Decent Espresso Machines (DE1) - Any thoughts?

[>decentespresso]

Ramping Up



Now that we've made one "final final" machine, and just about finished with 10 machines for the beta testers, we're moving to doing building 100 espresso machines at once. Attached are some photos of the factory as we prep for that. You'll notice a lot more boxes as all parts now need to be available on the assembly floor in larger quantities.


OUR PLAN FOR THE NEXT SEVEN MONTHS

Our immediate goal is to achieve 2.5 machines shipped per day through March and April, so that we ship 100 espresso machines in those two months.

By May, I hope that we'll be able to achieve 5 machines shipped per day so that in May/June we'll have concluded building all 300 of the DE1 v1.0 machines (which have all been pre-ordered).

This will coincide with parts arriving in July for the next batch of espresso machines, in quantity=1000. In March, we'll be taking orders for that batch, which I hope will sell out.

We'll need to have sped up to 12 machines shipped per day, at that point. This should be achievable because about a third of the DE1 v1.1 assembly work will be done for us by our parts suppliers (that's what why we're working on manuals so assiduously) and we'll have hired about 10 more people to build machines.

If we manage to pull this all off, we'll be able to start shipping in larger quantities (20 machines per day, to equate 4000 espresso machines per year) starting in October. That assumes, of course, that we've received good reviews and that we have enough customers to warrant that level of production.

In about a month, the first reviews of our espresso machines will start being published, and we'll learn how people feel about what we've spent the past 3 years making.

-john

[>decentespresso]

How to Assemble a Grinder



We're close to shipping our "own-brand" coffee grinder, and this week Alex made IKEA-style "wordless" assembly instructions. Since we have a global customer base, that's a clean approach, and I'm hoping we can avoid our instructions being as occasionally confusing as IKEA's instructions can be.

The reason for the instructions is that we've "modded" a well-made-but-boring grinder to have a 3-axis moveable portafilter stand and scale on it. The goal is both to weigh and to be able to center the grounds properly in the basket.

For our grinder and espresso machines, we'll also be making CGI-and-real-footage based Youtube video instructions, which will have English talking on them with subtitles for our major customer languages.

Personally, I think the videos are the way to go, as they're so much clearer and easier to follow than reading diagrams.

A comical aside: we had a lot of trouble modeling coffee beans in CGI, as most of our attempts looked more like "black bean soup". Each bean is a separate object, so we had to use particle-based software to model them somewhat realistically.

Our current priority is shipping the grinders out to our pre-order customers and gathering feedback from them.

I'm aware that many people see Decent as a "high tech, disruptive company" but that's not how I see us. I like to think of everything we do as "practical" and "focussed on drink quality".

A good quality grinder, at a reasonable price, with weighed doses and a well-centered coffee mound, is something I think needs to exist. You'll be able to choose the grind quality you like, from "Super Jolly" to "K30", depending on the burr set you choose.

[>decentespresso]

Calibrating to Reality



Attached above is what I'm currently working on. Each DE1 and DE1+ needs to have its sensor calibrated by a separate device (typically a Scace 2 portafilter for pressure and temperature, and a 0.1g accurate scale for flow) so that manufacturing variations in the sensors we use are taken into account.

As a customer, you'll also have access to this page, via Settings->Machine->Calibrate.

To calibrate your machine, you'll run a shot and then enter in what your test equipment said actually happened. Then, rerun the shot and see how close the Decent espresso machine now is to the test equipment.

Separately, Ray and I are doing the final debugging of our "upgrade firmware over Bluetooth, using the tablet" process. In the past, upgrades have been done via a USB cable, which I don't think is great for customers. Ray has the firmware side (which receives the upgrade) of upgrading working, and he found a bug in my tablet code (which sends the upgrade), and I'll fix that when I'm back at the office on Monday.

After calibrating and firmware upgrading, Ray and I have a few remaining todos, which we hope to finish this week, so we can send the machines out to beta testers.

---

Still to do on the software side:
- the new flow profiling code which counts pumps strikes to measure flow, but which recalibrates itself regularly using the flow meter. This will give us very-low-latency flow measurements, which is useful both for more accurate flow profiling shots but also to be able to show you short live channeling defects in your espresso shots.
- noise: there are a few spots in making espresso where the two pumps are on 100% and in sync, and that causes a rattling noise.
- implement cleaning and descaling cycles
- implement the "flush water through the group head" button

None of the above are very time consuming, and we're pretty confident we'll have it wrapped up by the end of the week.

[Bob_Weis]

I am waiting for my DE+ and haven't received an update otherf than not yet

[>decentespresso]

I am waiting for my DE+ and haven't received an update otherf than not yet

Hi Bob,

In this post 3 days ago, I laid out the schedule:
http://coffeesnobs.com.au/brewing-eq...tml#post625627

which has machines going out to AU during May and June.

The main variable is CE safety compliance certification, after which AU compliance is very fast. I will continue to keep everyone abreast on developments on that front, on this (and other) fora.

-john

[>decentespresso]

Drip Drip



Our first beta tester found a problem with our espresso machine, that we have replicated here.

Sometimes at the end of an espresso, the flushed water comes out briefly with enough force that it sprays out in a 120º cone (top left photo), with some of that water hitting the back of the mirrored lip, and then dripping between the drip tray and water tank, onto the table (top right photo).

Result: some water on your kitchen table.

This "flush burst" only happens for the first second of the flush. For the rest of the flush (about 5 seconds) the water fills up inside the metal box we call the "flush diffuser" (bottom photo) and then drains out at a slower speed.

This fast start to flushing "shouldn't be happening" based on our understanding of what's going on: obviously, our understanding of the first second of the flush is incomplete.

We should be able to improve the situation with a firmware change in the future, by ensuring that the flush path has little remaining water or pressure in it. But, that will take time to make sure we've correctly understood why there was pressure in the flush path in the first place, and that we've truly removed it.

So: I asked my engineers to develop a hardware fix for this problem. I'd rather fix it now "for certain", and then also plan on a later firmware revision to make the fix unnecessary.

By putting a small length of silicone tubing on the exit point, the burst of water comes out at 120º and then hits the walls of the tube, slowing down, and then draining into the drip tray (bottom photo).

You can pull the silicone tube off in a few months if it bothers you, once/if we fix this problem with a firmware update.

---

A Clue in Solving our Wet Puck Mystery?

A side note: after making an espresso, our pucks have always been a bit wetter than I like (wetter than other espresso machines).

As of today, I now suspect that the wet puck and the "flush begins with high force" might be linked.

Here's my theory: if the flush path had some resistance to it, that would hamper the force at which the group head decompresses, and thus allow a bit more water to remain on the puck. If that's true, this should be fixable by changing our valve timings in firmware.

[HBexile2]

The exhaust fitting should be under the top of the grate, you should have a hole cut out or a feature in the grate that allows the exhaust fitting to go through. All machines have had this since the 1960's and it's a very basic feature. On a lot of your machine you have tried to reinvent the wheel on so many things for no reason. The flush box you designed is a prime example, why not look at existing machines for guidance? they all work perfectly well and manage to slow the water flow very well. Anyway good news for the "300" who won't have wet counter tops.

[>decentespresso]

The exhaust fitting should be under the top of the grate, you should have a hole cut out or a feature in the grate that allows the exhaust fitting to go through. All machines have had this since the 1960's and it's a very basic feature. On a lot of your machine you have tried to reinvent the wheel on so many things for no reason. The flush box you designed is a prime example, why not look at existing machines for guidance? they all work perfectly well and manage to slow the water flow very well. Anyway good news for the "300" who won't have wet counter tops.

Regarding the flush diffuser, there was a long discussion some time ago on Home Barista, with people posting other flush diffuser designs. A lot of E61 machines still don't slow the flush down and splash a bit. In the end, our design is very similar to the Simonelli Black Eagle design, except that theirs is accessible from behind the drip tray, and thus easier to clean (we don't have the space: in ours you have to take the top cover off to get to the flush diffuser).

As to "you should have a hole cut out or a feature in the grate that allows the exhaust fitting to go through" could you post a photo of what you mean? I'm not sure what part the "grate" refers to, is that the "shower screen" you mean?

We do have a dedicated flush path, with forward flushing (rather than back flushing), but we diverted from common practice for a reason. With backflushing, coffee oils find themselves gradually sticking to the tube which is also used for water into the puck, leading to a situation where the water path into the coffee is not clean (or easily cleaned). We flush dirty water out a dedicated path that is never used for clean water, as a small potential improvement to drink quality.

-john

[>decentespresso]

The exhaust fitting should be under the top of the grate, you should have a hole cut out or a feature in the grate that allows the exhaust fitting to go through. All machines have had this since the 1960's and it's a very basic feature. On a lot of your machine you have tried to reinvent the wheel on so many things for no reason. The flush box you designed is a prime example, why not look at existing machines for guidance? they all work perfectly well and manage to slow the water flow very well. Anyway good news for the "300" who won't have wet counter tops.

Someone PMed me that you probably mean this feature in drip trays:



If so: yes, we did have that feature in a much earlier version of our drip tray cover.

However, we found that once there were a few millimeters of water built up in the drip tray, that a splash occurred during the flush phase.

Looking at E61 machines with this feature, we also found that the high speed flush burst didn't contain itself tidily to that hole. On HB people with this sort of machine complained that they get splashed, which was our experience with our own implementation of this as well.

So, that's why we went with a "flush diffuser" box, which is something I saw on the Black Eagle and thought was well executed by them.

-john

[>decentespresso]

Josephine's First Latte

Her latte art might need a bit of practice, but she pulled a good espresso and made acceptable microfoam.

[>decentespresso]

Worldwide Standards for Plumbed Water



We're finalizing our design for the Plumbing Kit, which allows you to plug a DE1PRO into your cities' water supply.

The valve that our suppliers offers for this is perfect, except that it only comes with "NPSM G3/4" threads, which don't seem to be widely used anywhere. So, we're going to CNC our own thread adaptor to BSPP G1/2, which seem more common.

I'd like to hear from you about this: what kind of fittings do you have in your country, for plumbing an appliance in (such as a dishwasher)? We'd like to make our own adaptors for the common standards, to make plumbed-in installation easier for you.

5 custom PC Boards and their components came in today. We're hand-soldering the components on, in order to do a final test that we've not made any mistakes before we order 1000 PC boards to be made for us. We're building the pump-version today.



The same PC board is being used for our two Refill Kit variations: the Plumbing Kit (pressurized city water, using a valve) and the Catering Kit (water sucked out of a tank with a pump). The chassis is the same in both cases, but with different holes drilled into them.

I had an idea this week to try to make the chassis "latchable" onto either the DE1PRO and/or your table. The reason: at the Korea Coffee Show I twice had the Refill Kit fall off the table, eventually breaking one of the cable connectors. We're still finalizing how that latching mechanism will work. We've prototyped something ugly in our shop using bent sheet metal, and we'll be asking our prototyping company to make a not-ugly version for us shortly.

Attached above is a cross-section render of our Plumbing Kit, as well as the thread adaptor we're having made.

[>decentespresso]

Intertek Decides



Intertek has given me their final opinion, which is that our tablet will be considered a "remote control". Unfortunately, this goes against what Intertek's consultants have told us for the past two years (which was that Bluetooth would be treated as if it were a cable), and this is a bit of a frustrating turn.

From a safety compliance standpoint, if the tablet is a remote control, then every action has to be re-OKed on the device itself, which would not be a very nice user experience.

I had prepared for the worst news, and we've discussed here my planned counter-proposal, which is to physically tether the tablet to our espresso machine. I'm pretty sure that will meet their requirements.

Today, I sent Intertek the photo above as an example of this idea. The actual implementation will be more discrete, but I sent them this photo (bottom right) because the tethering idea is clear to see and understand.

In a past meeting, they indicated that a "Torx screw" was considered "not user-removable". Therefore, my intention is to connect the tablet stand to the DE1 via a cable, secured by a Torx screw.

We thus will ship our espresso machine in a UL compliant manner (tethered). If the end user removes the tether (using a Torx screwdriver and wire clippers), it is no longer UL compliant.

Another possibility for the future would be for us to do away with the Steelie stand, and instead, move to the custom-made bent-sheet metal stand that we drafted in an earlier incarnation of the DE1CAFE. (top photo)

I wonder if the Steampunk Coffee machine had the same compliance issue, as they encase and tether their included Android tablet. Steampunk | Craft Brewing Machine | Alpha Dominche LTD | Alpha Dominche LTD (bottom left photo)

For now, though, my focus is on shipping espresso machines that are UL and CE compliant. We can revisit this issue in 6 months when people haven't been waiting so long for machines from us.

[gc]

So John, does the "final Intertek determination" mean that everything else about the machine is approved? Or are their still other hurdles to jump for certification?

[>decentespresso]

So John, does the "final Intertek determination" mean that everything else about the machine is approved? Or are their still other hurdles to jump for certification?

It's progress, but it's not the last hurdle. They still haven't opined on the internals of our water heater and group head heater, and on how the group head is wired. They promised to close that out by tomorrow. That will conclude the "construction review".

We then will need to submit two machines to them, with all the requested changes, for a "final UL review". I got the feeling that this would not be onerous, but it ain't over 'til it's over.

Then, there'll be a CE review, which will also be in two phases (construction and full) with 6 machines submitted (and destroyed). Then, we'll find out if Intertek's CE consultant, who we've been seeing for the past two years, gave us good advice.

-john

[gc]

Many thanks for that update John. Patience and peace required by us all through that I guess. Good luck with it all.