There's a few threads about Londinium I and I've posted some messages there of a quest to find the best "modus operandi" to get the finest cup the easiest way. Over the months it's become difficult even for me to find back these messages so I sum them up compactly here, leaving out earlier paths I went down that were maybe helpful but not essential in getting that maximum stability shot after shot.

The best parts of this aren't my 'invention' of course, it's just a result of the things I've learned and picked up from those smarter than I.


OPERATION

Quoting Kfir: "Optimum boiler pressure is ~ 1.2 bars and it works well for room temperature up to 28°C."

If you set a timer, the L-1 will be about ready to make espressos after an hour but it can be done quicker.

Starting up with the steam valve open, closing it again when a firm plume of steam comes from it also makes sure there is as little air in the boiler as possible.

Quoting Kfir: "To warm up the machine faster in the morning 4-5 small open port flushes in 2-3 minutes intervals work great (also for clearing air from the TS after maintenance)."

Recently (copying what Kfir and others are doing) I've started to do pulls with just 3 seconds pre-infusion and a very short flush (just 60-70ml) after every espresso. This short flush not only cleans the screen of any coffee grains sticking to it, it also ensures the thermosiphon will continue to flow with the shortest interruption. I've pulled many espressos with no need to do this flush as there was nothing to correct, but sometimes, like after a long idle time or a longer pre-infusion, the TS tends to need time to get to work again. So now I just do it always, as it doesn't hurt any process anyway.


BACKGROUND INFORMATION

During pre-infusion, the lever is down, the piston is up and the thermosiphon flow that keeps the group warm is stopped. So when pre-infusion is stretched more than (for instance) 3 seconds, the group cools down more and afterwards the TS pipes that have also cooled down will have lots of work to do. They do this at a slow pace, which is good in a way because it prevents the group overheating if you pull many shots. But in such cases the short flush helps to get the group temp up a bit and the TS flowing.

To illustrate, I've collected a few pictures and graphs from my earlier postings.

A recent one of 6 shots, each followed by the 60-70ml short flush:
("BT" in the graph in this case is the probe connected to the outside of the group, about halfway up the neck and about 2cm away from it to the side. You also see the values of 2 probes connected to the TS pipes. While roasting with this software Artisan doing the logging, BT means Bean Temperature)



An example of a pull after a longer idle time, the TS needing some time to pull itself together and a flush (this time using the Scace to do it, just testing it out) intervening to get it all tight and flowing immediately again. When always doing a short flush after a pull, it's not necessary to monitor this and occasionally do one.



A number of Scace shots. The PF started off cold but most other shots had practically identical temperatures. At minute 225 you can see the TS pipes swapping their flow direction and it's clear that it's basically inessential which way it flows. I agree with Eric S here, who has stated before how the direction of flow is inconsequential.



A few more shots where one TS reversal was observed without any problem.
(A static disturbance temporarily stopping the TS probes communicating visible in the 'flat lines' mins. 140-145):



Machine warmup and 8 pulls, 2 months ago:
(Note the probe was not in exactly the same spot then, on the outside of the group and it shows a higher temp than elsewhere. The exact temperature doesn't say much as the graph is about the stability, the tendency to keep coming back to the same temperature position in the shortest time and remain there even during idling. Every pull having a temperature peak at the same peak is also indicative of stability over time.)



An example of the machine warming up, the lower TS pipe sitting a little closer to the warming boiler and heating up a little quicker, then one flush 'correcting' this and inducing the upper pipe to be the 'hotter' one:



A graph from April this year with lots of espressos in a short time. It demonstrates how the group won't overheat easily.



Three pictures of the hot water jets inside the group doing the pre-infusion: as soon as the lever is pulled down and the piston rises, these four water inlets are opened and the jets meet in the middle of the group, evening out their propulsion and then raining down evenly on the shower screen and next on the puck.

A Scace device measures the temperature of that pre-infusion rain.







An example of how I positioned the probes when I had two probes on the group. In the graphs used here I had just one probe on the group and that probe was a little closer to the neck and a little higher than in this picture:



And a picture of the Amprobes used with the Scace in between:

Frans, this is most generous of you, and I've "Command D'd" this thread for safe keeping!

Excellent idea and thanks for the consolidation

Thank you, perfect timing as my londinium should be here next week.

It seems Reiss has had another thought on this subject ..
http://londiniumespresso.com/blogs/lond ... from-day-1

And Rick Bond also claims to have resolved all the temperature stability issues without the need to flush.
..but i will let him explain !!

Thanks for the link, it's good to see Reiss and Frans agree:

frans wrote:Recently (copying what Kfir and others are doing) I've started to do pulls with just 3 seconds pre-infusion and a very short flush (just 60-70ml) after every espresso. This short flush not only cleans the screen of any coffee grains sticking to it, it also ensures the thermosiphon will continue to flow with the shortest interruption.

Reiss wrote:After each shot it is recommended that you pull a short flush of approximately 60mL. This ensures the necessary thermal imbalance is reintroduced to the system, resulting in the thermosiphon promptly recommencing.

Basket probe
I built a little basket probe, based on a modification that Peter from http://www.kafko.nl made me a while ago. It originally held a shielded probe but that one responded a little too slow. The Scace probe wins there and I wish I could take that one out to use in my own basket setup.



The K-type wire probe is sealed using Loxeal 55-03 which is food safe. It can withstand 150°C (300°F).

A close up of the tiny "Loch Ness monster" sticking out of the lake blue Loxeal:



My basket probe is not a standardized one so someone else who also builds one will probably have slightly different results. I also noticed, as I've seen before, that although the Scace shows consistent results, these may not be reflecting what happens in an actual coffee puck.

Bottom view:


Measurements with real coffee in a puck packed around the probe differ over shots. A faster flow gives a higher temp reading, for instance. So it's best to run a series without changing too much in your dose, grind and tamp routine, and draw any conclusions from the overall trend.

The coffee puck neatly covers the probe, also after tamping:



Pressurestat
I've also played a bit with the p-stat setting, to see what could be an optimum for brew temp. Over the past weeks I had the p-stat clicking off just before 1.2 bar and during today's testing I turned it slightly upwards, a tiny bit over 1.2 and the spot where I usually measure 86°C (186°F) I now have 87.5°C. This lifted the taste out of the fruity flavors more towards sweetness, but I liked all of the shots I tasted. Most shots peaked around 93-94°C (199-201°F) in the puck with outliers of 92°C and 96°C. I will try out this slightly higher group temp some more and decide later if I want to keep it.

One typical graph:



Conclusion (for now)
My impression is that the p-stat setting and thus group / brew temperature can be tweaked a little to play and learn, but none of it was necessary to get it "right" for me.

Against my own advice to not change any variable during puck temp measurement, I did change the p-stat a little and I also changed grinders in the end.

Grinding
I consistently used 17g of Brazilian beans, Fazenda Passeio roasted by Londinium and I mostly used the Mahlkönig Vario, switching over to the HG One when I was particularly happy with the taste / temperature combination. But then the flow was a bit too fast and I needed to change the HG One setting a bit... et cetera.

The reason to use the Vario is that my right "grinder turning" elbow hurts a bit after many recent HG One doses for previous measurements

First drops, then flow:






In the cup:





Puckology view, the probe hasn't broken the puck:

As usual, great work Frans.

I also found that a faster flow gives a higher temp reading (during my very rudimentary measures). When I was using the same coffee/dose/grind, I was getting 197-200f very consistent.

When I changed coffee which resulted in a faster flow, I got readings as high of 205f. This really puzzled me. Why is this? I believe this point is critical in understating the machine and being able to get an optimal brew temp and results. Is this a measurement anomaly or the reality of what's going on in the cup? I believe it is reality.

In summary:
- fine grind, slower pour = lower brew temp (pstat of 1.2b recommended?)
- coarse grind, faster pour = higher brew temp (pstat of 1-1.1b recommended?)

Miami_AJ wrote:When I changed coffee which resulted in a faster flow, I got readings as high of 205f. This really puzzled me. Why is this?

Try this:

With the tap on hot water, run it very very slowly and put your hand under it. Observe you want to take it away to prevent burning but you can tolerate it a bit.

Then with the same heat, and lots of water flow, put your hand there again. OUCH!

Lots more energy is transferred.

Also, when the flow in the puck is slow, some of the heat / energy has time to go elsewhere, like into the metal mass of the PF and group.

In some pulls, it was more noticeable than in others that the peak temp was reached as the first water drops reached the puck & probe, then subsided a bit as the puck was slowly soaking up water and peak temp was touched again (a little higher even) as the lever was released and water under more pressure passed the probe (the seconds around the minute 13 mark below).

Thank you Frans, I love the pictures...

Few small observations of my own:

Optimum boiler pressure is ~ 1.2 bars and it works well for room temperature up to 28c.

Group idle temperature in the range of 83-85c gives out the best results, I would not recommend going beyond 86c and adjusting the PSTAT accordingly.

From measurements I did I got about 95c peak brew temp at 1.3 bars so ~94c for 1.2 bars makes sense.
The length of the pre infusion also imapcts the brew temperature.

To warm up the machine faster in the morning 4-5 small open port flushes in 2-3 minutes intervals work great (also for clearing air from the TS after maintenance).

Kfir.