Control for a roaster

Roasters and roasting

Moderators: GreenBean, Gouezeri, bruceb, CakeBoy

Control for a roaster

Postby Neo » Sun May 25, 2008 5:48 pm

Yes,instead of contributions I have again another question.
I'm planning a micro-roaster, something like Joris' but mine will be a semi. The problem is that I dont really know how to control it. I was planning a PID+SSR+electromagnetic valve to make the temperature curve steady and a straight line. I've checked some products and the price is way too high for me.
I considered PLC, which is expensive, so no luck. The PIDs allowing changing set points are pretty expensive. I wonder if someone can suggest some cheap PIDs that can do the job or some alternate methods?


Edit: I once though about using an I/O port of my computer to send signals to a normal PID and use to signals to change the set point instead of pushing buttons. It sounds not that possible ,right? My knowledge on electronics is pathetic :oops:
Anfim supercaimano
Gaggia Baby
Cloth filter
Syphon
Lovely coffee from a great roaster
User avatar
Neo
 
Posts: 618
Joined: Mon Sep 10, 2007 5:45 pm
Location: Hong Kong

Postby JonS » Mon May 26, 2008 10:48 am

Depends how good your software / electronics knowledge is to be honest. If you can write software, then I would consider a combination of the thermocouple and interface chip from this article and an arduino to make a standalone (or computer driven) control board.

There's also some good starting code for PID-like control for arduino, here

Obviously, piecing all this together is a big task, but doable if you have the appropriate skills, or a willingness to learn them along the way.

Alternatively, depending on the design of the roaster, is this level of control what you really need ? I would consider starting off full-manual and seeing how different roasts actually behave. Total repeatability from roast to roast, with different ambient conditions, is actually harder than it seems (imho)

I think the trouble with a lot of this, is that it can end up detracting from the process itself. Automating roasts seems like a cool idea, but more and more I find myself patiently watching over every roast, keeping notes about what's going on, and generally immersing myself in the process. Embark on a project like this, and you might find you spend more time tweaking the control system for its own sake, than you do simply roasting for pleasure and enjoying the results ! I'm not trying to put you off the idea, though :)

Jon
Iberital L'Adri 1grp semi-auto | Macap MXKR Auto | Mazzer Super Jolly | Hottop Programmable
Isomac Zaffiro w/PID | Isomac Gran Macinino | Mazzer Mini
Scace Thermofilter
Custom-built grinder timer
JonS
 
Posts: 488
Joined: Wed May 09, 2007 12:18 pm
Location: Chesterfield

Postby dsc » Mon May 26, 2008 11:06 am

Hi Neo,

PID on it's own is not a complicated algorithm, so it shouldn't be too hard to program it. To make a working PID controller on your own is a different thing though and I wouldn't go that way if I were you. Of course you can use programmable uC, but you will need some control, electronic and programming knowledge to do it. In the end you will probably spend more than you would when buying a reasonably priced PID, so think twice before you start your project.

Depending on the roaster temperature profile you can use a simple On-Off histerisis algorithm, which is pretty much what a thermostat does. It all depends how fast the chamber looses temperature and if that's quite slow you can use your heater to occasionally turn on to keep the temps above a certain level (and below a high level as well).

As you said PLCs are expensive and you need a bunch of stuff to start, PSU, main unit and probably an I/O module to plug in the TC/temp probe.

You can buy PIDs for around 60$ from Auber, throw in a fairly good TC, which should be around 15-20$ and you can put the whole together for less than 100$ probably.

Cheers,
dsc.
"This week I are mostly...trying to build a grinder..."

Mahlkonig Guatemala for FP/drip, Elektra T1 for sale, Gaggia Classic in progress
User avatar
dsc
 
Posts: 1753
Joined: Fri Feb 02, 2007 11:43 pm
Location: UK

Postby Neo » Mon May 26, 2008 11:32 am

Thanks for the input. I am not actually making an automated roaster, but instead i wish to use the PID to help me control the heat, hence allowing me to fine-tune in some stages like the exo- and
endothermic ones.
I will look at the articles and think more thoroughly.
Anfim supercaimano
Gaggia Baby
Cloth filter
Syphon
Lovely coffee from a great roaster
User avatar
Neo
 
Posts: 618
Joined: Mon Sep 10, 2007 5:45 pm
Location: Hong Kong

Postby JonS » Mon May 26, 2008 2:30 pm

dsc makes a fair point for the other end of the spectrum, especially if you are not looking to actually automate the roast in its entirety.

I have to throw in this article as an old, but interesting discussion of roast profile control.

Common experience and the more usual attempts at roaster control centre around electric roasters. These tend to need powerful elements, with a large thermal inertia such that they take a while to heat up in response to a control signal. They also are still propagating heat when they are turned off, so can overshoot. PID control is ideal for this kind of scenario, either to maintain a fixed temperature, or control a ramp.

A gas roaster is a somewhat different proposition as far as I can tell and I personally don't have any experience with one. I think building your roaster and spending some time understanding how it behaves under manual operation, using only bean mass and possibly drum ambient air thermometry will give you the knowledge you need to pick the right control system.

Jon
Iberital L'Adri 1grp semi-auto | Macap MXKR Auto | Mazzer Super Jolly | Hottop Programmable
Isomac Zaffiro w/PID | Isomac Gran Macinino | Mazzer Mini
Scace Thermofilter
Custom-built grinder timer
JonS
 
Posts: 488
Joined: Wed May 09, 2007 12:18 pm
Location: Chesterfield

Postby Neo » Mon May 26, 2008 3:37 pm

Do you think a 200g drum will be too small? I dont want to use a large one because I hope to keep the roasted beans fresh and to save some money in the trial period -.-
On the heating method, I chose gas over electric heater is also due to a quicker response. It might be a bit difficult to control at first, without PID, but I might want to worry later. :P
Anfim supercaimano
Gaggia Baby
Cloth filter
Syphon
Lovely coffee from a great roaster
User avatar
Neo
 
Posts: 618
Joined: Mon Sep 10, 2007 5:45 pm
Location: Hong Kong

Postby Neo » Mon May 26, 2008 3:56 pm

Oh btw, that article helped quite a bit. At least now I know the english name of Ramp/soak PID :D I know there's such kind of thing out there but couldnt figure out the name :) Bad english hehe. Any ideas on using such a PID, which is rather unsuccessful in the article?
Anfim supercaimano
Gaggia Baby
Cloth filter
Syphon
Lovely coffee from a great roaster
User avatar
Neo
 
Posts: 618
Joined: Mon Sep 10, 2007 5:45 pm
Location: Hong Kong

Postby JonS » Tue May 27, 2008 10:19 pm

Off the top of my head, I think you would aim to use a PID with multiple ramp/soak segments, and you would setup each segment with a ramp time and target temperature, to achieve a steady ramp up of temperature (the ramp part of each segment) as each segment progressed. I'm not sure you'd need to make use of the soak part.

The problem as I see it (and, with the benefit of using a Hottop programmable for some time now, by means of a comparison) is that this level of pre-determined time/temperature ramping can end up being somewhat counter-intuitive.

Let's say you want to steadily ramp up to 1st crack, then you want to keep it ticking over through first, and get a nice pause before 2nd crack, then drive it through second to the dump point. Time and target temperature are hard to adjust accurately, primarily because different bean moisture, different beans, humidity and ambient temperature will affect how long it takes the beans to reach a given state. I think that's where the author of that article is coming from. Sometimes, you want to progress to the next phase of the roast, when the beans reach a certain state, not necessarily after a given time.

This doesn't mean that you can't use a multi-segment ramp/soak PID to drive a roast, it's just that I think you'll get less control than you'd ideally like.

With a gas roaster, I would think you would have the benefit of lower thermal inertia and power in reserve for raising the temperature on demand. With an electric element, turning off the element because you've reached the target temperature is not ideal, because it will take a while to come back on when you decide you need to heat again. Consequently, I see a lot of people setting the hottop programmable up to basically under-achieve and never reach the target temperature for each segment, so that the element is always on to some degree. This is how I use mine, too.

One side issue is the difference between drum air temperature and bean mass temperature. The former is probably more easy to use as the control loop for applying heat, the latter gives a better indication of the progress of the roast. If you use the PID, you probably limit yourself to control via one or the other, but not both.

Like I said, off the top of my head (and I am pretty frazzled today). But, it very much depends on the open-loop behaviour of the roaster.

I can't really comment on the 200g part, to be honest, other than you'll end up with maybe 10-15% less than that post roast. That may suit you fine, as long as you are happy to make the time to roast. I enjoy roasting, but I have to make a concerted effort to roast 3 times a week in my hottop, which has a 250g batch.

Jon
Iberital L'Adri 1grp semi-auto | Macap MXKR Auto | Mazzer Super Jolly | Hottop Programmable
Isomac Zaffiro w/PID | Isomac Gran Macinino | Mazzer Mini
Scace Thermofilter
Custom-built grinder timer
JonS
 
Posts: 488
Joined: Wed May 09, 2007 12:18 pm
Location: Chesterfield

Postby Neo » Wed May 28, 2008 7:55 am

What I wish to achieve is a straight line, ideally. There might be a few curves but that's inevitable since different beans have different content that I need to get used to. From what I've heard/learned, flavour profile is determined by evaporation of different kinds of compounds. Then, with a straight line, it's possible to express the original flavour of a certain bean. Of course, if you soak at a certain part, a flavour may be enhanced or diminished.
Speaking of temperature control with PID, I prefer taking bean mass temperature as the variable to be observed. Besides, I found out that there's a ramp function in my current cheapish PID, but the details have to be confirmed.
I understand that there will be a loss in weight, however as a student, I cant afford to make a large drum or dump too many beans in failed roasts. 50g is not a big problem, but I will talk with my friend to get an agreement.
Thanks for the input.
Anfim supercaimano
Gaggia Baby
Cloth filter
Syphon
Lovely coffee from a great roaster
User avatar
Neo
 
Posts: 618
Joined: Mon Sep 10, 2007 5:45 pm
Location: Hong Kong

Postby grib » Thu May 29, 2008 8:08 am

I don't know if its helpful but I've cam across this recently. It's a microcontroller with 9 digital inputs & outputs and 6 analog, it also has an ethernet port and runs a builtin webserver. It also has temperature/humidity/pressure sensors as an addon. That means, with very little web programing you could setup profiles which are controlled via a web page, which monitor and adjust your temperature and timing. I'm planning on using this on a drip coffee machine I'm building and I know (for me anyway) writing web type code is a lot easier than full on C programing.

Andrew

PS Did I mention it's dirt cheap? Only €40
User avatar
grib
 
Posts: 120
Joined: Wed Aug 15, 2007 7:48 am
Location: Dromore, Northern Ireland

Postby ivdp » Thu May 29, 2008 9:41 am

Controling heat means in most cases, lowering the heat.
This could be done by letting heat escape through a valve in/on the roasting chamber.
KvdW Idro - M K30
ivdp
 
Posts: 384
Joined: Sat Sep 25, 2004 4:56 pm
Location: Lieren, The Netherlands

Postby Joris » Mon Jun 02, 2008 5:31 am

Intresting thread, keep that data coming ;)
Veni, Vidi, Imbibi Coffea !


Bezzera Lever Machine (customized)
Doserless Mazzer Major
"Juggernaut" Homebuilt Roaster
Aeropress, Vacpots, presspots and mokkapots
User avatar
Joris
 
Posts: 1786
Joined: Wed Dec 14, 2005 1:56 pm
Location: The Hague, Netherlands


Return to Roasting - Equipment and Techniques

Who is online

Users browsing this forum: No registered users and 21 guests