Bret wrote: I grind fine and do what I will call a "high-pressure low-flow preinfusion" which yields 30-40 seconds of long slow dripping syrupy emulsion goodness (to my taste). I've found that low pressure preinfusion delivers significantly less viscous deliciousness.

There's another: "high pressure, low flow." A handful of these general approaches along with the time to n-grams in the cup could cover a lot of ground.

Tonefish wrote:The 4-9-6 was 4bar for the initial phase rather than 4ml/s @ 0 bar. With levers you can hit your desired initial pressure within 1-2 seconds.

This can still be done on a DE1. 8ml/s until 4 bar, pause, rise, decline. Advanced mode let's you do anything.

it seems we are in want for some surrogate parameters that can qualify a roast, same for the first part of extraction (formerly known as Pre infusion).

Surely the benefit of the proposed database of shots or pulls is that it device independent to the extent of the range of espresso machines and could thus generate useful information for a wide range of users, rather than only those with particular flow/pressure systems.
We may choose or have no choice but to exercise our p.i. control via grind or dose.
All the options inside the p.i. phase remain individual and that just allows engagement with the process. So there needs to be just enough "granularity" or detail in the database to allow valid comparisons and it may best to avoid adding extra detail, if it limits the user to a certain procedure.
In this case, we are defining pre infusion as the phase which loads the puck with water and possibly aiming to do some blooming, soaking or channeling, while that proceeds.
It ends where the next phase begins. That next phase is likely to be establishment of a small flow, as opposed to individual drips. So although Jims proposal of 5 grams is hard for me to beat, I'd like to suggest transition from drip to first flow is a reliable measure of preinfusion being over. In practice, it would be when individual discrete drips merge into a tail or continuity of water outflow.
At first sight this is inexact, but then so is the end of pre-infusion. Worse the judgement is also subjective, but maybe thats ok, reflecting the reality of the process?

Tonefish wrote:Assaf, How would you find this time to 5g useful? Did you just vote for your own idea?

That is simple. If someone says:

PI 15 seconds until first drops; pull 55 seconds. It may hide a 10 second stall a-la scott rao.

So PI 26 seconds to 5 grams and then 44 seconds makes more sense to me. Now the 10 second stall is part of PI. 44 second (rather slow) flow.

I don't know if I'd pick 5 gram. that is quite a lot. I'd say 1-2grams would be closer to the goal. (you want to be sure to be past PI & stall - but not take a big chunk out of flow).

It also works with 5 second PI since 1-2 grams would be an additional second or so. So a rather low margin of error.

DE crowd seems to want to share curves which I can also do (both flow and pressure). I think the PI&FLOW pairing makes more sense as it describes the though behind the pull. It doesn't have to be exact.

In other words the n-gram delimiter allows the recipe to be descriptive rather than prescriptive.

JayBeck wrote:This can still be done on a DE1. 8ml/s until 4 bar, pause, rise, decline. Advanced mode let's you do anything.

Yeah, I figured as much. Thanks Jayson!

Marcelnl wrote:it seems we are in want for some surrogate parameters that can qualify a roast, same for the first part of extraction (formerly known as Pre infusion).

Agreed!

Graham J wrote:... So there needs to be just enough "granularity" or detail in the database to allow valid comparisons and it may best to avoid adding extra detail, if it limits the user to a certain procedure.

Yes! Some general sense of the type of Preinfusion and it's duration is about as granular as you'd want to get to keep it simple.

AssafL wrote:That is simple. If someone says:

PI 15 seconds until first drops; pull 55 seconds. It may hide a 10 second stall a-la scott rao.

So PI 26 seconds to 5 grams and then 44 seconds makes more sense to me. Now the 10 second stall is part of PI. 44 second (rather slow) flow.

I don't know if I'd pick 5 gram. that is quite a lot. I'd say 1-2grams would be closer to the goal. (you want to be sure to be past PI & stall - but not take a big chunk out of flow).

It also works with 5 second PI since 1-2 grams would be an additional second or so. So a rather low margin of error.

DE crowd seems to want to share curves which I can also do (both flow and pressure). I think the PI&FLOW pairing makes more sense as it describes the though behind the pull. It doesn't have to be exact.

In other words the n-gram delimiter allows the recipe to be descriptive rather than prescriptive.

In your example you'd still never know that there was a stall and where it was if you only know n-grams in some duration. That was my point with the question to you. You and Jim resonated with your comment regarding "wholly repeatable" yet the chance of being repeatable across multiple users with only the n-grams duration conveyed is mighty slim. So how does the next user use your 26 second PI when all they know is 26 seconds to n-grams? So far we have the common options of:
- low pressure, high flow (lever, slayer)
- low pressure, low flow (trickle, blooming) <= closest to your stall example
- line pressure (3-4 bar like many machines, and an E61 that could be bumped to hold pressure beyond the 8-10 second chamber filling)
- high pressure, low flow (choking)

Even without the further customized options "wholly repeatable" is out the window with knowledge of only duration for n-grams.

I also agree that 5g is to big for the n-grams. Cheers!!

Tonefish wrote:This has really got me thinking. Recalling from your awesome thread that on any of the longer preinfusions, and especially with a ristretto pull, the coffee going into the cup is largely from the water that was introduced to the coffee during what we call "preinfusion." With that being the case, the preinfusion is really the "infusion" and the rest is the extraction. It's almost as though preinfusion is really pre-extraction infusion.

To put it another way, the water entering the coffee during what we call preinfusion is the majority of water that ends up in the cup. Again, this is for longer preinfusions and not long shots. As you notionally stated in your thread IIRC, it seems the water introduced during the extraction really just pushes the infusion (done during "preinfusion") out into the cup (along with additional simultaneous infusion although maybe not the majority), and most of that post preinfusion water remains in the basket after the pull.

This observation may explain why those who pull long-infusion shots sometimes report that the result tastes more like brewed coffee than espresso.

Tonefish wrote:In your example you'd still never know that there was a stall and where it was if you only know n-grams in some duration. That was my point with the question to you. You and Jim resonated with your comment regarding "wholly repeatable" yet the chance of being repeatable across multiple users with only the n-grams duration conveyed is mighty slim. So how does the next user use your 26 second PI when all they know is 26 seconds to n-grams?

My terminology "wholly repeatable" is sloppy. Not the Wholly, which I fully stand behind. It is the repeatable that I think is a sloppy word.

I personally do not believe coffee is very much repeatable. Which batch of a roast did you buy? How many days after roast? Ambient conditions (Over in the Static thread it turns out coffee absorbs water from the air - how much depends on RH and time). Coffee machines are built rather crudely to standards that are not metrology grade (hence the Scace devices, Fluke Thermometers, Erics telemetry stuff, etc. Even Acaia). Debits change between the same model machines (who knows why - maybe gicleurs clogged by swarf...). On many you need to open them up and adjust the pressure -from the manufacturer.

In the context of the above the n-gram makes sense to me. I make an effort to try Scott Rao recipes. To try Slayer's. Jim's Turkish Espresso. It won't be identical but I have a place to start my dialing in. It is a starting point.

CBR + Crude Timing gives me a coordinate on the map to explore. One can add to that special instructions - like stall.

(BTW - Is there a difference between a stall and a slower elongated flow? Tried it a few times and couldn't decide.)

AssafL wrote:I personally do not believe coffee is very much repeatable. Which batch of a roast did you buy? How many days after roast? Ambient conditions (Over in the Static thread it turns out coffee absorbs water from the air - how much depends on RH and time). Coffee machines are built rather crudely to standards that are not metrology grade (hence the Scace devices, Fluke Thermometers, Erics telemetry stuff, etc. Even Acaia). Debits change between the same model machines (who knows why - maybe gicleurs clogged by swarf...). On many you need to open them up and adjust the pressure -from the manufacturer.

I can appreciate all of this.

In the context of the above the n-gram makes sense to me. I make an effort to try Scott Rao recipes. To try Slayer's. Jim's Turkish Espresso. It won't be identical but I have a place to start my dialing in. It is a starting point.

Heck Yea! Between the new recipes and discoveries from my "accidents" it really expands the workspace and enjoyment factor.

CBR + Crude Timing gives me a coordinate on the map to explore. One can add to that special instructions - like stall.

(BTW - Is there a difference between a stall and a slower elongated flow? Tried it a few times and couldn't decide.)

I'm struggling with what CBR means? As for stall versus elongated flow, that's a good question. I figure both give time for blooming and because of the nature of a relatively fine espresso grind and tamped puck, the water is permeating relatively slowly into the puck in both cases as well. It seems that either way there is water available above the puck waiting for its time to be absorbed, so from that perspective they seem very similar.

I'm finding that the 5g time and the time at which drips become flow are nearly the same: when I see 5g on the scale, look to check the timer, and look back, the flow is already established. If this turns out to be similar across machines, 5g is easier to time than the transition to flow.

Since most of us are pulling doubles with similar doses, 5g out might be consistent enough to use as a starting point. If transition to flow times are varying off that point (later, or perhaps even before) then recording that time might also be warranted.

I assume that it almost certainly won't look the same on a low dose single, or a high dose triple, and the 5g metric might need to change, but the double is standard enough that trying this across machines, roasts, grinders, etc. might be both useful and interesting.