PID Controllers..?

[Adrian Brentnall]

HI All The metalwork for my homebrew glass kiln has arrived - so I'm looking at the electrics / electronics to control it.

The temperature control is being handled by a dedicated controller ('cos I've got a spare one..!)

I'm looking to implement an independent 'over-temperature' circuit, just in case the main controller has a bad day, or the power relays stick 'on'. With 6kw of heat in a well-insulated enclosure, the temperature could get fairly high, fairly quickly...

A while back I bought two of these

- on the grounds that, even though I didn't want to use them as controllers, they did include an 'out of limits' alarm and a digital display of temperature.

I have contacted the seller to ask for more details - but wondered if anybody here has used a similar piece of kit - and could advise. I looked up PID control on Wiki - and I'm afraid I ended up more confused!

This controller will let me enter values for P, I and D, also for CycleTime, heating rate and AL.

Quoting from the (Chinese) data sheet "AL value is set for the value in which PV-SV is bigger than the AL (PV-SV>AL)".

The example they give is SV = 128 AL = 3 High limit alarm PV=> 131 - alarm on Low limit alarm PV=>125 - Alarm is on otherwise it's off

I want to trip the alarm if the temperature goes above 850c - so far my best guess is to program a set value of (say) 400 and AL of

450... - to have alarm off between -50 and +849 ? Anybody clued up on these things - and does that sound about right ?

Thanks Adrian

[Martin Brown]

The PID bit is to determine the feedback loops for hitting the set target temperature(s) optimally with a given known thermal inertia. It also keeps it more closely regulated when done right. The manual ought to tell you how to choose the right numbers but then again...

This bit is about setting alarm levels.

Wouldn't you also want the alarm to trigger if the kiln was off from the programmed temperature cycle by more than a few tens of degrees?

The emergency stop might be better off being an independent failsafe. Modern PID controllers are likely to be microprocessor controlled and switching multi kW resistive loads might glitch them. I have a lot more faith in dumb thermocouples and analogue electronics not to latch up.

You will be amazed how "simple" failsafe systems can fail dangerous if you don't pay attention to all the possible modes of failure.

[Adrian Brentnall]

The manual's somewhat brief - and assumes a certain amount of prior knowledge! Besides - I'm not actually wanting it to do any controlling - just raise an alarm for me if the temperature goes too high.

Yes - that's what I'm wanting to do

In an ideal world - but 'the other' dedicated controller is looking after that for me.

Well - it's independent of the main controller.

Fair point - but I was hoping to use this 'off-the-shelf' solution - which includes a power supply, thermocouple input, relay output and temperature display, plus (hopefully) the logic to control that alarm.

I've worked in railway signalling...... but I think a 'Q - relay' may be a little OTT in this application!

I'm designing for a circumstance that (hopefully) will never occur - and the commercial kiln that this homebrew one is replacing didn't even have a 'master kill relay' (as I found out when the thermocouple leads developed a short and the controller gaily poured power into the kiln for a couple of hours while waiting for the thing to heat up!

So - in a sense, whatever I do in my kiln will be an improvement...

Adrian

[Roger Mills]

This device is a controller in its own right - PID = Preportional, Integral, Derivative

In other words, it computes the error (difference) between the target value and the current value of the controlled parameter and seeks to make it zero by varying its output in a way which depends on the magnitude of the error (proportional), the rate of change of the error (derivative) and the cumulative value of the error integrated over a period of time (integral)

As a bi-product, it appears to have an alarm function - presumably warning you when the error is too large - but it's difficult to see how it would do this if it's not actually controlling anything.

Wouldn't you be better off with a simple over-temperature thermostat which cuts the power and/or sounds an alarm if a safe temperature is exceeded?

The PID controller looks like a solution looking for a problem!

[Adrian Brentnall]

Yes - I understand that bit... but it did seem to provide 'for free' a digital temperature display, thermocouple interface, relay output etc....

I guess it depends on how 'clever' it is - whether it'd object to just sitting & watching..?

Yes - that's where I'm trying to get to!

Perhaps the KISS approach would be the right way in this case! (?perhaps?!) Working on the principle of 'what you don't fit can't go wrong'

- and bearing in mind that I'm not looking for tremendous accuracy - as a temperature of 'more than about 850c' means that Something's Gone Wrong' and it'd be a good plan to cut the power...

Going back to 1st principles ... there's always something like this

The 'comparator' scheme above needs a bit more 'building' - but would be kind of elegantly simple..

Thanks for the reality check! Adrian

[Bob Minchin]

I fitted one to a kiln for a friend a couple of years back and it looks pretty similar and had similarly chinglish terse instructions. I ended up setting everything to low temperatures to confirm operation ( and to speed things up) and then changing to realistic settings.

Good Luck

[Adrian Brentnall]

Thanks Bob! I may need to rethink the plan - maybe time to dig out the soldering iron and veroboard!

[meow2222]

Set I & D coefficients to zero, then you're just looking at temp with the P term. A PID controller is a semi-smart analogue controller, way OTT for th is job. You could also use thermocouple plus opamp, or opamp plus constant current source plus a bit of heating element wire in the kiln.

BTW I'd be interested in the design/build of a minimal kiln from scrap mate rials for 3rd world use, to enable people stuck in poverty to make a living . I dont know if thats something you have the skills for.

NT

[Adrian Brentnall]

HI Thanks for the comments. I'm rapidly coming to the conclusion that the PID is a sledgehammer for this particular walnut!

Looking instead at the AD595 ic - which takes a thermocouple input, generates its own 'cold junction' reference, and can be wired as a comparator - so it'll switch at a setpoint defined by a preset voltage input. They cost ?14 each - but looks like they're meant for the job.

What size of kiln do you want to build - and what's it for ? I'm building for glass fusing / shaping - which needs fairly fine control of ramp / hold (which implies fairly hi-tech) - but if you're thinking ceramics then there are quite a lot of people doing 'backwoods'-style kiln-building, which can be wood-fired.

A friend of mine back in the UK is part-way through building a wood-fired pottery kiln - he had a day's consultancy from this man

- and said it was very useful. His kiln uses some modern ceramic fibre insulation (same as I'm using in my new glass kiln) - not sure what the 'scrapyard' alternative would be...

An alternative approach seems to be a pit kiln

- after all - people were making pottery hundreds / thousands of years before our modern materials became available.

Adrian

[meow2222]

OTT. All you need is a low offset comparator or opamp.

Its really a question of what types can one build. Anything would be good.

I like the pit kiln. Only problem is we cant reuse a copyright article, and I cant write one due to lack of experience with them. Something for glassw ork would be good too, but same problems apply. A bit of electronics isnt a barrier to everyone.

thanks.

NT

[Adrian Brentnall]

I'm no expert in thermocouples - but the references I read seemed to 'big up' the need for a 'cold junction' reference...

Are you saying that's not required - given that my 'over-temperature' setpoint will be 850c (or 850 to 830 above ambient ?)

Also - I'm only building one off - so I'm happy to trade a few Euro for (relative) ease of use...

Well - if it's for ceramics use then it's just a matter of 'heat until it's hot enough and then let it cool down' - where 'hot enough' can be judged by peeping at the colours....

..but for glass, you're looking at more complex ramp/hold control

- both on the way 'up' and on the way 'down' - which also suggests that you'll need to fire with something more controllable (like gas or electric).

At its simplest, a kiln's just a box that gets hot. My local supplier of refactory materials has a little home-built ceramics test kiln that's simply an 18" dexion cube, lined inside with firebrick, resistance wire element tucked into a groove around the sides and a slab of kiln-shelf for a lid. Not pretty - but it works..

As I say - depends on purpose..

Thanks Adrian

[meow2222]

If you want accuracy to 0.1C then sure, use a temp calibrated cold junction , and connectors of the right metals. If 10C is near enough, forget about i t

Yup. I dont see how the 595 would get you any real advantage, unless you wa nted the extra precision.

Yes. I'm thinking maybe one could put a little solid fuel in there between the pieces, and let electric supply a lot of the power. Most such people do nt have electricity, but a minority do. The electric feeds I've seen are to o crude to run a kiln unless its very well insulated, theyre basically any old appliance flex strung from post to post.

Thank you. Maybe I'll email you

NT

[Adrian Brentnall]

The 595 claims +/- 3c, there's a more expensive version that's +/- 1c

Long-term, I'd like to build my own controller, with PC-link for programming - so I might go for the dedicated 595 chip, which will also allow a 10mV per degree c output for 'future expansion'. I'm still undecided - but, either way, sounds like it's a lot more practical than the PID controller I started out with!

My current glass kiln is a smallish one (14" square) and that pulls 12A. The new one has 2 x 15A elements. With glass you need to be able to put lots of heat in quickly - there's a certain range of temperatures where 'devitrification' occurs (glass goes crystalline & milky) so you need to go through that temperature range as fast as posible. Hence - lots of heat.

You're welcome - email address works. Thanks Adrian

[meow2222]

any opamp will do that

12A through any old appliance flex is doable. I've done it. 30A not. Sounds possible.

Turns out google gripes wont reveal the email addy. Oh joy. My addy's long dead. Do you want to drop me a note to nt @mailinator.com?

thanks.

NT

[Onetap]

BMS experience, i.e., PID controllers with thermistor (usually) temperature sensors.

Pv is Process variable, i.e., the temperature measured by the sensor contro lling the device.

Sv (not a standard abbreviation, so I believe it..) is Setpoint value, i.e. , the temperature you want the process at.

This seems to work for the example above, i.e., Alarm condition is On if P v > Sv+3 or if Pv < Sv -3.

Your best guess seems a bit odd; I's suggest you re-think it with the abov e.

Electric heaters and BMS/PID controllers are/were fairly common, typicall y on air handling unit heaters. 0-10V output from the controller, controlli ng a 415V heater battery via a solid state relay. The control panels were s upplied by others, so I'm a bit vague about the latter device.

[Adrian Brentnall]

Thanks for the comments. I think I've moved on to using a smaller sledgehammer to crack this particular walnut - currently thinking AD595 to interface with the thermocouple and provide 'thermocouple fail' and 'over-temperature' outputs....