TC4 roaster monitoring and controlling

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
User avatar
farmroast

#1: Post by farmroast »

Just about a year and a half ago I asked if it was possible to create some custom hardware to measure the BT (bean temp.)RoR (Rate of Rise/min.) JimG built a nifty circuit that would do so using a DMM. Since then a flurry of activity has yielded some incredible custom hardware, arduino based TC4 and a variety of software, bourbon, kona, gesha and Artisan. Each with different abilities. At present I'm just using monitoring hardware and software.

I recently have my TC4 up and running using the gesha program. This program monitors time, BT RoR, voltage output of my variac to the heating element, BT and ET. It incorporates a stand alone lcd display and also connects to my pc creating graphs and csv data files.

I've done a few roasts using this setup and find it amazingly useful. There will continue to be improvements and enhancements made and really appreciate all the efforts made to develop these tools for home roasters.
links:
TC4 discussion-development thread
kona: An Arduino PID roasting program
TC4 and software
Artisan beta roasting software


Here are a few pictures.

The Arduino TC4, 16x2 lcd display and "juice box"(measures voltage from variac)


Not the best driving due to fiddling with things but shows how detailed the BT RoR can be. The voltage(juice) readings are a little off at points, work in progress.


The whole setup on my DreamRoast homebuilt roaster. The hardware will soon be put into a project box.
LMWDP #167 "with coffee we create with wine we celebrate"

User avatar
rama

#2: Post by rama »

Nice setup Ed, I'm jealous of your 'juice' instrumentation.

I'll second that JimG's TC4 is super useful, as I'm sure the others mentioned are too (although I haven't tried them). I'm a data junkie. If you're not capturing BT over time, you'll have a really hard time creating reproducible results and knowing what changes resulted in what.

Can you do this with a simple BT probe, stopwatch, and clipboard? Yup- I did just that for a year. It turned roasting into a chore. But if you're put off by the seeming complexity of the TC4 and others, I'd suggest you at least start with this. It'll make you appreciate the TC4 even more when you eventually decide to build one. :)

User avatar
arriflex

#3: Post by arriflex »

Ed, if you had to pick two voltages to run your heating elements at, what would they be?

Your dreamroast is a great model to follow, I have acquired an 1800 watt convective fan and element assembly out of a residential oven that I hope to have two heat settings on. Assuming high is at line voltage, the real question is how much below line voltage to go for when you find yourself turning down the Variac or do you believe the Variac is a necessary tool to adjust the voltage in a more analog fashion?

Are you still liking the tc4 configuration? I am anticipating using a similar TC configuration to yours, any advice is welcome.

OffTopic:
Regarding agitation, would you recommend a flat or curved bottom pan? Thick or thin? Are you still happy with the 200 to 300 rpm I think I remember reading somewhere?

Thanks for your thoughts (or anybody's for that matter)!

arri

User avatar
doubleOsoul

#4: Post by doubleOsoul »

Farmroast, I'm always in serious awe with all the rigs you come up with.

OO

randytsuch

#5: Post by randytsuch »

arriflex wrote:Ed, if you had to pick two voltages to run your heating elements at, what would they be?

Your dreamroast is a great model to follow, I have acquired an 1800 watt convective fan and element assembly out of a residential oven that I hope to have two heat settings on. Assuming high is at line voltage, the real question is how much below line voltage to go for when you find yourself turning down the Variac or do you believe the Variac is a necessary tool to adjust the voltage in a more analog fashion?

Are you still liking the tc4 configuration? I am anticipating using a similar TC configuration to yours, any advice is welcome.

OffTopic:
Regarding agitation, would you recommend a flat or curved bottom pan? Thick or thin? Are you still happy with the 200 to 300 rpm I think I remember reading somewhere?

Thanks for your thoughts (or anybody's for that matter)!

arri
Another option is Catuai, is briefly described here
http://homeroasters.org/php/forum/viewt ... ad_id=2222

I know Jim G uses Catuai to roast with. It basically allows you to use a pot input to control a PWM output from the TC4. This output is connected to a SSR, which you would add in series with the AC hot to your heater. Jim likes this better then a standard PID control for roasting, and I plan to try it sometime and see how it works. For 1800 watts, you need a pretty large SSR, and then make sure you heat sink it.

Randy

User avatar
arriflex

#6: Post by arriflex »

I considered controlling the wattage via a sort of low frequency PWM via any of the various setups on an Arduino and most certainly plan to leave that option open through a "Hand off Auto" switch on the element. 25 amp ssr's (rated with a heatsink) are readily available at reasonable prices. That could be a "cleaner" option than the transformer I was considering using for voltages other than 120.

Do any of the software packages have the ability to manually drive the output? I could easily pot one of the analog inputs on the Arduino for a nice dial controller...

randytsuch

#7: Post by randytsuch replying to arriflex »

The Catuai program that I mentioned allows you to connect a POT to an analog input on an Arduino board, to control the PWM output available when you add a TC4 to the Arduino. Note that I have never used Catuai, I just know some of what it does.

JimG also wrote a low frequency PWM program, which I am pretty sure is part of Catuai. It ranges from 1/4 to 64 hz output frequencies.

Randy

User avatar
farmroast (original poster)

#8: Post by farmroast (original poster) »

arriflex wrote:Ed, if you had to pick two voltages to run your heating elements at, what would they be?

Your dreamroast is a great model to follow, I have acquired an 1800 watt convective fan and element assembly out of a residential oven that I hope to have two heat settings on. Assuming high is at line voltage, the real question is how much below line voltage to go for when you find yourself turning down the Variac or do you believe the Variac is a necessary tool to adjust the voltage in a more analog fashion?
I don't think you'd be happy with just 2 settings. I have been using a variac from the beginning but what Randy is talking about should certainly be considered as an alternative to a variac.
You should also realize that added wattage may not mean you can roast larger batches unless the heat transfer abilities of your roaster can utilize it without your MET (maximum environmental temp. hitting the beans) becoming too high. This will correlate with the size of your roast chamber, agitation, convection, wattage, batch size, etc.
Are you still liking the tc4 configuration? I am anticipating using a similar TC configuration to yours, any advice is welcome.
It works great for me! I work with JimG and Bill and they created exactly what I was looking for. Though again what Randy speaks of sounds like a really neat and similar option if one is starting from scratch.
OffTopic:
Regarding agitation, would you recommend a flat or curved bottom pan? Thick or thin? Are you still happy with the 200 to 300 rpm I think I remember reading somewhere?
The concept of my roaster is a bit different from a SC/TO. A SC/TO stirs the beans that lay on the bottom of the Stir Crazy. The stir mechanism must churn the beans/bean mass to allow exposure to the heat source and do it evenly to each bean. Efficiency is limited by the beans being massed where convection is unable to easily reach the bean surfaces consistently. My design lifts the beans and whirls them around elevated around the sides of the pot at a higher speed. The higher speed increases convection and increases the churning of the beans. This increases the exposure to heat and heat transfer ability and the higher speed increases the replenishment of heat(increased convection).

The concept of my design is similar to the large high tech, industrial roasters such as the Probat Saturn Series. I like my design better than the Saturn because my heat out of the TO top flows down the sides an d up the middle where the Saturn works the other way. In the Saturn the heat is concentrated in a smaller flow that's deflected with the cone at the bottom with more risk of high MET exposure and reduced efficiency.

For my design around 200-220rpm creates the best splay of beans and heat transfer. I uses a speed control motor with the appropriate gearhead ratio made by Oriental Motor. In a typical drum roaster rpms are limited by centrifugal force that eventually just glues the beans to the drum walls. This is why the Loring drum roaster was created with a fixed horizontal drum and moving paddles that allow higher speed agitation and increased convection.

The shape of the pot is important for the bean mass pattern achieved by the bean bats. The bats whirl around striking the beans as gravity drops them down to the bottom. If the sides are sloped the beans will be batted to the slope and fly too far up to the top. I chose a square bottom that allowed an even splay along the sides without shooting up to the top.

I went with a thin walled pot so as to not hold too much heat when I need to slow the roast around 1st crack. The insulation keeps from losing too much heat or having a cooling pot surface when my room temp. might be low.

Roaster design and dynamics are quite complicated. This is what makes some commercial roasters better than others. Off the shelf home roasters tend to be simple to keep cost down. For example, I saw the fins to be lacking in a hottop limiting efficiency. When users improved the fin size and arrangement the efficiency and roasting evenness improved.

It took me alot of thinking about what is happening in the roaster and to the beans and adjusting everything from trial and error to find the sweet spots.
hope this helps get a start,
farm

PS I'm thinking it might be useful to have a thread on "roaster dynamics" with group contributions. This would help for those thinking of designing/building a roaster or modifying an off the shelf roaster.
LMWDP #167 "with coffee we create with wine we celebrate"

User avatar
arriflex

#9: Post by arriflex »

Thanks for your thoughts, Ed; I started a new topic with that in mind: Roaster Dynamics

I suspected that a two speed heater wouldn't really be ample for repeat ability, thank you for affirming my thoughts.

Regarding drum/pot shape, your early posts on choosing higher speed agitation were inspiring to me. So much so that I've even considered a pot that tapers slightly inward at the top like a Turkish shape might. Anyway, I'll leave further discussion to the "dynamics" thread.

arri

JimG

#10: Post by JimG »

arriflex wrote:Do any of the software packages have the ability to manually drive the output? I could easily pot one of the analog inputs on the Arduino for a nice dial controller...
As Randy said, the Catuai application provides this function. As currently configured, Catuai expects to receive a 0-5VDC signal on the ANLG1 port of the TC4. The ANLG1 port consists of a 3-pin header that was designed to be connected to the 3 pins of a 10K pot (middle pin goes to the swipe).

Depending on the voltage on the middle pin of ANLG1, a 1 Hz PWM duty cycle of between 0 and 100% is selected and output on OT1 of the TC4 board. This is all intended to drive a heater through an SSR using digital manual control.

Optionally, a second pot can be similarly connected to ANLG2 on the TC4 board. By default, the Catuai application uses ANLG2 to control a 490Hz PWM output on the I/O3 port of the TC4. Since my roaster (Hottop) has a DC fan, I use ANLG2 to drive the fan through a Darlington transistor and 12V PSU.

With a trivial change in the source code, ANLG2 could instead be tied to OT2, which is another 1 Hz PWM output designed to drive a second SSR.

Catuai interfaces very well with pBourbon, a Processing language program for real time plotting. When connected to a TC4/Arduino running Catuai, pBourbon will plot ET, BT, BT-RoR, heater output, and fan output on Linux/Windows/OS-X.

Catuai can also be used as a standalone controller with a 16 x 2 LCD panel attached either to the TC4's I2C header or its LCD header.

Catuai, along with pBourbon and necessary libraries (all open source under the BSD license), can be downloaded here:
http://tc4-shield.googlecode.com/files/ ... REL100.zip

Jim