Decent Espresso Machines (DE1) - Any thoughts?

[>decentespresso]

Florence : Café ditta artigianale (stories from the road)



These past few weeks, I've been on the road with my DE1+, starting in London, then 3 cities in France, then the "World of Coffee" show in Budapest, and finally a week in Italy. I've got 10 days to myself now in France, so I can focus on finishing the programming of the DE1+ tablet software and then I'm back to Hong Kong. My partner Bugs Harpley can then take a short holiday (she'll also be exhibiting at the Berlin coffee show) while I run things.

This past wednesday, I got to spend a few hours with Francesco Sanapo Francesco Sanapo | Il blog di Francesco Sanapo the owner of cafe/roaster Ditta Artigianale, seen as one of (if not the) top cafe in Italy. The other Francesco (Masciullo) FRANCESCO MASCIULLO is Itay's top barista and Italy's candidate for this year's WBC in Seoul.

Francesco Sanapo was quite taken with flow control, as he told me that he regularly talks to La Marzocco and Nuova Simonelli to add user control over flow. They tell him that it's possible with hardware changes, but not as an end user. He invited me to share his booth with two DE1+ machines at HOST Milan http://host.fieramilano.it/en - the most important coffee trade show in Italy. Very flattered, I accepted, and if booth space and schedules permit, we'll do it!

At Ditta they use a La Marzocco Strada with a paddle to make their shots, and Scott Rao had visited a few weeks earlier to tweak their profile for even flow. Scott had created a rise-to-9-bar-then-lower-to-7-bar profile by visually trying to maintain constant flow. It turns out he did a good job, because with my DE1+ was set to create a constant-flow profile, and we watched what kind of pressure curve was needed. In the shot below, the pressure rose to ~8.9 bar and decline to ~7.4 bar : very close to Scott's settings.



In the café they use a Mythos grinder, which is my favorite, but for this demo I had to make do with a spare Mahlkonig K30 grinder. Every shot we pulled that used the K30 grinds had a characteristic jitter of pressure variation between 13 and 20 seconds, which you can see above (this shot also had a channel briefly opening at 23 seconds). The pressure rampup is smooth, though, with the K30. My apologies that I forgot to take screen pictures this past week when I had access to a Mythos, so you'll have to take my word for it, for the time being, that curves of shots pulled from Mythos grounds are very smooth.

My way of comparison, here's what a flow profile shot looks like when using grounds from my Lyn Weber EG-1. Quite smooth.

[>decentespresso]

Why are we Silicone Dipping our Heaters?

Steve Single asked about our choice of Silicone as an insulator, and our interns are confused about how to test the physics of the thing, so I wrote the message below for them, and also for anyone especially interested in why we're dipping our heaters in silicone, something which has not been done in any other espresso machine that we know of.

All insulators (except for a pure vacuum) will eventually reach temperature equilibrium with the thing they are insulating.

The piece of data we want is the "rate of heat loss" that silicone provides, and this is called the "r value". Read about this concept here: https://en.wikipedia.org/wiki/R-value_(insulation)

Silicone is sold as an insulating material, and its r-value compared to other insulators is given here: GEK Wiki / Insulative Products

I attach below an excellent academic paper ("New Advances in Silicone-based Thermal Insulation") that discusses various kinds of types of Silicone products for insulation. Here is an important consideration:
"Thermal conductivity of silicone foam is about half the value of pure silicone rubber. ". We have not yet found a source for foaming silicone that we can use, as we have a very strict requirement that the silicone be odorless even when heated to 160ºC.

Thus, if we can find a foaming silicone product that "doesn't stink", we will double the insulation and lower the weight and material use. We've got 9 interns here, all engineering students, so we've manpower to throw at this! At the moment we are using non-foaming silicone, which works well, but it's heavier and less insulating than if we could get a foamed variant.

I also attach a PDF explaining all the properties of silicone, but here is the important part for our use:
"Silicone rubber withstands high and low temperatures far better than organic rubbers. Silicone rubber can be used indefinitely at 150°C with almost no change in its properties. It withstands use even at 200°C for 10,000 hours or more, and some products can withstand heat of 350°C for short periods. Silicone rubbers are thus suitable as a material for rubber components used in high temperature environments."

From a safety perspective, we also like that silicone is an electrical insulator. Should a water leak be spraying inside our espresso machine, silicone will not absorb the water and cause an electrical short. Our double-electrical failover components on both heaters are prewired and totally embedded inside the silicone. Chemically, even fully submerged, silicone will absorb no more than 1% of its weight in water.

PDFs:
https://imageserv5.team-logic.com/mediaLibrary/99/InsilThin_Technical_Paper_1.pdf
https://www.shinetsusilicone-global.com/catalog/pdf/rubber_e.pdf

[gc]

I have purchased a DE1+ and have a question about steam tips:

I usually make piccolo's in a 300ml jug and had to modify my existing steam tip down from 4 nozzles to one to ensure enough time for proper texturing of milk. Will there be optional steam tips available with different steam jets on the DE1+ ? (I wasn't able to source an alternative with my existing HX machine with the same thread size.)

[>decentespresso]

I have purchased a DE1+ and have a question about steam tips:

I usually make piccolo's in a 300ml jug and had to modify my existing steam tip down from 4 nozzles to one to ensure enough time for proper texturing of milk. Will there be optional steam tips available with different steam jets on the DE1+ ? (I wasn't able to source an alternative with my existing HX machine with the same thread size.)

The steam tip that comes with the DE1 models is a single hole, as that works really well with the existing amount of pressure we generate. We also have a four hole tip available to us. While it's overkill for the DE1, I expect that we'll use it for the DE1CAFE and have it as an option for the DE1PRO.

[>decentespresso]

Building the de1 : What we're working on

Back in Hong Kong at our factory, we're building two "final release" machines (110V and 220V) in order to catch any last remaining mechanical issues before building the 300 "early access" machines. Having Interns doing the work has actually been helpful, as they build things "naively" and have caught some remaining problems we've been blind to.

The list of problems is not so long, but we do need to solve them.

---

PROBLEM #1: the temperature probe in the group head requires too much skill to mount perfectly, resulting in imprecise temperature readings. This was not caught previously because our machines were built by Jeffrey, and "he's too good". Our interns aren't as good as him.

SOLUTION: stop using a temperature bead and our own design mounting, and switch to using the same ready-made temperature probe we use everywhere else in the DE1.



We haven't yet nailed this, as putting the probe there does alter water flow characteristic, and even water flow on the coffee puck is an important goal for us. We may have as much as two more weeks design work on this before we can have a sample made for real testing.

---

PROBLEM #2 : the water flow meter connectors are about 2mm thicker than we expected. This makes it quite hard to get the tubing on. Again, Jeffrey managed but the interns had trouble. This was not caught before because the supplier had not provided us with technical drawings.

SOLUTION: switch to a new flow meter, from our favorite supplier (ODE, an Italian company), who also make our pumps and valves. Their parts are found everywhere in professional Italian espresso machines.



We've also gone through several iterations of our mixing chamber (the brown-colored part in the drawing above), because the one we had designed turned out to be very expensive ($85 each) to manufacture. The size is now much reduced, so we can injection mould the part, thus greatly lowering the per-piece price. We are awaiting a sample of our new design so we can test it before paying for the mould to be made.

---

PROBLEM #3: sharp bends required of the tubing, which could cause failure:

SOLUTION: enlarge the hole and the size of the rubber gromet as well.



---

That's what we're currently working on, so we can make our first batch of machines.

[gc]

The steam tip that comes with the DE1 models is a single hole, as that works really well with the existing amount of pressure we generate. We also have a four hole tip available to us. While it's overkill for the DE1, I expect that we'll use it for the DE1CAFE and have it as an option for the DE1PRO.

Great. Thanks.

[MrJack]

[ATTACH=CONFIG]16441[/ATTACH]

At Ditta they use a La Marzocco Strada with a paddle to make their shots, and Scott Rao had visited a few weeks earlier to tweak their profile for even flow. Scott had created a rise-to-9-bar-then-lower-to-7-bar profile by visually trying to maintain constant flow. It turns out he did a good job, because with my DE1+ was set to create a constant-flow profile, and we watched what kind of pressure curve was needed. In the shot below, the pressure rose to ~8.9 bar and decline to ~7.4 bar : very close to Scott's settings.

[ATTACH=CONFIG]16440[/ATTACH]

In the café they use a Mythos grinder, which is my favorite, but for this demo I had to make do with a spare Mahlkonig K30 grinder. Every shot we pulled that used the K30 grinds had a characteristic jitter of pressure variation between 13 and 20 seconds, which you can see above (this shot also had a channel briefly opening at 23 seconds). The pressure rampup is smooth, though, with the K30. My apologies that I forgot to take screen pictures this past week when I had access to a Mythos, so you'll have to take my word for it, for the time being, that curves of shots pulled from Mythos grounds are very smooth.

My way of comparison, here's what a flow profile shot looks like when using grounds from my Lyn Weber EG-1. Quite smooth.

[ATTACH=CONFIG]16439[/ATTACH]

Added the new shot profile to my previous analysis. The chart below shows the shot with Scott's settings (in red) versus the previous data (including the second image above, which is shown in orange below).

[>decentespresso]

Added the new shot profile to my previous analysis. The chart below shows the shot with Scott's settings (in red) versus the previous data (including the second image above, which is shown in orange below). [ATTACH=CONFIG]16464[/ATTACH]

Hey Mr Jack,

I'm in France for 10 days, getting some tablet coding time in and a slight rest. I should be able to finish the "save every shot to a data file" feature this week, and on Friday, Adrien from https://cafeism.fr/ is pulling shots all day as he wants to get Flow Profiling under his belt. I'll try to get my software to work by then, and can send you a whole lot of espresso shot data (well, a few dozen) as text files.

[MrJack]

Hey Mr Jack,

I'm in France for 10 days, getting some tablet coding time in and a slight rest. I should be able to finish the "save every shot to a data file" feature this week, and on Friday, Adrien from https://cafeism.fr/ is pulling shots all day as he wants to get Flow Profiling under his belt. I'll try to get my software to work by then, and can send you a whole lot of espresso shot data (well, a few dozen) as text files.

I look forward to it!

[>decentespresso]

Continuing reports below on our progress. Today there's a batch of more teeny-tiny things that need to be corrected, to make assembly and repair a bit easier.



We've had problems getting the less expensive flow meter to work acceptably in the DE1, so I made the executive decision to put the same $40 swiss-made precision flow meter in both the DE1 and DE1+, at least for these 300 machines we're making. This reduces by half the number of hardware platforms (DE1 and DE1PRO, at 110V and 220V) so that we now have 4 different models, instead of 8. This also means that this batch of DE1s will actually be hardware identical to the DE1+, and that an extra-cost firmware upgrade would be possible to migrate a DE1 to the DE1+. This makes the DE1 more expensive for us to make, but simplifies things for now. We'll need to decide, in the future, whether to simplify the DE1 hardware to make it less expensive to manufacture, or increase its price, or something else.

So, to summarize, this means that for the time being, we're only building DE1+ (and DE1PRO+) machines, and that people who order a DE1 will be actually getting a hardware DE1+ platform, with DE1 firmware in it.

[Dimal]

Gotta love prototyping...

Mal.

[>decentespresso]

Bad news about the Decent Scale : beta testers feel that it's too tall.

We made a half dozen samples of our scale for beta testing, and feedback from about half the beta users was that our design is too tall for espresso machine use. It works fine for pour-over user, or with shot glasses, but enough people use double-spouted portafilters (which remove a lot of height) that our scale, at 3.5cm (1.4") is just too tall to use with many glasses and mugs.

Because we're looking at USD$86,000 in initial costs to make the first 1000 scales (mould fees are expensive) I've made the decision to NOT manufacture this current design. Instead, we're going to take another pass at the mechanical engineer and use a half-height load cell, much as Acaia did in their transition from the Pearl to the Lunar scale. We'll also likely switch from four AA batteries to something smaller. From two segmented LCD strips on top of each other, we'll switch to one longer one, saving on height.

While we're at it, we'll also add a weight-over-USB (not just USB charging) feature because we had 3 companies evaluate our scale for embedding in their product line, and Bluetooth was deemed as not appropriate for that sort of embedded use.

I'll be emailing the 31 people who already ordered our scale and offering them a refund, unless they're willing to wait for us to revisit the design. Currently, I'm estimating the wait to be at least 6 months, as our engineering resources are all dedicated toward the DE1.

[>decentespresso]

silicone insulation tests

Our summer intern Lee has programmed an Arduino and set of temperature sensors to measure the insulating effect of the silicone dip on our water heaters.

At 170ºC (for steam) and 15 minutes later, the air temperature inside the DE1 is 40ºC (we want to stay under 60ºC), and 3 of the 4 temperature sensors show a ~95º reduction in temperature due to the silicone. One of the sensors (on the bottom of the heater), in one of the test, is 53ºC hotter, which likely indicates we need put more silicone thickness at that point.

I've asked Lee to do another test to narrow in on the hotter sensor, and also why it wasn't a case with the 100ºC heater. Also, we need to turn both heaters on (not just one) to adequately measure internal temperature.

Note that for the DE1/DE1+ we let the heaters cool after making an espresso, for home energy conservation. On the PRO and upcoming CAFE models, the heaters will be on all the time.

[>decentespresso]

Exporting shot history

Every espresso shot now automatically saves everything about itself to a text file, and with a little bit of fiddling, can be shown in Excel. Pressure, flow, temperature as well as all the settings (and profiles) to create (or recreate) the shot.

I need to write something to make the export-to-excel more friendly, but the foundation is there, and the shot saving will be automatic for all DE1 users, so that I as I program features to make the history more useful, your entire espresso history will be there to work with.

I also need to program viewing-historical-shots, viewing-many-shots-on-one-graph, god-shot graphing in the background, restoring settings from historical shots, and lots of other idea… All for the future!

The Excel spreadsheet below is a quick copy/paste job I did of 3 low pressure shots that occurred from a 2.5 ml/s constant flow rate. This opens up lots of interesting possibilities for data analysis.

[MrJack]

Every espresso shot now automatically saves everything about itself to a text file, and with a little bit of fiddling, can be shown in Excel. Pressure, flow, temperature as well as all the settings (and profiles) to create (or recreate) the shot.

I need to write something to make the export-to-excel more friendly, but the foundation is there, and the shot saving will be automatic for all DE1 users, so that I as I program features to make the history more useful, your entire espresso history will be there to work with.

I also need to program viewing-historical-shots, viewing-many-shots-on-one-graph, god-shot graphing in the background, restoring settings from historical shots, and lots of other idea… All for the future!

The Excel spreadsheet below is a quick copy/paste job I did of 3 low pressure shots that occurred from a 2.5 ml/s constant flow rate. This opens up lots of interesting possibilities for data analysis.

[ATTACH=CONFIG]16516[/ATTACH]

A game changer!

I can see some espresso myths shaking in their boots already...