Talk to me about ACS Evo Leva - Page 2

Recommendations for buyers and upgraders from the site's members.
Starguru (original poster)
Posts: 79
Joined: 9 months ago

#11: Post by Starguru (original poster) »

How does the Bezzera Strega fit into this conversation?

Primacog
Posts: 893
Joined: 2 years ago

#12: Post by Primacog replying to Starguru »

Someone else with actual user knowledge of the strega ahould chime in soon but as one who has only read about it, it appears to me that you cannot set at will the temperature of the grouphead and brew boiler independently nor can u set preinfusion pressure level either. The strega can be modified apparently but the full capahilities of a hybrid lever don't come with the product itself.
LMWDP #729

mgrayson
Supporter ♡
Posts: 655
Joined: 17 years ago

#13: Post by mgrayson »

So a general question about multi-parameter machines: How do you set them in practice? Optimizing functions in high dimensional spaces is ferociously hard, and if every function evaluation requires tasting an espresso, it would take decades. Optimizing one parameter at a time is a famously bad way to do it.

I mean, grind, dose, and volume are hard enough!

User avatar
baldheadracing
Team HB
Posts: 6281
Joined: 9 years ago

#14: Post by baldheadracing »

Starguru wrote:How does the Bezzera Strega fit into this conversation?
The Strega has:
- the Bezzera group; essentially an Astoria/CMA group without the sleeve so a single inlet hole;
- a horizontal cold-fed HX;
- group heaters controlled by thermostat; and
- is built on the same chassis as many other machines

What that means to me is:
- much lower cost without sacrificing build quality because many, many parts and components are shared with other Bezzera machines - and, although not public knowledge, other manufacturer's machines as well. Bezzera is one of the few companies that does a lot of manufacturing in-house instead of buying parts/assemblies from outside suppliers, which is another reason why their machines are less expensive than other companies. OTOH, if you want something other than a chrome box ... ; (As an aside, Bezzera has supplied everything from parts to complete machines to other companies.)
- cold-fed HX means you can control the brew temperature profile by flushing. In my experience, I only flush my Strega for five+ seconds to lower the brew temperature for Southern Italian (e.g., Naples) roasts. Northern Italian and medium roasts get a quick (two-second) flush. Anything lighter I walk-up-and-pull. Note flushing routines/times may vary between machines. (I have a Scace so I measured mine.) My usual roasts are medium-light and lighter;
- not being a dual boiler means steaming characteristics affect the HX. For example, in the ACS you can raise/lower the steam temperature (and thus pressure) independently of the brew boiler. In the Strega, raising the steam pressure (and thus temperature) will raise the brew temperature. However, as stated in the previous point, one can control brew temperature with flushing;
- the Evo Leva has both brew temperature adjustability and group heater adjustability; by separate PID's if I understand correctly. This lets you change the brew temperature profile independently of the (average) brew temperature. The Strega's group heater thermostat is non-adjustable; however it is a common mod to change the thermostat for a PID.
- the Strega does not have a brew pressure gauge, nor is it trivial to add one due to the design of the group (unlike the LSM group in the Evo Leva/Nurri/etc.). However, it is a common mod - and not difficult - to add a pre-infusion pressure gauge so one can see different pre-infusion pressures. I'd consider this a highly desirable mod if one drinks a wide variety of roast levels. (One can also do the dimmer mod to control the pump's flow.)

... and the big question: How does the Strega compare on taste? I would say that they are going to taste different, even when differences due to basket diameter (and thus bed depth) are removed. I have zero blinded experience with LSM group machines, and have never used an Evo Leva or a Nurri. The LSM group certainly has a strong reputation for excellent espresso from dark Neapolitan roasts/blends, so if those are what you enjoy, then that would be a consideration for me.

In summary, I would say that the Strega - and the Evo Leva and Nurri - are flexible machines that can handle a wide variety of coffees and roast levels. Of the three and independent of budget, I would go with the Nurri just because I think that it looks best. With budget in mind, I have a Strega that I bought used :D.
-"Good quality brings happiness as you use it" - Nobuho Miya, Kamasada

Starguru (original poster)
Posts: 79
Joined: 9 months ago

#15: Post by Starguru (original poster) replying to baldheadracing »

Very helpful, thank you. Trying to work thru whether I need this much complexity for the 1-2 shots I make on a normal day. Maybe a CT2/Export/la Pavoni would be more appropriate....

DaveC
Posts: 1774
Joined: 17 years ago

#16: Post by DaveC »

mgrayson wrote:So a general question about multi-parameter machines:

1.How do you set them in practice?
2. Optimizing functions in high dimensional spaces is ferociously hard,
3. and if every function evaluation requires tasting an espresso, it will take decades.
4. Optimizing one parameter at a time is a famously bad way to do it.

I mean, grind, dose, and volume are hard enough!
1. It requires the person to
  • Not be a complete beginner
    Understand the objectives they want to achieve (based on experience of different coffees)
    Understand how the functions work and then using common sense
    If you find grind dose and volume "hard enough", stick to a basic machine, I and presumably a lot of experienced people don't find it difficult at all
2. I don't understand what this means, sound impressive but conveys no inherent meaning as to the problem you envisage?

3. Decades implies 20+ years, or minimum 21,900 espressos @ 3 espresso shots a day?

4. Take the game of chess, there are 20 legal moves at the opening, of which probably only 3 or 4 are worth doing. Something called the branching factor comes into play. This is about 35 per ply (move). This gives on average 50-100 legal moves per turn in the middle game. This means if you are looking ahead 2 or 3 ply, the number of potential positions and moves you have to consider for each of them is extremely large. A Master or grand master, or any other player for that mater never evaluates the potential 100M positions (if I've done my math right) possible. That doesn't make sense, instead they reject clearly bad plays rapidly and concentrate on the few things likely to yield a good position

It's the same with setting parameters on an espresso machine. If I use the example of temperature, although there are 100 temperature points, generally we can use 5 or 6 of them only. If it's a dark roast, or a light roast, that probably drops to 3.

Setting a PID can be done initially using knowledge and then tweaked fairly rapidly, especially as often they are set up by people like me on machines for different companies.

It's a real life problem not a pure math problem, consisting of parametric and non-parametric behaviour. The boundaries however are not the extremes of each function, but a rather narrow set of valid parameters (like the first "good" moves in a game of chess).

The 3 main objectives I believe we should be looking for from a machine to optimise the flavour gained from particular coffees are:
  • Consistency
    Repeatability
    Control
The "dynamic" parameters to achieve this are usually only:
  • brew temperature
    brew pressure
    Preinfusion pressure
    preinfusion time
    Shot time
With the above there are still quite narrow regions where the parameter will be valid e.g. a preinfusion time longer than 15 seconds and shorter than 4 seconds...even if the range is potentially infinite, or a shot time longer than 40 seconds. PID parameters are not dynamic and when set should remain set correctly. I am however doing some new work around this particular area

So although your question cannot be answered in a single forum post...I think the problem is no where near as difficult as you "appear" to believe. It's also one of the reasons forums exist...the whole process can be simplified when many owners do the work, then share the results. It's also as complicated as you wish to make it...and as you learn you can tweak more things.

User avatar
baldheadracing
Team HB
Posts: 6281
Joined: 9 years ago

#17: Post by baldheadracing »

Starguru wrote:Very helpful, thank you. Trying to work thru whether I need this much complexity for the 1-2 shots I make on a normal day. Maybe a CT2/Export/la Pavoni would be more appropriate....
Those are three quite different machines - and all are quite different from an Evo Leva.

There's a spectrum - I guess two spectrums as they can overlap. At one end there are folks who would be happy if one could just push a button and the machine makes espresso (and milk drinks). Let's call that end, "convenience." At the other end are folks who want to try all kinds of different coffees and different brew variables. Let's call that end, "experiment."

Now take that to machines: I'm sure many would want a machine that offers push-button "convenience," but one that you can "experiment" with when one desires ... but how much "convenience" do you want, how much do you want to "experiment," and then there are other machine characteristics; aesthetics, size, capacity, ease of use, reliability, support, maintenance, availability ... and, of course, cost.

I think that it helps to consider those factors and narrowing them down before considering a technology solution.

Good luck!
-"Good quality brings happiness as you use it" - Nobuho Miya, Kamasada

mgrayson
Supporter ♡
Posts: 655
Joined: 17 years ago

#18: Post by mgrayson »

DaveC wrote:1. It requires the person to
  • Not be a complete beginner
    Understand the objectives they want to achieve (based on experience of different coffees)
    Understand how the functions work and then using common sense
    If you find grind dose and volume "hard enough", stick to a basic machine, I and presumably a lot of experienced people don't find it difficult at all
2. I don't understand what this means, sound impressive but conveys no inherent meaning as to the problem you envisage?

3. Decades implies 20+ years, or minimum 21,900 espressos @ 3 espresso shots a day?

4. Take the game of chess, there are 20 legal moves at the opening, of which probably only 3 or 4 are worth doing. Something called the branching factor comes into play. This is about 35 per ply (move). This gives on average 50-100 legal moves per turn in the middle game. This means if you are looking ahead 2 or 3 ply, the number of potential positions and moves you have to consider for each of them is extremely large. A Master or grand master, or any other player for that mater never evaluates the potential 100M positions (if I've done my math right) possible. That doesn't make sense, instead they reject clearly bad plays rapidly and concentrate on the few things likely to yield a good position

It's the same with setting parameters on an espresso machine. If I use the example of temperature, although there are 100 temperature points, generally we can use 5 or 6 of them only. If it's a dark roast, or a light roast, that probably drops to 3.

Setting a PID can be done initially using knowledge and then tweaked fairly rapidly, especially as often they are set up by people like me on machines for different companies.

It's a real life problem not a pure math problem, consisting of parametric and non-parametric behaviour. The boundaries however are not the extremes of each function, but a rather narrow set of valid parameters (like the first "good" moves in a game of chess).

The 3 main objectives I believe we should be looking for from a machine to optimise the flavour gained from particular coffees are:
  • Consistency
    Repeatability
    Control
The "dynamic" parameters to achieve this are usually only:
  • brew temperature
    brew pressure
    Preinfusion pressure
    preinfusion time
    Shot time
With the above there are still quite narrow regions where the parameter will be valid e.g. a preinfusion time longer than 15 seconds and shorter than 4 seconds...even if the range is potentially infinite, or a shot time longer than 40 seconds. PID parameters are not dynamic and when set should remain set correctly. I am however doing some new work around this particular area

So although your question cannot be answered in a single forum post...I think the problem is no where near as difficult as you "appear" to believe. It's also one of the reasons forums exist...the whole process can be simplified when many owners do the work, then share the results. It's also as complicated as you wish to make it...and as you learn you can tweak more things.
I was obviously unclear. I was asking about the real life problem and not the mathematical problem. I know that it requires judgement. I'll withdraw my question and continue to try to figure out why I am unenthusiastic about most of my shots.

DaveC
Posts: 1774
Joined: 17 years ago

#19: Post by DaveC replying to mgrayson »

You were completely unclear, and I answered each specific point you made. Of course it appears your question is a completely different one, and not relevant to the Evo leva.

Are you still making coffee on a LM GS3 as it states in your profile?
What specifically don't you like about your shots?
Is there a difference with the same coffee on another machine, if so what is the difference?

Starguru (original poster)
Posts: 79
Joined: 9 months ago

#20: Post by Starguru (original poster) »

It's difficult for me to know what I want here. There are too many options. Part of me likes the simplicity of the boiler/lever machines. They are simple, there are few options, the reportedly make good espresso. By removing options they simplify everything.


Part of me likes the idea of a "it can do it all" machine. I imagine they have a "just pull the lever and let it go" option, as well as allow control for when I want to play.

How can I reconcile that?

baldheadracing wrote:Those are three quite different machines - and all are quite different from an Evo Leva.

There's a spectrum - I guess two spectrums as they can overlap. At one end there are folks who would be happy if one could just push a button and the machine makes espresso (and milk drinks). Let's call that end, "convenience." At the other end are folks who want to try all kinds of different coffees and different brew variables. Let's call that end, "experiment."

Now take that to machines: I'm sure many would want a machine that offers push-button "convenience," but one that you can "experiment" with when one desires ... but how much "convenience" do you want, how much do you want to "experiment," and then there are other machine characteristics; aesthetics, size, capacity, ease of use, reliability, support, maintenance, availability ... and, of course, cost.

I think that it helps to consider those factors and narrowing them down before considering a technology solution.

Good luck!