Anyone work with process control, specifically water temp?

[ajlelectronics]

Sounds like another Arduino application. Use PWM for the heaters and control a small motor with a paddle to circulate the water.

 

[Kayos]

Sounds like another Arduino application. Use PWM for the heaters and control a small motor with a paddle to circulate the water.

I considered arduino's, PIC,s Atmels, Pi's, Basic STAMPS (do they still exist?) and decided I didn't want it to be an electronics project, ive enough of them on the go already, like my Pi based SQ audio install in my van, optical out into a MiniDSP, feeding old skool Alpine amps and DLS speakers, PI2 is giving me crazy ideas

I decided "ready built" industrial stuff was the way to go

Ive ordered a controller, plus a PT100 probe and a few connectors, once it lands from the far east I will see if it can actually maintain the temp without me having to constantly monitor and adjust it constantly

 

[brewdexta]

If you can't reach the accuracy or precision you need from the plethora of cheap PT100 probes out there then try these guys. I bought a couple of probes from them for my fermentation vessel. They have vast range and also do calibration plus they can make to order but that may be expensive, having said that a mate of mine got a couple made very cheaply comparatively after speaking to them nicely

 

[Kayos]

If you can't reach the accuracy or precision you need from the plethora of cheap PT100 probes out there then try these guys. I bought a couple of probes from them for my fermentation vessel. They have vast range and also do calibration plus they can make to order but that may be expensive, having said that a mate of mine got a couple made very cheaply comparatively after speaking to them nicely

Thanks for that

The controller I have ordered is actually from UK, its the same as the one you have as per the other thread as it was cheaper to buy that with a K type and an SSR than a unit on its own, also printed the instructions out for bedtime reading. The PT100 is coming from afar but I was being picky with the probe style.

I think the new controller will have the biggest impact, but if I need more accuracy I will certainly look them up, do you know if they have a "downslope" as they reach temp? as another problem I have is it going a degree or 2 over, but due to using a bigger bath to counteract stratification it takes a while to cool down again

 

[brewdexta]

The PIDS tend to learn the heat loss of the system and only put in what's needed so they don't overshoot, at least that's my experience. I made a 6L heat exchanger with a 3KW heater in it, the probe is on the output of the coil and it works fine.
You can see the PID pulsing the SSR less frequently as it approaches temp.

 

[voipio]

Some of the PID controllers have an auto tune function which, when enabled, does a single run up to temperature, after which the controller firmware calculates and memorises the settings for the proportional, integral and derivative terms. Other PID controllers can continuously autotune whilst on yet other PID controllers, the settings can be done manually, usually based on some instructions on how to do it, e.g. apply power for a set period, measure the temperature rise, and the time it takes to peak, then do some simple calculations. An overshoot of 2 degrees is unlikely since, once the temperature reached is in the proportional band, the power applied to the heater will start to back off. As Brewdexta said, as the temperature approaches the set temperature, you will see the heater power pulsing, usually with a decrease in the duty cycle.

 

[daleyd]

You could always use some of the fancy methods like Ziegler Nichols etc however I've found life is generally too short and autotune does the job much better than I could do anyway!

I did a job for a glass fibre manufacturer with a platinum bushing worth 1 million quid, and was told not to use autotune as in the past it had caused the bushing to be ruined. After a day of fiddling and trying to get decent settings in the controller I waited until the customer had gone for dinner and turned on autotune. Within 5 minutes I had the perfect settings, I think they suspected something but they never let on

 

[Kayos]

I don't work with stuff that expensive, used to do stuff with boilers with £50k control valves (chp boilers), when the client turned his back whipped off the do not remove seal and adjust the CO ratio, then put the seals on I got from the manufacturer, boilers started working so no more was said

 

[daleyd]

Yes that was a bit of a one off! It was apparently a cast platinum bushing that the glass flowed through to make the glass strands.

The heating was done by our temp controller which in turn controlled one of our thyristor units which controller the current through a transformer. The secondary current through the bushing was something like 1000 amps.

The worst part about it was it wasn't really our job, but the company that did it couldn't get it to work properly so we (I !) got called in to make it work, but without knowing how the whole thing actually was supposed to work!! (If it's a project you do from scratch you generally get to know it inside out rather than being dropped in at the deep end).

Boiler/ gas stuff is always a bit nervy, if those go wrong it's probably going to be more damage than a popped fuse!! We do a fair bit of controls work for hamworthy on the chp side.

 

[Kayos]

I suppose with boilers tho it's all designed (in theory) to fail to safe, 12 checks on a combi before it fires springs to mind

The chp stuff was all experimental as it was a few years ago, I got the job of replacing the prototype domestic version with the release version on a new build site, the ***** had moved the flue just enough so it didn't line up but was a pain to knock out neatly on brand new £750k houses (10 years ago)!

 

[daleyd]

pain to knock out neatly on brand new £750k houses (10 years ago)!

On the plus side it's unlikely I'll ever buy that house with it's dodgy flue then!!

 

[Kayos]

On the plus side it's unlikely I'll ever buy that house with it's dodgy flue then!!

It was a pain cos I had to do it, on about 20 plots

 

[Kayos]

The PID turned up today, decided to build a quick test setup with the K type until the PT100 turns up, already had to dial in 4 deg of adjustment to get it somewhere near!

As I'm measuring bottle temp but heating the water bath I decided to use my existing controller to monitor the bath temp, and if it gets excessively high (55 deg) it kills the power to the other PID and therefore the heaters

It's interesting to watch the difference in the bath and the bottle temp, and hopefully auto tune will be able to make it heat up quickly without boiling the water bath

 

[daleyd]

Does it have a different setup to use the k type - normally you have to tell the controller what sensor you are using.

It'll be I interesting to see what differences you find, hopefully it will be better than what you have now!

 

[Kayos]

It can be setup for a wide range of probes, from what I've read the K types are unreliable under 100 deg, I've ended up with minus 6 to get it close, alltho the supplied probe didn't like the developer

I've also got a 2kw "kettle type" heater element on route, it remains to be seen if it's better than the 2 aquarium heaters I'm using now

 

[voipio]

The thermal path forming the part of the control loop between heater and measuring point needs to have the different components proportioned correctly, else the system will oscillate horribly at one end of the scale or produce very poor and sluggish control at the other. Putting the PID controller's temperature sensor in the bottle, whilst heating the large tank that the bottle is in, is not a good idea, since the tank is likely to get very hot and then take a long time to cool, with the bottle temperature then continuing to rise well beyond the setpoint. If, on the other hand, the bottle was placed in a very small tank i.e. low thermal mass, you might get away with it. A 2kW heater in a 30 litre bath, with the sensor in the bath, and everything well-stirred, should work well and get the bath up to temperature in reasonable time - about 10 minutes for a 10 degree rise in temp.

Incidentally, I bought one of these: http://sansaire.com/ - said it was for cooking steak to perfection....

 

[daleyd]

6 degrees sounds a lot, though it could well be that being a budget controller it May not have any cold junction compensation built in (though it might). K type are not the most accurate but shouldn't be that far out!

When using thermocouples this can cause issues, with rtd you don't get the same effects as you are only measuring resistance.

 

[Kayos]

I broke the K type whatever it's made out of didn't like C41 developer and started dissolving, thankfully the new probe is stainless

 

[daleyd]

Oops!! Must be some fairly nasty stuff in there

 

[brewdexta]

I broke the K type whatever it's made out of didn't like C41 developer and started dissolving, thankfully the new probe is stainless

if you get bored waiting for it to be delivered, I have a couple of spare PT100s. The probe is stainless but the cable strain relief is just chromed so they tend to go a rusty without a bit of oil if its like the one below.