------------------------------------------ ----------------------------------------- Le café doit être noir comme le diable, chaud comme l'enfer, pur comme un ange, et doux comme l'amour.
"There is no right answer with coffee. There is only the elixir in your cup at the moment you partake."
"...I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind;..." - Lord Kelvin RECIPES thread => http://www.coffeegeek.com/forums/coffee/machines/585708
Posted Mon Dec 10, 2012, 12:29pm Subject: Re: Coffee Extraction Discussion, Questions for the membership:
And finally, the issue/discussion about predicting the produced coffee.
I use the simplified model of:
P = W - (A * C), where A is the constant dependent primarily on the brewing device/method.
Some people use LRR which has a small factor with extraction added, but in modeling the brewing process, this creates a dependency on an implicit variable (reminder - we CANNOT measure extraction directly. It is back calculated from strength, produced coffee (or a brew chart with assumed absorption) and the amount of brew coffee).
Absorption is useful in predicting the amount of produced coffee you will yield with the process as you adjust the brew ratio. It is determined by the difference of P and W, and is normalized by the amount of brew coffee.
Attached is the predicted produced coffee based on an assumed constant for absorption, defined as A = (W-P)/C. A is established at 1.16, and the predicted produced coffee is simply P=W-(A*C)
LRR is established by adding a second extraction term: A+E
Then, the question is exactly what extraction should be used? If the yield-based extraction is used, there is more variation in the prediction of produced coffee.
There is also the problem of having to predict the explicit variable before you make the coffee (or assume one) so you can determine how much coffee you will get, and if the amount varies by a percent one questions the validity of adding this variable.
Either way, I included LRR = A (constant) + E (based on immersion calculation, the same E that is determined by "Re: Coffee Extraction Discussion, Questions for the membership:") in the prediction of produced coffee. It has more spread around a linear fit (and therefore more variation in the prediction).
(Click for larger image)
------------------------------------------ ----------------------------------------- Le café doit être noir comme le diable, chaud comme l'enfer, pur comme un ange, et doux comme l'amour.
"There is no right answer with coffee. There is only the elixir in your cup at the moment you partake."
"...I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind;..." - Lord Kelvin RECIPES thread => http://www.coffeegeek.com/forums/coffee/machines/585708
The problem I have with your notion of "absorption", as I've said before, is that it involves taking the difference between two non-comparable things - the brew water and the brewed coffee. As such, it doesn't make sense as a physical definition, no matter how well it happens to correlate with observations. It's a bit like subtracting 200g of oranges from 300g of apples to leave 100g of fruit. The problem goes back to your model of brewing, which has the brewed coffee being produced by taking something away from the brew water. That something, the "absorption" has to have negative coffee solids in it for this to work, which is non-physical again.
Posted Mon Dec 10, 2012, 8:29pm Subject: Re: Coffee Extraction Discussion, Questions for the membership:
GlennV Said:
Yes, this is back to front.
The problem I have with your notion of "absorption", as I've said before, is that it involves taking the difference between two non-comparable things - the brew water and the brewed coffee. As such, it doesn't make sense as a physical definition, no matter how well it happens to correlate with observations. It's a bit like subtracting 200g of oranges from 300g of apples to leave 100g of fruit. The problem goes back to your model of brewing, which has the brewed coffee being produced by taking something away from the brew water. That something, the "absorption" has to have negative coffee solids in it for this to work, which is non-physical again.
It doesn't matter whether it makes sense as a physical definition. It adequately describes the process - and as such is probably the earliest model for brewing.
Adding something that makes physical sense, but does not show in the data doesn't make sense either. I know, for example, more fines get through into the coffee with finer grind. There is some mass transferred - but I haven't added a "grind fineness" factor. Mostly because there is no need. Residual error on prediction of produced coffee is small - so any effect due to grind fines, extraction level, phase of the moon, etc. is overshadowed by the simplified model of a constant.
As I said, absorption is merely the prediction of produced coffee as a factor of the ingredients (brew water and brew coffee). That's it. If there was a significant effect with extraction, then we would see it in the brewing control chart, and we would see a caution every time someone says "coffee absorbs 2ml of brew water for every gram of coffee".
For whatever reason, it does seem to work. We're talking pretty miniscule amounts of mass differences here, relatively speaking.
Again, the ONLY reason we need this is because of the importance of the produced coffee amount in the calculation of "extraction" for yield-based calculation. We parse the brew ratio in pre-brewing ingredients, but we really care about the yield ratio (for percolation methods). Except for espresso (where the actual produced coffee is controlled) we just need to know how to get what we want in the pot.
This is what works with my brewing model - and it fits with the conventional wisdom. I see no other significant effects to a constant as a substitution for absorption.
Now, this whole entire experiment is all with IMMERSION/INFUSION/STEEP brewing. It very well may be that the absorption for percolation methods DOES have an extraction component. There's a difference "in the grounds" of TDS of maybe a half a gram between what's in the grounds in percolation vs. what's in the grounds with immersion on about 25g of coffee. Is it significant? I dunno. I probably don't care.
I've only run a handful of brew ratios with pourovers and with my BCM-4C. Initially, it doesn't look like there's an effect - but strong brew ratios are very difficult to deal with on an auto-drip brewer, because I can't control the delivery rate very well. Running an 18 brew ratio vs. 16 brew ratio vs. 14 brew ratio (where I had to pulse the brewer and it took me about 10 minutes to brew) they still came out right around 1.7-1.8 for my calculation of absorption. The 14 brew ratio had two grind levels, the fine one was the really troublesome one - pulse, wait 1 minute for the delivered water to trickle out, turn on for 1 minute, then off once it started delivering.
With my BCM-4C (A=1.7) or my CCD (A=2.25), or my AeroPress (A=1.1-1.4), or my black and decker 10-cup brewer (about 2.0) - and a simple model, I can predict how much coffee I will yield solely on my brew water and the brew coffee mass. I don't need any more. I feel I can sufficiently predict how much coffee gets into the pot once I know the system absorption constant, and it works for all brew ratios and extraction levels.
Here's the other weird thing - espresso also seems to follow this rule. You can have a 17% yield, and an overdone overextracted lungo at 24% extraction (just pull a shot for over a minute... it will be overextracted LOL) and the process STILL seems to be an absorption of right around 1.1
------------------------------------------ ----------------------------------------- Le café doit être noir comme le diable, chaud comme l'enfer, pur comme un ange, et doux comme l'amour.
"There is no right answer with coffee. There is only the elixir in your cup at the moment you partake."
"...I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind;..." - Lord Kelvin RECIPES thread => http://www.coffeegeek.com/forums/coffee/machines/585708
jpender Senior Member Joined: 11 Jul 2011 Posts: 394 Location: California Expertise: I like coffee
Grinder: Kyocera CM-50 Vac Pot: S/S Moka Pot Drip: Aeropress
Posted Tue Dec 11, 2012, 2:31pm Subject: Re: Coffee Extraction Discussion, Questions for the membership:
Steve,
Although I sympathize with GlennV, I personally don't care how you count your fruit. I appreciate that there is practical merit to your approach. I was just amused that you turned the defintion on its head, so to speak, as LRR is defined merely as water retained per unit mass of coffee.
We really don't know in advance precisely the value of A or LRR, so any subtle variation due to extraction or moon phase is not useful when planning a brew, at least not a normal one. On this point I think we can all happily agree.
How to formulate the brewing equation is more a matter of preference. Your approach, being simpler, lends itself to back-of-the-napkin calculations. If using a computer this is no longer an advantage and I would expect Mojo to predict brewing output as well as you can. Using A or LRR should not matter.
And yet unless I'm mistaken it appears that with your graph above you are saying that it does matter. That surprises me given how little the difference should be.
A minor point: LRR and A are related by the yield, not the immersion extraction.
Ah, so if I apply the yield-calculated extraction, this would be closer?
It hadn't occurred to me because once I figured out the yield-calc extraction had NO correlation to taste, I knew it wasn't applicable to immersion brewing.
You have a point - if the yield-calculated extraction goes down as a function of brew ratio strength (which the chart does support) then it might have a small influence on LRR.
There's now the matter of tracking both yield-extraction (the TDS that makes it to the cup at the strength given by the immersion calculation) and immersion "diffusion" (gonna start to try and separate the terms).
Aye aye... LOL
Now I gotta go back and look at these with the yield-calculated extraction. It has no bearing on flavor profile, but maybe it has a small improvement in L (call that liquid retention ratio).
Just to be sure, you're saying that LRR = A (some constant) + E (calculated by produced coffee TDS), right?
------------------------------------------ ----------------------------------------- Le café doit être noir comme le diable, chaud comme l'enfer, pur comme un ange, et doux comme l'amour.
"There is no right answer with coffee. There is only the elixir in your cup at the moment you partake."
"...I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind;..." - Lord Kelvin RECIPES thread => http://www.coffeegeek.com/forums/coffee/machines/585708
l = r/c = (c+w-g-t-v)/c = ((c-g) + w - (t+v))/c = (t + w - p)/c, where p=t+v = t/c + (w-p)/c = extraction yield + "absorption"
Usually the immersion extraction is a couple of points higher than yield, right? But when you push the brew ratio they can diverge by quite a bit more.
Posted Wed Dec 12, 2012, 6:17am Subject: Re: Coffee Extraction Discussion, Questions for the membership:
I think I'm getting there.
So, in tracking "where the mass goes", LRR is referring to only the water portion of the retained liquid in the grounds. My concept of absorption is based on produced coffee, with the balance of the combination of undissolved grounds and the coffee solution in the grounds, with the original mass of the grounds removed.
Every brew process looks like this (even espresso - even though you don't know W)
Ingredients:
W=Brew Water C = Brew Coffee
==>> A brewing miracle occurs, you contact water and coffee together, then physically separate (or "extract") the coffee solution from the grounds.
you end up with
G=Wet spent grounds mass P=Produced Coffee
(this neglects any losses or evaporation of W during brewing)
The products of brewing coffee can be furthermore broken down as follows:
G consists of Pg, the amount of produced "coffee solution" but stuck in the grounds, made up of: -Tg, the TDS of any coffee solution in the grounds. This can be estimated by knowing the Sg, strength of the coffee solution in the grounds. -Wg, the water component in solution of the coffee solution in the grounds. This can be estimated as the water component of the coffee solution in the grounds, given Sg and then there's Ug, any undissolved coffee solids in the grounds. This can be calculated by G, minus Wg, minus Tg, minus T - it is the balance of the mass makeup once you subtract the extracted coffee, and any coffee solution.
P consists of T, TDS of the produced coffee, as defined by S, the strength of the produced coffee multiplied by P Wp, the water portion in solution of the produced coffee solution, (P-T) Up, any undissolved coffee solids that end up in the cup.
In the system model, mass preservation says:
C = Tg + Ug + T + Up and W = Wg + Wp
I've gone ahead and defined A as:
A = ((Tg + Wg + Ug) - C)/C = (Tg+Wg+Ug)/C
Stated another way using this expanded mass transfer model, (W-P)/C=A
P=T+Wp+Up, but usually Up is negligible but technically there, so we'll keep it in the model and just make it zero when we neglect it) A=(W - (T+Wp+Up))/C
L (the Liquid Retention Ratio) is, if I understand this right, an attempt to understand just the water component of the grounds as a function of the brew coffee, right?
L = Wg/C
Wg can be found by knowing Sg (and the rest of the mass). Sg can be measured, -for immersion it's =S, -for standard auto percolation (delivery rate matched to percolation rate) it's about S/3, -and for a manual pourover it could be somewhere in between depending on how much pooling you do when you're extracting (delivery rate>percolation rate gets the Sg closer to S). -If you have a situation where you extract (separate the grounds and the coffee) before the dissolution is past the steep part of the curve, it's possible to have Sg > S.
The difference between A and L
((Tg+Wg+Ug)/C) - (Wg/C) = (Tg+Ug)/C
Yep. That's a term based on yield-calculated extraction (Tg and Ug are directly dependent on yield-calculated extraction).
Interesting. I'll have to re-vamp my model, but there's the basis of the thinking. Fairly sure this is small by the numbers (on the order of 0.5g or less in a P=300g system).
------------------------------------------ ----------------------------------------- Le café doit être noir comme le diable, chaud comme l'enfer, pur comme un ange, et doux comme l'amour.
"There is no right answer with coffee. There is only the elixir in your cup at the moment you partake."
"...I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind;..." - Lord Kelvin RECIPES thread => http://www.coffeegeek.com/forums/coffee/machines/585708
jpender Senior Member Joined: 11 Jul 2011 Posts: 394 Location: California Expertise: I like coffee
Grinder: Kyocera CM-50 Vac Pot: S/S Moka Pot Drip: Aeropress
Posted Wed Dec 12, 2012, 1:04pm Subject: Re: Coffee Extraction Discussion, Questions for the membership:
Netphilosopher Said:
My concept of absorption is based on produced coffee, with the balance of the combination of undissolved grounds and the coffee solution in the grounds, with the original mass of the grounds removed.
That's what Glenn was objecting to, the way these quantities are mixed together. It works just fine but it's not the most intuitive approach. And while it isn't explicit in the definition your method assumes that there is a slight positive correlation between extraction yield and water absorbed. That could be true, who knows. But it looks like you've tacked that on in order to make the math simpler, not because of a physical model of absorption or data that demonstrate this.
Symbols: = New Posts since your last visit = No New Posts since last visit = Newest post
Forum Rules: No profanity, illegal acts or personal attacks will be tolerated in these discussion boards. No commercial posting of any nature will be tolerated; only private sales by private individuals, in the "Buy and Sell" forum. No cross posting allowed - do not post your topic to more than one forum, nor repost a topic to the same forum. Who Can Read The Forum? Anyone can read posts in these discussion boards. Who Can Post New Topics? Any registered CoffeeGeek member can post new topics. Who Can Post Replies? Any registered CoffeeGeek member can post replies. Can Photos be posted? Anyone can post photos in their new topics or replies. Who can change or delete posts? Any CoffeeGeek member can edit their own posts. Only moderators can delete posts. Probationary Period: If you are a new signup for CoffeeGeek, you cannot promote, endorse, criticise or otherwise post an unsolicited endorsement for any company, product or service in your first five postings.
![Netphilosopher: Immersion Extraction for All Runs vs R.jpg](javascript:ImagePopup("/images/59607/ImmersionExtractionforAllRunsvsR.jpg", 481, 289, "Netphilosopher: Immersion Extraction for All Runs vs R.jpg");)
![Netphilosopher: Predicted Produced using two retention models for all runs.jpg](javascript:ImagePopup("/images/59608/PredictedProducedusingtworetentionmodelsforallruns.jpg", 481, 289, "Netphilosopher: Predicted Produced using two retention models for all runs.jpg");)
= No New Posts since last visit 