Automating A Brewery

[roger mellie]

Stumbled across this site whislt I was looking for a cheap PLC for a project I am working on. US259 for 8 Analog IO and 16 Digital IO - this would seem to be any amount for a brewery.

http://www.lt-automation.com/T100%28e%29.htm

The PLC monitoring ability is provided by Modbus - which allows communicatios to any number of HMI type SCADA packages - some of which allow free access to some of their functions - or a limited number of IO points or 'tags'.

Whilst 'dabbling' in VB to get interation with your process might seem 'do-able' to some - an interface to a PC based HMI package would definitely be more mainstream.

RM

Wow, great find RM. The price rocks, as do the features. I do like the idea of wireless control, and remote internet control.

I used PLCs in industry nearly 20 years ago. They certainly have come a long way. Makes me feel a bit nostalgic looking at ladder diagrams though.

Wortgames, I think this might be more suitable than a labjack. Worth reading up on at least. I've just read the circuit cellar review and I want one ... Pity I've invested so much time and effort into a custom system. :blink:

keep inventing, and brewing great beer,
Arnie

It doesnt get much simpler than ladder. The ability to write little basic scripts should make communicating with a HMI package pretty simple. I would imagine that analog comparators would make up the bulk of the code - certainly seems to have that functionality available.

PM me if you are going to go down this road - the main windows HMI package that I use has a free demo available with 32 tags - I could let you have a copy unoffically - the Modbus Driver is available also.

RM

 

[Wortgames]

I sent them a quick question by email and got a full price list by what looks like auto-reply:

View attachment PriceList.pdf


Can anyone see anywhere on the website that it clearly states how interactive it can be? Not being an expert I'm not really sure what sort of spec I'm looking for. I'm assuming that you can mess about with it in real-time via ethernet (or internet!) but I'd hate to spend all that money on an assumption!

 

[roger mellie]

I sent them a quick question by email and got a full price list by what looks like auto-reply:

View attachment 9355


Can anyone see anywhere on the website that it clearly states how interactive it can be? Not being an expert I'm not really sure what sort of spec I'm looking for. I'm assuming that you can mess about with it in real-time via ethernet (or internet!) but I'd hate to spend all that money on an assumption!

From what I can gather - the main method of interacting with the PLC would be via serial means - RS-232 via a comms port on a PC. This appears to be inbuilt.

The TCP/IP route requires additional hardware (199 by the looks of things) and is more of a nice to have.

If all you want to do is have a PLC doing stuff automatically, and you then want to interact with that - like changing a temperature setpoint, animate a process page to see whether a valve is open or closed (or indeed open or close a valve), have a sequence that allows you to start , stop, pause etc... then it looks like this little PLC coupled with a windows based HMI (human machine interface) through a Modbus driver (server) would do the trick.

It would certainly be worthwhile asking the questions to the website to confirm this.
RM

 

[Wortgames]

OK, I sent them an email last night and had this response in my inbox this morning:

-----

From: Latin-Tech Inc
Subject: Our PLC in your Brewery

Yes, the T100MD2424 can do the job ( http://www.lt-automation.com/T100(e).htm ).
You can do on-line programming or monitoring, either connected to a PC or through internet.
You can make a program that contains an Automatic and a Manual way to run. Using a software key ( from a SCADA or an EXCEL's button) you can switch fom Automatic to Manual, overriding any part of the process.
You can make on-line (real time) changes for temperature setpoints, timer values or any other parameters without stopping the normal operation.
In general you can do the following;
1) Develop a PLC program that allows you to work in automatic (no intervention) or manual modes.
2) Using a SCADA or EXCELL ( http://www.lt-automation.com/Scadas(ing).htm ), you can monitor or change operation parameters.
3) If you have a PC with internet connection and you connect the PLC to this PC, then you are able to access the PLC from any other place, through internet, at no extra cost.
4) Our PLC can be programmed in ladder +Basic, so you can develope several "recipes", simply by changing variables in BASIC
5) With our voice modules ( http://lt-automation.com/VoiceModule.htm ) you can add voice messages for alarms, process functions,etc

If you need any further assistance, please do no hesitate in contacting us again

-----

So apart from the shock of having a prompt and thoughtful reply from a real human being, is this all good news?!


(edit: tidy)

 

[ArnieW]

So apart from the shock of having a prompt and thoughtful reply from a real human being, is this all good news?!

Getting a reply quickly from a real human being is excellent news Wortgames. ... Oh, and the details seem excellent too.

Did you say their ladder+basic was available via download for a trial period? If that is the case, I'd suggest you play around with it and see if you can get it to do what you want. I vaguely remember that you can "simulate" having a PLC without actually forking out the cash for one.

I believe they come with PID capability also. I'd suggest this is fabulous news - PID will quickly become your friend if you are attempting to reach and maintain a target temperature.

The PLC side of it seems easy (at least in my head). The other bit you will eventually need to think about is interfacing between the PLC and your brewery. It won't be able to switch 10-20A heaters without some additional work, for example. I can help guide you with this stuff though, and I'm sure there are many others who would offer.

cheers, Arnie :beer:
(enjoying a nice cold Helles in warm and sunny Melbourne - ahh, holidays are a great invention)

 

[Wortgames]

It's glorious here today isn't it?!

I'm still flip-flopping betwen the PLC and the LabJack to tell the truth.

I've exchanged a couple of PMs with Roger Mellie, he kind of likes the PLC but then he's a bit of a PLC-whisperer it seems so it is familiar territory for him.

The PLC's certainly look like they have heaps of potential but also some drawbacks. Not least of which is if I need to get the speccy software, that is an extra $200 on top of the $350 purchase cost of the small PLC. It's still not entirely clear how much I can pause a 'script' or edit upcoming lines either, although as RM points out the actual processes probably wouldn't change much, just variables like temp and duration, and it shouldn't be too hard to create a pause function.

If PID means what I think it means, ie basic processes like temp control running by themselves, then I can certainly see that it would be useful (essential?!)

There are demo versions of the software and a huge amount of documentation for the PLCs here: http://www.lt-automation.com/Download.htm

I haven't felt brave enough to go near it yet, as I'm still a bit overwhelmed by it all and it will probably take me a month of Sundays to figure out what I was even looking at.

The LabJack is the great unknown, but at $200 to purchase it is significantly cheaper, and it looks like with a bit of effort I (we? ) can probably rustle up a pretty cool interface for it. Even if it proves necessary to get the top software, ProfiLab, that's only another $165 making a total cost of $365 versus $550.

Another important issue to consider is the probes, as I would probably want at least 3 or 4 of them (and the more the merrier I reckon, subject to cost and available inputs). It looks like the PLC may require some additional circuitry to gear the probe output voltage to suit the PLC input. Not sure about the LabJack, but I'm guessing this could probably be done in the software, thereby reducing hardware costs?

Output-wise I only really need to control one heating element and one pump at this stage, but the future is limited only by my imagination and available outputs. Ultimately it could control any number of valves and robotic features. Again, the LabJack looks attractive here as I could theoretically just add on additional units as I run out of outputs (or inputs for that matter). Both the LabJack and the PLC would obviously require relays on the outputs, but I reckon I can probably figure that out, having a pseudo-electrical background.

I'm dead keen to have a look at your current system though Arnie, and I'd love to hear any more advice regarding the use of temp probes.

 

[goatherder]

If PID means what I think it means, ie basic processes like temp control running by themselves, then I can certainly see that it would be useful (essential?!)

There's no reason you can't do PID with the labjack - you'll just have to code the PID loop yourself. It shouldn't be big deal to do this. Google just got me this: http://learncontrol.com/pid/code2.html which has pseudocode for a PID loop.

 

[roger mellie]

If PID means what I think it means, ie basic processes like temp control running by themselves, then I can certainly see that it would be useful (essential?!)

There's no reason you can't do PID with the labjack - you'll just have to code the PID loop yourself. It shouldn't be big deal to do this. Google just got me this: http://learncontrol.com/pid/code2.html which has pseudocode for a PID loop.

Maybe I am missing something here - I dont see the point in PID control coupled with Digital (On/Off) final control elements. There isnt anything to modulate. A step controller with Ramp/Soak functionality - now that would be handy - but PID?

I would like to see if Zizzle's way of temp measurement using the DM???? chip could be modified to become 0-5V DC - then the PLC method becomes very cost effective.

RM

 

[ArnieW]

I'm still flip-flopping betwen the PLC and the LabJack to tell the truth....

If PID means what I think it means, ie basic processes like temp control running by themselves, then I can certainly see that it would be useful (essential?!)...

Another important issue to consider is the probes, as I would probably want at least 3 or 4 of them (and the more the merrier I reckon, subject to cost and available inputs). It looks like the PLC may require some additional circuitry to gear the probe output voltage to suit the PLC input. Not sure about the LabJack, but I'm guessing this could probably be done in the software, thereby reducing hardware costs?

Output-wise I only really need to control one heating element and one pump at this stage, but the future is limited only by my imagination and available outputs. Ultimately it could control any number of valves and robotic features. Again, the LabJack looks attractive here as I could theoretically just add on additional units as I run out of outputs (or inputs for that matter). Both the LabJack and the PLC would obviously require relays on the outputs, but I reckon I can probably figure that out, having a pseudo-electrical background.

I'm dead keen to have a look at your current system though Arnie, and I'd love to hear any more advice regarding the use of temp probes.

Hi Wortgames,

I can't remember how accessible things were with the labjack in real time, but if that is not an issue then I can see why you are drawn to it.

Goatherder is right about PID. In fact I've coded PID in my machine - it's not quite perfect yet, but it does it better than other systems I've experimented with.

I'm guessing you may need some input interfacing with both systems to get the best out of them. As an example, if you are using the LM35 temp probes, and you have a 10bit analog to digital converter (you'd have to check what each system is capable of), then you get essentially a 0.5 deg C resolution with a direct connection. If you want better than that you need to amplify the input with op-amps.

OK, so temp probes. My own experience (probably not exhaustive) is that you can get either analog or digital probes - analog are at times easier to use (depends on your system) but are prone to noise and you may need to amplify as mentioned above. Digital are simple if matched with the right system, are not as prone to noise, and will potentially give better resolution/accuracy than an analog one.

My analog probe of choice is the LM35. My digital probe of choice is the DS18B20. I use both in my system - the LM35 gives virtually an instant reading, where the DS18B20 currently takes about 3/4 of a second with the picaxe.

The other main analog probe you can get is a thermister. They are very common, but require a bit of extra calibration, and I don't think they have a linear response which can be a headache.

I believe that you can connect a DS18B20 directly to your PC serial port and read temps with it (given the right VB or C code). I'm sure I read somewhere about a global temperature montoring project that encouraged this, but I couldn't find reference to it with a google search.

I use the picaxe as a simple interface that does the same job as well as give expansion for many more inputs and outputs.

cheers, Arnie

 

[Gulf Brewery]

Hi everyone in this thread

I have sort of been following this thread. I wrote some code in VB.net to monitor temperatures and control step mash (PM if you want the source code). I used the DS1820's as sensors.

If you want code for reading / writing to the DS1820's, have a look at Digitemp and also Martybugs.net

The Gulf Brewery uses the DS1820's for monitoring the brewery - all of the actual temperature controls are done dedicated temperature controllers - would love to automate them, but I can't justify the cost (yet)


Cheers
Pedro

 

[ArnieW]

Maybe I am missing something here - I dont see the point in PID control coupled with Digital (On/Off) final control elements. There isnt anything to modulate. A step controller with Ramp/Soak functionality - now that would be handy - but PID?

RM

I've found PID to be very useful. I use a PID algorithm to vary the power percentage on my heater. The output is a straight digital on/off, but varies with a duty cycle. In my case the system will go on/off via a solid state relay for a percentage of a 1 second period depending on the requested power. 50% power is on for 500mS, off for 500mS. Because it is a heater, it averages out the power nicely. It would be a bit nauseating to watch a light globe do this because it is quite slow.

Anyway the result of this has convinced me that it is very useful. I did it without PID control for a long time, and PID is certainly not essential, but it makes for more stable temps than my previous experiments.

cheers, Arnie

 

[ArnieW]

I made a soluniod from a 1/4 turn ball valve and a central locking actuator years ago.

It worked awsome as a fuel switch-over in my custom built landcruiser.

best thing about the central locking units they are pretty strong and you just have to reverse the polarity to push the other way...

Just cut the handle off about 15-20mm up drill ahole and position the central locker so on full extention its fully closed and on retraction the valve is fully open..or viseaversa..

Takes a little mucking with the initial setup but works a charm..



Dunno if this helps with anything... :huh:

Hi Sqyre,

do you have any more details or pics of this?

I gave it a bash the other day but the results were not spectacular. To get the 1/4 turn, the actuator was too close to the pivot point to get enough leverage. To get it to move the ball valve, it was only reliable for about 1/8th turn (half the ball valve travel). This was with a new ball valve which might be a bit tighter than a used one, but as they heat they get tighter anyway.

cheers, Arnie

 

[ArnieW]

Hi everyone in this thread

I have sort of been following this thread. I wrote some code in VB.net to monitor temperatures and control step mash (PM if you want the source code). I used the DS1820's as sensors.

If you want code for reading / writing to the DS1820's, have a look at Digitemp and also Martybugs.net

The Gulf Brewery uses the DS1820's for monitoring the brewery - all of the actual temperature controls are done dedicated temperature controllers - would love to automate them, but I can't justify the cost (yet)


Cheers
Pedro

Hi gulf brewery and gidday zizzle,

thanks for posting the details on the DS1820's. I personally love them for their resolution, ease of use, accuracy etc.

Zizzle, I missed your post earlier in this thread, but good info about what can be done simply.

Arnie

 

[sqyre]

Hi Sqyre,

do you have any more details or pics of this?

I gave it a bash the other day but the results were not spectacular. To get the 1/4 turn, the actuator was too close to the pivot point to get enough leverage. To get it to move the ball valve, it was only reliable for about 1/8th turn (half the ball valve travel). This was with a new ball valve which might be a bit tighter than a used one, but as they heat they get tighter anyway.

Sorry mate, all i have is a badly blured memory from 10 years ago...

I admit there was teething problems that took a while to work out..i remember the actuator was very close to the pivot point of the valve and i have a recolection of replacing the original cut handle with one set at around 45 degree's. But cant be 100% certain... Also the valve was used for fuel swithching so it possibly lubricated the ball BUT i also have a very vague thought of loosening the internal ball holding assembly so it moved freely.

The parts i used were all pretty much second hand and it might just be a shear fluke it worked at all but it did work in the end, and quite well at that..

SO...the problem your experiencing is the fact the actuator is not strong enough to move the valve the full distance??? or you havent got enough travel to achieve a full 1/4 turn?..or both...

If not strong enough the only thing i would suggest is possibly 1. Make sure your power supply is big enough as i remember mine worked a lot better when hooked up to a car battery than my 2amp supply..

2. maybe explore the option of running 2 actuators with a "T" piece handle with actuators either side, 1 pushing the other pulling. be a bit of an eye-saw hanging off the side of a brew-rig but....it might work.

If your not getting enough travel and your as close to the pivot point as you can get then i dont like your chances.

Sorry to get everyones hopes up....i may have to break out some old ball valves and central lockers and try and find a solution. Also it would depend on the size of the ball valve...Once again cant recall what size it was.

My apologies..

sqyre.

 

[ArnieW]

Sorry mate, all i have is a badly blured memory from 10 years ago...

My apologies..

sqyre.

Well I'm a happy little brew au-tow-mater

I've just cracked the auto ball valve thingy by thinking a bit outside the square.

It was sqyre who got me thinking about actuators ...
Which got me thinking about the more simple up/down motion of gate valves and if I could use an acutator with that ...

Which got me pulling apart a gate valve and thinking that it would be mechanically difficult to maintain a good seal and use an up/down actuator (instead of the three turn thread to open/close it).

When I noticed a stepper motor I had and got to thinking about rotating the handle of a gate valve ...

Which got me thinking that you need a big/expensive stepper to do that properly ...

Which drew my attention to some geared motors I had lying around, and then a quick trip to Jaycar for a pully system and presto ... a working motorised gate valve that (at the moment) takes about 8 seconds to open or close.



This is the top view - on the left is the pully that is in place of the gate valve handle - on the right is the geared output pully from the motor. It has more than enough grunt to do the job.

The gate valve was off the shelf from Bunnings. The motor is one of the Jaycar geared thingys as per:
Geared motor

Now all I need to do is work out some more permanent mounting and then sort out the control gear.

cheers, Arnie

 

[Wortgames]

Arnie, that is very cool!

Hi Pedro, welcome to the beerporn peepshow!


I heard back from the folks at LabJack. Apparently the LabJack U3 would be the go:
http://www.labjack.com/labjack_u3.php?prodId=25

It has 16 analog/digital IOs plus 4 digital IOs, and sells for US$99.

They seem pretty confident that the free DaqFactory Express software that comes with it will do the trick nicely.

Probe-wise, the options are:

EI-1034, sensor in stainless probe, US$49
EI-1022, sensor in plastic probe, US$29.

However they suggested I could look at the data sheets for the probes, source the core sensors, and make up the probes myself.

It looks like the probes can hook up directly to the LabJack, I just need some wotsits to go between the labjack and the mains relays for the outputs.

The ProfiLab software that Ocean Controls sells is nothing to do with the LabJack folks by the looks of it, which suggests that at least this LabJack stuff is a fairly open standard.

Ocean Controls also sells a similar device, not LabJack brand but I would imagine very similar, called iUSBDAQ:
http://oceancontrols.com.au/data_acquisiti...usb/iusbdaq.htm.
They've got it for $149 plus GST but apparently it sells for $79 in the states, 20% less than the LJ U3.

There's more information (and some cool screenshots) on the iUSBDAQ here in a 1.8Mb PDF:
http://www.hytekautomation.com/Downloads/i...6_UserGuide.pdf

Does anyone fancy sussing that one out for value / features?

 

[Zizzle]

Hey Arnie, looks great. How do you know when to stop the motor?

I also found another old post of mine:

But probably the easiest way is to go down to your local jaycar and buy a QM1538

http://www1.tpgi.com.au/users/p8king/qm1538.htm

They cost AU$50 come with a temp probe, and will plug into your computer (it features RS-232 output of readings).

Another alternative is a RS232 based 10-bit data logger for AU$25. DI-194 But you would need to find a wire up some thermo couples.

For me, the DS1820s really are great. You can see a link to a ~$40 Ozitronics kit from Groggy's site if you are not handy with a soldering iron. Otherwise you can build up a DB9 plug and order some free samples from Dallas Semi and away you go.

 

[ArnieW]

Hey Arnie, looks great. How do you know when to stop the motor?

I also found another old post of mine:

But probably the easiest way is to go down to your local jaycar and buy a QM1538

http://www1.tpgi.com.au/users/p8king/qm1538.htm

They cost AU$50 come with a temp probe, and will plug into your computer (it features RS-232 output of readings).

Another alternative is a RS232 based 10-bit data logger for AU$25. DI-194 But you would need to find a wire up some thermo couples.

For me, the DS1820s really are great. You can see a link to a ~$40 Ozitronics kit from Groggy's site if you are not handy with a soldering iron. Otherwise you can build up a DB9 plug and order some free samples from Dallas Semi and away you go.

Hi Zizzle,

I'd not come across those data loggers before - a good price and good entry point for automation.

A good question about knowing when to stop the motor. What I tested the other day was that there was enough torque to drive the gate - and I have no doubt that mechanically this is a good solution.

The next bit of R&D is working out the control of it. The simplest solution would be to turn on the motor for a set amount of time (at the moment about 8 secs) to fully open or close the gate valve. Because it is driven by a belt, my guess is that you could actually drive it for a longer time (say 8.5 secs) to ensure it is fully open or closed. Once the valve reaches a limit, the belt slips (a natural clutch). This effectively calibrates the valve every time you reach a limit (fully open or fully closed).

Then if you want to open it 50%, you turn the motor on for 4 seconds. You only recalibrate every time you close it fully or open it fully.

If that fails, the next step would be to put a pot on the gate valve stem. I've attached a picture with the concept.



The pot (on left) could have a pulley with rubber wheel that is driven directly from the stem of the gate valve. It might need to be recalibrated from time to time, but hopefully would be quite accurate. The pot position would indicate the gate valve position.

If there is too much slippage (ie. requires recalibration too often to be useful), the next step would be to have a gear or chain linkage between stem and pot to ensure there is no slippage.

I think I'll get the chance to test this on Monday. I'll let you know how it goes.

Arnie

 

[kirem]

Hey arnie,

Have you thought of using a servo motor in your valve experiments?

Kirk

 

[ArnieW]

Hey arnie,

Have you thought of using a servo motor in your valve experiments?

Kirk

Hi Kirk,

My first experiments ages ago were with servos. There are two main problems with servos and the system I'm working on - firstly, servos generally have about a 180 deg range and the gate valve requires about 3 revolutions; secondly your average servo does not have enough torque for this kind of job. After that, you actually get more bang for your buck with a good geared motor like the one I've got pictured. It is about 1/3 of the price of a basic servo and does the job much better. If you need to hack the servo to get it to do 3 revs, you might as well start from scratch I figure.

cheers, Arnie