"Espresso! My Espresso!"
An Ongoing Internet Novelette
by Randy Glass - Copyright 2002 - All rights reserved
E-mail me at firstname.lastname@example.org
Time Surfing vs. Temp Surfing
The "Temp Surf" (which would be more accurately referred to as a time surf) as is being used by so many alties, is a method that allows the Silvia user (and other owners of machines that have a boiler) to pump water into the boiler and then wait a pre-determined, user-selected amount of time before pulling their shot. This is designed to create an environment that starts the machine at approximately the same temperature each time and hopefully holds a reasonably constant temp throughout the pull. Testing by Greg Scace, Andy Schecter, and to a much smaller extent myself have shown this method effective. Without somewhat complicated, and relatively costly add-ons to the machines, it's about the best thing an owner can do in an attempt to eliminate, or at least limit, one of the variables in the making of espresso- water temperature.
I had been using this method (and had documented it as it appears on my website) for quite some time now with good effect. I have pretty much gotten consistent with all the other variables and was still getting a bit of difference between pulls- most notably I could see a difference between the first and second doubles I pull each morning (I normally only pull two doubles a day). The taste was similar, but I could occasionally smell a difference in the aroma of the espresso in the cup.
Andy Schecter contacted me and offered me the use of an inexpensive digital thermometer he had laying about. It's the type that is used in the kitchen for monitoring the temperature of cooking meat. It features a probe at the end of a long electrical wire. The probe is inserted into the meat and the other end is plugged into the display head which features an LCD readout of the temperature. These are available for around $15 to 25 from various retail sources (I saw a similar one at target the other week for about $16). "Send it on!" I replied.
When it arrived I could see that it had been modified. The probe, usually a 4" to 6" long metal probe about 3mm in diameter had its original probe material removed and replaced with a thin, short piece stainless steel tubing of a minimal length for easier placement in an espresso machine and faster response. He also sent along a bit of insulating wool to assist in mounting the probe.
As soon as it arrived, the top of Silvia came off and the probe was fitted on top of the boiler between the two ends of the heating element where they protrude from the boiler. The wire was fed down through the vent hole near the steam wand and then back around the brew head to the top of the machine so that the display could be easily read while working.
IMPORTANT: Before reading any further, please be aware that the following figures are of the external, metal surface of the boiler itself and not of the water hitting the coffee. Because of this, The temperatures reported here are an indirect measurement and are being used here as a mere indication of what is going on in the PF. Beyond that, it could easily be argued that this is the hottest surface of the boiler as well. The device being used to measure these temperatures is an inexpensive device and by no means can be considered accurate, although it does seem to repeat its measurements quite well. All temperatures, unless otherwise indicated, are in Fahrenheit. Beyond that, if you attempt anything like this be sure to unplug the machine before attempting to work on its interior and if your probe has any metal parts including, but not limited to, a metal sheath over the wire be aware that it is conductive and can cause a short in the machine if it touches any live wires. This first set of figures I recorded were for the idling temperatures. These were taken before I had made any coffee with the machine (on four separate morning sessions) and after the machine had idled for about an hour or maybe a bit more:
With one exception, the temps were fairly consistent, enough to say that the T-stat was operating in a dependable manner and the thermometer was reading in a way that was repeatable and dependable. My machine has the 110C. Thermostat (230F).
The temp surf and temperature recording is done this way:
- Open hot water wand and turn on hot water switch.
Each vertical column below represents one double shot. The temperature readings I recorded for four sessions, with two double shots each, were as follows:
After examining these figures it was clear that there was an inconsistency between the first shot and the second, with the second being hotter nearly every time. To combat this I decided to pull a blank shot before pulling the first shot of the day. I wanted to get to a point where my first and second shots (and hopefully subsequent shots when that happens) to be as equal in temperature as possible. For the next few days I would let the machine warm up as usually (for about an hour or so- sometimes more), and then when I was about ready to "go to work" I would just hit the brew switch and bleed about three ounces of water through the PF. After that I would proceed as usual. Here I used my normal temp surf timing method of about 40 seconds. Here's what I got:
The figures above were fairly typical of what I was seeing. I was noticing that the indicated temperature at the beginning of the pull were a good indicator of the temperatures later in the pull. I decided to try to see if i could get more consistency in temperatures, from one pull to the next, by surfing to a temperature instead of a time. Here's what I got:
In the pairs of numbers above I tried to start the pull when the temperature was indicated at about 211 or 212 or so. In both those cases the readout was fairly steady throughout the pull, only rising near the end when the resistance in the puck began to grow and the flow of water decreased. Further experimentation showed that 210 or 211 yielded the most consistent temperature through the pull, rising only about two or three indicated degrees.
From what I have been able to see (and smell and to some extent, taste) it seems that around 211 to 212 is a very good starting place for me. Based on that I experimented with "surfing to a temp" instead of surfing to a time. I further experimented with running the pump as usual until the light came on, but then let it run about five seconds longer. Dose, and tamp as usual, lock and load, but instead of using the clock to determine the starting point of the shot I used the temperature read-out. I started my pulls at about 212 and they ended at about 217 or so. Later I tried starting at about 210 and the shots ended at about 213.
Maybe more interestingly, I noticed that when the machine was choked a bit and the flow-volume of the espresso was slower than would be expected (and thus a longer pull in time than normal) the ending temperature was a number of degrees higher then usual, presumably due to the constricted flow of cold water into the boiler adn the greater amount of time the water was exposed to the heating element.
CONCLUSIONMy testing is still in the early stages, but from what I have seen, our old "time" surf is an effective method, but not as effective as having some sort of temperature read-out and surfing to that- the real "Temp Surf". The best solution would be a PID to accurately control the heating element and keep the boiler at one temperature without having to worry about other factors. Next would be a temp sensor installed in the brewhead, just above or below the showerscreen, or in the boiler (or all three!) to monitor the temperature of the water itself. Since these are beyond the ability and/or desire of many users to install (and expensive, in the realm of around $150-175 for the PID setup) the nest best thing would seem to be what I have been experimenting with here- Temp Surf to the boiler temperature and pull the shot when the boiler is at a user-determined temperature.
I will again remind you that the figures shown in this chapter are of a relative nature only as this thermometer is sensing the temperature at the top of the boiler's metal surface and this is by no means a calibrated instrument. As the sensor is not a standard one (it has been modified by Andy) and it is not mine, I hesitate to immerse it in boiling water or to freeze it to check its accuracy.
I am going to continue to test with this device, but for now these findings would indicate that using some sort of fast-reacting temperature reporting device could be an effective and inexpensive tool in controlling one more in what seems to be a never-ending chain of variables when making espresso. The actual start temp will take some experimenting to find the best taste point, but temp surfing seems to be a better solution than time surfing.
AddendumI have been using the digital thermometer for about two weeks now, and I am still learning how to use it most effectively. One of the more interesting things I have found is this:
As we now have learned, surfing to a temperature appears to be more effective than surfing to a time. For example, I have learned that I can make some pretty darn good espresso when I surf to about 212 degrees as stated above. But what I have learned in the last couple of days is that there are two ways to go about this. As before, I bleed the boiler through the steam wand using the hot water switch the light comes on indicating that the boiler is heating. The temperature begins to drop until the heating element catches up with the temperature of the water in the boiler (the hot water + the newly-introduced cold water). The temperature bottoms out at about 207 to 209 or so, and then begins to rise again. But if we graph the temperature changes we find that the indicated temperature is passed twice- once while the indicated temperature is on the downward trend, it bottoms out, then rises again, to once more pass through the target temperature.
So, when do you begin the shot? Do you hit the brew switch while the temperature is rising through your set point, or do you start the pull while the temperature is on the downward trend? Over the last few days I have been playing with both and what I have found is that if the pull is begun while the temperature is rising, it more often than not overshoots and you end up with about a five to seven degree difference between the beginning of the shot and the end (a range of 211 to 217 is common). On the other hand, if the pull is begun during the indicated decrease of temperature then the temperature stays very stable through the shot, often within three degrees or less.
Let's take a look at this chart (which represents a fictional set of temperature data). In this example the user has chosen 209 degrees as the temperature at which to begin the pull. But as we we see here, the temperature curve crosses through 209 degrees twice- once on the way down and once on the way back up. My observations seem to indicate that a more stable temperature can be found as the boiler is on the way down.