La Marzocco GS3 Prototype

Vital Stats

Please note that the below stats are all for the prototype machine - any and all could, in fact, change in the final production version. These details are provided primarily for a reference point for this current state of the machine and nothing more. Check with ESI and/or LMI before committing to any decisions based on these details.

Brew Temperature Stability

Early on in testing this machine I had the chance to evaluate its temperature stability using a Scace Thermofilter and Fluke setup. Given what I was tasting, my initial results were shocking and yet almost expected. Meanwhile, Internet coffee enthusiasts were clamoring for hard numbers. Given how "out there" my results were as compared to other espresso machines, I hesitated to share them. I decided I needed to reconfirm my results. I had various other people come by, with their own equipment, to do their own tests to see if I was making mistakes. Terry Z came down from Espresso Parts with his rig. But none of this removed the fear (and none of them seemed willing to put their reputations on the line with the numbers we were getting).

Thankfully, soon after this happened I was relieved of this duty and freed of the risk of being ridiculed. Bill Crossland sent Greg Scace (developer of the Scace Thermofilter, scientist and general coffee guru) one of the GS3 prototypes and Greg then took on the task of quantifying the performance of the machine. Rather than using my results, I am merely going to re-purpose (with Greg's permission) his data and his results and clarify based on my data.

Some initial thoughts on the topic first: As noted earlier in the review, the value of intra-shot temperature stability is a hotly debated topic. The reality, however, is that the GS3 is designed to deliver a flat-line brew temp profile. Rather than debating the value of this profile, it is probably wise to simply take it as it is. If you're looking for a different kind of brew profile in your espresso machine, you should simply eliminate the GS3 from consideration.

The prototype machine has a nearly untuned PID. The initial programming is not optimized. LMI is aware of this and there is a process underway to develop better tuning for the PID. It is very likely that this will result in noticeable improvement in temperature stability.

One thing that in important to know is that the blind taste tests that I ran indicated that even the average coffee consumer seems to be able to differentiate between shots down to the current 0.3F granularity. With some coffees (the more finicky ones) differences of 0.3F can be profound.

In any event, the following are Greg's results - with notes and clarifications based on my tests. For full details on the results Greg reached, you should read La Marzocco GS3 Thermal Performance Quantified in the forums.

Greg primarily compared his prototype GS3 to his PID’d Linea 2AV. Note that this is not a stock Linea and has dramatically improved temperature stability over the stock Linea. I would love to see a comparison like this between the GS3 and a stock Linea EE as well as, of course, between the GS3 and a Synesso. In any event, the PID'd Linea provides an excellent benchmark to compare the GS3 with. Most professional baristas would dream of being able to work on such a machine.

The first graph shows brew temperature over a session, showing inter-shot stability. Greg made these measurements using the WBC protocol, which doesn't use any cooling or heating flushes. This clearly demonstrates not only the incredible inter-shot stability of the GS3, but also the fact that it doesn't actually need any flushing to achieve that stability. The Linea, on the other hand, requires significant work to initially get to temp and then to manage temp once there. For the home user, this is profoundly important and demonstrates the incredible "first or intermittent" shot performance of the GS3 (the more applicable session model for home users).

The second graph shows one (randomly selected) shot from each of the machines. These were selected from a session of fairly continuous duty samples. This shows the brew profile of each of the machine. As you can see, the GS3 has a similar profile, though missing the initial "hump" that seems to be indicative of the AV version of the Linea. Total variance is incredibly low.

In his tests, Greg found that:

These numbers are consistent with my initial findings though they show slightly higher numbers for the results of the GS3 temperature stability number than what I found on my prototype (or what Terry saw testing my prototype with his rig).

These results seem to indicate that the prototype GS3 delivers the highest degree of any (stock) espresso machine that has currently been independently tested and verified using the Scace Thermofilter and WBC protocol (the emerging standards for this sort of measurement).

Programming Options

Following are some highlights (not inclusive) of the options available through the various programming modes.

Dose, Distribute, Tamp. Repeat.

This section will cover the basic methods used for making espresso with this espresso machine. If you are a beginning barista, I would strongly suggest you first read Espresso Machines 101 and Jim Schulman's excellent Home Barista's Guide to Espresso.

The key to being a good barista is consistency. Everything the exact same way each and every time. Grind and dose the exact same amount, distribute in the exact same way and manner, tamp with the exact same pressure—each and every time. To do this, you need to practice and pay attention.

Verify the portafilter basket is dry, as water will follow the path of least resistance. If some of the coffee grounds in the basket are wet, the water extracting your coffee will pass more quickly though it than the drier surrounding area. Consequently the wet coffee will be over-extracted while the remaining coffee remains under-extracted, producing a classic unevenly extracted shot.

When dosing coffee, the keys are to grind immediately before you dose and to be consistent with your dose. In addition, consider experimenting with the amount of coffee you dose. I've found that each coffee has an ideal dose for a desired flavor profile. To be consistent with these varying doses, I prefer to rap the portafilter on the grinder fork zero to three times depending on the desired dose. With a La Marzocco ridged double basket, I find that this allows me to dose from 17 to 20.5 grams.

Distribute the coffee evenly throughout the basket. This is not only the least appreciated concept in espresso-making, it may well be the most important factor in getting a good cup. The idea is to create an even density of coffee grounds throughout the entire basket. You're not just leveling the surface, you are literally distributing the coffee in the basket. There are various methods for achieving this goal. The two used in this test were the Schomer method (often described as the "NSEW" method) and the Stockfleths move (sometimes described as the "rotational" method). In both cases, you use your finger to drag the pile of grounds around the basket while providing downwards pressure. One of the greatest training innovations of the last decade has been the so-called naked portafilter. This product will quickly allow you to identify flaws in technique. These errors create soft spots in the distribution and the subsequent uneven extraction and channeling.

The goal of tamping is to preserve that even-density bed of coffee that you have just created. If done correctly, the coffee puck will provide sufficient resistance to the pressurized water to extract the liquid we call espresso. There are a couple of keys to getting a good tamp. First, work on a counter of the correct height. It's surprising how often this important detail is overlooked. The counter needs to be low enough so that you can get sufficient leverage and high enough so that you have stability when tamping. Second, have the correct grip on the tamper. Hold it as you would a doorknob. Use your thumb and index finger to provide the downward pressure (not the palm of your hand). Third, have the correct arm orientation. Your forearm should be vertical and your shoulder should provide the downward force through that rigid and vertical forearm. Fourth, tamp with at least 30 pounds of pressure on your second tamp if you want to be consistent. Fifth, if you choose to rap the portafilter with the tamper, do so with as little force as is possible, and only do so once. Doing otherwise is likely to create gaps between the coffee and the wall of the basket, which will result in channeling. Finally, never rotate your tamper under pressure. Doing so will "tear" the surface of the coffee, creating weak spots that will result in uneven extraction.

Brew Temperature Management

As noted earlier in the review, it may not be necessary to flush the group for true stability. Nonetheless, I decided it would be wise to flush one ounce of water through the group before pulling each shot. I did this by programming the "single short shot" button to dose one ounce of water with no pre-infusion. After building the shot, I would first hit this button, wait a moment, then insert the portafilter and extract the shot.

If, upon tasting the shot, it seemed like the brew temperature was wrong I would then go into the programming mode and alter the brew temperature. I found that changes to the boiler temperature manifested themselves accurately at the group quite quickly. Even very large changes were fast (+5F = 1:05; -5F = 2:55 [with 4x four ounce flush]).

Finally, I have two key tricks/hints for you. First, whenever you buy coffee, ask what the target brew temperature is for espresso with that coffee. This is going to save you a lot of heartache. Second, trust your taste. If it tastes good to you, it's good.


A lot of people helped me put this together. And I couldn't believe how much coffee and equipment was used for this write up. If I were allowed to have more fun, it would have been illegal. If you want to read more about the process of reviewing the GS3, check out the appropriate thread in The Bench in the forums.

Personal Acknowledgements

Special thanks go out to Bronwen Serna, Kyle Larson, Duane Sorenson, Stephen Vick, Billy Wilson, Phuong Tran, Mike McKoffee, James Cooper, Amanda Bryan and all the other tasters and testers. Thanks to Duane also for sharing thoughts and insight. Thanks to Joel, Jana, Trey, Mason, Matt, Doug, John, Bronwen, David, Vince, Terry, Tracy, Trish, Tonx, Andrew, Terry and Tony for the coffees. Thanks to all the La Marzocco folks who provided guidance, feedback and questions along the way - with special thanks to Bill, Jacob and Kent. Thanks to Terry and Dave at Espresso Parts for the help with the plumbing. Big huge thanks to Terry and Ken for helping with my testing and for Greg for all his work. Thanks to all the various GS3 testers out there (Greg, Mark, Andy et al) for their feedback and collaboration. Thanks to the members who asked questions and made suggestions along the way. Special thanks to Dan for his fantastic copyediting that turned my usual drivel into a readable piece of prose. And most of all, special thanks to the wonderful Valerie Hoecke for providing assistance and support and tolerance in all possible ways.

Sponsorship Acknowledgements would like to gratefully acknowledge the generous support of the following cafés, coffee roasters, and equipment providers. Without their help, this review would not have been possible.

Equipment Coffee

La Marzocco International
- Prototype La Marzocco GS3
- Naked and spouted portafilters
- Various baskets

Stumptown Coffee Roasters
- Puro Caff cleanser
- Pressure and temperature test rigs

Espresso Parts NW
- FloJet Pump System
- Everpure Water Filtration
- Espresso Tamp EP5
- Scace Thermofilter
- Fluke Multimeter

Other test equipment
- Cimbali Junior Grinder
- Fluke and probe
- Portafilter-fitted pressure gauge
- Espresso Vivace Ergo Packer with 58mm piston
- Black Deluxe Lava Tamp with 58mm piston
- CDN timer
- Digital gram scale
- Pitchers, cups, shotglasses, etc.
- Rinza Dairy cleanser
- 4mm Allen Wrench
- Flathead screw driver
- Philips screw driver
- Naked Portafilter
- Triple basket
- Scotchbrite green scrubbies
- Pällo Coffee Tool

Stumptown Coffee Roasters
- Hairbender
- Ethiopia Sidamo
- El Salvador Finca Kilimanjaro

Intelligentsia Coffee & Tea
- Black Cat
- Nicargua Los Delerios

Ecco Caffe
- Espresso Reserve
- Brazil Cerrado Daterra Reserve
- Ethiopia Yirgacheffe
- Ethiopia Yirgacheffe Natural Hama Co-op
- Brazil Fazenda Cachoeira Pulped Natural
- Various experimental espressos

Olympia Coffee Roasters
- Peru Penachi

Victrola Coffee
- Streamline Espresso
- Ethiopia Sidamo

Zoka Coffee Roasters and Tea Company
- Espresso Paladino
- Sumatra Lake Tawar

- Espresso

Doma Coffee
- Vito's Blend
- Ruby's Blend

Espresso Vivace
- Dolce
- Vita

Hines Public Market Coffee
- Espresso

Caffe Fresco
- Daterra Reserve

Comparative Equipment

La Marzocco Linea EE 3 group - semi-automatic, dual boiler commercial machine. Generally considered the standard for high-quality commercial machines. The Linea is a plumbed-in 200 plus pound machine with an external Procon rotary pump. The brew boiler has a volume of five liters and 1900W heating element. The steam boiler has a volume of 12.5 liter and 4000W heating element

Modified La Marzocco Linea EE 2 group - another semi-automatic, dual boiler commercial machine, but in this case modified by David Schomer of Espresso Vivace in order to deliver greater temperature stability and consistency. This is the smaller version of the 3 group machine above, with a 3.4 liter brew boiler and a 8 liter steam boiler (1400W and 3000W heating elements, respectively).

Mistral - a custom made, semi-automatic dual boiler commercial machine made by Kees van der Westen. The Mistral is loosely based on the Linea in that it uses many of the raw components of the Linea as starting points and builds a machine that looks like a sculpture. There are, however, differences beneath the skin that are often dramatic. These include pre-infusion and some significantly altered plumbing.

  La Marzocco Linea Mistral

La Marzocco Linea