It's long been my delusion that scale is tastable. By that I mean that when scale gets beyond a certain point, the cup changes.

To test this hypothesis, I borrowed a Scace II and took a series of measurements before and after a full boiler and HX descale. Time elapsed since last descale was about 4 months. Water is Crystal Geyser ~100 to ~115ppm. On 3 hours per day.

All shots : 5:30 interval, ~70ml in 30 seconds @ 122 PSI.
Scace remains in group.
EDIT : Shots replotted as series to better illustrate.
PID Tuning for Maximum Stability : Boiler temp 228±0.3°F



Shot temperature very even on the Pre series, but with more rise during shot.
Shot temperatures rise with subsequent shots on the Post series

=============================================

Post Descale. PID Tuning to mimic pStat : Boiler temp 228-0.7/+2°F



Shot temperature falls after first shot, then tightly grouped.


Data plotted using http://www.ielogical.com/coffee/ShotPlot.xls, an Excel workbook that takes the work out of plotting layered and series graphs.

Interesting, but lacking a control. Do a few successive days of measurement so that you can be sure the jump is outside the usual day to day variation.

All series done the same afternoon:
Pre Descale the machine had been on for about 4 hours, 1 hour post the last real shot.
Post Descale the machine idled for an hour with a half hour between the PID parameter change.

In another 14 weeks, I can borrow the Scace again and repeat the test.

This series after 1 year without descaling

The parameters :
Vibiemme HX PID
Boiler 227°F, PID set for maximum stability. Boiler typically idles within ±0.3°F
Ambient 78.5°F , no air con, no windows open for still air
Inlet water 78.5°F in a container outside the machine, OPV not returned
Scace and TC adapter in group 1.25 hours before measurements start
Pull 60ml every 10 minutes for 5 shots pre and post descale
Scace not removed between shots





The first pull after the long idle period are close, but the descaled boiler and HX appear to transfer heat more efficiently, as expected.
Probably the only information is that PID control on an e61 HX boiler does not guarantee double-OH-naught control on the business end of the group.

NOTES :
- The intershot interval is suppressed.
- The post descale jet from the scace was much more violent and volume was difficult to judge on the beaker. The Scace is dried after use with compressed air to ensure the orifice and filter remain clean.

It looks like the heat transfer to the HX is more efficient post-descale, which is to be expected. So instead of the water cooling off slightly shot to shot, it now heats up slightly. Has the tuning on the PID changed? If lime was removed from the sensor, you could run a larger proportional gain and lower integral & derivative times. That way the faster loop may keep up with the faster heat transfer.

Other than the 1°F change in the set point the tuning is the same.

As far as the tuning change suggestion, that's essentially what the thermostat On/Off mode in the first post does. Effectively, P is 100%, I & D are 0. What's interesting is the boiler oscillates several degrees, but the group is more stable. A change in the shot interval could exhibit different results.

I'm not certain, but I think the thermosyhon flow, at least on the Vibiemme HX with its restrictor, is less efficient when there is little variation in the boiler temperature. With tight control, there is probably less turbulence and a more gradual thermal gradient. With the heater full On / Off, the top of the thermosyphon tube is hotter relative to the bottom and flows better. With less gradient and reduced flow water over heats in the thermosyphon and get blasted out as a lump, heating the group.