Tweet
This is going to be a long one, folks. Bear with me.
So for those that are wondering what became of Andy's Londinium - it made it's way to my workshop shortly after the L1 Luxe First Look thread descended into slander and wild accusations of ulterior motives, in the hope that we could turn it into something that performed as good if not better than was advertised.
That review was contrary to most of what was being written about the machine in the early days, but as I can now see, was pretty much on the money. As soon as I got it onto the bench I had to agree, the thing was just not that well finished for an object that was sold as a 'Luxe' with all that the word implies, and all that was advertised.
First of all, let me preface by saying, I have no particular interest in slagging off Londinium, Reiss or anyone else that extols the virtues of the machine. After all, he just inadvertently gave me a source of income.
He has done really well to bring his creation to market and it's very easy to sit on the sidelines and criticise his efforts, but as Chopinhauer I think said in the other thread, it just reinforces that it's no trivial task to design, construct, market and sell a high end espresso machine. I wish I could do it. Maybe I will one day, but not any day soon.
My one major gripe is that many things have been said publicly on the Londinium blog which are just plain silly, full of hyperbole and pseudo-science that in my opinion is designed to bamboozle the non-technical reader.
In no particular order, some below issues that Andy and Chokkidog observed and I agree-
-Dings in boiler, leaky fittings and poorly applied lacquer
-Poor finish of laser-perforated vents etc, ie. not finished post-cut to remove scorch marks and burrs
-Over-tightened fasteners that have permanently distorted panels
-Inattention to design issues with poor ergonomics and internal heat buildup (switch placement, vent behind the water tank)
-High tolerances in panel folds and poor alignment during assembly leading to user annoyances (drip tray impossible to remove when full)
In all fairness some of these issues are not unique to the L1, and many Italian machines are poorly finished if you look hard enough, but they are not labeled 'Luxe'. The bottom line is that this machine is not even close to being in the same league as a Rocket in terms of quality fit and finish and attention to detail.
So, to get on with it. Andy brought it round and gave me carte blanche to do whatever I thought needed to be done, and invoice him for the parts and labour accordingly, within reason! My first task was to figure out what was going on with the seemingly inconsistent results in the cup. By this stage there was plenty of data coming in from customers in the States with mixed reports about temperature stability - manifesting itself in some machines as a seemingly random drop in idle temperature measured at the exterior surface of the group, with correspondingly cool shots.
At this stage, the erroneous measurements were attributed to the way those users were pulling shots, somehow delaying the downward pull and pushing air into the thermosyphon. Bollocks, says I, and carried on logging measurements. I was definitely seeing the dip in the group temps. In normal circumstances, the thermosyphon works really efficiently and as soon as there is steam pressure in the boiler the group temperature comes up really quickly and reaches an idle temp in the low 80's C. However, sometimes it would not recover after a shot, and keep falling over until a quick flush was performed, or self-righting after a longer delay.
In short, I found that the group temps and resulting shot temps were perfect, as long as you could see the group temp starting to drop and reverse it with a quick flush. Occasionally it worked flawlessly by itself, often it didn't. So, the question became, why does this thermosyphon / heat exchanger have issues that conventional ones don't seem to have? As far as I am concerned, the answer lies in the source of water to the HX loop - namely the boiler. In a standard HX, the injection pipe that feeds the HX tube comes from pressurised cold mains water, or a vibe pump / tank, with a check valve in place to prevent backflow. Thus, when cold water is heated inside the HX, the thermal expansion of the water creates pressure as the system is a closed loop. Pressure increases up to the limit of the expansion valve, commonly 11 bar.
In the case of the Londinium with it's open thermosyphon fed from the boiler, there is no expansion and associated pressure increase as the boiler pressure and the thermosyphon pressure are always equal - there are no intervening valves, just a short length of copper pipe. This is not the first machine to work this way - my guess is that it's a Faema design (the Lambro for instance) but I have also seen pictures of Pavonis, Astorias and of course the QM Achille that work in the same way.
So, my theory was that if the thermosyphon loop was turned into a true HX (ie a separate hydraulic system), everything should work properly and the operating pressure in the HX would increase due to the thermal expansion of cold water entering the system. So I blanked off the boiler connection and fed mains water into the HX injector, with a one-way valve, expansion valve and gauge. And guess what, it worked flawlessly with rock solid group idle temperatures shot after shot, contrary to a statement by Reiss about how that configuration would be a total disaster.
So at that point I could have gone back to Andy and said you have to plumb the machine directly into the HX to get it to work, but clearly that wasn't going to work as Andy paid for a tank machine, and damn it thats what he wanted. By this stage Reiss had got his act together and offered a retrofit kit to add a one-way valve between the boiler and the TS tube to turn the TS into a closed loop. The key detail he missed out on was that in such a setup you need to be able to let the excess pressure escape other than by doing potential damage to pipes and seals... this is why we have expansion valves. He got there in the end and the current configuration is a combination Fluid-o-tech one-way and expansion valve.
At this point, the only viable solution seemed to be to drastically reconfigure the way the thermosyphon worked, and gain effective control of the group temperature. A cartridge heating setup similar to the Strega was investigated, but unfortunately there wasn't enough room in the group casting to get away with it.
Eventually I settled on shoehorning a small volume brew boiler directly behind the group, with short, rigid pipes attaching it to the TS ports. The one-way valve and expansion valve was retained, as well as the system of feeding it via boiler pressure. The control loop is very effective as the heat source is only about 60mm from the group, and water is directly circulated to the group by straight 12mm brass pipes with no restriction, and in operation the brewing temperature is completely discrete from the steam boiler, as it is in conventional dual boiler machines.
While I was redesigning the innards, I also did the following:
-Replaced Sirai pressurestat with an Expobar with a much tighter deadband (this is now a moot point except for the space saving)
-Moved mains switch to a more sensible location below the gauge
-Re-aligned the front panel and modified the drip tray to fit properly
-Machined the group sleeve to fit the E61 screen properly, so it bottoms out on the screen, freeing up about 2 mm of clearance in the basket
-Vented the anti-vac valve into the tank so it doesn't fog up the glass panels
-Stopped all leaks
-Replaced stock tiny steam wand with cool-touch standard length unit
There is still more work to do on the exterior, we plan to add a cup tray and vent the top panel but I am in the process of locating a stainless steel specialist to undertake this work, it goes well beyond my skill level/patience.
I'm pleased to report the machine is behaving dare I say it as advertised, with a few added niceties and the wonderful ability to compensate for changes in ambient temperature or bean choice by adjusting the PID setpoint - Reiss got it right when he said that these groups are EXTREMELY sensitive due to the enormous surface area. Even walking past the machine can cause a movement of air that will suck one degree out of the group's thermal mass. Quite amazing that the TS works at all!
Pictures of the work done so far to follow in the next post, I think this one's long enough, but I bet I've missed a few points that I wanted to expound on... oh well. I'll think of them soon enough.
So for those that are wondering what became of Andy's Londinium - it made it's way to my workshop shortly after the L1 Luxe First Look thread descended into slander and wild accusations of ulterior motives, in the hope that we could turn it into something that performed as good if not better than was advertised.
That review was contrary to most of what was being written about the machine in the early days, but as I can now see, was pretty much on the money. As soon as I got it onto the bench I had to agree, the thing was just not that well finished for an object that was sold as a 'Luxe' with all that the word implies, and all that was advertised.
First of all, let me preface by saying, I have no particular interest in slagging off Londinium, Reiss or anyone else that extols the virtues of the machine. After all, he just inadvertently gave me a source of income.
He has done really well to bring his creation to market and it's very easy to sit on the sidelines and criticise his efforts, but as Chopinhauer I think said in the other thread, it just reinforces that it's no trivial task to design, construct, market and sell a high end espresso machine. I wish I could do it. Maybe I will one day, but not any day soon.
My one major gripe is that many things have been said publicly on the Londinium blog which are just plain silly, full of hyperbole and pseudo-science that in my opinion is designed to bamboozle the non-technical reader.
In no particular order, some below issues that Andy and Chokkidog observed and I agree-
-Dings in boiler, leaky fittings and poorly applied lacquer
-Poor finish of laser-perforated vents etc, ie. not finished post-cut to remove scorch marks and burrs
-Over-tightened fasteners that have permanently distorted panels
-Inattention to design issues with poor ergonomics and internal heat buildup (switch placement, vent behind the water tank)
-High tolerances in panel folds and poor alignment during assembly leading to user annoyances (drip tray impossible to remove when full)
In all fairness some of these issues are not unique to the L1, and many Italian machines are poorly finished if you look hard enough, but they are not labeled 'Luxe'. The bottom line is that this machine is not even close to being in the same league as a Rocket in terms of quality fit and finish and attention to detail.
So, to get on with it. Andy brought it round and gave me carte blanche to do whatever I thought needed to be done, and invoice him for the parts and labour accordingly, within reason! My first task was to figure out what was going on with the seemingly inconsistent results in the cup. By this stage there was plenty of data coming in from customers in the States with mixed reports about temperature stability - manifesting itself in some machines as a seemingly random drop in idle temperature measured at the exterior surface of the group, with correspondingly cool shots.
At this stage, the erroneous measurements were attributed to the way those users were pulling shots, somehow delaying the downward pull and pushing air into the thermosyphon. Bollocks, says I, and carried on logging measurements. I was definitely seeing the dip in the group temps. In normal circumstances, the thermosyphon works really efficiently and as soon as there is steam pressure in the boiler the group temperature comes up really quickly and reaches an idle temp in the low 80's C. However, sometimes it would not recover after a shot, and keep falling over until a quick flush was performed, or self-righting after a longer delay.
In short, I found that the group temps and resulting shot temps were perfect, as long as you could see the group temp starting to drop and reverse it with a quick flush. Occasionally it worked flawlessly by itself, often it didn't. So, the question became, why does this thermosyphon / heat exchanger have issues that conventional ones don't seem to have? As far as I am concerned, the answer lies in the source of water to the HX loop - namely the boiler. In a standard HX, the injection pipe that feeds the HX tube comes from pressurised cold mains water, or a vibe pump / tank, with a check valve in place to prevent backflow. Thus, when cold water is heated inside the HX, the thermal expansion of the water creates pressure as the system is a closed loop. Pressure increases up to the limit of the expansion valve, commonly 11 bar.
In the case of the Londinium with it's open thermosyphon fed from the boiler, there is no expansion and associated pressure increase as the boiler pressure and the thermosyphon pressure are always equal - there are no intervening valves, just a short length of copper pipe. This is not the first machine to work this way - my guess is that it's a Faema design (the Lambro for instance) but I have also seen pictures of Pavonis, Astorias and of course the QM Achille that work in the same way.
So, my theory was that if the thermosyphon loop was turned into a true HX (ie a separate hydraulic system), everything should work properly and the operating pressure in the HX would increase due to the thermal expansion of cold water entering the system. So I blanked off the boiler connection and fed mains water into the HX injector, with a one-way valve, expansion valve and gauge. And guess what, it worked flawlessly with rock solid group idle temperatures shot after shot, contrary to a statement by Reiss about how that configuration would be a total disaster.
So at that point I could have gone back to Andy and said you have to plumb the machine directly into the HX to get it to work, but clearly that wasn't going to work as Andy paid for a tank machine, and damn it thats what he wanted. By this stage Reiss had got his act together and offered a retrofit kit to add a one-way valve between the boiler and the TS tube to turn the TS into a closed loop. The key detail he missed out on was that in such a setup you need to be able to let the excess pressure escape other than by doing potential damage to pipes and seals... this is why we have expansion valves. He got there in the end and the current configuration is a combination Fluid-o-tech one-way and expansion valve.
At this point, the only viable solution seemed to be to drastically reconfigure the way the thermosyphon worked, and gain effective control of the group temperature. A cartridge heating setup similar to the Strega was investigated, but unfortunately there wasn't enough room in the group casting to get away with it.
Eventually I settled on shoehorning a small volume brew boiler directly behind the group, with short, rigid pipes attaching it to the TS ports. The one-way valve and expansion valve was retained, as well as the system of feeding it via boiler pressure. The control loop is very effective as the heat source is only about 60mm from the group, and water is directly circulated to the group by straight 12mm brass pipes with no restriction, and in operation the brewing temperature is completely discrete from the steam boiler, as it is in conventional dual boiler machines.
While I was redesigning the innards, I also did the following:
-Replaced Sirai pressurestat with an Expobar with a much tighter deadband (this is now a moot point except for the space saving)
-Moved mains switch to a more sensible location below the gauge
-Re-aligned the front panel and modified the drip tray to fit properly
-Machined the group sleeve to fit the E61 screen properly, so it bottoms out on the screen, freeing up about 2 mm of clearance in the basket
-Vented the anti-vac valve into the tank so it doesn't fog up the glass panels
-Stopped all leaks
-Replaced stock tiny steam wand with cool-touch standard length unit
There is still more work to do on the exterior, we plan to add a cup tray and vent the top panel but I am in the process of locating a stainless steel specialist to undertake this work, it goes well beyond my skill level/patience.
I'm pleased to report the machine is behaving dare I say it as advertised, with a few added niceties and the wonderful ability to compensate for changes in ambient temperature or bean choice by adjusting the PID setpoint - Reiss got it right when he said that these groups are EXTREMELY sensitive due to the enormous surface area. Even walking past the machine can cause a movement of air that will suck one degree out of the group's thermal mass. Quite amazing that the TS works at all!
Pictures of the work done so far to follow in the next post, I think this one's long enough, but I bet I've missed a few points that I wanted to expound on... oh well. I'll think of them soon enough.
Comment