[Tips/Theory] Translating an Espresso Recipe across different basket size/diameter

Beginner and pro baristas share tips and tricks.
Team HB

Postby samuellaw178 » Mar 20, 2017, 1:30 am

Just like our resident chemist expert (rpavlis) has stated many times, making espresso is really just a form of liquid chromatography. As someone who has done both liquid chromatography and espresso, I find that is such an apt description. The significance of it really, is some of the insights gained in the chromatography field can be applied directly onto espresso making.

In making espresso, one often encounters the need of using different baskets, or just plainly confused by the plethora of baskets available on the market. It's not that clear (even for advanced baristas sometimes) what the effect of basket choice has on the flavors, nor how to translate an optimized recipe onto another machine with different basket size (ie 58mm to 49mm basket). Sometimes, the basket size (58mm or not) even becomes the main consideration factor when deciding on an espresso machine purchase, which is likely due to the prevalence of 58mm basket and its assumed superiority.

So the question is, what is the difference between the 58mm basket and the smaller baskets (43mm, 49mm, 51mm, 54m etc)? Can you achieve the same result using either? Is there anything magical about the 58mm diameter basket?

Thus, in this discussion, let's assume our goal is to keep the same extraction/flavor (obtained on a 58mm basket), but translates it to a different basket diameter.

[Background information]
When changing a basket, you have multiple considerations (some of the main ones):

- Basket shape
- Filter basket total hole area
- Grind size/fineness/particle distribution
- Coffee dry dose

If you have a different basket shape (ie conical shaped single vs cylinder shaped double), you can bet that the extraction dynamics will be different. Exploring different basket shape is beyond the scope of this discussion, as we're going to consider only cylindrical, non-tapered double baskets like the VST/IMS/some other double baskets. That's because unfortunately no one really uses conical-shaped column for chromatography so not much is known about that. :oops:

If you're interested though, what I can share anecdotally is I find that conical-shaped single basket is less efficient from the perspective of Extraction Yield (measured by a refractometer). A typical extraction yield on a conical-shaped basket using brew ratio of 1:2 is 12-14% max. Whereas, a cylindrical-shaped double basket is often 19-21% for the same brew ratio.

Basket total hole area is inter-connected with grind fineness and coffee dose. The goal is to achieve a constant puck resistance for a regular espresso flow rate. The relationship can be explained by the following crude equation/description (which may not be mathematically correct but mainly to depict the relationship).


So, to achieve the same puck resistance, you can either use a larger coffee dose, or use a smaller/finer grind, or a smaller/lesser basket hole area. Unfortunately. the basket hole area is typically fixed for a particular basket so you can't change that easily without changing to another basket.

The gotcha here is - changing the grind fineness will not only change the puck resistance, but also affect the extraction dynamics (flavors) due to the different particle distribution. ie. conical grinder will be more bimodal (creating more boulders) when grinding at a coarser setting. So, if you really want to replicate an existing espresso recipe, fixing the grind particle distribution by using the same grind fineness is crucial.

And then there is of course the usual espresso parameters/variables, such as brew water temp, puck temp, pressure profile, preinfusion, tamping etc which we are not going into, and will all be assumed to be constant.

In pharmaceutical industries, column chromatography techniques are often employed to separate biomolecules. R&D is typically carried out first at the small lab scale, then moved on to a larger pilot scale and eventually the commercial scale. Thus, predicting and translating the separation efficiency of a small lab column onto a larger column (50-100 times larger) is critical for what they do.

Here's one of the many articles on the scaling up process (if you wish to read).
http://images.alfresco.advanstar.com/al ... -43408.pdf

There are quite a bit of technical & irrelevant details there. But what really matters to us here is the Figure 3 on Page 6(not copied here for copyright reason).

Note that in Figure 3, when scaling up, there are some parameters to keep constant and only one parameter is allowed to be changed.

Comparing the Figure 3 to our espresso process:

To keep the efficiency while scaling up/down, you'd want to keep these constant:

[chromatography parameters] <->[parameters in espresso]
Linear flow velocity <----------> flow rate as dictated by basket diameter ( I will expound on this later)
Stationary Mobile Phase <------> use the same coffee (stationary phase) and water (mobile phase). Duh.
Protein Loading<----------------> non relevant
Feed conditions <---------------> non relevant
Gradient <------------------------> non relevant
Bed Height <---------------------> Puck height, determined by packed dry coffee dose (will be explained as well)

You'd change the following:
Column Diameter <-------------> Basket Diameter

So the gist is, you can expect to keep the separation/extraction efficiency the same, by keeping all the relevant parameters the same, except for the basket diameter. In another word, you can expect different basket size to extract similarly.

So how are you going to apply this knowledge? Obviously it's not going to change your espresso routine a whole lot. But hopefully, it will help you to understand, what would be the proper way to alter the recipe (as much as you can within the constraints), when you're trying to compare a small diameter basket (Mini Vivaldi, small levers) vs the bigger baskets (58mm),

So, you'd want to keep the linear flow velocity and puck height constant. How would you go about it?

The linear flow velocity is a description of how fast the water is flowing through the puck vertically. So it is related to the flow rate, but not the flow rate itself. It can be measured as centimeter/milimeter per second, and is a function of flow rate & surface area (basket diameter) as follow:


So how do you keep the linear flow velocity constant with different basket diameters? Fortunately, there's no complicated calculation or major change that you need to do. The simplest way to maintain the linear flow velocity, is to (i) maintain the brew ratio, (ii) maintain the shot time & (iii) alter your dry dose accordingly (next paragraph). Just by doing these three, the linear flow velocity & flow rate will fall into place without you doing much.

Coffee dose is the only variable you should change without affecting the extraction dynamics. To normalize the puck height, you need to reduce the coffee dry dose by a calculated factor (based on surface area/basket diameter).

For example, if you want to scale down a 20g dose on a 58mm basket onto a 51mm basket, the dose required to maintain the same puck height is 15.5 gram . This is assuming dry ground coffee has the same density (weight/volume) after tamping so weight & volume are proportional/interchangeable. ie. 20g of tamped coffee will be double in volume compared to 10g tamped coffee.


So, scaling down a 58mm recipe to a 51mm basket would look like this:
Dry dose: 20g ->15.5g
Output : 40g ->31.0g
Shot time : 30s -> 30s

In conclusion, you should be able to replicate any recipe from a 58mm basket onto a smaller basket diameter. The underlying requirement is, you will need to use the same grind size, the same brew ratio, the same shot time, and the same puck height. While this may or may not be feasible with the current available baskets (due to lack of basket varieties, as smaller baskets tend to have less total hole area), it is achievable from technical point of view.

So it seems apparent that when a 58mm portafilter is called a commercial-sized portafilter, it really is literally just that - it's part of the scale-up scheme so you can use a larger dose in a commercial setting to achieve the same extraction achieved on a smaller basket with a smaller dose (confused yet? :lol: ) . Of course it may just be another happy Italian coincidence. :D

The implication is, if you really like the high dose extraction on 58mm portafilter, you have no chance of replicating that accurately using a smaller dose on the same machine, unless you start using a smaller diameter basket. AFAIK that is not possible as current machine manufacturers don't offer that flexibility (understandably so because you need a different tamper) but hopefully this is something that will be explored eventually.

User avatar

Postby Peppersass » Mar 21, 2017, 2:15 am

Really cool analysis.

Jim Schulman explored a similar concept in this post about converting a dialed-in double to a single without changing the grind setting (because a grind change would drastically alter the flavor). Note CafeIke's post right after Jim's about the formula working for some baskets and not others due to height variation.

Here's another of Jim's posts on baskets you might find interesting.