New Giveaway - Wort Monster Conical Fermenter!

Home Brew Forums > Home Brewing Beer > Automated Brewing Forum > Any interest in starting new open source automated brewing project...for propane?




View Poll Results: Does this seem like a good idea?
Yes, this sounds awesome! 28 82.35%
No, this sounds idiotic! 4 11.76%
You'll shoot your eye out, kid!!! 2 5.88%
Voters: 34. You may not vote on this poll

Reply
 
LinkBack Thread Tools Display Modes
Old 01-17-2013, 12:00 AM   #21
crane
Feedback Score: 0 reviews
Recipes 
 
Join Date: Sep 2011
Location: San Diego, CA
Posts: 248
Liked 20 Times on 17 Posts
Likes Given: 1

Default

Quote:
Originally Posted by jrubins

So, I mentioned earlier that smarter folks than I made the statement about derivative control. In that case I'm referring to a conversation with friends who have actual engineering degrees, so I'm merely restating their input. I do, however, know a bit about audio, and understand the principle of a low-pass filter. I suppose what you are thinking is that a large mass of water, which takes some time to change temp will naturally act as a low pass filter due to the systems lag time. However, there is still quite a lot of noise with a PID controlled heating system using electricity to heat the water. I'm not sure if the relatively high volume:energy input would mitigate that.

Check out the temperature variation on this guy's coffee roaster project.
The red line is quite 'jagged' compared to the set point
Yes the large mass of water will act as a low pass filter. With a properly designed analog front end for your temp sensor you will not see a very jagged signal. In my system my thermistors only vary by +/- 0.1F so I am seeing very little noise in my signal. Prior to refining my analog front end I was experiencing quite a bit of noise that reached into the +/- 1F range. I also take 5 simultaneous readings and average them out in firmware. I too have an engineering degree and studied feedback control theory in college. In my feedback control lab I experienced first hand how digital PWM control loops can outperform their analog equivalent by a long shot. When the system being controlled has a natural low pass filter in it like we do the PWM output is no different than an analog output and therefore you still can achieve derivative control due to the lag if the system.


__________________
crane is offline
 
Reply With Quote Quick reply to this message
Old 01-17-2013, 12:23 AM   #22
jrubins
Feedback Score: 0 reviews
Recipes 
 
Join Date: Aug 2012
Location: San Jose, CA
Posts: 78
Liked 5 Times on 3 Posts
Likes Given: 3

Default

Quote:
Originally Posted by crane View Post
Yes the large mass of water will act as a low pass filter. With a properly designed analog front end for your temp sensor you will not see a very jagged signal. In my system my thermistors only vary by +/- 0.1F so I am seeing very little noise in my signal. Prior to refining my analog front end I was experiencing quite a bit of noise that reached into the +/- 1F range. I also take 5 simultaneous readings and average them out in firmware. I too have an engineering degree and studied feedback control theory in college. In my feedback control lab I experienced first hand how digital PWM control loops can outperform their analog equivalent by a long shot. When the system being controlled has a natural low pass filter in it like we do the PWM output is no different than an analog output and therefore you still can achieve derivative control due to the lag if the system.

Ok, that makes sense, however, a new electric system >$ than my current propane system, so I'm willing to forge ahead!

Any thoughts on temp measurement devices. You mention thermistors , how about a nice long thermocouple - would that work to give accurate readings without averaging, or am I better off handling that in code?

http://www.alltronics.com/cgi-bin/item/22U013/search/%22J%22-Type-Thermocouple


__________________

In bottles: Last Day IPA, AAPL American Pale Ale
On deck: Some sort of porter type thing

The answer to the ultimate question of life, the universe, and everything? 42. The number of bottles in my first batch of homebrew? 42. Coincidence? I think not.

jrubins is offline
 
Reply With Quote Quick reply to this message
Old 01-17-2013, 01:12 AM   #23
crane
Feedback Score: 0 reviews
Recipes 
 
Join Date: Sep 2011
Location: San Diego, CA
Posts: 248
Liked 20 Times on 17 Posts
Likes Given: 1

Default

Quote:
Originally Posted by jrubins

Ok, that makes sense, however, a new electric system >$ than my current propane system, so I'm willing to forge ahead!

Any thoughts on temp measurement devices. You mention thermistors , how about a nice long thermocouple - would that work to give accurate readings without averaging, or am I better off handling that in code?

http://www.alltronics.com/cgi-bin/item/22U013/search/%22J%22-Type-Thermocouple
I am not a huge fan of thermocouples for a few reasons.

1 voltage they produce is extremely small in the microvolt range and therefore they are a lot more susceptible to noise.
2 because of the extremely small voltage they produce you need a very high gain and high precision amplifier to get the signal up into the range where an ADC can measure their output. At my last job we had a few products with thermocouple inputs and we had to use a low drift low offset opamp along with 0.1% resistors. Even with that we had to calibrate each amplifier circuit.
3 the way thermocouples work requires you to need a second temp sensor (we used a thermistor for this) to measure the temperature where the thermocouple wire ends. Now any inaccuracies in your cold junction compensation temp sensor cause inaccuracies in your thermocouple measurement.

I prefer thermistors due to their ease of implementation. They can also be accurate enough for our application. I test all of mine in an ice bath and in boiling water and all of them have been within 0.1F. No need for any calibration.

Averaging in code is trivial so I wouldn't worry about it.
__________________
crane is offline
 
Reply With Quote Quick reply to this message
Old 01-17-2013, 05:06 AM   #24
tob77
Feedback Score: 0 reviews
Recipes 
 
Join Date: Sep 2012
Location: Napa, CA
Posts: 60
Liked 6 Times on 4 Posts
Likes Given: 14

Default

Quote:
Originally Posted by crane View Post
Yes the large mass of water will act as a low pass filter. With a properly designed analog front end for your temp sensor you will not see a very jagged signal. In my system my thermistors only vary by +/- 0.1F so I am seeing very little noise in my signal. Prior to refining my analog front end I was experiencing quite a bit of noise that reached into the +/- 1F range. I also take 5 simultaneous readings and average them out in firmware. I too have an engineering degree and studied feedback control theory in college. In my feedback control lab I experienced first hand how digital PWM control loops can outperform their analog equivalent by a long shot. When the system being controlled has a natural low pass filter in it like we do the PWM output is no different than an analog output and therefore you still can achieve derivative control due to the lag if the system.
PWM is an approximation of an analog signal that can work very well as long as the PWM cycle frequency is very fast when compared to the system's lag. In theory, PWM can only be as good as analog control, not better. However, we often do not have true linear analog control. For example, the flow through most valves at 50% open is not half of what it is at 100% open. It may be possible for PWM to outperform an analog equivalent because of this nonlinearity in some systems.

Sorry for sounding a little bit like a textbook. I too am an engineer, and I love this sh!t
__________________
tob77 is offline
 
Reply With Quote Quick reply to this message
Old 01-17-2013, 09:55 AM   #25
alien
Feedback Score: 1 reviews
Recipes 
 
Join Date: Apr 2012
Location: Philadelphia, PA
Posts: 1,229
Liked 61 Times on 55 Posts
Likes Given: 61

Default

Probably most people doing this sort of project would use DS18B20s ("one wire probes") for temperature measurement, e.g.

https://www.oscsys.com/projects/brewtroller/system-design/direct-fired

The digital output is easy to interface and they don't require calibration. Also you can buy probes off the shelf from places like Brewer's Hardware.

But for an open source project you maybe don't have to worry about this at all. People can add on their own temperature measurement routines adapted to whatever sensors they have.

alien is offline
 
Reply With Quote Quick reply to this message
Old 01-17-2013, 10:47 AM   #26
alien
Feedback Score: 1 reviews
Recipes 
 
Join Date: Apr 2012
Location: Philadelphia, PA
Posts: 1,229
Liked 61 Times on 55 Posts
Likes Given: 61

Default

Here's a bunch of resources for you on gas valves and controls.

http://wickedstone.com/brewing/Html/gas%20valves.htm

The basic set up for intermittent (on-off) control is this:



The off-the-shelf PID controller can be replaced with a Brewtroller or similar Arduino type set up controlling a relay.

alien is offline
 
Reply With Quote Quick reply to this message
Old 01-19-2013, 07:42 PM   #27
kladue
Senior Member
HBT_LIFETIMESUPPORTER.png
Feedback Score: 0 reviews
Recipes 
 
Join Date: Dec 2006
Location: Turner, Oregon, Oregon most of the time now
Posts: 2,296
Liked 43 Times on 40 Posts

Default

If you are going the Arduino route, then consider an alternative method, PWM a single solenoid with restriction, and a surge dampening chamber tee'd into gas line to burner down stream to reduce flame fluctuation.
Most solenoids have 100,000 cycle lifetime so I would not expect them to die before the rest of the brew system does and the PWM duty cycle should be in 1/2 - 1 second range which will limit total cycles. For the surge chamber, 2" iron pipe and fittings from the big box building supply should work, start with 12" of pipe and increase/decrease as needed to get acceptable flame fluctuation. Setup would be to adjust restrictor / valve to give max fire with solenoid open 100% of the time and then vary off time to reduce flame level. You should not need a solenoid larger than 1/4" (opening in burner orifice is usually smaller than 3/16"), larger solenoid would be a waste of time and money.

__________________
kladue is offline
 
Reply With Quote Quick reply to this message
Old 01-23-2013, 10:27 AM   #28
cuz_man
Feedback Score: 0 reviews
Recipes 
 
Join Date: Jan 2013
Posts: 4
Default

Sorry if this has been covered in a previous response, but with regard to the electric vs gas discussion, have you considered using phase angle control? This would allow you to use an analogue control signal to modulate the AC waveform to a heating element. Do a search on eBay for SSRMAN-1P for a solid state relay mounted version.

Having tuned a lot of heating PID loops, I question the need for a derivative term. Depending on the thermal load, temperature loops tend to be proportional gain heavy with a relatively small integral term depending on how well insulated your vessel is. A noisy temperature measuring device will give you hell with a derivative term.

My 2 cents.

__________________
cuz_man is offline
 
Reply With Quote Quick reply to this message
Old 01-23-2013, 07:04 PM   #29
alien
Feedback Score: 1 reviews
Recipes 
 
Join Date: Apr 2012
Location: Philadelphia, PA
Posts: 1,229
Liked 61 Times on 55 Posts
Likes Given: 61

Default

I think you are right about dropping the D term in general. I do wonder if there are some different issues around tuning PIDs for gas as opposed to electric.

For instance, gas heats the bottom of the pot not the liquid it contains so perhaps it is more important to consider mixing inside the vessel. Poor mixing creates lag and increases the tendency to overshoot, which would make me think about a more conservative setting (i.e. less P). Placement of the thermometer might also be a bit more tricky.

__________________
Magic Smoke brewhouse and grill

Arduino PID controller - Automatic Smoker - Cooler MLT - Counterflow Chiller - SSR demo - PWM timer - Folding Brew Table
alien is offline
 
Reply With Quote Quick reply to this message
Old 01-24-2013, 04:57 PM   #30
jrubins
Feedback Score: 0 reviews
Recipes 
 
Join Date: Aug 2012
Location: San Jose, CA
Posts: 78
Liked 5 Times on 3 Posts
Likes Given: 3

Default

Quote:
Originally Posted by alien View Post
I think you are right about dropping the D term in general. I do wonder if there are some different issues around tuning PIDs for gas as opposed to electric.

For instance, gas heats the bottom of the pot not the liquid it contains so perhaps it is more important to consider mixing inside the vessel. Poor mixing creates lag and increases the tendency to overshoot, which would make me think about a more conservative setting (i.e. less P). Placement of the thermometer might also be a bit more tricky.
Definitely, getting the heat measured correctly and uniform throughout the water or wort is important, but that's not unique to gas heat.
I'm thinking of using one of these guys http://www.homedepot.com/Paint-Paint-Tools-Paint-Mixers-Stirrers/h_d1/N-5yc1vZbtta/R-202251543/h_d2/ProductDisplay?catalogId=1&langId=-1&storeId=10051#.UQFy6UpxerU to keep the temperature uniform. I'm also thinking of using 1 or 2 probes with 3 sensors each to measure the temp at the top,middle, and bottom of the pot.

Quote:
Originally Posted by cuz_man View Post
Sorry if this has been covered in a previous response, but with regard to the electric vs gas discussion, have you considered using phase angle control? This would allow you to use an analogue control signal to modulate the AC waveform to a heating element. Do a search on eBay for SSRMAN-1P for a solid state relay mounted version.

Having tuned a lot of heating PID loops, I question the need for a derivative term. Depending on the thermal load, temperature loops tend to be proportional gain heavy with a relatively small integral term depending on how well insulated your vessel is. A noisy temperature measuring device will give you hell with a derivative term.

My 2 cents.
Phase angle control... that's brilliant! You could potentially just use the arduino to PWM output the angle control without the need for a separate PAC box on the SSR.

Unfortunately, I think that i'm stuck using propane due to my inability or lack of desire to build an electric infrastructure for brewing in my (rented) house.

So, given all that, I'm still thinking servo control of an adjustable regulator is the way to go. I'll order one of these
http://www.ebay.com/itm/330833209709?ssPageName=STRK:MEWAX:IT&_trksid=p398 4.m1438.l2649
this weekend
and 3 of these
http://www.adafruit.com/products/374
or would these be better
http://www.adafruit.com/products/165

I suspect the digital sensor is better since I can put 3 on a single pin, but it requires the 1wire library, which will take up precious memory on the arduino.
Any thoughts on the tradeoff between physical pin space and memory space for the 'duino?


__________________

In bottles: Last Day IPA, AAPL American Pale Ale
On deck: Some sort of porter type thing

The answer to the ultimate question of life, the universe, and everything? 42. The number of bottles in my first batch of homebrew? 42. Coincidence? I think not.

jrubins is offline
 
Reply With Quote Quick reply to this message
Reply


Quick Reply
Message:
Options
Thread Tools
Display Modes


Similar Threads
Thread Thread Starter Forum Replies Last Post
Brewtarget 1.2.3 - Free Open Source Brewing Software rocketman768 Brewing Software 260 04-11-2013 04:36 AM
Brewtarget 1.2 - Free open source brewing software rocketman768 Brewing Software 294 02-22-2013 05:48 AM
Open Source Brewing PowellAvenueBrew General Beer Discussion 20 09-21-2011 04:30 PM
New free open source brewing software rocketman768 General Beer Discussion 23 01-19-2009 11:01 PM
Open Source Beer Project rbenvin Recipes/Ingredients 3 06-04-2007 09:27 PM