Ideal brew temperature management by HX espresso machine type

Need help with equipment usage or want to share your latest discovery?
User avatar

#1: Post by HB »

As an addendum to the article How I Stopped Worrying and Learned to Love HXs, I would like to share my observations of how "heat exchanger-centric" a particular espresso machine affects the ideal brew temperature management regime. These categories of heat exchanger espresso machines help answer important questions like "How often do I need to flush? How much? Does the flush amount change depending on how long the espresso machine has been idle?"

My shorthand for these distinctions are:
  • Dragon - key characteristics are lots of flash boiling, fast recovery, nearly zero thermal memory, and slowly rising brew temperature profile. Simply stated, after the cooling flush, the heat exchanger output is the brew temperature. Examples include the Elektra Semiautomatica, Bezzera BZ07, and the Olympia Maximatic.
    • Mixer - key characteristics are modest flush, medium to slow recovery, considerable thermal memory, and initial rising then falling brew profile. Unlike the Dragon, the Mixer's brew temperature isn't determined solely by the output of the heat exchanger. Other factors, such as cool water mixing via an heat exchanger injector, backflow from a thermosyphon, and the attenuating effect of a heavy grouphead temper the final brew temperature. Examples include HX E61 espresso machines like the Vibiemme Domobar Super and Quickmill Vetrano.
      • Agnostic - key characteristics are small, fixed volume flush or none at all, and long thermal memory. Careful tuning of a Mixer with tweaks in the design can produce an espresso machine that is heat exchanger in name only. Examples include the Cimbali Junior and Nuova Simonelli Aurelia.
      As the last entry suggests, these categories are not immutable. With minor modifications or boiler pressure adjustments coupled with barista techniques, an espresso machine that naturally fits in one category can morph into one of the other categories (e.g, Ian's HX Heaven or 1½ Boiler).

      The practical benefit of recognizing the characteristics of heat exchangers is the time saved learning the correct brew temperature management scheme. For example, I recognized the Olympia Maximatic as a Dragon by flushing the group until the water stopped flash boiling, waiting a minute or so, then repeating; it was fully recovered. The Elektra Semiautomatica is also a Dragon with a slightly heavier grouphead, but they share the same flush-n-go technique for targeting the brew temperature.

      The video below demonstrates this point:
      Hint: Don't watch the video, LISTEN to the video!

      Note: The above is a modified version of my comments from Olympia Maximatic - Second Look, reposted for easier reference.
      Dan Kehn


      #2: Post by dman777 »

      HB wrote:As an addendum to the article How I Stopped Worrying and Learned to Love HXs
      Great follow up, thanks!

      I have a question concerning the BZ07. I was studying the Temperature measurement Bezzera BZ07 PID thread and I was wondering...

      Would this still be considered a Dragon machine or a Mixer instead? The reason why I ask is because it seems the thermal design/properties of the BZ07 might of changed:
      another_jim wrote:Something has changed. I got only a few seconds of flash boil at 1.4 bar. Maybe they've goosed the grouphead heater. The BZ02, where the group is directly connected to the boiler, ran hotter and seemed happier with longer flushes and lower boiler pressures.
      And 2 users reports that the flushing/no flushing does not always influence the brew temp(and leaves me confused since there is no logic to it but complete random since the word 'sometimes' for shot temp not influenced from flushing):
      bas wrote:yes, 9, 11 or 13 seconds total flush time (e.g. 2, 4 or 6 more seconds after flash boil ends)...

      if I don't flush at all sometimes shots are hot and sometimes shots are cool; with some blend it doesn't matter but with most I assume this has something to do with the temperature swing in the group head...
      cpreston wrote:This is exactly the same behavior as my BZ07 has, FWIW.

      User avatar
      HB (original poster)

      #3: Post by HB (original poster) »

      dman777 wrote:Would [the Bezzera BZ07] still be considered a Dragon machine or a Mixer instead?
      With its short thermal memory and clear indication of end-of-flush, I consider the BZ07 a Dragon.
      Dan Kehn

      User avatar
      Team HB

      #4: Post by another_jim »

      The old school HX machines Dan calls dragons were designed to be used with no flush after idling -- the head was cool, the water was hot, they collided, and you got an about right shot. Trouble is that it isn't as good as a shot with a shortish flush that heats the head, empties the HX, and gets the machine close to its proper equilibrium -- the shot profile you get after pulling shots at one a minute to one every 90 seconds for a long period.

      Having that equilibrium state from the get go is the goal of machines with tuned thermosyphon loops and cold water injectors into the HX. The Aurelia succeeds at this, as do the carefully tuned and jetted commercial E61 machines like the Legend or the Mistral. The Cimbalis (which use a tunable steam plus water HX/brewhead combo instead of a thermosyphon) and the more economic commercial E61s like those by Brasilia or Wega get close and can presumably be tuned to be perfect by a techs who are trained in this black art (one that seems to be the province of Italian espresso makers, and used only for some of their customers or machine models).

      The home E61s lack all these refinements and therefore need very long flush routines. This is irking, since the home E61s see much more use from idle than commercial ones. I would like to see a successful E61 tuning project take place. I think Eric may be closing in on this.
      Jim Schulman


      #5: Post by mhborstad »

      Jim beat me to the punch, sneaking in during final preview :D. Here's what I had written when his post popped up:

      I've started looking at software-modeling the thermodynamics of my machine, and it seems like most Hx machines can be broken down to a model consisting of two heat exchangers. Three modes of operation might be useful to look at: a walk up shot, Jim's continuous-shot standard, and a steady-state where the machine is allowed to run until output temp stabilizes.

      The Bezzera (and other "dragons") have a small boiler Hx and a slightly cool grouphead (electrically heated in the case of the Bezzeras). Brew temperature after flushing the superheated water is mosty a function of boiler pressure(temperature) and volume flow. The brew temperature is the Hx output, moderated a bit by the group. A slower single should be a tiny bit hotter than a double*. Whether the temperature profile is flat/rising/declining for a given flow depends on boiler setpoint (and grouphead once I have control). Flushing makes very little difference, but tweaking the boiler pressure was immediately noticeable.

      The flow rate dependence can explain why a no-flush shot works on these machines. The flush volume is basically a double, and you'd think that that would give you burnt espresso, but when it's slowed to extraction speed, the cool group has time to work. The factory settings (on mine) were cool out of the Hx, and the Bezzera manuals make no mention of flush: insert PF, push button. Maybe we're defeating a very clever design?

      I think that the fast Hx combined with control of the grouphead could lead to better ways to run the dragons. Accurate control of boiler and group gives all sorts of options to tune either a lock-and-go or flushed shot, with excellent repeatability and some control over profile.

      Reported temperatures from Eric's GH thermometer show E61s reading significantly hotter. A larger/slower boiler Hx matches a mode where the output of the Hx is both cooler, and an Hx not fully emptied during the flush. The brew water is being warmed by the group during the first part of the shot at least, and the temperature during the shot is influenced by the "mixing" of remaining hot and incoming cold water in the Hx. Having a "shot-rate" temperature below target temp means that the flush is a more effective (and more critical) aspect of temperature management.

      What are typical "leave the pump running" steady-state temperatures of the different classes both in unrestricted flush mode and at typical shot flow rates?

      *That single-double difference might go away if run with no flush... the slower flow allows the GH more time to draw down the Hx water?


      #6: Post by Livin »

      another_jim wrote:I would like to see a successful E61 tuning project take place. I think Eric may be closing in on this.
      Do you have a link to what you are referring here to?

      I have an Isomac Millennium v2 and ended up lowering the boiler pressure to 1 bar (after becomes thermally stable - about 45 mins). This now produces almost exactly 200F water at the group. And since I steam milk only about 2-4 times a week, waiting an extra 20-30 seconds each time is no big deal. I also found that extending the steaming time allows me to stretch the milk slower to produce better foam.

      User avatar
      Team HB

      #7: Post by another_jim »

      I'm speculating (and gently nudging Eric). Eric makes the group thermometer for the E61s and has been posting on their internals for years. He has not developed a kit for tuning the home boxes; but I hope it does at sometime.
      Jim Schulman


      #8: Post by hmdavis »

      I'm curious as to how you'd classify the Salvatore Famosa.

      I've tried temp surfing--sound method (as illustrated in the video) and a digital thermometer, but I'm still having a devil of a time figuring out when/how to pull a shot, which isn't bitter or sour

      It's a great machine, but I'm flummoxed--at the moment.


      User avatar
      HB (original poster)

      #9: Post by HB (original poster) »

      Sorry, I haven't used any of Salvatore's espresso machines, but I'll offer some general comments:

      It's easy to recognize a Dragon because the flush time/sound will be the same every time (e.g., 90 seconds later). The second flush of a Mixer will be much less vigorous and shorter than the first. An Agnostic will be short the first time and zero the second time. The rebound time for Dragons is very short, usually less than 10 seconds. Mixers require more time or the tail end of the temperature profile will plummet; 10 seconds is a typical minimum and 30 seconds is a typical maximum. An Agnostic will typically have a short rebound time similar to a Dragon.
      Dan Kehn