BC600 roaster review  Page 13
Noted, thanks. I was thinking keeping it right next to me would allow me to cut off the gas supply immediately if I had a fire. Thanks for the heads up.CarefreeBuzzBuzz wrote:2 quick notes for you and others. Putting the propane outside is a wee bit safer. I have a 20 foot hose and leave my outside. On really cold days I'll bring it in the garage first thing on a roast day to warm up (garage is heated) and then back outside it goes.
On your Artisan use, you mentioned optimal smoothing post roast. If you select that, then you are overriding your Smooth Deltas settings.
Basically its an additional algorithm that works post roast as it looks back and forward and can only do that post roast. Your post roast curve won't look like your during roast curve if it is checked. You can check and uncheck it after a roast is done to see the difference.
See the descriptions of the settings at:
https://artisanscope.org/docs/curves/
Thanks Michael
I'll check out the optimal smoothing when I get home and post the nonoptimal smoothing versions of the same roasts I originally posted. Thanks for the link to the info!

 Supporter ♡
Plus over 9000 on this.CarefreeBuzzBuzz wrote:... 2 quick notes for you and others. Putting the propane outside is a wee bit safer. I have a 20 foot hose and leave my outside. ...
A few hours ago I went outside and checked my propane tanks out of boredom (this is what happens when you're retired) ... and smelled that chemical that they put in propane . One of the hoses was leaking where it joins the coupler to its tank. I had lifted the tanks on Tuesday to gauge how much propane was left after roasting, and I guess that was enough for that hose to start leaking. Hoses were approximately five years old.
Good thinking. It is good practice to have a gas cutoff valve located close  but not too close  to the roaster.mojoit wrote:Noted, thanks. I was thinking keeping it right next to me would allow me to cut off the gas supply immediately if I had a fire. Thanks for the heads up ...
 CarefreeBuzzBuzz
That's why I use quick disconnects and keep the hose inside. Anyone that wants my parts list send a DM.baldheadracing wrote:Plus over 9000 on this.
A few hours ago I went outside and checked my propane tanks out of boredom (this is what happens when you're retired) ... and smelled that chemical that they put in propane . One of the hoses was leaking where it joins the coupler to its tank. I had lifted the tanks on Tuesday to gauge how much propane was left after roasting, and I guess that was enough for that hose to start leaking. Hoses were approximately five years old.
 mkane
 Supporter ♡
My tank's right under my roaster and I roast in the garage. One day I might go out with a bang.
 CarefreeBuzzBuzz
Just like the rest of California!
Had a look at checking and unchecking the optimal smoothing box for some roast curves. No change in ROR curve, even after saving, closing Artisan, and reopening to check that the setting still stood. I wonder if checking/unchecking to see the difference in curve is only available before the .alog file is originally saved? Unsure. At any rate (no pun intended), I'll try unchecking it before my next roast session and see if that's the case.CarefreeBuzzBuzz wrote:On your Artisan use, you mentioned optimal smoothing post roast. If you select that, then you are overriding your Smooth Deltas settings.
Basically its an additional algorithm that works post roast as it looks back and forward and can only do that post roast. Your post roast curve won't look like your during roast curve if it is checked. You can check and uncheck it after a roast is done to see the difference.
 CarefreeBuzzBuzz
I'll inquire as well. And update the info if needed. Always learning.
 CarefreeBuzzBuzz
So Optimal depends on the Smooth Deltas setting and shouldn't have a time shift. So if Smooth Deltas is 0 you won't see a difference checking it and then unchecking it. Give it a try. Set Smoothing to 20 and then check/uncheck Optimal. Now set Smoothing=0 and check/uncheck Optimal.
 MaKoMo
 Delta Span: the period in seconds used to calculate basic/core RoR values by dividing the deltatemp/deltatime (so deltatime is the deltaspan used for this). A Delta Span smaller than twice the sampling interval has no effect, larger Delta Spans lead to time shifts (also in case "Optimal Smoothing Post Roast" is ticked).
 Smooth Deltas: proportional to the number of basic/core RoR values to be averaged over to compute the final RoR values used for predictions and rendering. This smoothing process produces a shift on the time axis if "Optimal Smoothing Post Roast" is not ticked. Not that a linear decay smoothing is applied here over all the readings considered.
 "Optimal Smoothing Post Roast": as noted in the tooltip of this element, if ticked an optimal smoothing algorithm is used that does not produce any (additionally to that of the Delta Span thing) shift on the time axis. This one is only available offline and not during the recording. Thus if you tick this, you will have different/shifted renderings of RoR during vs after roasting.
 "Smooth Curves": curves, temperature as well as delta curves, are average smoothed by the same algorithm used for "Smoothed Deltas" (decay smoothing). Here it is applied to the curves, incl. the BT/ET before the core/basic Deltas are computed. Thus applying curve smoothing results in smoother deltas and also in a shift on the time axis if "Optimal Smoothing Post Roast" is not ticked. However, Curve Smoothing is only applied post roast and this adding Curve Smoothing larger than 0 results in different renderings in the offline and online case.
 "Smooth Spikes" is similar to "Smooth Curves", but using a specific algorithm to remove tiny spikes and noise.
Best settings depend on the noise produced by your meter, probes and overall system as well as the sampling interval. In general you should apply only as much smoothing as strictly necessary (especially also because any online smoothing results in time shifts). If you want to have the same smoothing applied for offline and online (e.g. to track exactly a background RoR), don't apply any curve smoothing and don't tick the "Optimal Smoothing Post Roast".
I still vote for fighting noise before the roasting software, by selecting/configuring a meter to produce minimal noise, with high precision (against digital noise resulting from digital sampling of signals with low resolution), selecting probes that produce minimal noise (note that probes with small diameter produce a lot more noise and that different probes of the same diameter can produce a different amount of noise) and protecting the system against groundloop noise and electrical inferences. I recently did some experiments on this comparing different meters and probes to select my next setup.
I see that I have to write a blog post on this at some point.
 Smooth Deltas: proportional to the number of basic/core RoR values to be averaged over to compute the final RoR values used for predictions and rendering. This smoothing process produces a shift on the time axis if "Optimal Smoothing Post Roast" is not ticked. Not that a linear decay smoothing is applied here over all the readings considered.
 "Optimal Smoothing Post Roast": as noted in the tooltip of this element, if ticked an optimal smoothing algorithm is used that does not produce any (additionally to that of the Delta Span thing) shift on the time axis. This one is only available offline and not during the recording. Thus if you tick this, you will have different/shifted renderings of RoR during vs after roasting.
 "Smooth Curves": curves, temperature as well as delta curves, are average smoothed by the same algorithm used for "Smoothed Deltas" (decay smoothing). Here it is applied to the curves, incl. the BT/ET before the core/basic Deltas are computed. Thus applying curve smoothing results in smoother deltas and also in a shift on the time axis if "Optimal Smoothing Post Roast" is not ticked. However, Curve Smoothing is only applied post roast and this adding Curve Smoothing larger than 0 results in different renderings in the offline and online case.
 "Smooth Spikes" is similar to "Smooth Curves", but using a specific algorithm to remove tiny spikes and noise.
Best settings depend on the noise produced by your meter, probes and overall system as well as the sampling interval. In general you should apply only as much smoothing as strictly necessary (especially also because any online smoothing results in time shifts). If you want to have the same smoothing applied for offline and online (e.g. to track exactly a background RoR), don't apply any curve smoothing and don't tick the "Optimal Smoothing Post Roast".
I still vote for fighting noise before the roasting software, by selecting/configuring a meter to produce minimal noise, with high precision (against digital noise resulting from digital sampling of signals with low resolution), selecting probes that produce minimal noise (note that probes with small diameter produce a lot more noise and that different probes of the same diameter can produce a different amount of noise) and protecting the system against groundloop noise and electrical inferences. I recently did some experiments on this comparing different meters and probes to select my next setup.
I see that I have to write a blog post on this at some point.
LMWDP #360, https://artisanscope.org