"What is Baked Coffee?" - Rao

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
User avatar
Almico
Supporter ♡

#1: Post by Almico » Feb 27, 2018, 11:54 am

I read through Scott's latest blog posts and was intrigued by the one on baked coffee.

https://www.scottrao.com/blog/2018/2/24 ... -dont-know

This morning I made a concerted effort to keep my RoR steadily declining and want to see the result in the cup. Here is a decaf roast (won't be cupping this one):

Image

and a Sumatra:

Image

This Sumatra's curve does dip at 1C but recovers after a minute. My practice has been to ignore the crash rather than try and compensate for it. The coffee tastes wonderful. This coffee also likes a little push through 1C, hence the flat section pre 1C.

As far as the blog post: I'm having a hard time getting my brain around why a dipping RoR (crash) and a rising RoR (flick) could both yield the same defect in the cup - baked flavors.

I'm also having a hard time reconciling the TC probe dimension allegation:

Image

If the 10mm probe is reacting enough to track all those little spikes, how is it not sensitive enough to follow the same macro curve?

I use a 6mm probe that dances all over the place during the roast. Here is the decaf graph before turning Artisan off:

Image

User avatar
cimarronEric
Sponsor

#2: Post by cimarronEric » Feb 27, 2018, 12:40 pm

Almico wrote:This Sumatra's curve does dip at 1C but recovers after a minute. My practice has been to ignore the crash rather than try and compensate for it. The coffee tastes wonderful.
As Rao asserts in the blog post "...many readers are thinking 'my RORs crash but MY roasts aren't baked.' All I can say is that if you ever master avoiding the crashes, you'll change your mind."

As I have worked the crashes out of my roast, the juiciness has exploded, the development has evened out significantly and I have started noticing the baked notes as a roast defect, because I am shooting for clean, sweet and juicy as an end product. This is not to say that baked notes taste "bad". Many coffee drinkers like them.
Almico wrote:As far as the blog post: I'm having a hard time getting my brain around why a dipping RoR (crash) and a rising RoR (flick) could both yield the same defect in the cup - baked flavors.
It is the crash that yields the baked flavors, the flick is simply a byproduct of the crash.[/quote]
Almico wrote:I'm also having a hard time reconciling the TC probe dimension allegation:

If the 10mm probe is reacting enough to track all those little spikes, how is it not sensitive enough to follow the same macro curve?

I use a 6mm probe that dances all over the place during the roast. Here is the decaf graph before turning Artisan off:
I've never used Artisan so I'm guessing you mean before adding some kind of smoothing. The same thing holds with Cropster, except that I'd guess they were using a different moving average for the 2 probes. Even a 10mm probe will record little RoR spikes if the averaging is short enough.
Cimarron Coffee Roasters
www.cimarronroasters.com

User avatar
Almico
Supporter ♡

#3: Post by Almico » Feb 27, 2018, 1:26 pm

cimarronEric wrote:As Rao asserts in the blog post "...many readers are thinking 'my RORs crash but MY roasts aren't baked.' All I can say is that if you ever master avoiding the crashes, you'll change your mind."
I don't get a crash with any other coffee except this Sumatra, but I'll try another roast this afternoon and kick the heat to remove the roast.

But we need to remember, the curve represents the temp of the probe and is not necessarily indicative of what's going on inside the bean. I would have to add quite a bit of heat to this coffee to maintain the curve structure. My guess is that would be worse for the coffee. But I'll try.

User avatar
MaKoMo

#4: Post by MaKoMo » Feb 27, 2018, 1:29 pm

Those RoR spikes might also have some thing artificial relating to the resolution of the temperature signal. To illustrate I have here a signal just recorded using Artisan from a Hottop 2k+. That machine returns its temperature values without any decimals. Here we have the 2sec sampling interval recording given as raw data without any smoothing. Note that here oversampling is activated which results in requesting two readings per interval (so one per second) and averaging those two readings resulting in some decimals.

Image

and zoomed in a bit

Image

You can see clearly the measurements and how they increase by one full C. Each such increase results in a spike of the RoR signal while after that increase its zeroing as the temperature in between the steps does not increase. You also see well the effect of the oversampling that rounds the edges somewhat.

Just activating the spike filter of Artisan results in this one

Image

Now adding smoothing (factor 2) to the temperature signals makes also the derived RoR signal more smooth, but there is still some oscillation visible, resulting from the pur temperature resolution.

Image

Finally adding smoothing (factor 10) to the RoR signal and the oscillations are gone.

Image

Not sure about baked or not, but not only your probe, but also the resolution of your meter counts. Even if higher resolutions 2 decimals as provided by Probat roasters returns mostly noise, this noise can be smoothed way easier without any oscillation into a nice signal. Not working with decimals is "killing the RoR".

User avatar
cimarronEric
Sponsor

#5: Post by cimarronEric » Feb 27, 2018, 1:42 pm

Almico wrote:I don't get a crash with any other coffee except this Sumatra, but I'll try another roast this afternoon and kick the heat to remove the roast.

But we need to remember, the curve represents the temp of the probe and is not necessarily indicative of what's going on inside the bean. I would have to add quite a bit of heat to this coffee to maintain the curve structure. My guess is that would be worse for the coffee. But I'll try.
Some coffees are "crashier" than others for sure. Given the flatness of the curve before 1C I would guess there's virtually no amount of heat that could be added at that point to avoid a crash and it'd take some adjustment of the entire curve. This being a 2C roast falls outside my wheelhouse although I take one of our blends right to the edge. And Scott readily says his theories are geared towards roasts that end before 2C.

I don't disagree with what the readings are vs what is happening in the bean. I only know what I know from roasting and tasting.
Cimarron Coffee Roasters
www.cimarronroasters.com

mfortin

#6: Post by mfortin » Feb 27, 2018, 2:28 pm

In Artisan, the RoR is for 1 minute. In Cropster like in Rao's graph, the RoR is for 30 seconds. Does that change anything to the steepness of the slope (crash) of the RoR?

mfortin

#7: Post by mfortin » Feb 27, 2018, 2:31 pm

cimarronEric wrote:
As I have worked the crashes out of my roast, the juiciness has exploded, the development has evened out significantly and I have started noticing the baked notes as a roast defect, because I am shooting for clean, sweet and juicy as an end product. This is not to say that baked notes taste "bad". Many coffee drinkers like them.

How have you worked the crashes?

User avatar
Almico
Supporter ♡

#8: Post by Almico » Feb 27, 2018, 2:41 pm

cimarronEric wrote:Some coffees are "crashier" than others for sure. Given the flatness of the curve before 1C I would guess there's virtually no amount of heat that could be added at that point to avoid a crash and it'd take some adjustment of the entire curve. This being a 2C roast falls outside my wheelhouse although I take one of our blends right to the edge. And Scott readily says his theories are geared towards roasts that end before 2C.

I don't disagree with what the readings are vs what is happening in the bean. I only know what I know from roasting and tasting.
I could definitely add enough heat to eliminate the crash. I use an air roaster and "roast by wire", steering it all along the way. I just can't imagine that would be better for the roast.

User avatar
MaKoMo

#9: Post by MaKoMo » Feb 27, 2018, 2:48 pm

mfortin wrote:In Artisan, the RoR is for 1 minute. In Cropster like in Rao's graph, the RoR is for 30 seconds. Does that change anything to the steepness of the slope (crash) of the RoR?
No

User avatar
cimarronEric
Sponsor

#10: Post by cimarronEric » Feb 27, 2018, 3:36 pm

mfortin wrote:How have you worked the crashes?
I've worked with Scott on a consultative basis. Lots of roasting and tasting and tweaking.
Cimarron Coffee Roasters
www.cimarronroasters.com