Geeks only - Raspberry PI controlled kiln

[rondoc]

On 2/2/2022 at 11:19 AM, cadenrank said:

Disregard this. After a checking again, they were both matching each other this time, and then when I gave the kiln some power to heat it up, they both got up to about the same point as well. 

I did learn that the connectors I bought fit backwards or normally into my pyrometer, so I'll have to keep an eye on the negatives and positives when I plug into my pyrometer.

I have accurate tc readings but every so often the reading gets lost so reads 32F and is only recovered if I run

kiln-controller again. Not a good omen for a reliable controller. Trying to trace why this occurs. It is not due, I feel, to mechanical connection problems as it happens with no movement. It could be the crappy pin to socket connections from gpio to breakout tc board. The pins on the Raspberry pi are smaller than standard pins.  Soldering direct to the boards might be a more reliable method.

 

[Bill Kielb]

1 hour ago, rondoc said:

Trying to trace why this occurs. It is not due, I feel, to mechanical connection problems as it happens with no movement.

Can you measure the noise? Scope? Or maybe apply 1 mv steady state and see if the reading is stable for for a long enough time for things to heat up. If it still disappears then on to freeze spray …… not a big fan, as these assemblies are not industry proven / hardened.

Edited February 4, 2022 by Bill Kielb

[cadenrank]

11 hours ago, rondoc said:

The pins on the Raspberry pi are smaller than standard pins.  Soldering direct to the boards might be a more reliable method.

I opted to solder direct from the 31856 to the GPIO pins, and haven't had any issues to loosing the thermocouple readout. I really only opted to do this because I'm comfortable with a soldering iron, and I didn't think about buying the pins and such to make pin to pin connections, mostly because as I mentioned, I only intend to use this for monitoring for now, and I was trying to keep it as cheap as possible. I think I used 22gauge stranded wire, twisted it tight, stuck it through and soldered. 

[High Bridge Pottery]

17 hours ago, rondoc said:

I have accurate tc readings but every so often the reading gets lost so reads 32F and is only recovered if I run

kiln-controller again. Not a good omen for a reliable controller. Trying to trace why this occurs. It is not due, I feel, to mechanical connection problems as it happens with no movement. It could be the crappy pin to socket connections from gpio to breakout tc board. The pins on the Raspberry pi are smaller than standard pins.  Soldering direct to the boards might be a more reliable method.

 

Does it not spit out "emergency!!! too many errors in a short period, shutting down" in the terminal? The default code is taking 40 readings every 4 seconds and should hit that warning if more than 30% of them are 'bad_count'.

I think the log.info level is actually too low to spit anything out in terminal and probably just puts it into the log file.

Edited February 4, 2022 by High Bridge Pottery

[silfrsmithr]

Well, I got my '55 amplifier today. Hooked it all up and things looked good. Mt thermocouple read 75F, so I put my lighter on it, and it dropped and went into negatives very quickly. Made me think of reversed polarity. I switched my thermocouple wires and the thing runs perfectly. My TC block had red and yellow marked on it, and I had always assumed red as positive. It turns out the yellow is actually positive. I had my TC wires crossed this entire time, and I suspect this caused the problems with the '56 amplifier as well.

Now, is a triple SSR bank reasonable protection against an ssr sticking closed? What about snap discs (on climb)?

 

Thanks,

Ken

[Bill Kielb]

19 hours ago, silfrsmithr said:

Made me think of reversed polarity. I switched my thermocouple wires and the thing runs perfectly.

Good catch! I think I read something about that here on Jan 27th?

19 hours ago, silfrsmithr said:

Now, is a triple SSR bank reasonable protection against an ssr sticking closed? What about snap discs (on climb)?

 

The normal way is single SSR per leg and add a definite purpose safety contactor feeding the SSR’s...

Of late Kiln manufactures have been adding standard 12v relays in front of each leg instead of the clunky contactor which ends up pretty economical and reasonably reliable as the relay turns on first (unloaded) and stays on - no mechanical cycling.

Matching fast fuses (Amps squared seconds) can protect the SSR from damage for top notch design. Most SSR manufactures tell you the rating not to exceed.

 

Edited February 6, 2022 by Bill Kielb

[rondoc]

42 minutes ago, Bill Kielb said:

Good catch! I think I read something about that here on Jan 27th?

The normal way is single SSR and add a definite purpose safety contactor feeding the SSR’s...

Of late Kiln manufactures have been adding standard 12v relays in front of each leg instead of the clunky contactor which ends up pretty economical and reasonably reliable as the relay turns on first (unloaded) and stays on - no mechanical cycling.

Matching fast fuses can protect the SSR from damage for top notch design. Most SSR manufactures tell you the rating not to exceed.

 

I use a back to back SCR block rated at 600 volts and 100 amps. I have had a few total shorts that blew my 100 Amp breakers but the SCRs were fine. I feed the SCRs with  MOC3060  chip that is a zero crossing switch so have made a pretty indestructible SSW. In the very rare event that my SSW did short out my SiC element would raise kiln temps to cone20!! so I always watch temps and cones when getting near cone10 for glaze firing or 1800F for a bisque..  I have made some hardware that I fired to cone20 so I know it gets that hot. 

[Bill Kielb]

17 hours ago, rondoc said:

I use a back to back SCR block rated at 600 volts and 100 amps. I have had a few total shorts that blew my 100 Amp breakers but the SCRs were fine. I feed the SCRs with  MOC3060  chip that is a zero crossing switch so have made a pretty indestructible SSW

SSRs are already zero crossing and optically isolated, just drive them with a few volts dc. Mechanical relays pretty much goof proof easy way to protect against a stuck SSR and he doesn’t need to worry about leakage killing him when loading the kiln. The mechanical relay also easy to interlock through an independent high limit as well if he chooses. The new stuff today is generally good.

What temperature is cone 20 and where do you get cones?

Edited February 6, 2022 by Bill Kielb

[rondoc]

On 2/5/2022 at 11:21 PM, Bill Kielb said:

SSRs are already zero crossing and optically isolated, just drive them with a few volts dc. Mechanical relays pretty much goof proof easy way to protect against a stuck SSR and he doesn’t need to worry about leakage killing him when loading the kiln. The mechanical relay also easy to interlock through an independent high limit as well if he chooses. The new stuff today is generally good.

What temperature is cone 20 and where do you get cones?

Cones from Orton. The parts were a dry mix of SiC. Kaolin and wollanstonite. The mix was then moistened and pressed into plaster molds. The resultant ceramic is very strong and can take cone10 with no change. I have the exact formula if anyone is interested. If partly fired it can be machined then post fired for strength. 

[Bill Kielb]

1 hour ago, rondoc said:

Cones from Orton.

What temperature?

[rondoc]

I have the cones 17 thru 20 still. I believe cone 20 is about 2700F. I had a visual disappearing filament pyrometer and a thermoelectric one  that I used as well as the cones.  I know it is about 2700 F as I melted steel in it by mistake! Steel melts around 2800F. ( BTW this is about 35 years ago so my memory is a bit rusty. I do have my lab note books for reference if anyone has interest. )

[rondoc]

On 2/3/2022 at 11:52 PM, Bill Kielb said:

Can you measure the noise? Scope? Or maybe apply 1 mv steady state and see if the reading is stable for for a long enough time for things to heat up. If it still disappears then on to freeze spray …… not a big fan, as these assemblies are not industry proven / hardened.

 

[rondoc]

A problem might exist using a SSW that might be because not all SSW are zero crossing . This means that they will switch on whenever the pi says to go on. A zero crossing SSW will go on the next time the 60 cycles goes to a zero voltage AND the pi says ON.  By switching at zero crossing the electrical noise is much reduced. That is also why relays are noisy because they switch on at any time the pi says go on. Relay contacts do also bounce giving intermittent sparking and much electrical noise. Relays can useful like cutting when off power with over temp. but not for a PID controller like the pi. 

 

 

[MarkTilles]

2 minutes ago, rondoc said:

A problem might exist using a SSW that might be because not all SSW are zero crossing . This means that they will switch on whenever the pi says to go on. A zero crossing SSW will go on the next time the 60 cycles goes to a zero voltage AND the pi says ON.  By switching at zero crossing the electrical noise is much reduced. That is also why relays are noisy because they switch on at any time the pi says go on. Relay contacts do also bounce giving intermittent sparking and much electrical noise. Relays can useful like cutting when off power with over temp. but not for a PID controller like the pi. 

 

 

I have cheap 40A SSRs from Banggood and Wish and am not having any off issues with my operations, even with 15 foot long cables strung all over the place. Very stable measurements, no operational errors I’ve seen. Just lucky maybe?

Edited March 18, 2022 by MarkTilles

[Bill Kielb]

44 minutes ago, MarkTilles said:

Very stable measurements, no operational errors I’ve seen. Just lucky maybe?

No, SSR’s for the most part are zero crossing and optically isolated. Can’t think of any that aren’t actually (I am sure they are somewhere), it’s sort of an important part of the package of a solid state relay, rather than build your own from SCR’s 

The off or on issue that most any device can have is a function of it being solid state.  They all have some leakage and when they go bad can stick on or off with greater likelyhood than a mechanical relay sticking on.

[Pyewackette]

On 4/15/2021 at 11:23 AM, neilestrick said:

I'm excited to see what sort of element life I get. I've got customers with the same model kiln that are getting 300 firings or more from their quad elements with standard relays. Sucks that I'll have to wait a couple years to find out!

Is a year and a half close enough?