Post By Steve82
Gaggia Classic temperature stability 'study'
I stumbled across this today, and thought it was interesting (haven't finished reading it yet). Don't know if it's been posted before.
With a change in the control methodology, the instability can be reduced to <2°C with a time between shots of less than two minutes.
I think wrapping the boiler feed around the boiler would seem easier.
Easier than what, fitting a PID and pulling a cooling flush before your first shot?
By changing the control methodology I mean intentionally running the boiler hotter than your desired brew temp and not allowing the boiler temp to recover (as distinct from controlling the boiler to the desired setpoint and waiting ages between shots for the boiler temp to recover).
That was the first thing I thought of when I read the initial report. It should reduce that sharp drop caused by the cold water inflow. I must find a source of 6mm or ¼ " copper tubing and give it a try.
Originally Posted by MrJack
I have a PID for my Classic. but I haven't got around to fitting it yet, partly because I'm pulling pretty good shots without one. I have a K-type TC clamped to the top of the boiler, reading out on a DMM , so I can watch the temperature rise/fall and start the shot at about the same point of a cycle. Some of my best shots lately have been started just before the thermostat kicks the heater elements on.
This may be because the thermostat in the machine I am currently using seems to run it fairly hot. It switches the heater elements on at about 98°C and the temperature climbs to about 110° before it peaks and starts to drop again. I have another Classic, and it cycles between 93° and 106°C. I had a spare thermostat, so I swapped it into one of the machines, and it was in between the other two, at about 95 - 108°C.
All these readings were taken when the machines had been on for at least 15 minutes and were well stabilised.
All three are labelled as 107° thermostats, and each of them is quite consistent, but they vary from one another, perhaps because they are all different ages.
This is a another good reason to install a PID, but for the moment my taste bud are just not prodding me into doing it.
That will help increase the average (and peak) shot temperature, but won't do anything for stability during the shot, which was the subject the study.
Originally Posted by Dragunov21
Yes it will, as it applies a greater amount of heat energy to the water flowing through the boiler, without increasing the peak temperature.
When you start the pull it's already heating, rather than waiting until the temp drops before kicking in the element. Temperature-surfing is a somewhat similar principle (if less consistent and effective).
I'll say again, I've tested mine with a calibrated probe and, at the puck, with a flow-rate equal to that when I pour a double, the temperature rises ~0.5°C above setpoint and falls <1.5°C below setpoint.
Last edited by Dragunov21; 14th August 2013 at 07:17 PM.
I agree, ive been running my PID classic a bit hotter SV, tweak the PID settings so that it actively hunts this temp, a quick flush for like a second or 2 before locking in. I am running the Auber RTD that screws directly into where the thermostat used to be.
Sure theres still some temp drop but nothing like in that article. Some coffees respond well to a slight drop off in temp, i prefer some shots i get out of the classic to BES900. Then again i prefer shots out of my LP lever better than both. Ive learnt that having a stable temp through the whole shot is not the be all end all.
I guess your milage may vary
I don't know if you have read the full article, but the author tested the effect of having the element on the entire shot with negligible effect.
He also found that the temperature of the liquid coming out of the puck was not closely coupled to the inflow temperature from the boiler, and thus measuring here gives a 'false' impression of stability (although this raises the question of where temperature stability is most important).
So, if you are measuring the flow out of your PF, you may not be seeing the full variation in boiler water temperature (and you may well be exposing the coffee to hotter than desirable temperatures).
Edit: Sorry, just noticed you measured at the puck. Upstream or downstream? (measurements of both would be interesting!)
My measurements were conducted with a bare probe, dead on top of the "puck".
The puck was made of sponge of sufficient thickness/compression to regulate flow (because I had to pull a tonne of shots and obviously an actual coffee puck doesn't react the same way in subsequent pulls as the first).
Admittedly this doesn't take into account the variation in flow that could occur while the puck saturates (I don't know whether a non-saturated puck is of greater, lesser or similar restriction to a saturated one).
I set the PID to auto-learn and played with setpoint, initial flush quantity and time between shots until I found a combination that provided stability I considered "as good as I could be bothered with".
I find the part about the element being on full not having any effect very interesting...
What I found is that with the element at 100% (as results initially from pulling a cooling flush) the puck temp still drops as the shot progresses, but at a much slower rate. Keep in mind that my boiler housing temperature is controlled to about 120°C so it transfers heat to the water much faster than a stock machine or PID'd machine with a standard control strategy.
After the shot (or cooling flush), it takes 1:30 for the water in the boiler to heat sufficiently to give the +0.5/-1.5 result. It's easy to tell when this point is reached because the heating load is indicated by the %OP required to maintain boiler housing temp. On my machine, the housing temp stabilises after a shot with the %OP sitting at ~9-15% (because the water is sucking heat out of it). This gradually drops to ~6-9% (as the water heats up and less energy is required to maintain boiler temp. As it sits further, the required %OP drops to ~1-4%, which indicates the water is far too hot (and is likely above 100°).