topbanner.gif
Login
Username

Password




Not a member yet?
Click here to register.

Forgotten your password?
Request a new one here.
Shoutbox
You must login to post a message.

renatoa
08/10/2022 1:56 AM
nguyencoffeesupply
and RoyB, Welcome

renatoa
08/06/2022 1:31 PM
Welcome, oak202, CoffeeNutZ and Mlcharlestonsc

Strangeworth
08/04/2022 9:34 PM
Thank you!

renatoa
08/03/2022 2:09 AM
Strangeworth and BigPalm, welcome to forum

Tamesino
08/02/2022 12:01 PM
Thank you guys and gals !! Exited to be here greenman

In Memory Of Ginny
Donations

Latest Donations
dmccallum - 10.00
JackH - 25.00
snwcmpr - 10.00
Anonymous - 2.00
Anonymous - 5.00
Users Online
Guests Online: 13

Members Online: 0

Total Members: 7,627
Newest Member: nguyencoffeesupply

View Thread

Who is here? 1 guest(s)
 Print Thread
The well tempered roaster
allenb
My point was that one should avoid large changes in air flow that result in slow minimal spouting at one point in the roast and towards the end, blasting them towards the top of the roast chamber in order to keep ET low enough to keep rate of rise in check. Most fluidbed designs force one to go to these extremes in air flow if maintaining a constant heat input but if you're able to get through a roast without this occurring then all should be good. All of the ones I've built or used required extreme variation if heat input was fixed.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
 
georgemvg
I am frustrated. I found the topic interesting, though I do not know if my question is related.
I am very familiar with controlling water temperature with a pid.
Though, coffee roasting with a fluidbed roaster, has two "items", one is the air and the other are the beans.
As you rise the ET, them the BT comes along and follows the ET, with an offset that tends to become zero, as roasting cycle comes to end.
What is the point of a pid, except from a preheating cycle?
You cannot set a stable desired ET and leave the beans to roast.
I have no roasting experience, but I am trying to understand the machine fundamentals to build my own roaster.
If you have two identical BT graphs, with identical bean density and moisture, wouldn't the roasted result be the same?
 
renatoa

Quote

georgemvg wrote:
...
1. Though, coffee roasting with a fluidbed roaster, has two "items", one is the air and the other are the beans.
2. As you rise the ET, them the BT comes along and follows the ET, with an offset that tends to become zero, as roasting cycle comes to end.
3. What is the point of a pid, except from a preheating cycle?
4. You cannot set a stable desired ET and leave the beans to roast.
...
5. If you have two identical BT graphs, with identical bean density and moisture, wouldn't the roasted result be the same?


1. Not always both are variables... there are various approaches. Some people could perform a roast keeping airflow constant the whole cycle, others the opposite, set a fixed power and drive the roast from airflow only. And both succeed.

2. Never the offset will be zero, even close... for a FB. To have an average of 5C degrees RoR for BT through the dev phase, ET have to be at least 10-20 C degrees higher than BT.

3. Even the preheating can be done without PID... finding by experiments the power level for your preheat target. You start with 100%, then lower to say 40% for 180 C, when you are at 170 C. And wait stabilize. Could be faster than a an out of tune PID.
But many people roast with FB without any preheat, there is no need for such big power focused in such small volume.
The 100 grams machines can reach 200C in a minute starting from ambient with less than 50%, if airflow is right.

4. This thread subject establish that you have at least 2 temperatures of interest during a roast.
Starting directly with the higher level could be too much for the beans. Conversely, the lower level could be not enough to drive them to FC.

5. Yes, this is the point of profile following, but so far the people seems mesmerized by the BT profiles, requiring very precise beans measurement. Very few ventures into mastering roasting based on ET profiles, and even less into power profiles.
Especially the last is the most recommended approach for a FB machine, imo.
For this purpose a new "PID" type was been introduced, called him "proportional on setpoint", that obtain heater power directly from BT profile data, without measuring anything, and no error estimating. And it works pretty accurate...
 
georgemvg

Quote

renatoa wrote:

Quote

georgemvg wrote:
...
1. Though, coffee roasting with a fluidbed roaster, has two "items", one is the air and the other are the beans.
2. As you rise the ET, them the BT comes along and follows the ET, with an offset that tends to become zero, as roasting cycle comes to end.
3. What is the point of a pid, except from a preheating cycle?
4. You cannot set a stable desired ET and leave the beans to roast.
...
5. If you have two identical BT graphs, with identical bean density and moisture, wouldn't the roasted result be the same?


1. Not always both are variables... there are various approaches. Some people could perform a roast keeping airflow constant the whole cycle, others the opposite, set a fixed power and drive the roast from airflow only. And both succeed.

2. Never the offset will be zero, even close... for a FB. To have an average of 5C degrees RoR for BT through the dev phase, ET have to be at least 10-20 C degrees higher than BT.

3. Even the preheating can be done without PID... finding by experiments the power level for your preheat target. You start with 100%, then lower to say 40% for 180 C, when you are at 170 C. And wait stabilize. Could be faster than a an out of tune PID.
But many people roast with FB without any preheat, there is no need for such big power focused in such small volume.
The 100 grams machines can reach 200C in a minute starting from ambient with less than 50%, if airflow is right.

4. This thread subject establish that you have at least 2 temperatures of interest during a roast.
Starting directly with the higher level could be too much for the beans. Conversely, the lower level could be not enough to drive them to FC.

5. Yes, this is the point of profile following, but so far the people seems mesmerized by the BT profiles, requiring very precise beans measurement. Very few ventures into mastering roasting based on ET profiles, and even less into power profiles.
Especially the last is the most recommended approach for a FB machine, imo.
For this purpose a new "PID" type was been introduced, called him "proportional on setpoint", that obtain heater power directly from BT profile data, without measuring anything, and no error estimating. And it works pretty accurate...


3. Μy main question is about pid. What is the point of all these poppers pid upgrades I see, if you cannot set a desired temperature, and you manually control airflow and power? For the initial sampling roasting?
When you set the power, either by a knob or by computer, lets say at 60 percent, the power is at 60 percent. No changes from pid there...
A common temperature graph of a pid control, is the effort of the pid to maintain as best as possible a temperature that is given to it.

5. Can you elaborate on that?
 
renatoa
As I wrote, the pid is necessary for the "follow the profile" dominant mantra.
Ask them why they don't limit the upgrades to the bare necessities required for a right roast... maybe the temptation of big boys complicate toys... obsession of control Grin the more the knobs, the happier I am... Grin

My perception is that not all the guys are jumping right to the most complicate update especially for the pid reason... if you mean the TC4, most popular upgrade solution, it's only a coincidence, the pid is there as a bonus, not as main attraction... only some of the sketches done for TC4 are featuring a pid, and a lot of people is not using it... imo.

If you mean the plastic black boxes, known as pid temperature controllers, they aren't used in FB roasters, they are too slow.
Their realm is rather in boilers, or big machines, with 30 seconds response time.

If you mean elaborate on proportional on setpoint control, please check the attached graph. This roast is done with such control method.
Basically the lower red graph line, the heater power, is derived directly from the desired BT graph, using a single proportionality factor, and not related with any actual temperature at all. There are no integrative, nor derivative corrections, so this approach can't be called a PID.
Sure, there was a measurement in that machine, as you can see, but is just for the record, not for control. And to check how close was the real roast to the background profile (the graphs with pale colors).

Using other wording, the power graph is a scaled down version of the BT graph, with a constant factor for the whole roast.
The formula of this control is below:
HTR% = MIN% + (BT-AT)/kP, where
MIN% is a minimum power at start, in this case 25%, visible on graphic too
BT is the desired beans temp value at a moment of roast
AT is ambient
kP is the process proportionality factor, value = 5 for this roast

For example, the power level at dry end, 150 C degrees, according to the formula above should be: 25% + (150-25)/5 = 50%
... and you can check from the graph it is exactly as this.

How I find the values of these two factors I wrote above, that controls the whole roast?
The proportionality factor can be very close approximate from the heater level where your machine stabilize empty at 200 C value, which was 48% for the machine use for the graph.
The minimum starting value is more empirical... I am using as starting point half of the power level for 200 C I wrote above.
It depends a lot of ambient temperature, the colder outside, the higher the MIN%
If you preheat the machine, then you can start with a much lower value, but not zero.

Obviously these values depends on your machine power and load... the above were for a 2000W machine and 100 grams load.
But if you find yours following the same logic, and you start a test roast this way, there are great chances to obtain a good roast, with DE/FC in reasonable time windows, i.e. 3-4 minutes for DE and 6-8 minutes for FC.
From there you can play easily with "profiles" taking into account that about 3-4% power percent changes shift the roast duration by about a minute.
renatoa attached the following image:
1216_cc_fb_pons_kp5_2_ki0_final_2.png

Edited by renatoa on 02/09/2022 8:48 AM
 
allenb
georgemvg wrote:

Quote

3. Μy main question is about pid. What is the point of all these poppers pid upgrades I see, if you cannot set a desired temperature, and you manually control airflow and power? For the initial sampling roasting?
When you set the power, either by a knob or by computer, lets say at 60 percent, the power is at 60 percent. No changes from pid there...
A common temperature graph of a pid control, is the effort of the pid to maintain as best as possible a temperature that is given to it.

5. Can you elaborate on that?


Ok, I'm sure as you've seen, temperature control and/or profile control with coffee roasting is taking a dive into the most confusing rabbit hole ever experienced if one does not already have this field in their background especially with the difficulties there are in trying to communicate the concepts clearly and succinctly.

To start off with your primary question, why use a fancy proportional/integral/derivative type control scheme if one will be manually setting various temperature setpoints for the hot air as the roast progresses?

Answer: Many will use 3 or 4 different escalating hot air setpoints as the roast progresses in a manual fashion. When a hot air setpoint is selected, the controller (whether it be Artisan and TC4 or a Fuji PXR PID ramp/soak controller as I use) will bring the hot air temperature up to that setpoint and hold it there without you needing to raise and lower the power level back and forth to try and keep it stable. This is why the need for PID, even when manually selecting hot air temperatures (no profile automation)

The reason for selecting 3 or more hot air setpoints during the course of the roast is to always keep a sufficient spread between hot air temp and bean temp to maintain needed rate of rise at the different phases of the roast.

Now, we can go one step fancier and use what they call ramp/soak profiles which Roastlogger (no longer supported) and Artisan can do. With this feature, you can preprogram the different hot air setpoints and how long to get from one setpoint to the next . With this feature, one is making a hot air temperature profile that the controller will automatically follow when you hit go. I used this feature with Roastlogger and worked very well. As renatoa pointed out, this does not use PID since it is not trying to bring under and overshoot down to a minimum and hold a setpoint constant. As each new setpoint is reached, it automatically ramps to the next without trying to hold any of them.

As renatoa mentioned, one does not have to have automated temperature control to run a fluidbed but it makes it much less of a challenge since ET will float all over the place without it. Having it also allows one to set up ramp soak profiles which is a cool feature and again, keeps you from chasing your tail trying to control things.

As renatoa mentioned, there's an option for power management through a roast using a proportionality factor which is a good way to go but requires one to do a deep dive into what makes their roaster tick which may be beyond what some will want to wade into if you aren't a committed control geek. Grin
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
 
allenb

Quote

renatoa posted:

If you mean the plastic black boxes, known as pid temperature controllers, they aren't used in FB roasters, they are too slow.
Their realm is rather in boilers, or big machines, with 30 seconds response time.


Curious as to what you are referring to as a "plastic black box known as PID temperature controller"? I and 100's of other roasters are using these with excellent results for PID control and ramp/soak profile control of a fluidbed roaster. 30 seconds response time? This may have been the case 40 years ago with some of them but even then, process control automation required immediate response times, not half a minute later. In fact, my most favored control scheme these days is my Fuji PXR4 PID ramp/soak controller as it allows easy setup without having to crank up my laptop and fiddle with lots of pages of selections and eliminates an error prone combination of PC and software.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
 
renatoa
Frankly, was just an enumeration of the possible meanings of the pid term our fellow georgemvg could refer in his post, when he wrote about "pid upgrades"
I don't remember any new build here in the last two years intending to use a "black box" pid controller, that why my erroneous conclusion.

Those cheap chinese models I know are using on-off slow PWM with two seconds cycle as default control scheme for the SSR heater.
I am not aware about units using phase angle or skip pulse techniques (known as ICC for TC4 users) No hands on experience with Fuji's, too expensive :(

Using such pulsing control for a FB is up to the user... how picky he is... but the variations in ET are there, visible and measurable... please check this video:
https://www.youtu...pCotdob-bU
...no idea about the effect on beans of this pulsing heat.
 
allenb

Quote

renatoa wrote:

Frankly, was just an enumeration of the possible meanings of the pid term our fellow georgemvg could refer in his post, when he wrote about "pid upgrades"
I don't remember any new build here in the last two years intending to use a "black box" pid controller, that why my erroneous conclusion.

Those cheap chinese models I know are using on-off slow PWM with two seconds cycle as default control scheme for the SSR heater.
I am not aware about units using phase angle or skip pulse techniques (known as ICC for TC4 users) No hands on experience with Fuji's, too expensive :(

Using such pulsing control for a FB is up to the user... how picky he is... but the variations in ET are there, visible and measurable... please check this video:
https://www.youtu...pCotdob-bU
...no idea about the effect on beans of this pulsing heat.


You are correct in that most of the low cost PID controllers (less than $100 USD) can't do a proportional cycle rate of less than 2 seconds and some will have a hard time forking out $250 to drop it to 1 second.

Another option is to use a low cost (less than $60.00 USD) PID controller that has 0-10vdc and 4-20 mA analog output
https://www.therm...uct_id=280

with one of these: https://www.wolfa...ate-relay/

It's output is phase angle control.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
 
Jump to Forum:

Similar Threads

Thread Forum Replies Last Post
My 1-2 KG drum roaster Drum Roasters 8 08/09/2022 11:57 PM
My U.A.F. hot air roaster (Dedicated to renatoa) Fluidbed Roaster 4 07/22/2022 3:03 PM
1-6 lbs Roaster Build - Nothing Fancy Drum Roasters 26 06/21/2022 3:52 AM
Waiting on my Roaster and beans. Roasting Coffee 4 06/21/2022 3:44 AM
Any roaster that doesn't trigger smoke alarm when used inside? Drum Roasters 3 06/20/2022 3:10 PM
Homeroasters Association Logo, and all Content, Images, and Icons © 2005-2016 Homeroasters Association - Logos are the property of their respective owners.
Powered by PHP-Fusion Copyright © 2022 PHP-Fusion Inc
Released as free software without warranties under GNU Affero GPL v3
Designed with by NetriX