Automating A Brewery

[randyrob]

Hello Arnie

Have you figured out in detail what you want the machine to do?

yes, i'm not looking at full automation in anyway just looking at shaving a few hours of the start of my brew day. so it would only have to perform a few simple tasks for me.

something like

1.) set a timer to switch on my 240v electric element on my HLT (say 6am)
2.) when the HLT reaches a desired temperature (say 75*C) open solenoid valve from the HLT to the MLT / start 'dough in' motor.
3.) sound air raid siren 1 1/2 hours after dough in to wake me up for the sparge

saving me 2 1/2 hours off my brewday

======

heck now i think of it i could even simplyfy it more by setting a manual timer for the HLT and have a temperature controller set to my desired strike temp.

then have another timer that opens the solanoid / switches motor on.

then just set my alarm clock to wake up??????????


====

There are many ways to approach this kind of thing, and the WASP is not unlike the DSE parallel port interface that I first started with.

i would definately give the parallel port interface a go, can it switch on and off 240v appliances?

My main reason for using the picaxe is that I can embed intelligent process control in the machine - and not have to rely on the PC. I've had a few nasty PC crashes at inconvenient times before. With my new rig it will be both stand alone and allow PC control.

wow looking foward to seeing that beast!

Thanks Rob.

 

[Sammus]

I should have quoted a few posts here, but i couldn't be bothered reading back through to find them, anywho, you can get pretty cheap (cheapest i've seen anyway) SSRs, PIDs and thermocouples from www.auberins.com

None of the thermocouples are particularly useful, theyre mainly designed for surface mount use on coffee machine boilers - plus the PID controllers come with the K type threaded t/c probe. which is actually nickel plated brass, and completely unsuitable for screwing into the side of a vessel containing and pressure (including the pressure exerted my liquid, say). I know this from experience The PID controller seems to work well though, I prefer my Fuji PXR3, but thu auber instruments ones are roughly 1/4 the price just thought someone might be interested (no affiliation)

PS I like that labjack thing... one day maybe

 

[kirem]

I use Auber PID controllers and really like them. I use the screw in threaded probe that came with the unit, but it screws into a plastic tee piece in the wort path exiting the heat exchanger.

I had to seal the probe with food grade sealant once it was screwed into place.

I notice that they have a unit now that has a ramp/soak function, might be good for programmed step mashing or automated mashing then mashout.

 

[ArnieW]

Little by little I have been developing a VBExpress (VB2005, or whatever) user interface platform for automating home breweries... Some recent screenshots are at the end of this blog entry:

http://firststatebrewers.com/blog/2007/05/...day-activities/

It still has a long way to go before it is functional, but once completed I intend on making it open source and letting anyone use it. Right now its mostly just the component rendering code, but I have slowly been developing the framework for controlling events and sequencing them as well. My hope is that it'll be easy enough to use that anyone can do it - so long as they can handle the electronics side of things

Ah Garrett, now that you've come out of lurker mode you will have to keep us informed of all sorts of developments like your proposed pressure sensors and experiments with picaxes

[OT]

Although I had to wince when I read back through your "snip, snip" blog - ouch - been there ... :blink:

[/OT]

Your rendering code is looking good. I'm sticking to developing the picaxe end of things for the moment, but one day I will cut my teeth on VB2005.

 

[ArnieW]

i would definately give the parallel port interface a go, can it switch on and off 240v appliances?
wow looking foward to seeing that beast!

Hi Rob,

The parallel interface does not switch 240V without adding relay drivers to do the work (or SSRs). Most cards will be the same or else be limited to lower current applications than we might be using.

I actually have the DSE kit all wired up but surplus to requirements, and have done VB6 coding to get it to work as well. I'd be happy to let it go to a good home - the documentation for it must be around here somewhere.

There is a link for it here:

http://www.dse.com.au/cgi-bin/dse.storefro...duct/View/K2805

As for working out what the machine will do - I find the temptation is like you describe - figure out something and then figure out the next feature would be good - then figure out the next fancy thing you want it to do - then ...

cheers, Arnie

 

[ArnieW]

I should have quoted a few posts here, but i couldn't be bothered reading back through to find them, anywho, you can get pretty cheap (cheapest i've seen anyway) SSRs, PIDs and thermocouples from www.auberins.com

None of the thermocouples are particularly useful, theyre mainly designed for surface mount use on coffee machine boilers - plus the PID controllers come with the K type threaded t/c probe. which is actually nickel plated brass, and completely unsuitable for screwing into the side of a vessel containing and pressure (including the pressure exerted my liquid, say). I know this from experience The PID controller seems to work well though, I prefer my Fuji PXR3, but thu auber instruments ones are roughly 1/4 the price just thought someone might be interested (no affiliation)

PS I like that labjack thing... one day maybe

The auberins stuff looks pretty good - they even include 'beer brewing system' as one of the possible functions for the PID I looked at!

The other possible approach to a new machine would be to use or hack X10 controllers. There is some open source software that allows internet control and is designed for remote house control functions but I'm sure could be used for brewing.

http://misterhouse.sourceforge.net/

cheers, Arnie

 

[Sammus]

I googled LM35 and found that a LM35DZ (quite cheap) is a sensor with a linear output of 10mV/degC and range of 0-100C.

Does this mean that the signal output will read 0V at 0C and 1000mV at 100C?

If this is the case, then to work out the accuracy of the sensor through a 12bit analogue input (like in the LabJack) would be something like:

voltage resolution of the labjack is about 0.596mV. So there is about 1000/0.596 ~ 1677 discrete steps in the 0-1V range, which equates to 100/1677 ~ 0.06 degC accuracy...

since the range of analogue inputs is 0-2.44V, using an appropriate amp one can gain accuracy ~ 0.02 degC

is there some other intrinsic sensitivity limitation of that little chip? surely you cant get that much accuracy for $1.95 USD

 

[ArnieW]

I googled LM35 and found that a LM35DZ (quite cheap) is a sensor with a linear output of 10mV/degC and range of 0-100C.

Does this mean that the signal output will read 0V at 0C and 1000mV at 100C?

If this is the case, then to work out the accuracy of the sensor through a 12bit analogue input (like in the LabJack) would be something like:

voltage resolution of the labjack is about 0.596mV. So there is about 1000/0.596 ~ 1677 discrete steps in the 0-1V range, which equates to 100/1677 ~ 0.06 degC accuracy...

since the range of analogue inputs is 0-2.44V, using an appropriate amp one can gain accuracy ~ 0.02 degC

is there some other intrinsic sensitivity limitation of that little chip? surely you cant get that much accuracy for $1.95 USD

The LM35 does work as you say Sammus, although the LM35 has a greater range than the LM35D series (-55 deg to + 150 deg C).

To get any meaningful results from a 12 bit analog input you need to have a very clean power supply, and if you want to amplify you again need some really good design like an instrument amp to make it worthwhile. Personally, I'm not sure anything greater than 0.1 deg resolution is worthwhile or even meaningful given other inaccuracies, beginning with the LM35 itself. If I had a 12 bit input, I'd make sure the supply was good and settle for the LM35 without an amp.

Just my 2c worth.

 

[alexbrand]

Hey Folks!

A brewing colleague of mine developed a software to control the brewing devices. It is available in German language at the moment only (translation into English soon). It uses the Dallas semiconductor DS18S20 as temperature measurement unit at the COM-port and the parallel-port with a relay card to control the mash tun (heating (gas/electric), stirring device, etc.). It's still a kind of pre-release, butit works. Just ask Zwickel...

For those of you speaking German I could provide a download link to the programme, just send me a PM.

Greetings,

Alex

 

[Sammus]

The LM35 does work as you say Sammus, although the LM35 has a greater range than the LM35D series (-55 deg to + 150 deg C).

To get any meaningful results from a 12 bit analog input you need to have a very clean power supply, and if you want to amplify you again need some really good design like an instrument amp to make it worthwhile. Personally, I'm not sure anything greater than 0.1 deg resolution is worthwhile or even meaningful given other inaccuracies, beginning with the LM35 itself. If I had a 12 bit input, I'd make sure the supply was good and settle for the LM35 without an amp.

Just my 2c worth.

Well I'm considering using a labjack which is controlled by a commputer via a USB port. I'm pretty certain the 5V rail on a computer PSU is pretty clean isnt it?

Also I've done some testing and the 5V rail on the PSU always seem to be out by up to 0.2V - but still stable. Will this mess up the various reference voltages and such in a controller like the labjack?

 

[bugwan]

Hey Folks!

A brewing colleague of mine developed a software to control the brewing devices. It is available in German language at the moment only (translation into English soon). It uses the Dallas semiconductor DS18S20 as temperature measurement unit at the COM-port and the parallel-port with a relay card to control the mash tun (heating (gas/electric), stirring device, etc.). It's still a kind of pre-release, butit works. Just ask Zwickel...

For those of you speaking German I could provide a download link to the programme, just send me a PM.

Greetings,

Alex

Guten Tag Alex,

I was going to ask if you know Zwickel, but I noticed you mentioned him in your post..! Great to see some international brewers on the site. I've downloaded the software you mentioned from Zwickel and my broken German got me through
Interesting software, I think there are a couple of brewers here working on a similar product.

My problem is waterproofing my DS18S20s for liquid temp measurement. I guess a stainless probe would be best.

My main aim (in terms of automation) is to be able to monitor temps in the brewery on a web page - (not hard in theory, using Digitemp etc, but they're all Linux solutions and I'm running Windoze). That way, I always know where I'm at. In the future, I'd like to rig up some relays to control temps manually if necessary, so that diacetyl rests etc could be done remotely. Probably not necessary, as it's not hard to do with a timer...but hey, boys and their toys....!

 

[Zizzle]

The problem with software is that everyone's hardware setup is different, so you can't really make a simple generic brewery controller software package.

To make matters worse, people are using different platforms. I for example, wouldn't waste my time on Windows based PCs. Just too unreliable. People practically throw out old computers that are very capable of running linux. Just grab one and start playing.

A wanted to play around with microcontrollers so I chose to go that way, but backup would be an old PC laptop with Linux on it. You can interface the DS1820s to the serial port and use the parallel port for controlling relays.

I wouldn't use any other temp sensors than the DS1820s. It's easy to interface with. No need to worry about analog signal conditioning, supply conditioning, correction tables, ADCs etc.

I've been working as a software engineer for the past few years mostly doing embedded work, so for me the software is the easy part.

Reliable mechatronics and sensors is the harder part.

 

[newguy]

I'm pretty certain the 5V rail on a computer PSU is pretty clean isnt it?

Also I've done some testing and the 5V rail on the PSU always seem to be out by up to 0.2V - but still stable. Will this mess up the various reference voltages and such in a controller like the labjack?

I've never 'scoped a computer 5V supply rail, but I doubt that it is clean. Standard practice when mixing analog and digital is to have separate supply rails and sometimes separate grounds. Digital circuitry creates ugly power supply noise due to its inherent switching (high/low) nature. This switching creates current 'bursts' that play hell with analog electronics.

I wouldn't worry about the 0.2V supply difference. Any well designed electronics will never depend on an external supply for critical internal references - the labjack most likely falls under the 'well designed' category.

 

[Sammus]

hmm may just go with digital... if I ever get this thing happening

Ive been reading through the labjack manual online and it seems way over my head... maybe one day though

 

[alexbrand]

Good morning!

My problem is waterproofing my DS18S20s for liquid temp measurement. I guess a stainless probe would be best.

@ bugwan:
I used a capped and salt filled copper tube. It is fast and big... :


But this is not my definitive solution, at the moment I am working around to get a better "thermometer". I drilled a hole into a SS screw wich will function as a housing for the DS18S20 in the future. This screw then will find its place in the middle of the side wall of my mash tun:



This idea was published some weeks ago by a good metal worker in a German brewing forum.

@Zizzle:

Right, different hardware means different software. The reason why I've chosen the DS18S20 are its costs. They are for free if you order a sample at Maxim-IC. Even if you have to pay for them including the interface for the COM port they cost less than 4 USD.

Using DigiTemp by Brian Lane a log file with the temperatures measured is created. And this logfile is read by Thomas' software.



Gut Sud!

Alex

 

[Zwickel]

My problem is waterproofing my DS18S20s for liquid temp measurement. I guess a stainless probe would be best.

Hi bugwan,
at a holliday in oz my wife bought a couple of SS chopsticks. Always when I need something for the brewery, Im going to search the kitchen
So I found the chopsticks there and got an idea:

filled inside with epoxy.

Cheers

 

[bugwan]

Hi bugwan,
at a holliday in oz my wife bought a couple of SS chopsticks. Always when I need something for the brewery, Im going to search the kitchen
So I found the chopsticks there and got an idea:

filled inside with epoxy.

Cheers

Morgen Alex and Zwickel, you've inspired me with your ideas. Zwickel, never has a piece of cutlery been put to better use!

Does the epoxy interfere with temp measurements - I mean does it insulate the DS18S20 too much??

I'm just setting up a Linux box now, so I should be able to post some web-based temp measurements in the coming days...

 

[Zwickel]

Does the epoxy interfere with temp measurements - I mean does it insulate the DS18S20 too much??

No, not at all. If I put the stick into hot liquid, within seconds it shows the actual temp, very quick.

 

[ant]

A question for all you automation gear-head types... can someone pls give me some advice on whether this unit would be suitable?

I'm looking for a couple of solenoid valves, in the first instance I'm wanting one to run a HERMS unit, where a low temp reading in the mash tun will trigger the solenoid valve to be opened and recirculate through the HERMS unit. To my simple mind, this unit is s/s, it's 1/2" like the rest of the system, it's normally closed and is temp rated to 80C - a safe temp margin more than I will be running from the mash tun. :huh:

 

[bugwan]

No, not at all. If I put the stick into hot liquid, within seconds it shows the actual temp, very quick.

Great work Zwickel, thanks again. I'll be in the kitchen tonight....looking for stainless steel chopsticks.