In the Smackdown thread timo wrote:
I'd say rather that with domestic lever machines in general and especially the lower brew pressure domestic spring lever, using freshly roasted (though degassed) beans and a good micro-adjustable or stepless grinder, you are more likely to get consistently good extractions with a very light tamp or no tamp at all. One should not be amazed by this. That you are amazed shows how very effective the tampanista propaganda has been.
He actually
should have written:
"In my opinion...you are more likely to get consistently good extractions with a very light tamp or no tamp at all."
Because, this is his opinion. I've been getting good extractions with the Ponte Vecchio for two years, across a broad degree of grind/dose/tamp parameters.
I will admit to once having read that a 30# tamp is standard, and that I followed this advice early in my espresso days, as I'm sure many have. A good extraction can be had on the Ponte Vecchio if one does a handstand tamp, provided the grind is suitably coarse. And, with a suitably coarse grind and the proper dose, one can take multiple pulls on the Ponte Vecchio and get good extractions WITHOUT DISTURBING THE PUCK. This is a FACT (although, of course, the "good extraction" part is subjective.).
I will NOT say that using a hard tamp and a suitably coarse grind/proper dose that one is "MORE LIKELY to get consistently good extractions", because I could provide no evidence for this. But neither has timo provided any to support his claim to the reverse.
I hope to have time to prepare some illustrations of this in the future, but these competing ideas can be "modled" using physical principles (that I am only vaguely familiar with).
1. GRIND FINENESS: Consider that if one does not to tamp, the grind will need to be suitably fine to produce the pressure necessary for an ~30 sec extraction (we will leave dose out of this for the moment, but of course this is also very important). The finely ground coffee will be extracted over a 30-sec period, and water will carry away X amount of solids from each grain of coffee. The relatively small grain size will lead to a greater proportion of those solids being extracted than they will from a larger grain over the same period of time. Conversly, because of smaller interstitial (between-grain) spaces, less water eill be flowing by any particular grain than it would if the grain size were larger. This could lead to relatively LESS solid being extracted from a given grain. Until someone has given this a lot more thought than I currently have time for, timo cannot make the claim that one is "MORE LIKELY to get consistently good extractions" with a finer grind.
2. FLAVOR PROFILES: Added to the fluid dynamic question just briefly introduced above, is the question of flavor profile of extraction as a grain of coffee is exposed to water. Which flavors come out first, and which last? Do all bean types follow the same pattern? Does roast degree inlfuence this? How about bean freshness? . Timo cannot make the claim that one is "MORE LIKELY to get consistently good extractions" with a finer grind unitl he presents some evidence re: flavor profiles during extraction.
3. DOSE: Added to the two above is the question of dose, because it will influece them both. A greater dose increases pressure - one therefore needs to decrease grind fineness appropriately to compensate. Water flowing through a thicker puck will behave differently than it will though a thinner puck - and the flavors will be extracted differently. This will also need to be addressed.
Lastly, as every scientist is woefully aware of, there are the INTERACTIONS between factors in any experiment, such as the three above terms (and probably a suite of others). Simply put, this means that the effect of GRIND FINENESS will not be the same at Dose 1 as it is a Dose 2 (or Flavor 1 vs. Flavor 2), etc. etc. etc.
I've been giving a lot of thought to these ideas the past several years. This is why I am RIGHTLY amazed that one can get such great results from a spring lever ACROSS such a range of variables.
I hope others will put forward some ideas that might, in the long-term, be testable to some degree. For now I'll just shut up and drink the coffee.
PC