Talking About Pressure Profiling



Is it me or is everyone else a little surprised at the lack of discussion of pressure profiling?

Most of the interesting discussion, if not just about all of it, has been over at home barista where various people are building pressure profiling units for their one group machines.  Still much of that discussion remains more about how to build it, than about desirable profiles.

Pressure profiling is undoubtedly going to become more readily available.  Though the capacities of the Slayer, the Strada and Cimbali’s new machine are all different, they are all chasing the ability to manipulate pump pressure to improve espresso.  A lot of this desire comes from the profile of espresso produced by lever machines, which have a very different pressure profile compared to a pump driven machine.

Perhaps I shouldn’t be surprised, given my own experiences with pressure profiling.  I was asked, a little while back, about why I hadn’t posted anything about it and I thought I would take this opportunity to explain:

John Ermacoff very kindly leant me a pressure profiler that he had built for a while.  I am not going to go into exactly how it worked, but we hooked it up to one of the groups in my 2 group Synesso.  This meant that I could only use one group at a time because then I had a Cyncra, not a Hydra, so the changes in pressure would influence both groups if they were open.

The machine allowed for almost any profile, and included a manual and automatic mode.  I played with the manual mode for a while, but prefered the automatic mode because it was repeatable and allowed me to make specific adjustments.

The entire time I used it I felt lost.  Juggling dose, grind size, flow rate, volume of espresso and time is enough.  Suddenly introducing pressure variations meant that it was a lot harder to isolate the problem.  If the shot was underextracted then keeping the profile but changing one variable often resulted in chasing my own tale and very quickly burning out my palate.

Changing the pressure even slightly undoubtedly changed the cup.  The impact was always obvious, but rarely did I ever feel as if I had genuinely improved the cup.  By giving me so much control I ended up lost in a sea of variables from which I never really found my way out.

I confess that I didn’t approach this as clinically as I should have.  I didn’t record enough data and often ended up defaulting back to a very simple profile – which seems to be quite a popular one online:

  • - A relatively lower pressure preinfusion period, lasting long enough to soak the cake (5-8s depending on dose).
  • - A smooth, relatively quick rise to 9 bars over 2-3s.
  • - A period of 9 bars lasting 7-10s
  • - A declining pressure in the remaining section, from 9 bars down to 8 or perhaps lower.

This is by no means what I would claim to be the perfect profile.  It is one, however, that I could rationalise in my head.  The period of preinfusion at the start would increase the likelihood of an even extraction.  The period of declining pressure in the final third meant that we were extracting the nearly-spent coffee less aggressively, yet still adding volume to the shot – aiding balance, and reducing negative flavours.  Of course you could easily tear this theory to pieces – I just couldn’t find a better way to explain it. 1

One thing that surprised me was how a lever machine acts, as I had initially thought this profile would be quite close to it.  I was quite wrong – though my experiences comes from the use of a Victoria Arduino Adonis and a Scace 2.

In terms of pressure the lever allowed line pressure through when pulled down, as the machine has a heat exchanger for brew water.  Upon releasing the lever the pressure shot up to around 12 bars before almost immediately starting to drop down towards 0 in a fairly linear fashion.  I only found this out after I had returned the profiler, and I wish I could have tried it on the Synesso.  That way I could have dose two identical baskets had a better idea of the influence of temperature on our love affair with lever espresso.  The Scace 2 showed that initially the brew water was very stable – in the first 12-15 second after pulling the lever the temperature barely moved – maybe 0.2/0.3°C.  Then the water gradually began to decrease in temperature, losing 4 or 5°C by the end of the shot.  This makes sense.  When the lever is pulled water fills the chamber above the coffee.  Only once it starts to pass through the coffee, and lose thermal mass, dose it really start to cool off. 2

This profile is a very long way away from what most pressure profiling machines out there are doing – or at least what people have said they are doing.  Then again, as I said at the start, I can’t find many people talking much about it.  I guess I had hoped to see more on places like coffeed, coffeegeek, twitter or on various blogs.  Personally I’d be as interested in finding out what isn’t working for people, as much as what is.  When I asked the Cimbali engineer which profiles he liked best, he was a little evasive – in fact very few people are willing to really get behind any one profile.  I know these may change coffee to coffee, but with enough data out there I hope we’ll eventually be able to make a pretty good educated guess, the way we can with things like temperature and dose, based on the density of a particular coffee/blend.

I do have great hopes for pressure profiling, but at the same time realistic ones.  I don’t expect pressure profiling to discover sweetness, or flavours that have previously remained undiscovered in coffees.  I hope they can improve the clarity and presentation of them and help reduce detracting tastes and flavours.  It won’t make average greens taste great, or bad roasts acceptable.

Right now it seems that engineering is, for once, way ahead of the barista.  The pressure profiling machines coming can do almost anything we want – we just need to work out what that is…

I will close this post by saying thank you again to John Ermacoff – I learned a great deal through his incredibly generosity, and can’t wait to see what he comes up with next.

  1. If anyone has any info on the effect of pressure at that stage in the shot – on caffeine extraction, oil emulsion etc etc then I would love to see it!  ↩︎
  2. Again, correct me if I am wrong here  ↩︎

33 Comments

  1. I’ve only recently been reading about pressure profiling and it seems very exciting. Coming from Leeds, UK, the best coffee around is from a little espresso bar called Opposite, where they use a Synesso Cyncra (extremely well). However there is no place to experience pressure profiling around here, so all my excitement is generated by videos such as this from slayer espresso:

    http://vimeo.com/3372757

    Here the shots look so syrupy its unbelievable, and from their reactions it looks like it tastes pretty good. I presume that the increased viscosity of the shot is due to an increased extraction of oils from the cake creating a denser emulsion, however it could be due to more carbohydrates in the extraction which would make the cup sweeter. (note I know nothing about coffee chemistry only emulsions and dispersions etc.)

    From the videos by Slayer they seem to a have a few pretty solid profiles down, however I imagine they are experimenting and refining all the time.

    -Ben

  2. I’m one of those folks who think pressure profiling isn’t worth the investment or development money being thrown at it. This is partially based on having some of the most experienced “pressure profiler” baristas pulling me shots on tuned machines, but also based on rudimentary pressure profiling via machines that can do it in limited fashion manually (paddle group LMs, GS/3 paddle group, Speedster). Like you James, I just don’t see a tangible improvement in the cup, and we barely know anything about espresso as it is – throwing a brand new paradigm – pressure profiling – into the mix just confuses.

    However, as nascent and infant stage as we all are in espresso, one thing we do know a bit about is temperature. We know, via temperature stability, that sometimes even a 2 degree C drop in temps can greatly influence the taste of the espresso. We also know, through lots of coffee chemistry science (for eg, the Illy Chem of Quality book) that higher temperatures towards the end of a shot of espresso result in more bitter extraction, more caffeine extraction (caffeine’s the world’s most bitter food substance) and other undesirable elements.

    Which leads me to believe that TEMPERATURE profiling is not only something we, as a collective barista, espresso and coffee community, can better wrap our brain cells around, but will provide much more tangible, tasting results in the cup that can be evaluated and improved on.

    I’ve long felt that a declining temperature curve – say starting at 202F, maintaining that for the initial saturation / brew, and perhaps starting a decline at 15 seconds in down to a finishing temperature of 190F, would result in some stellar coffee (starting temp relative to the coffee used, level of roast, etc).

    Why? Because to this day, the best shot of espresso I’ve probably ever had has come from an Elektra Micro Casa a Leva spring lever espresso machine – and it probably saw similar declining temperatures (and, caveat, also declining pressure lol – but I think temps had more to do with it).

  3. Jsut about every resource I’ve read has described caffeine as being bitter, which is unsurprising as it is an alkaloid.

    That said – I should probably taste it myself to be sure.

  4. The virtues brewing with pressure are undeniably unquantified.

    I think your point about trying to match the pressure profile of lever is fantastic. Conventional wisdom certainly tells us that we’re experimenting with pressure for that reason. The available tools don’t perform that way at all, but, they may offer the flexibility to.

    When we’re really going good on the Slayer, the pre infusion times can be up to 30 or 35 seconds. I wonder, is that saturation time a function of puck depth and grind rather than dose? I’ve been brewing with different baskets to try and learn something about that.

    To your point about the thermal mass and footnote number two; We know there’s a relationship between pressure and temperature, so as the chamber empties and pressure reduces.. I bet there is a thermal change.

    This is all very interesting and exciting, thanks for posting and I look forward to when some research becomes presentable. This will be a learning endeavor for all of us.

  5. Funny, I’ve always heard that caffeine is tasteless too.. but as an alkaloid, that makes total sense. I had powder form caffeine at defcon 16, but, we mixed it in drinks, so I didn’t taste it.

  6. @ Joe ~ Caffeine is definitely not tasteless; while it doesn’t have much in the way of flavor, it is very distinctly bitter. Pick up a small sample of pure powdered caffeine from your local chemistry lab, put a dab on your tongue, and you’ll see what I mean.

  7. Hi James:
    I, too, experiment with pressure profiling. But flavorwise, I find it only a small benefit. From time to time I still use a declining profile ending up ~6 bars. This doesn’t produce spectacularly better shots, but does seem to reduce bitterness. Perhaps the best benefit is that it slows down the flow rate at the end, so that one has a little more time to decide exactly when to cut the shot off. This makes pulling shots more relaxing and — if you can believe it — less stressful. Nice!

  8. Here’s an example of a pressure based experiment that I’ve been a party to. On this day we made some awesome coffee. What really struck me was the fat content. It was so oily, the mouth feel was intense. This is the tip of the iceberg. This is just the first day of tests. http://www.pouredover.com/2010/01/herkimers-newly-modified-synesso.html

    I think James is right about the lack of discussion. I myself am not sure what to say. My blog post was really intended as a conversation starter and I think James’ was too. So, I’m going to put out some thoughts from my notes from another example:

    (Brewing on Slayer)
    In a pre infusion phase of around 3-5 bar the puck saturates slowly (15 seconds sometimes 30 or more), in my opinion, this seriously reduces the opportunity for channeling. A neglected potted plant is analogous to this situation. After the soil has become dry, a can of water dumped on top will only run out the bottom. The careful application of water over a duration of time allows for even distribution. I’ve tried to initiate channeling with ugly, uneven tamping. Tamping that caused squirters at work on 9 bar(bottomless portafilter). With the slow saturation I was unable to create that kind of flow.

    After I increase to 9 bar, the espresso behaves very much as I’m used to. First, drip drip and then a stream right down the middle. The shot runs for 25 or 27 seconds more before its colour starts to change.

    I’ve noticed a huge difference in the visual cues. Normally, looking up under the bottomless portafilter, I would expect to see some striping. Even on a really good shot there is lighter and darker interspersed. With this kind of brewing the extraction is really even. I see more of what I’d think of as the ideal color.

    If anyone has had the opportunity to work on this, I’d love to hear about how you’re brewing it.

  9. @Joe ~ Indeed, although not overtly flavorful in concentrations used in colas and the like, caffeine is so universally perceived as bitter that it is used to demonstrate functional bitter gustatory capacity in patients.

  10. we have been playing for about 2 weeks now on ours.

    I think its about right what your saying, the 15 sec preinfuse helps give a more even colour but I’m not sure if that equals better taste yet. I roasted a batch of an el salvador on wednesday and gave 2kg to a friend with an FB70 and kept the rest for us. The shots they were pulling seemed to have more body but less complex flavours then we were with the 15 sec preinfuse. Temps were relatively the same ~203f. Shot times excluding preinfuse were also similar.

    Temperature changes on the slayer also seem to make a bigger impact on the coffee than they did when I used an FB80.

    I’d also say the preinfuse is better for fresher coffee. I’m loving it at 3 days where I used to prefer around 7 – 10 on the marzocco. I’m getting great colour and body from very fresh coffee.

    We need a slayer forum.

  11. Josh- Slayer and FB80 do have some differences in temperature control. I know just who bug about a Slayer forum…

  12. i understand the offset. What else do you mean by differences?

    another thing we tried today –
    setting the preinfuse pressure is a little fiddly we have found, so we did it by putting a cup under each group, running preinfuse and then weighing the cups after 1min to see how close we had each group. From what I understand the flow rate might be a more accurate measurement than running a shot and seeing what it comes up to.

    Please correct me if this is way off.

  13. I think my biggest hopes for pressure profiling is that I will be able to use lighter roasts for espresso. I think espresso roasts are a huge compromise when it comes to getting the most of the coffee’s “personality” out – You have to let go of so much complexity and finer notes just to get enough sweetness and body for it to be balanced as an espresso brew (There are of course exceptions, i.e. a lot of natural ethiopians can pull off at a lighter roast and keep the wonderful structure in the espresso cup). I wish someone would write something about this, as I probably wont be able to play with it enough myself in a while. But the way you and others have described it flavourwise, it sounds like it might have a positive impact on those very features that make us roast darker.

  14. Josh, I love that, weighing the cups. I haven’t tried it. Take this off James blog? tom[at]pouredover[dot]com

  15. I hadn’t really thought about that. Considering how finicky some espresso set ups are when it comes to hitting the correct point of extraction, this could have real benefits.

    It will be interesting, if people continue to find less than they hoped for in pressure profiles, to see how the manufacturers spending a lot of R&D money will respond.

  16. I’ve tried both the slow ramp up and fast ramp up and I have to say I’m a fan of:

    – fast ramp up to 9bar, probably even without engaging the pump controller, simply set it to 8.5bar at the beginning and engage the pump, let the gicleur do it’s thing

    – leave it at 8.5bar for around 10s

    – slow ramp down to 6bar or so

    [quite close to a lever shot]

    I’ve found this reduces the harsh/bitter flavours that normally appear at the end of the shot, but I’m not sure if this is due to lower flow pulling less out of the coffee, or simply less obvious blonding, thus pulling a slightly longer shot which means a bit more diluted ‘taste’?

    I’ve been messing about with low pressure pre-brewing (around 5-6bar), but for some reason it works worse than fast rump up to 8.5bar. I thought low pressure in the initial phase of the extraction would slow it down, but it actually made it faster. This is probably cause by fines being pushed down less than on full 8.5bar and thus not blocking the water path so much. It’s usually ok if you leave it at 5-6bar for a while, but when you crank it up to 8.5bar it tends to go bad, usually channelling or at least showing signs of uneven extraction. I’ve noticed a similar thing with long preinfusion, it did more bad than good on my machine (Elektra T1).

    One way to control the ramp up is the gicluer size, I’m using a 1.0-1.1mm one now and it takes 3-4s to get full pressure on the group. I will have to experiment a bit more with larger jets and see what it brings.

    Regarding different baskets and the depth of the puck I think the amount of space above the puck matters. Try comparing a normal double in a double basket and the same dose and grind in a triple basket (it will have loads of space).

    Regards,
    dsc.

  17. Man, this is truly interesting !!!… In my experience, I’ve always obtained better espressos with a lever machine than with a pump machine. As many of you said, the espresso with the lever has more oils in it, resulting in a ragged or speckled beautiful crema. This was harder to achieve in a Marzocco or a Cimballi with a pump, but much easier in older Faemas or even cheap mexican Nacional. Reading the article, it makes me think it is directly related to the pressure profile: if the pressure reaches 12 bars at the begining, when the grounded coffee is fresh, and the pressure diminishes as the coffee lets go the oils and flavors, instead of the linear pressure of the pump machines, the pressure profile has something to do for sure… Man, we have to work on this…

  18. I started a thread for new LM GS/3 paddle owners on CoffeeGeek, but it didn’t produce very much: http://coffeegeek.com/forums/espresso/machines/457185 . So, consumers seem pretty silent about their experiences, too.

    At this point I’ve settled into a routine of line pressure until first drips (8-12 seconds), then pump until the color starts to turn, and finishing with a few seconds of line pressure. I’ve tried bypassing the initial line pressure and going straight to pump, but this seems to sacrifice some complexity in the cup.

    Of course, I have no rush of customers in front of me and can play with the paddle as I chose.

    Meanwhile, in talking with real baristas, I am getting a clear message that pressure profiling will not work in a commercial setting, unless it is simple to control and easily repeatable.

  19. “Meanwhile, in talking with real baristas, I am getting a clear message that pressure profiling will not work in a commercial setting, unless it is simple to control and easily repeatable.”

    I’ve heard a lot of this attitude and I think that it is just an attitude. Why is pulling a 50 second shot suddenly looked on as impractical. Are we looking to serve the best coffee or just the quickest and easiest? I think its not always easy to be watching and adjusting a shot while dosing another or steaming milk but its not impossible.

  20. Not every lever has the same pressure profile. It depends on the spring and on the pretension applied. If the pressure drops linearly to zero, you’ve got zero pretension. La Marzocco’s lever started at 14 bar and dropped to 7 according to the Slayer guys. So the La Marzocco lever has quite a bit of pretension.
    http://www.slayerespresso.com/2007/12/10/lever-espresso-extraction/
    And the HX levers have a preinfusion pressure which is the same as line pressure, while the dipper groups have boiler pressure preinfusion. Then you’ve got the Fellini preinfusion which is also something else.
    So ‘your’ profile is probably quite similar to another lever machine out there. And wildly different from a lot of others.
    Also, the cimbali guy probably didn’t answer because it would be embarrassing to admit that the machines they made 50 years ago did basically the same thing, but automatically >:)

  21. Surely every lever machine has to return to 0? Otherwise, as they have no three way valve, there would be quite horrific portafilter sneeze when you take the handle out.

  22. They do return to zero, but not linearly. For example, a lever could concievably start at 12 bar, drop linearly to 6 and then hit the block. The spring never is in relaxed state in most lever machines. Check the graphs in the Slayer link. The LaMarz starts at 11, drops linearly to a bit less than 5 and then hits the block.

  23. But as far as the coffee is concerned – it would experience a linear drop to zero.

    Its resistance determines the rate of decline in pressure, and even though the spring may no longer be expanding it has created pressure that will continue in a linear fashion to drop to zero as the pressure is released through the coffee itself.

    I figure this is the only way to get a linear decline that I see on my Scace 2 – as if the spring completely extended then the last part of the pressure profile would likely be a slower drop in pressure.

  24. That residual pressure would be gone after a few drips because water isn’t compressible. Maybe some more because the puck expands as it isn’t compressed anymore, but nothing serious. The flow stops very soon after the lever bottoms out on my La Peppina. The drips after a lever bottoms out are not something you want in your cup anyways.
    Maybe the VA groups don’t have pretensioning. Is the first few cm of pulling down the lever very easy?

  25. James — Interesting that your default pressure profile is fairly similar to what an E61 group does automatically.

  26. RE: Footnote #1 The period of declining pressure in the final third meant that we were extracting the nearly-spent coffee less aggressively, yet still adding volume to the shot – aiding balance, and reducing negative flavours. Of course you could easily tear this theory to pieces – I just couldn’t find a better way to explain it.

    Why do you think declining pressure equals less aggressive extraction? I think it’s just the flow rate that is “less aggressive”? This would imply the contact time is greater in that final third. Did you compare this with a shot with no declining pressure in the final third, implying “normal flow rate”, and therefore shorter time to yield the same volume?
    I had thought the brew pressure only overcomes the hydraulic resistance produced by the puck and it was flow rate and hence contact time which affects the extraction. I think that when brewing at a constant pressure there is no significant pressure gradient across a coffee particle so there is no extraction of stuff out of a coffee particle due to pressure. I do think pressure does other things (eg emulsification) to the coffee extract but essentially espresso is produced by washing the surface of specifically sized coffee particles with a specific volume of water in a specific time.
    However, maybe it’s possible that changing the brew pressure during brewing does produce a pressure gradient across a coffee particle. This would result from the pressure “front” moving across the particle. Maybe then there will be extraction of stuff from inside, or at least deeper inside, the coffee particle? This effect may produce the difference observed in flavour.
    What do you reckon?

  27. Less pressure indeed means lower flow rate, but from what I’ve experienced on my rig it also means steadier more even extraction from the whole puck ie. late blonding. Perhaps at that stage of the extraction the coffee gave away most of it’s ‘value’ and increasing the contact time is needed to pull something more out of it. Whether the ‘more’ is good or not is the real question (something that probably depends on the type of coffee used).

    Regards,
    Tom

  28. *Comment removed due to violation of comments policy*

    May I remind people – no personal attacks that contribute nothing but animosity.

    -JH

Submit a comment