know why but I think my first reply disappeared, so ...)
Yes, Jason Bruce and I have been corresponding. I am still waiting for my 31856 card to show up in the mail, and once it does I will wire it up and my Python programmer buddy will help me with the code installation and setup.  My lasty programming experience was PERL about 15 years ago LOL. I would be interested to hear more about your own enhancements, I want and need an extra relay trigger for overheating protection to shut down the source to the main heating circuit (can use the original relays in the kilns themselves) if any of my solid state relays should get stuck in the ON state. 
// Mark 

[newps]

I have started to put together a word document to explain some of my  choices and the reasons behind them together with how I achieved the result and will share it when done. It needs some photos to help explain how it all fits together. Then will share it as a PDF on here

[leof]

I registered to this forum because of the interesting thread about using a Raspberry to control a kiln.

Up till now I used my own program on a Raspberry to control my kiln. I made it in Windows IoT with a Windows Universal App (see the interface of the app in the attached file). Although that works fine I was on the lookout for something that would last for the future: The Raspberry isn't supported by Microsoft since version 3B. So I had begun porting my app to Python3, until I discovered this nice kiln-controller app of jbruce.

Also the controller is much more refined than the one I had made. In a few days I will fire my wife's kiln for the first time with this app! Excited!

On the first attached photo of the back of the kiln you see the main relay on the DIN-rail and on the right of that there is another relay. That one I wanted to introduce to you because it's a safety net. This relay switches off the main relay after a set time, about an hour after the expected firing period. The relay is added to the circuitry that signals an open door. I even go to sleep while firing!

[MarkTilles]

Cool, I’m still waiting for my MAX31856 thermocouple interface card! I ordered it from a company in Holland, it was shipped and made it to the local post office, but I never got a notification so it went back to the seller. They wrote me after six weeks that it came back to them.  So they have shipped it again… argh!

Otherwise I have my raspberry pi zero w all set up and ready to go. I have set up a three-phase 240 V solid-state relay bank with a big aluminum tube heat sink and cooling fan. I built it as a modular extension cord so I can plug it together with either of my two 3-phase kilns, since I can only run one at a time. All I need to do is swap over the thermocouple connections between the kilns and I can use the one raspberry controller for either.

Then I bought some of the proper type of compensating cable for my secondary S-type thermocouples I will be using with a separate REX-type PID controller for overheating protection (that’s what will handle the original main 3-phase magnetic relay into each kiln). This controller will also give me a secondary parallel view of the temperature in the kiln.

Looks like you also use solid-state relays what I can see in your picture. Here is my three phase solid-state relay extension cord which uses 40 amp SSRs from China, but since my circuits are only 16 amp max I’m not worried about them not containing true 40a triacs inside them (as I have read on the Internet that they are often under overrated). Yes it will be inside a metal chassis, and yes I have a ground fault interrupter on all circuits in the garage

Edited April 14, 2021 by MarkTilles

[leof]

1 hour ago, MarkTilles said:

Looks like you also use solid-state relays what I can see in your picture. Here is my three phase solid-state relay extension cord which uses 40 amp SSRs from China, but since my circuits are only 16 amp max I’m not worried about them not containing true 40a triacs inside them (as I have read on the Internet that they are often under overrated). Yes it will be inside a metal chassis, and yes I have a ground fault interrupter on all circuits in the garage

 

I use 1 SSR . I used the AliExpress ones, but wasn't impressed. On this one I have not only mounted it on a heatsink as you can see in the picture but also installed a fan that's triggered by a simple bimetal contact. The fan is mounted on the back cover of the kiln, that isn't in the picture.

Hope you get your MAX13856 amplifier soon!

Edited April 14, 2021 by leof

[Bill Kielb]

2 hours ago, MarkTilles said:

Looks like you also use solid-state relays what I can see in your picture. Here is my three phase solid-state relay extension cord which uses 40 amp SSRs from China, but since my circuits are only 16 amp max I’m not worried about them not containing true 40a triacs inside them (as I have read on the Internet that they are often under overrated). Yes it will be inside a metal chassis, and yes I have a ground fault interrupter on all circuits in the garage

 

1 hour ago, leof said:

I use 1 SSR . I used the AliExpress ones, but wasn't impressed. On this one I have not only mounted it on a heatsink as you can see in the picture but also installed a fan that's triggered by a simple bimetal contact. 

These both look well thought out.  Some potential ideas for you

Protecting SSRs

Protection for SSRs (If you are interested in protecting) generally requires fast fuses which the manufacture will list the requirements for based on their design components.  
 

Cooling: 

When we design with SSRs we try and use passive cooling rather than a fan. To That end in the states I have always been able to fulfill the need economically and in a custom fashion at https://www.heatsinkusa.com/  I am sure there are plenty of other extruded aluminum sites around though.  An interesting thing to watch on mating surfaces and thermal transfer is here: https://youtu.be/8MOTMq9g8Nk  

Interesting  in that thermal paste is not all that good in reality and good surfaces are truly the best.

Economical safety or Dp relay. Take a look at a skutt  SSR schematic where they have integrated a simple 12v relay for safety / lid switch. It’s pretty economical approved idea.

 

Edited April 14, 2021 by Bill Kielb

[neilestrick]

2 hours ago, MarkTilles said:

All I need to do is swap over the thermocouple connections between the kilns and I can use the one raspberry controller for either.

You can put a simple toggle switch on your control box to switch between thermocouples.

[MarkTilles]

Great idea!  Double pole - double throw?

[MarkTilles]

12 minutes ago, Bill Kielb said:

 

These both look well thought out.  Some potential ideas for you

Protecting SSRs

Protection for SSRs (If you are interested in protecting) generally requires fast fuses which the manufacture will list the requirements for based on their design components.  
 

Cooling: 

When we design with SSRs we try and use passive cooling rather than a fan. To That end in the states I have always been able to fulfill the need economically and in a custom fashion at https://www.heatsinkusa.com/  I am sure there are plenty of other extruded aluminum sites around though.  An interesting thing to watch on mating surfaces and thermal transfer is here: https://youtu.be/8MOTMq9g8Nk  

Interesting  in that thermal paste is not all that good in reality and good surfaces are truly the best.

Economical safety or Dp relay. Take a look at a skutt  SSR schematic where they have integrated a simple 12v relay for safety / lid switch. It’s pretty economical approved idea.

 

Thanks Bill, great ideas and information. Regarding the fan for my three phase power supply, I have already have a k-type thermocouple and PID attached to the heat sink to monitor its temperature, and then I will see whether I really even need it. I have a speed controller on the fan so I can simply turn it up to the point that keeps things at a comfortable temperature, I heard a heat sink attached to SSRs shouldn’t exceed 50°C?

 

... and as far as protection for the solid-state relays, I’m not planning any, if any should fail I have spares. They are cheap. The heating coils shouldn’t suddenly short out and caused an overload on any of these relays. If they are going to fail they are going to fail :-)

Edited April 14, 2021 by MarkTilles

[leof]

9 minutes ago, Bill Kielb said:

When we design with SSRs we try and use passive cooling rather than a fan. 

Could you elaborate on that?

[leof]

2 minutes ago, MarkTilles said:

I heard a heat sink attached to SSRs shouldn’t exceed 50°C?

My SSR specification says: 80°C. I use a 50°C switch for the fan.

[Bill Kielb]

48 minutes ago, MarkTilles said:

Great idea!  Double pole - double throw?

This usually works but in thermocouple world is strictly not allowed as each connection becomes a couple because of the dissimilar metals. It all depends on the sensitivity of the equipment and of course the condition of the couples that form inside and outside the switch. In theory since at the same temperature they will be equal and offset. If your controller is sensitive, this might not work. A standard I type plug in might be just as easy and it is made for swapping. Just put male ends on the kiln tcouple and a female in your device that way everything is designed with tcouple matching metals.

 

44 minutes ago, leof said:

Could you elaborate on that?

The manufacture will provide graphs and design criteria and you pick a heatsink size with that or greater capacity. A 50 amp SSR’s performance or effective limit decreases rapidly with temperature so designers like to oversize and keep as cool as practical and Max the load at maybe 30 amps. Convection is free and deependable. Not hard to do for yourself just follow the graphs and use realistic ambient temps.

 

80c Max can derate your SSR by 50% or more. Your prox. Cfm fan required can be approximated by sensible heat formula and Max makeup air temp. Your variable speed setup is a great start and should reveal very quickly if you do a continuous run for a reasonable time. I would set it at 35 - 40c though to start.

Edited April 14, 2021 by Bill Kielb

[MarkTilles]

26 minutes ago, Bill Kielb said:

This usually works but in thermocouple world is strictly not allowed as each connection becomes a couple because of the dissimilar metals. It all depends on the sensitivity of the equipment and of course the condition of the couples that form inside and outside the switch. In theory since at the same temperature they will be equal and offset. If your controller is sensitive, this might not work.

Yes, about the extra connection that could cause reading inaccuracies, I thought about that briefly, I’ve been slowly learning about those things. I just ordered a 25 m spool of compensation cable to connect my thermocouple to my PID controller. I will do some experimentation and see how it looks.

About my SSRs in particular, they are surely over rated for my needs, as I’m only pulling about 10 A on each of the 3 phases.

It’s great to have experienced resources like you online here! Thank you for being involved and helping out all of us.

Edited April 14, 2021 by MarkTilles

[neilestrick]

At Alpine kilns we just used a basic simple toggle switch to go between the top and bottom thermocouples, since the controllers only had one TC input. It was done that way for decades and with several different pyrometers and controllers. The switch shouldn't affect the reading of the thermocouple as long as there isn't a temperature difference across the connections, which there won't be because they're right next to each other. It's the same reason the thermocouple blocks that connect the TC to the wires at the kiln are just simple brass connections with screws, not TC specific metals. If that connection isn't messing up the reading, then the toggle shouldn't either. Plus type K just aren't that precise to begin with.

[neilestrick]

My SSR setup has everything mounted in a large panel that sits about a foot away from the kiln (10 cu/ft), so it does get some heat from the kiln. I don't usually run a vent to pull heat out of the space since it's such a large room. The relays have heat sinks, and I also have a 120mm/110cfm fan in the panel that moves a lot of air from the bottom up. Best of both worlds IMO. It's set up to the L&L specs, which are basically the same as Skutt. I'm using a Genesis controller, and the fan is connected to output 4, which has several setup options for running fans. The controller tracks the circuit board temp, and it stays under 90F in a cone 6 firing. I'm considering putting a  panel on the outside of the control box to block the radiant heat from the kiln just to further protect everything. Probably not necessary, but an interesting project for the sake of comparison. I'll measure the temp of the relays the next time I fire.

[Bill Kielb]

1 hour ago, neilestrick said:

If that connection isn't messing up the reading, then the toggle shouldn't either. Plus type K just aren't that precise to begin with.

I would respectively disagree with this statement.

Extension wire, holders and connectors are made of dissimilar metals side for side to minimize the unwanted couple issue. It does seem to be a common sentiment by many folks though but everyone  also seems good with using the extension wire religiously, go figure. Anyway Proper holders are made the same way, one side one metal, the other side different. All these items are polarized for a reason to minimize adding new couples along the way.

Anyway in practice, it likely will work; however Rasberry Pi’s have had noise issues with different tcouple boards, and pid controllers generally favor conversion to current loop for noise and transmission reasons. So I definitely would not suggest it. Actually shielded tcouple wire is a thing  now since more electronic controllers are in use.

In industry is this allowed, (the toggle switch, dissimilar metals) the simple answer is no, absolutely not. Can it work -yes it might.

As far as accuracy, K type thermocouples although economical are impressive at .75% from 0 to a max of 1372 C -   2.2 degrees in over 1300C. Quite impressive actually in my view, not sure where they ever got a bad reputation.

Edited April 15, 2021 by Bill Kielb

[neilestrick]

Every pottery kiln manufacturer in the US uses basic porcelain thermocouple blocks with brass terminals, no TC specific metals, and they work very well.

I love type K TC's, and they work great for pottery kilns. Definitely not the top choice for industry, though. I'm just saying there are more accurate options out there.

[Bill Kielb]

25 minutes ago, neilestrick said:

Every pottery kiln manufacturer in the US uses basic porcelain thermocouple blocks with brass terminals, no TC specific metals, and they work very well.

Yes, the screws are brass, SS, You name it. The block terminal is the correct metal per the marking in the block. Brass screws seize up less than most especially in chromel and alumel. 

I didn’t know you got an SSR control. How do you like it?