PID temperature controllers (split topic)

[Tony]

Edit: Split from HPT thread, regarding Julian's consideration of adding a PID page to the wiki

That would be very timely Julian, as I've just bought a PID controller.

I used to do PID work of a different kind, for flying surface and thruster control for unmanned underwater vehicles.  These little control modules are a new one on me though!

[Julian]

Sounds like you're eminently more qualified than me to write stuff on PIDs!  The settings on these things baffles me, luckily most seem to come with an auto set-up these days.

Wouldn't feel happy about publishing the page unless it's been given the once over by someone who knows what hey are talking about i.e. Vince or similar.

[Tony]

My PID has turned up but I'm a little clueless as to what to set.

It's this - £16 delivered inc K type thermocouple:

http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=160550439764

Datasheet:

http://www.fmfranklin.com.au/products/data/rkc/c100inst.pdf

The variant I've been sent has part no REX-C100FK02-M* AN-NN

This has single K type input, one alarm, and relay output.

I'm planning on adding a 12v supply to my control box, which passes through a pressure switch, then the PID relay for controlling a SSR on the immersion heater.

My understanding from the limited reading I've done is that if set right, the PID will effectively PWM cycle the SSR as the set temperature approaches, in order not to overshoot, and to avoid having the hysteresis around the set point that other types of controller have.  The duty of this cycle appears to be set by the "T" parameter (Proportional cycle 1-100 seconds).

I'm guessing we need to set Proportional band "P" but not "I" or "D".

From my PID days I remember that I allows for loading offsets preventing the set point being reached by P alone, but can accumulate over time if the set point is never reached aka windup).  Imagine an aircraft control surface - if the set point is beyond the physical range of movement then the I term will build up and up in an attempt to hit the set point, so the next time it is requested to move back the other way, it'll take a long time to respond as the I term needs to drain the other way.  In some ways demething might be seen the same way - struggling to reach the set temp while the Meth is boiled off, and accumilating I with the potential for massive overshoot if all the Meth boils off.  On the other hand, heat loss from the system is a loading offset that might be taken into account by a small I, with windup limited by the Ar (anti-reset windup term).  Tricky.

D is effectively a damper to prevent overshoot from I.

How you would choose an appropriate values for P and I and Ar though... not a clue.

[Tony]

Actually there is a good Wikipedia page on the subject.

http://en.wikipedia.org/wiki/PID_controller

[Julian]

Good find, great price!

The spec on ebay says it's self tuning.  If I were you I'd set in that mode and let it get on with it.

Think you are correct about not needing all the parameters, I've vague recollections of a couple not having much effect.

If I were you, I'd ditch that TC and use a simple exposed junction jobbie.  Typically like this ... http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=250317881490&ssPageName=STRK:MEWAX:IT  much more versatile, faster acting and longer.

Make pockets out of 6mm copper tube with the end crimped and soldered and just poke the TC to the end of the tube.  You can then have several pockets in various places and move the TC around as required ie tank temp for processing and still head for distilling.

Yes, great PID page, but the theory and maths behind it is way over my head.  I set the non auto one I've got by reading some Noddy speak instructions coupled with much trial and error!

The wiki page I was contemplating was more on integrating the PID with the processor and various safety interlocks etc, although a section on set up would be useful.  I've got a basic diagram drawn out but was criticised on the forum for using 230V controls.  Probably a fair comment from a safety point of view but I was trying to keep it simple so nubies could understand.  If it's any help I'll willingly send you a copy.

I have read (wikipedia, I think) that SSRs only fail open.  This is defiantly not my experience, I've had one fail closed as I believe someone else on the forum has. So an interlock between the Hi temp alarm on the PID and the power to the heater via a contactor or suitably rated relay is a must in my eyes.

[Tony]

I'd be interested to see your design.

Yes my concern over SSRs is the failing closed possibility.  There needs to be a mains side tell-tail somewhere at least.

I'm thnking that having the pump controlled by the level switch too would be a good idea (with an override for pumping the bio out).

12V frame PSUs rated 1A can be got (delivered) for a few quid so no problems on the low voltage side.

[Bio Dervotee]

Hi folks, its me again. Subject to what you are saying,re ssr`s  ,  my thought`s would be that all semiconductor components will fail in a short circuit condition.   I have rarely seen semiconductor parts fail in open circuit unless  the sc caused excess current flow. there has got to be a simple solution to this, ssr`s that I looked at drop 1.6volts and if voltage across ssr drops below this by 0.3 volts then we could use a crowbar type trip circuit to kill heater. Just something to shut it down and set an alarm notice. Think it might be a good thing to have pump still running if  temp.  gets too high, suppose it depends what mode your system ix in
Cheers folks
Will

[Bio Dervotee]

Hello, me again,
there is the option of using a 12volt car battery charger as a power source, with a little help from a 12volt regulator chip, very simple to build and easy to make.
Keep all your 12volt relays in a sealed plastic box feeding your processor needs near to it, take an umbilical to your remote unit and have cutouts where heater cannot run unless pump is running and pressure level  is ok ,[barometric sensor]  temp. is ok and whatever else can be done to ensure a wee bit of safety.
Maplin do voltage regulators with circuit diagrams to suit
So hope this may be of some use to you all
Cheers
Will
ps  misssed some bits
pissssssssed
Will

[Rotary-Motion]

I'd be interested to see your design.

Yes my concern over SSRs is the failing closed possibility.  There needs to be a mains side tell-tail somewhere at least.

I'm thnking that having the pump controlled by the level switch too would be a good idea (with an override for pumping the bio out).

12V frame PSUs rated 1A can be got (delivered) for a few quid so no problems on the low voltage side.

pc psu's are 12 volt got an old pc you can rape?

[Tony]

Yes but already butchered to run the drywash pump 

I like Julian's system that uses a contractor in series with the ssr that only switches when there is a failure. I do like SSRs with automatic zero crossing switching, best way to control an emmersion.