Brew pressure profiling update 3

Beginner and pro baristas share tips and tricks.
gscace

Postby gscace » Mar 29, 2007, 5:47 pm

Pressure Profiling Update 3

I think I'm homing in on some concrete information about pressure profiling that is worth blabbing on about, so here goes:

First thing is that there seems to be a practical maximum pressure value beyond which things go to hell in a handbasket tastebudswise. I found this out when I was experimenting with pressure profiles that included a high-pressure hump at the beginning of the extraction. Values for the hump were as high as 165 psi (11.3 bar). My thinking was that if sweetness was extracted early in the brewing process, then maybe I could emphasize it by increasing pressure to a high value early on, reducing to more conventional values later. The result was that I extracted crema that was markedly bitter. Bitterness was muted by reducing the magnitude of the hump. Bitterness was removed once the max pressure value was reduced to 9.5 bars at the coffee cake, which adds credence to the conventional 9-bar wisdom. I learned later that Jim Schulman has also observed this effect. I also successfully replicated the effect for Peter Lynagh, of Terroir Coffee, when he came down to visit.

Once I learned that there was a maximum practical brewing pressure, I began to think of pressure profiling in terms of the minimization of undesirable tastes, rather than in terms of super-extracting desired compounds. Taste tests with Nick Cho demonstrated to me that the sweet tastes and mouthfeel are developed early in the brew process, with dilution occurring later. Unfortunately, in the tests mit Cho (gesundheit!) extraction of bitter compounds also occurred during the dilution phase, meaning that a balance needed to be struck between dilution of the drink to a desired volume and introduction of negative taste components. Reducing the brewing pressure as brewing progresses seems to be beneficial in reducing bitter tastes extracted during the dilution phase. Tests with Nick pointed to accentuated sweetness in Counter Culture Toscana, which in retrospect was really a subtraction of bitter components added in the last third of the brew process.

The most successful general profile that I have come up with to date (I've now got an excel file with a bazillion profiles mapped out in it) is a pressure profile in which the pressure rapidly increases from some nominal start value to 9 bars at the group (elapsed time of 1 second from start to full pressure) with a short 9 bar soak, then ramps downward in a slow linear fashion (straight line degradation) to around 7 ½ bar, with a more rapid, second-order (curved downward) slope over the last few seconds to around 6 bar, arriving at 6 bar 30 seconds after initiation of the brew cycle. This general shape can be used with or without a pre-infusion step. The pre-infusion that I've been using is to soak the cake at around 2 - 3 bar for 3 seconds, then quickly ramping to maximum value with an exponentially increasing slope. Full pressure is attained 3 seconds after the end of the low-pressure soak. This combination produces liquid evenly over the bottom of the filter basket almost as soon as the pressure begins its rapid increase. Regardless of whether or not pre-infusion is used, I've been using a very similar profile toward the end of the brewing period. If pre-infusion is not used, I increase the time of the high-pressure soak by a few seconds.

It seems to me that the efficacy of preinfusion is coffee specific. There are three basic coffees that I've inflicted pressure profiling upon and with which I can comment. I've been drinking dry-processed Ethiopian SOs, and Ethiopian-based simple blends that have a lot of mouthfeel and body. My tests with Nick used Counterculture Toscano. I don't really know what is in it, and for the purposes of this discussion I don't think I really need to know. And last weekend, Peter Lynagh and I concentrated on a very lightly roasted Brazil. Here's a link to Terroir's information about the coffee:

http://info.terroircoffee.com/content/view/17/2/

For brewing temps, the Ethipian DP and Toscano seem to like around 201F. The SO Brazil was brewed at 195 F. The blends seem at first blush to benefit from preinfusion. I think that the soak and subsequent rapid pressure ramp may produce more mouthfeel, but I need to revisit this as I haven't been systematic enough, particularly in light of our results with the Brazil. We found that pre-infusion was a waste of extraction time when brewing the SO Brazil. Unlike the blends, the SO Brazil produced two predominate tastes with great clarity - nuts and sweetness reminiscent of dried figs. Both nuttiness and sweetness were enhanced when the pre-infusion step was removed. The clarity of the Brazil made differences between brewing at constant pressure and brewing with profiled pressure easy to discern. Bitter tastes were evident with constant-pressure brewing and demonstrably removed by profiling.

Recently folks have come up to me on the street, shoving their pudgy fingers in my puny chest, demanding to know that if pressure profiling meant deemphasizing the last portion of the brewing process, why not just stop brewing early? After I imagine breaking their finger with a deft marital arts-type motion, I counter with the argument that early brew termination is different. Compare two shots of the same volume, but with one brewed in the style of brewtus interruptus, and the other brewed to satisfying completion with pressure profiling. If one terminates early, for example at 18 secs in a 27 sec extraction, the volumetric flow rate during the 18 second period is much faster than the flow rate for an extraction taking place over the entire period. The extractions that we are observing with profiled pressure have more or less constant flow rate throughout the entire time period, which means that the volumetric flow rate is less by 50%, but over a longer time period. This changes the taste.

As I try different coffees and gain more experience I'm getting more confident that pressure is worth exploring as a brew parameter. There is still a lot to learn here, but the tests with Terroir indicate that variable pressure is useful when one is brewing clean SO espressos in which one or two tastes are showcased. I'm not sure which cart drives which horse when it comes to blends. I don't know if various widely used pre-infusion schemes were developed because they work with traditional blends, or if blends end up being developed to mask machine deficiencies. My cynical self thinks it's the latter. I have a lot more to learn about this, and it gets more difficult when I'm using coffees I've munged together. I don't feel like I'm a good enough roaster or blending dude to come up with confident conclusions, other than to go back with what I learned about the Brazil SO and see if I can make any hay with the stuff I usually drink. That means I'm not close to closure, which means that if you all don't behave, rather than dislocate your digits I'll spring Pressure Profiling Update 4 on you.

User avatar
jesawdy

Postby jesawdy » Mar 29, 2007, 6:52 pm

Greg-

Now that you've been at this awhile longer, I am curious to know your thoughts on whether pressure profiling based on volume rather than time is of more or less interest to you.
Jeff Sawdy

gscace

Postby gscace » Mar 29, 2007, 9:52 pm

jesawdy wrote:Greg-

Now that you've been at this awhile longer, I am curious to know your thoughts on whether pressure profiling based on volume rather than time is of more or less interest to you.



I would be interested in linking pressure to pulses from flowmeter sensor just to check it out. I can see arguments that extraction is both flow-based and time-based. It's prolly a bit of both. I've got my hands full with time-based profiling and it's interesting enough and works well enough that I'm not planning on fooling with flow-based profiling anytime soon.

-Greg

User avatar
HB
Admin

Postby HB » Mar 29, 2007, 10:16 pm

gscace wrote:The most successful general profile that I have come up with to date (I've now got an excel file with a bazillion profiles mapped out in it) is a pressure profile in which the pressure rapidly increases from some nominal start value to 9 bars at the group (elapsed time of 1 second from start to full pressure) with a short 9 bar soak, then ramps downward in a slow linear fashion (straight line degradation) to around 7 ½ bar, with a more rapid, second-order (curved downward) slope over the last few seconds to around 6 bar, arriving at 6 bar 30 seconds after initiation of the brew cycle.


Interesting, this sounds like brew pressure profile of a commercial spring-driven lever espresso machine.
Dan Kehn

User avatar
another_jim
Team HB

Postby another_jim » Mar 29, 2007, 10:53 pm

Yeah.

A commercial spring will be 9 to 6 bar, declining linearly by volume (the spring force being determined by the remaining water in the cylinder. Given that the flow is increasing (linearly?), this means the time versus pressure profile is roughly a quadratic curve. It could be quite close to the profile you synthesized, even to the final declining curve.

It may sound like an anticlimax to "reinvent" the lever profile, but it's not. You've been able to explore a lot of possibilities, and do it with far more stable equipment, especially on temperature. It could be that the outcome of the work is not that commercial machines get full profiling motors, but motors with simpler controls that produce the optimal profile you discover, perhaps with a few simple adjustments.
Jim Schulman

gscace

Postby gscace » Mar 29, 2007, 11:11 pm

HB wrote:Interesting, this sounds like brew pressure profile of a commercial spring-driven lever espresso machine.



Actually the profile is continously curved. Lever machines are linear for the most part and look like a sawtooth. Maybe I'll get off my lazy yass and post an excel graph if I can figure out how to make the damn thing a picture for cryin out loud.

-Greg

gscace

Postby gscace » Mar 29, 2007, 11:16 pm

another_jim wrote:Yeah.

A commercial spring will be 9 to 6 bar, declining linearly by volume (the spring force being determined by the remaining water in the cylinder. Given that the flow is increasing (linearly?), this means the time versus pressure profile is roughly a quadratic curve. It could be quite close to the profile you synthesized, even to the final declining curve.

It may sound like an anticlimax to "reinvent" the lever profile, but it's not. You've been able to explore a lot of possibilities, and do it with far more stable equipment, especially on temperature. It could be that the outcome of the work is not that commercial machines get full profiling motors, but motors with simpler controls that produce the optimal profile you discover, perhaps with a few simple adjustments.



Actually I don't care what it ends up looking like. If it's like a lever machine then that is what it is, but we learn that there is benefit to it. So it's not anti-climactic at all. Plus in a lever machine you don't get to optimize it.

User avatar
cannonfodder
Team HB

Postby cannonfodder » Mar 31, 2007, 11:20 pm

gscace wrote:Actually the profile is continously curved. Lever machines are linear for the most part and look like a sawtooth. Maybe I'll get off my lazy yass and post an excel graph if I can figure out how to make the damn thing a picture for cryin out loud.

-Greg




You essentially do a screen capture, paste it into an editor, crop it down, compress the image to the size you want and save it in the desired format.

Or Email me the spreadsheet and I can convert the graph. It is relatively simple but takes a couple different programs.
Dave Stephens

User avatar
cannonfodder
Team HB

Postby cannonfodder » Mar 31, 2007, 11:29 pm

gscace wrote:Recently folks have come up to me on the street, shoving their pudgy fingers in my puny chest, demanding to know that if pressure profiling meant deemphasizing the last portion of the brewing process, why not just stop brewing early? After I imagine breaking their finger with a deft marital arts-type motion...



Smack them in the forehead with a hot portafilter then exclaim 'No Spresso for YOU!'

Your work is appreciated. One interesting observation, for all our advances in technology, it is starting to sound like the ideal extraction profile is that of an old school spring lever machine.
Dave Stephens

User avatar
hbuchtel

Postby hbuchtel » Apr 01, 2007, 3:38 am

HB wrote:Interesting, this sounds like brew pressure profile of a commercial spring-driven lever espresso machine.



Part of the spring lever profile is the pre-infusion at boiler pressure, the ramp up to which is determined by the size of the hole where the water comes into the piston-chamber.

Henry
LMWDP #53