The attraction of PIDed single and double boiler machines is that you get the shot temperature you see; no fiddling, no flushing. If the shots are too bitter or too sour, you can change the temperature with a few button pushes, and get a predictable result. If you walk up to another PIDed machine, and use the same temperature and brew recipe, you will get the same result. Roasters are taking advantage of this cross-platform consistency to publish brew recipes that are actually useful and meaningful.
There is a catch. The temperature sensor controlling the brew boiler must be set close to the cold water inlet to work properly. At that spot, the brew boiler temperature needs to be several degrees higher than the temperature the puck will experience. This is the offset. The size of that difference varies by machine, and is sensitive enough so that it can vary among the individual machines of the same model. You may get a new machine where the offset is wrong by several degrees, and as the machine ages and scales, the offset may drift. Bottom line, you need to be able to calibrate the offset.
In the past, this calibration required the use of accurate portafilter thermometry like the Scace thermofilter or the earlier Schomer portafilter. Judgment was required using these devices, since their thermal mass delayed the readings for about 7 to 15 seconds. Here is a method that requires no thermometry and no judgement calls, since it relies on the liquid to vapor phase change, which is easily seen as flash boiling.
This technique has been used for at least ten years; but has usually been posted as instructions on how to use for particular machines. I think it should be published as technique that is widely applicable to many PIDeed espresso machines. Here is DaveC's 2008 instruction for the Izzo Alex DB.
Set your brew temp to 100C (or your local boiling point), let the machine stabilize. Flush the group without a portafilter. Do you see flash boiling? Now set it to 99C and let the machine stabilize. Flush the group. Do you see no flash boiling? Congratulations, your offset is perfectly calibrated.
Here is a video of a properly calibrated machine showing the flow at 98, 99, 100, and 101 ceontigrade. There is weak boiling at 99C that takes time to develop; to me this indicates I'm not quite there yet. But if you pick this point for your offset, your calibration is still going to be pretty good. The 98C is clearly to low; and the 101C flow is boiling no harder than the 100C.
The idea is to find the **lowest**
temperature at which you see flash boiling, the LFBT
(lowest flash boiling temperature). If it is at the local boiling point in your preferred temperature unit, you are set. If the LFBT is X degrees higher than you local boiling point, your offset is X degrees too low and needs to raised by X degrees. If the LFBT is X degrees lower than the local boiling point, your offset is too high, and needs to be lowered by X degrees. (There is a gotcha for Fahrenheit users, the offset is usually in Centigrade even if the panel reading is in Farenheit -- check your manual, or always do this calibration in Centigrade).
Please post your thoughts, criticisms and observations. I tried this technique on the Bianca I'm reviewing. With a few minutes work (along with a few hours of stabilizing time), it arrived at the same offset I got doing several days of thermometry. So I'm quite enthusiastic about how easy and physically elegant this is. However, there may be problems or ambiguities I didn't experience.