by ianh on Sat Oct 29, 2011 10:46 pm Firstly, pressurization is fast, and there is often an initial flow surge at the very start of pulling a shot, which seemed to flush through fines. This varied with different coffees and grind, and initial thoughts were to try a smaller gicleur. However the flow rate for the rest of the shot was good. Pre-infusion was tried as an alternative, and gave great results. I had intended to purchase the semi-automatic as reviewed; however production of this model has been discontinued, with only the model T1, the automatic equivalent with programmable dosing buttons, being made.
In retrospect, I am glad I got a T1; the programmable dosing is useful. But it also made it very easy to add pre-infusion: the hole where the manually operated switch used to be had been covered with a blanking plate, and the lever and switch assembly were available as a spare part. It was easy to fit, and the switch, wired in series with the pump, allows a shot to started without the pump, and then once pre-infused, to be finished off at full pressure. The supply pressure, which provides pre-infusion, is 3 to 3.5 bar. Initial flow rate is improved, and the shots are more consistent. And the installation looks good using the original Elektra part as shown in the photo.
The second issue involved the boiler filling. We drink mainly 'long blacks', which involve some water being added to the cup before pulling the shot. Due to the vagaries of the fill control system, shots were often being interrupted by the fill pump solenoid cutting in during a shot, noticeably affecting the results. The vagaries are common to most machines: that is, rather than using a dual probe to exactly control upper and lower levels, they use a single probe level control, and electronic timer to determine fill level. There is a delay in starting the fill once the probe is above water level, and a delay in stopping the fill once the water level is above the probe. The result is unpredictable, and sometimes disruptive, filling.
The Elektra A3 very conveniently has both delivery and fill indicator lights which are easy to break into for wiring modifications. A relay was added to isolate the fill solenoid when delivery was happening. This solved the initial problem; however, due to the design of the 'brain box' it resulted in 60 volt across the fill indicator light. The brain box seems to be using the fill solenoid coil to hold it low when delivery is occurring. A second modification using two relays as shown below, where the solenoid valve is only isolated when both fill and delivery are occurring simultaneously, avoids this problem.
At about the time these modifications were being done, 'false pressure' was very occasionally experienced, and the first thoughts were that they were some side effect of the modifications. All modifications were removed, reverting to factory default, but still they occasionally occurred. Before reinstating the modifications, an effort was made to identify the cause, which is not easy as they are intermittent and infrequent. A lot of time was spent measuring temperatures and pressures. Much has been written about this phenomenon with lengthy explanations including Dalton's Law of Partial Pressures, Boyles Law, the function of vacuum break valves, and many other ideas some of which are obviously correct, and others quite spurious.
My conclusion is that: false pressure only occurs on startup from cold, and results from normal operating pressure being obtained without a saturated atmosphere present in the boiler. The normal pressure stat controls temperature by switching the heating element on and off at a saturated pressure that is equivalent to the required temperature. If for some reason this pressure is achieved at a lower temperature without a saturated atmosphere, a temperature to achieve saturated atmosphere is never reached. By observations, this pressure can be achieved with a combination of circumstances: probably due to contraction of the water volume during previous cooling the level can be below the filling probe when first turned on. During previous cooling, the vacuum break valve admits air and most moisture will condense, more so when the tank is colder. During the startup, the combination of water being heated and expanded, the air being heated and expanded, and the the water volume being increased due to filling action, operating pressure is achieved below boiling point - by my observations, usually around 95 to 98 degC.
The problem is so infrequent as to not warrant concern. However, the answer is probably to fit a thermostat (ideally PID controlled) rather than a pressurestat. Interestingly, some of the Elektra machines do have thermostats fitted.
Fit in the nifty Arduino controller than another member has fiddled with and voila. 
