No announcement yet.

The Science of Espresso Machine Boilers

  • Filter
  • Time
  • Show
Clear All
new posts

  • The Science of Espresso Machine Boilers


    Whilst I certainly agree aluminium is a better thermal conductor than brass (almost twice as good in fact) for a given volume of each metal, brass holds far more heat.

    For a given volume, aluminium and brass hold about the same amount of heat (al is slightly better -dependent on alloy)
    Whilst the specific heat of Al is a bit over twice that of brass, brass is 3.36 times as dense (in both cases comparing "casting" grade versions of Al and Brass....)
    So for a given volume - brass has a "thermal mass" approximately 1.6 times that of Aluminium....

    And then there is the mass (300 grams) of the water in the boiler with a specific heat 4.5 times that of aluminium.... compared to a few grams of water in the thermoblock (and that few grams is absorbing energy rather than acting as a thermal mass stabilising the temperature - which is the case in the boiler.)

  • #2
    Re: Detailed comparison - EM6910 vs Silvia

    JavaB, I dont understand why you are looking at specific heat capacities(?) Specific heat from what I remember is how much energy is required to raise the substance by 1 deg C. It has nothing to do with thermal mass. Specific heat is an intensive property.

    edit: comment retracted...i should read the post properly.


    • #3
      Re: Detailed comparison - EM6910 vs Silvia


      Yeah, I dont really like the term "thermal mass" (hence the quotes here and in my post above).... its really about how much energy is held in the system - more like thermal inertia I guess....

      And the greater the specific heat the greater the energy transferred in (or out) for a one degree change.... So if you can get a material (like water) with a high specific heat and a reasonable density then it makes a really good thermal stabiliser....

      Aluminium has pretty high specific heat but lacks density, whereas brass, although lower in specific heat is far more dense.... and more effective for a given volume..... water has the best of both worlds...

      And as Greg says above, the reason that Miss Silvia takes so long to heat up is this whole mass of brass, water etc must be increased to brew temp.... whereas in the Sunbeam there is only a relatively smaller amount of energy to heat the thermoblock....

      The electronics in the Sunbeam should stabilise the thermoblock after initial heatup.... whereas the Silvia depends more on the mass of hot material (brass and water) - the "thermal mass" to keep things stable.


      • #4
        Re: Detailed comparison - EM6910 vs Silvia

        yep...go a page back...sorta explained all that in engineering terms...hahaha. Here it is for your convenience.

        Originally posted by Wushoes link=1177174323/0#14 date=1177332970
        Send it down to Melbourne Sparky...Id be more than happy to do it. I have a naked, 6910 and datalogging DMM. This will put to rest all the bs everyone is talking.

        I agree with you on the aluminium argument there. Im currently studying a mech engineering unit called heat transfer and thermodynamics (boy is it hard) and cannot fault the "In fact al is more thermally conductive, and so responds faster than brass to fluctuating temperatures." The heat transfer rate through Al will be greater than in brass. But maybe having the a boiler respond slower to thermal gradients is a good thing(?) [smiley=undecided.gif] Sort of like a thermal lag that allows the boiler to maintain temperature without much energy needed to keep it there. What it does mean though is that the thermoblock will respond to a re-programmed brew temp a lot faster than a boiler. 2 distinct advantages of each technology.

        My premise for this argument is this. We all know the conservation of energy law....basically its the 1st law of thermodynamics and states that "for all adiabatic processes between 2 specified states of a closed system, the net work done is the same regardless of the nature of the closed system and the details of the process". So in essence, the 1st law is a simple energy balance. {Total E in} - {Total E out} = {Change in E total of system}

        Aluminium as discussed above is a better conductor of heat. Applying Fouriers Law for conduction we have q" = -k dT/dx.
        Thermal conductivity [k{brass}] of Brass is 109 W/mK. Thermal conductivity [k{A}] of Aluminium is 250 W/mK.

        Take an instant in time where the rate of change of temperature with respect to the x-direction perpendicular to the direction of heat transfer does not change (i.e. dT/dx = same for Brass and Aluminium), it is easy to see the heat flux {q"} of the Aluminium is much larger than that of Brass. Which proves Sparkys point that Aluminium has superior thermal characteristics. Actually I really didnt need to go through all that to prove that point. All I needed to do was compare their k values (thermal conductivity values).

        But it sets up my next point. I discussed above the 1st law of thermo. If Aluminium is quick to heat and quick to cool compared to brass, it then requires more energy to keep it in thermal equilibrium (heat flux). The advantage of having a brass boiler is that the silvia element doesnt need to work as hard. There is a "thermal lag" or as explained above the heat flux of brass is lower....meaning it takes longer to heat up and longer to cool down. Aluminium is suited to tasks such as being heat sinks where the higher thermal conductivity is desired.

        Thermoblocks and boilers are two totally different things, so the two really cannot be compared.

        The test with the thermofilter will either confirm or deny Sunbeam is the shit or a piece of shit and:
        • runs at programmed temp
        • has good intrashot thermal stability
        • has good inter-shot recovery

        Would be very interesting to see how the Sunbeam stacks up against a PIDd Silvia in all the above points. My prediction is that it will fail test 1, do well in test 2 and even better in test 3.


        • #5
          Re: Detailed comparison - EM6910 vs Silvia

          Guess what my final year thesis/project will be on??? : Theres not much innovation going on in boiler lets see where it takes me


          • #6
            Re: Detailed comparison - EM6910 vs Silvia


            I missed that post of yours :-[..... must have been browsing elsewhere and saw Gregs above...

            Your post takes me back to 2nd year Uni physics.... soooooooo many years ago now : :

            I agree with your postulation at the end of your post with one slight reservation... Im not sure it will do all that well in test 2.... The reason for this is there will be a thermal gradient between the heater and the water in the stainless steel lined heat exchanger..... Now if they measured the actual temperature of the water as it leaves with something like a quality PID and used that to control the temperature (closing the feedback loop)... it should have good stability during the shot.

            But Id be surprised if they arent measuring the actual "block" temperature somewhere.... and that wont be exactly the same as the water temperature under dynamic conditions..... and I doubt the thermal regulator is the same quality as those in the fairly expensive PID kits.

            Ill be very interested in seeing the results.


            • #7
              Re: Detailed comparison - EM6910 vs Silvia

              Well my postulate for test 2 is totally independant of test 1. To have have good intrashot stability doesnt mean it needs to be at the specified brewing temp. It just needs to be as close to a horizontal line as possible. Take for example the attached graph. The machine was programmed to 96 deg....yet the brew temps are somewhere between 82 and 83 deg C.....but hold steady.

              I do agree there will be thermal gradients at the different junctions of materials. Should go something like convection from water to stainless (due to bulk motion), conduction through stainless to aluminium, conduction through aluminium to outside wall...and finally convection to T{infinity}....if what you say is true about measuring the outside wall of the thermoblock. In that case Sunbeam engineers should have easily accounted for this with an offset or delta value in the brains of the machine. Not hard to do as there wont be much change in the delta value between 92 and 96 degrees, and everything can be easily interpolated....even in a dynamic is only a 4 degree C change in brew temp in 2 degree increments.

              I drew on my knowledge from 2nd year thermodynamics and 3rd year heat transfer and thermo....physics yes...but more engineering. I doubt science students would cover the physics that engineering students study at the same depth. Yes....inter-faculty is crap ;D


              • #8
                Re: Detailed comparison - EM6910 vs Silvia

                Originally posted by Wushoes link=1177174323/15#22 date=1177339746

                The reason is this...even with my primitive tests where I restricted the flow to that of espresso using the foam cup method...I found the water temp was quite constant...albeit way under the desired brew temp.

                I drew on my knowledge from 2nd year thermodynamics and 3rd year heat transfer and thermo....physics yes...but more engineering. I doubt science students would cover the physics that engineering students study at the same depth. Yes....inter-faculty is crap ;D
                Yep, I remember seeing those results..... Will be a very interesting test....

                By the way, in my day engineering students (of which I was one) did the same physics which science students did (after all last time I checked the same laws of physics apply to both :) as well as subjects like "Engineering chemistry and materials" etc..

                But I wouldnt knock the science students.... they just look at things in slightly different ways....

                It doesnt matter if the coffee cup is half empty or half full - there still isnt enough damn coffee left ;D ;D


                • #9
                  Re: Detailed comparison - EM6910 vs Silvia

                  Well believe it or not...first year phyisics, chem and maths is all handled by other faculties other than engineering. All handled by science and maths by the maths department from science.

                  Lets just say science students ask why....and engineers ask how....we look at things totally differently....engineers dont care why it happened...we want to know how something happened so we can fix the damn thing. seems obscure but makes sense to me...scientists seem too stuck in theory.


                  • #10
                    Re: Detailed comparison - EM6910 vs Silvia

                    The plot thickens it seems


                    • #11
                      Re: Detailed comparison - EM6910 vs Silvia

                      I was a science student once...Dont have much against any other faculty.

                      However, my specialty was atmospheric science. Pre-Stern report, the whole climate debate was pretty much science vs. politics. Everyone leans towards the pollies because the science is too difficult to understand by your garden variety layman. Pollies just keep saying that when the time comes, technology will solve the worlds climate problems. Sadly, the technology will probably create another, possibly more sinister problem.

                      Without taking into account the temperature, test 1 plot looks nicer. With the other two, you run the risk of overshooting at the start of the brew cycle. While it may even out the sharpness in the first part of the shot, it may be detrimental to shot flavour of more sensitive SO.


                      • #12
                        Re: Detailed comparison - EM6910 vs Silvia

                        I see nunu.....that is the job of engineers in this generation. We not only need to be innovative, solve problems, create new things, we also need to make sure the materials we use will not pose any problems for future generations. Eco-friendly and recyclable is what were taught.

                        This is getting way off track....anyways...nunu...the only way to tell for sure is to use a proper thermofilter device that throttles the flow of water. i.e. Scace device...which is a very pricey thing. Im not even sure the Victorian Barista Championship machines were calibrated using the Scace which is what they do at the WBC. I remember Pete doing on the fly adjustments to the pump on his FB80 because one of the competitors called a tech timeout because the flow of water was not right.

                        So Sparky...where abouts are you in Australia? Is it possible to send your thermofilter down to Melbourne to conduct some tests?


                        • #13
                          Re: Detailed comparison - EM6910 vs Silvia

                          JavaB youre right, a factor of between 1.4 to 1.6 for an equivalent volume. However, people are fond of holding a heavy brew group and saying " there is a large amount of thermal mass in there", compared to this "light al equivalent" .... This is like comparing apples to carrots. For only a slight increase in volume, al will still be lighter but hold the same amount of energy AND respond quicker to thermal changes.

                          Wushoes, thats alot of technobabble but not saying much. The bulk material properties are not closely related to the heat loss of the material. You need to know the emissivity of the surface, the surface area and the contact resistance, before you can state that the 1st law of thermo says that it will take more energy to keep an al block hot.

                          Its true that Silvia has a lot more stored energy in the boiler. Water is an excellent heat storage medium. But it is a very poor thermal conductor. So much of that stored energy is no good for anything else but to store energy long after the machine is switched off.

                          Just be aware that I have a Sunbeam thermoblock. The thermoblock actually forms part of the group (ie the top of the group). The shower screen then bolt straight into the thermoblock, and the group seal is embedded in the bottom of the thermoblock. So heating of the group is taken care of by the thermoblock. This is a very efficient design.
                          Ive also conducted a number of tests with my Sunbeam TB and find that the element wattage is more than enough to heat cold water up to brew temperature at proper brewing flow rates.

                          My concern is that these observations about thermal mass etc are all moot with respect to how this machine works. Aside from giving the portafilter time to get up to temperature, the Sunbeam can be run as soon as its ready.

                          The bottom line is that the Sunbeam thermoblock technology is pretty good. Its engineered to a performance standard, and I believe it meets this standard.

                          As for the Silvia, with all these thermal inertia arguements, it is still not cut and dry as to how it achieves its performance. The brew water is held far above 100 C in the boiler. During its transit from the top of the boiler to the group it has to lose this excess heat to reach brew temp. This is a dynamic process and will depend on many factors including flow rate. So Silvia is likely to brew hot at fast flow rates, such as the no resistance scenario. In contrast the Sunbeam will tend to run cold (as it cant supply enough heat at these fast flow rates). This again is an apples to artichoke comparison and means nothing.

                          Wushoes should look at thermodynamic modelling of a coffee boiler (maybe a HX vs a Boiler design) for his final year project. I mean a real dynamic model that deals with the brewing condition, not just the steady state model suitable for idling between shots. This will be a tough model requiring some real numerical grunt. It will require some careful simplifications. However, there are plenty of engineering packages around that can handle it.

                          Engineers use the discoveries and understanding that is developed by scientists to build things.




                          • #14
                            Re: Detailed comparison - EM6910 vs Silvia

                            Originally posted by Wushoes link=1177174323/15#27 date=1177384631
                            So Sparky...where abouts are you in Australia? Is it possible to send your thermofilter down to Melbourne to conduct some tests?
                            Im in Brisbane. Given its a one-off home-built device that I have calibrated using around 100 data points to determine a nonlinear calibration curve and it has other ideosyncracies that I know well, but will probably confound an engineering type, Id say no.

                            But I may be in Melbourne say June or July for couple of days...


                            • #15
                              Re: Detailed comparison - EM6910 vs Silvia

                              Heck, what a can of worms! Sorry if any terminology I used was at all misleading - Ill leave you guys as the experts to fight out the whys and wherefores of whats going on inside. Ultimately the physics I used, flawed or otherwise, was attempting to explain the possible reasons for the demonstrated temperature stability on the Silvia in both free flow and restricted flow modes.