Homeroasters - Home Roasting Coffee Community » DATALOGGERS - CONTROLLERS - RATE OF RISE METERS » Dataloggers/Controllers/Rate of Rise Meters
Who is here? 1 guest(s)
 Print Thread
Wonky ROR in Artisan when logging roasts
roastedgourd
Top  #1 Print Post
Posted on 06/17/2015 3:55 PM
I'm having trouble getting an accurate reading of BT ROR during roasting. I'm using a solid drum Huky with the Center 301 datalogger plugged into Artisan. During the roast, the ROR curve on the screen fluctuates wildly, with seemingly no rhyme or reason to its gyrations. The actual BT reading is nice and smooth, following the expected trajectory. After I end the roast session, however, the ROR curve smooths right out and follows the expected curve. I'm running Artisan 9.3. I assume I'm just missing a setting, but I can't figure it out. Any ideas?
-Ian
 
ginny
Top  #2 Print Post
Posted on 06/17/2015 4:35 PM

Quote

the ROR curve smooths right out and follows the expected curve


maybe your beans are just what it expects.

hardware/software/add ons...

what are you using here.


-g
 
roastedgourd
Top  #3 Print Post
Posted on 06/17/2015 5:46 PM
Not sure I follow... I listed my setup above, but here's a more concise list:

Huky 500
Stock type k thermocouple for BT
Center 301 data logger with usb connection to Artisan v9.3 for windows

What am I missing?
-Ian
 
ginny
Top  #4 Print Post
Posted on 06/17/2015 7:49 PM
Ian,


seems like I missed the info in your initial post but not sure if it makes a difference since I do not know what you are roasting.

it does make a difference.

let us know.


ginny
 
allenb
Top  #5 Print Post
Posted on 06/17/2015 10:00 PM
Ginny, bean temp is tracking fine. His issue has to do with the way Artisan is displaying RoR on the graph during the roast and doesn't have anything to do with the way his roast and the real RoR is progressing. I'm guessing it's as he is suspecting and is most likely a sampling time issue that needs to be adjusted in Artisan.

Allen
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
 
ginny
Top  #6 Print Post
Posted on 06/18/2015 6:23 AM
oh, ok guess I could not read the chart well either...


back t bed for me.


-g

party
 
JackH
Top  #7 Print Post
Posted on 06/18/2015 7:08 AM
I am not familiar with the Artisan application, but I found this post that may help with smoothing the Ror curve in Artisan:

http://forum.home...post_48758
Edited by JackH on 06/18/2015 8:28 AM
---Jack

KKTO Roaster.
 
roastedgourd
Top  #8 Print Post
Posted on 06/18/2015 8:09 PM
Thanks folks. I made some progress by piecing together info from several different sources. Curves are looking better.
-Ian
 
allenb
Top  #9 Print Post
Posted on 06/18/2015 8:19 PM
That's good news! Please let us know what you did to improve it.

Allen
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
 
roastedgourd
Top  #10 Print Post
Posted on 06/18/2015 8:33 PM
I set the sampling rate to 3 seconds instead of 2 and enabled oversampling. I tried adjusting smoothing, but it seems to only adjust the curve postroast. It seems that the issue (if I understand it correctly) is that the probe is subject to picking up lots of "noise" which is expressed as seemingly wild fluctuations in temperature on the ROR graph during the roast. After the roast, the curve is averaged out and smoothing can take place, which enables a more helpful curve. While it's not perfect, and sampling every 3 seconds allows more to happen in the roast than I'd prefer, I think once I get used to it, I'll be able to do what I want to. It's entirely possible that I'm still missing something that could make it better, so if there are any Artisan users out there, I'd still love your input. I've used Cropster and Roastlog and neither one of those platforms have this issue, so I know it can be done, but perhaps you get what you pay for. smile
-Ian
 
MaKoMo
Top  #11 Print Post
Posted on 06/19/2015 1:29 AM
You get what you contribute too!

Artisan implements two different smoothing algorithms. One is applied during recording that cannot know the future and the other one is applied in all other cases. The issue of not knowing the future is that there is no algorithm on this planet that does not has the negative effect to introduce a delay/shift of the signal (without knowing that future). Both algorithms can be configured under menu Tools >> Extras. "Smooth Curves" for the temperature curves, "Smooth Deltas" for the delta curves. Note that smoothing of the temperature curve does improve the smoothness of the delta curves as those are computed from the temperature data. Note further, that during recording there is no smoothing applied to the temperature curve, but only to the delta curves.

Due to the (negative) effect on quantification (the time signal has usually a way higher resolution than the temperature ones), an increase of the time interval (which results in an "increased" resolution of the temperature signal) the RoR is computed from has a positive effect on the smoothness. So larger sampling intervals or an increase of the "Delta Span" the RoR computation is based on has a positive effect.

However, one has always to balance. Filtering always results in information loss (by theory) and in the "not-knowing-the-future" case even in a shift in time. The more filtering the higher that effect.

I argued somewhere else that all this filtering in software should only be applied as last resort. It is way better to minimize the noise on the measuring system in the first place. This can be done by all kind of methods, but one is to use larger probes that react slower. Those act as a kind of physical filter very early (and therefore very effective) on in the measuring chain.

So using very thin probes to reach this high reactivity that I some times read is to me completely nonsense. Also the trend to run for close-to real-time sampling intervals does not make much sense to me for a process like roasting where those 3sec +- do not make any difference considering the hugh delay the thermal mass of those machines anyhow have.

Anyhow, if you need further support please choose one of the commercial roast-logging options
Marko
 
Turingalad
Top  #12 Print Post
Posted on 06/23/2015 9:04 PM
What does "Config > Oversampling" do?

(Dumb question) what are Delta ET/BT? Are they just other options for datalogging from a 3rd and 4th probe?

I'm a newb Artisan user. I wish I had the $$ to burn through a sack of beans figuring out how to use the software. :-P
 
rgrosz78
Top  #13 Print Post
Posted on 06/23/2015 9:38 PM

Quote

Turingalad wrote:

What does "Config > Oversampling" do?

(Dumb question) what are Delta ET/BT? Are they just other options for datalogging from a 3rd and 4th probe?

Not sure about Config > Oversampling.
The Deltas are NOT for additional thermocouples. Delta ET shows the change over time in the reading from the ET thermocouple. Delta BT shows the change over time in the reading from the BT thermocouple.
Life is too short to drink bad wine ... or bad coffee!
 
Web
Jump to Forum: