Decent Espresso Machines (DE1) - Any thoughts?

[>decentespresso]

Now that we have learned that BLE was a method that would fail UL, it's possible that in a future version we'll move to using USB to control the machine. BLE could then be used for configure the espresso machine using your own device and a download app, but the app would not turn coffee/steam on. My understanding is that approach would be UL compliant.

I attach several mockups we made 18 months ago, of "tray warmer" concepts for the DE1CAFE, if we went with the tablet-attached approach. Since we thought BLE was OK, we went with a removable stand design for the tablet, where there is not enough space for cups.

I'm curious as to what you guys think of the cup stand idea.

-john

[JMcCee]

I have always favoured that configuration, mainly but not only because of the nice design aspect of the bent metal glued tablet mount. I was disappointed with the generically available wobbly and ugly looking magnet thing. I suspect if the bent metal mount was originally adopted Intertek issues would not have been such a problem. Personally, I would not have a problem paying a little more to have that bent metal mount.

[bigdaddy]

I like the bent metal shape of the bracket which holds the tablet. As for the cup holder guard rail... I like the one that has the two supports in the middle back of the machine, as shown in the top picture or you could have one support on each of the corners. The reason I say that is, I think the rail which is supported will not bend down and will be stronger over time and in a commercial environment.

Cheers.

[>decentespresso]

Can you hear me?



We noticed that the sound of the pumps on the production DE1 espresso machines was quite a bit louder than in our pre-production models. We didn't jump on it because we should this was likely caused by how we're driving the pumps now, and that we could fix this in software. We were also expecting a bit louder sound because we switched to more powerful pumps in August.

Last week, we brought the pump power down to 30%, the sound still remained, so we moved to assuming that the assembly was to blame.

Sometimes, you just need a weekend off, because we couldn't find the problem on Thursday and Friday last week. On Monday, within 10 minutes I'd figured it out, just by tapping around with my finger, inside an unplugged machine. A rubber pad behind the metal box which holds the pumps and the rattling noise went away.

I did want to formalize the audio level of the machine, so first we tried a decibel meter. The problem is that the machine very briefly has an 88db sound when the valve opens up to start espresso. So, peak dB wouldn't help.

Then, we compared a history of the sound levels between the before vs after machine, and while it was different, it wasn't substantially different.

The issue is that particular frequencies are much more annoying than others. Louder bass sounds are actually somewhat pleasing, whereas higher pitched sounds get annoying fast.

So, I switched to an FFT analysis program for iOS, which shows me a running 10 second average of the sound level of all the frequencies.

The result of that analysis is in the attached chart.

The bottom line is the room when it's quiet. The middle line is the "after" espresso machine, and the top line is the "before" (rattly) espresso machine.

Men's speaking voices are typically between 1.5k and 8k, and there you can see about an 8dB reduction from the introduction of the rubber pad. On some frequencies in that range, the reduction is as much as 12dB.

Most significant is the 14dB reduction in sound level, around 10.2dB. I'm guessing that's the annoying higher-pitched sound that we noticed.

A few caveats:
- our tests were conducted with the case off, and with the microphone placed 2cm from the pumps
- this sample was the noisiest time of the espresso when there is no backpressure and the pump flow is at full.
- we expect to be able to improve the sound profile of our espresso machine over time, with firmware upgrades, as long as we haven't missed an obvious piece of metal rattling away. <grin>

-john

[MrJack]

Maybe v2.0 could have active noise cancelling...

[roastini]

Maybe v2.0 could have active noise cancelling...

Active noise cancellation for home appliances

[>decentespresso]

A firm beta launch



SOFTWARE FLOW METER: Today's latest espresso firmware arrived on my desk this morning, addressing problems we've been seeing with our new "software flow meter". This new algorithm counts pump strokes (so that it's very low latency) and recalibrates itself using the flow meter (so it's accurate). It's the only piece of major new technology we have in this final build (as compared to the version Intertek received for testing in December).

PUCK TEMPERATURE: we've been interested to see if the temperature sensor that is directly above the coffee grounds is affected by a cool portafilter, now that we're in production. It should be, and in these tests, it was. The portafilter-cooling effect lessened as expected as the it heated up with use.


Here's what the photos tell us.

Photo #1: this is a real espresso shot with too high a pressure (because the grind was much too fine). Result: temp regulation is pretty good, and the two flow rate curves (water vs weight) line up well. Flow is lower than the goal at 15s because the puck is "jammed" with grinds that are too fine. As the puck erodes with time, you can see the flow rate increasing and pressure decreasing.

Photo #2: a room temperature (not preheated) "sensor basket" reduced the puck temperature readings, as would be expected. This results in a "brew temperature" that is 4ºC cooler than desired, at the start of the shot. By 15 seconds into the shot, the hotter water the DE1+ has added has brought the brew temperature to the goal temperature of 88ºC. Flow control was quite good. However, there is too much flow rate data smoothing, causing a few seconds delay when we asked to drop the water flow from 4ml/s to 2 ml/s. Ignore the brown line here: I wasn't using a cup to hold the water.

Photo #3: same as previous, but I used a cup to measure gravimetric flow rate. Happily, the brown vs blue lines line up well. The temperature drop at the start of the shot is now much less, as would be expected by the 2nd shot since the portafilter has been warmed by the previous shot. We'll decrease the smoothing tomorrow, which should decrease the delay. Note that coffee grounds are clogging the exit hole, causing erratic pressure, which also shows up as violence on the brown line (caused by a water spurt).

Photo #4: after unclogging the sensor basket, it's cooled a bit from several minutes out, and the effect is noticeable at the start of the shot. Also, there's still some clogging going into the exit hole, which is why we're getting "interesting" pressure and violence on the brown line. Note the line-up between the blue/brown lines.

Takeaways:

1) the puck temperature sensor is working well, and as expected, a warm portafilter makes a big difference to the first 15 seconds of the brew temperature

2) flow control is good, but is overly "smoothed" and thus has a 3-second lag when big flow rate changes are requested.

---

MISCELLANY from today:

FLUSH: this crucial missing feature (flushing the group head) is now implemented. You can use this to (a) rinse your portafilter (b) rinse out the previous shot's grime on the shower screen, and (c) preheat your cups.

INTERTEK: has confirmed that the "chained tablet" approach I proposed to them is acceptable. However, now they are now requiring our tablets to also be UL certified (when they were not attached, it didn't matter). Our current tablets are CE tested and certified (we have all the tests). I prefer to buy a UL approved tablet from someone and skip the expense and delay (Intertek prefers to $$$ for another test). I'm still negotiating with them about this.

CLEANING: We still need to implement a cleaning cycle for the group head and steam wand, test the "rattling pumps" fix some more, and then...

....7 machines go off to the beta testers (hopefully this week).

[JMcCee]

Intertek and their obstructionist motives concern me. The inference is compliance depends on how much you pay them. . Is there an alternative to using Intertek? And if there is would changing compliance methods be prohibitively regressive, ie going back months or even years?

[>decentespresso]

Intertek and their obstructionist motives concern me. Is there an alternative to using Intertek? And if there is would changing compliance methods be prohibitively regressive, ie going back months or even years?

Yes, there are alternatives to Intertek.

However, every month we don't ship is:
a) a month that pre-payers get more annoyed
b) a month that people who want to buy from us get annoyed
c) another month where I get to write a check to cover payroll.

I prefer to make small compromises now, get the product out, and get people making coffee. We can then take 6 months to move the machine to a slightly different direction (ie, USB comm instead of bluetooth, fixed stand instead of Steelie) and do it in a calmer way.

There's considerable pressure on us to ship, and since the machine is very close to my being happy with doing that, that's my preferred course.

-john

[JMcCee]

You replied quicker than I got my edit in but okay, all clear. I look forward to getting my DE1 whenever.

[>decentespresso]

Help me Proof the Standard Profiles



I've cleaned up the set of default pressure and flow profiles that are included with the DE1 and DE1+. It would greatly help me if those of you with greater knowledge of coffee machine history were to check that what I've written is (a) true and (b) doesn't omit some important information.

In writing this post, I realized that we're very sparse on "volumetric" profiles because I've always used either a timer-based espresso machine (before I met Rao), or a scale (post-Rao).
Volumetric profiles are definitely more consistent than timed ones, and not all of you will have a scale, so I'll do a bit of work this coming week to create some useful profiles.

For "boiler temperature" I've generally with the rule of "if it's Italian-style, it's hot (92ºC to 94ºC) whereas the other styles of espresso are in the cooler range (88ºC to 91ºC). I'm not sure this generallisation is entirely defensible.

ps: the DE1/DE1+ app downloads for Windows/Mac/Linux have all these profiles in them. Seeing the curves might help.
https://decentespresso.com/downloads

-john

-----

Best overall pressure profile
{We recommend this pressure profile as the most likely to produce a good espresso in the most varied number of cases. The decreasing pressure will help reduce acidity.}

Classic Italian espresso
{This will imitate the espresso style of the majority of cafes around the world. It uses a short preinfusion with a flat 9 bar pressure profile.}

Flow profile for milky drinks
{John Buckman, the founder of Decent Espresso, finds that this flow profile produces the best espresso shots for milky drinks in the widest variety of circumstances. It is especially tolerant of not-yet-very-good barista technique.}

Flow profile for straight espresso
{John Buckman, the founder of Decent Espresso, finds that this flow profile produces the best espresso shots for straight espresso drinks in the widest variety of circumstances. It is especially tolerant of not-yet-very-good barista technique.}

Gentle flat 2.5 ml per second
{Created by French barista Adrien Senac of "Cafeism", this innovative technique allowed Adrien to produce excellent shots from very aromatic, lightly roasted beans that he had never before managed to successfully pull.}

Gentle preinfusion flow profile
{Created by a famously innovative Seattle-based espresso machine manufacturer, this technique works well with aromatic, lightly roasted coffee beans.}

Gentler but still traditional 8.4 bar
{Professional baristas worldwide have found that slightly lowering the pressure from 9 bar down to 8.4 produces great espresso more often. The gentler pressure still produces traditional flavors but requires less perfect preparation skills from the barista.}

Hybrid pour over espresso
{Another approach created by a famously innovative Seattle-based espresso machine manufacturer, this technique requires you to grind your coffee very finely. Try this with very lightly roasted coffee beans. The resulting flavor is a blend of pour-over and espresso techniques.}

Innovative long preinfusion
{Another approach created by a famously innovative Seattle-based espresso machine manufacturer, this technique requires you to grind your coffee very finely. Try this with very lightly roasted coffee beans. The resulting flavor is more like a pour-over than espresso.}

Low pressure lever machine at 6 bar {Lever espresso machines are why we refer to "pulling an espresso." Your arm muscle pulls a long level to create pressure on the coffee puck. This 6 bar espresso is a classic Italian technique, and will result in a very sweet and gentle espresso.}

Preinfuse then 45ml of water
{A famous Australian barista champion suggested this technique to us. The idea is to fully saturate your puck, and then squeeze the brewed coffee out using a calculated water volume. This results in consistently brewed espresso without needing to use a scale.}

Traditional lever machine
{Lever espresso machines are why we refer to "pulling an espresso", because your arm muscle pulls on a large level to create pressure on the coffee puck. This 9 bar espresso is the most common type of classic lever espresso you'll have, and many fans consider lever shots to be the best espresso they've ever had.}

Trendy 6 bar low pressure shot
{Some lightly roasted espresso beans smell great but resist being well extracted into a drink. Sometimes, the solution can be to lower the pressure to 6 bar. Try this technique if you're having trouble with a particularly sophisticated lightly roasted bean. }

Two spring lever machine to 9 bar
{A final evolution of the classic lever machine style was to add a second spring. This helped work around a deficiency with single springs, namely that the top pressure gave was reduced almost immediately. The second spring required more arm muscle but is often considered the final perfection of the lever espresso style.}

default
{This profile is gentle on the coffee puck and not too demanding on the barista. Produces a very acceptable espresso in a wide variety of settings.}

e61 classic at 9 bar
{The E61 machine was invented by Faema in 1961 and is still the most classic and most copied espresso machine design. It is likely what most people mean when they say "Professional Italian Espresso Machine"}

e61 classic gently up to 10 bar
{A high tech variation on the classic E61 10 bar shot, this version brings more acidity out at the end with its increasing pressure. This makes it more appropriate to drink without milk. It will be quite thick, so adding water to make an Americano is recommended.}

e61 rocketing up to 10 bar
{With a technician's help, you could sometimes boost the pressure on the classic E61 machine to 10 bar. You will need to grind finer, and your espresso will very thick. Excellent for medium roasted beans mixed with whole milk.}

e61 with fast preinfusion to 9 bar
{E61 machines evolved with time, with some models adding an optional, manually controlled preinfusion step. In most cases, preinfusion will produce a more evenly extracted shot.}

[Downunder55]

John,

Been having another play this morning with the DE1 and DE1+ software, it's so great to be able to do this.

One thing did occur to me again and that is the integration of the Bluetooth scales on the DE1+ and not on the DE1.

As you are finalising the software thought I would raise this.

Appreciate you are trying to differentiate the two models to support the different features and price points between the DE1 and DE1+, but if the base capability is already in the DE1 then are you just precluding the potential sale of an additional $ 136 product from your range for little trade-off.

I am assuming here that a DE1+ purchaser is after a lot more than just scale integration.

Anyway just a thought ...... Happy for a simple "Yes good idea" or "No, just too ^&*% busy" answer

[>decentespresso]

Been having another play this morning with the DE1 and DE1+ software, it's so great to be able to do this. One thing did occur to me again and that is the integration of the Bluetooth scales on the DE1+ and not on the DE1. As you are finalising the software thought I would raise this.
Appreciate you are trying to differentiate the two models to support the different features and price points between the DE1 and DE1+, but if the base capability is already in the DE1 then are you just precluding the potential sale of an additional $ 136 product from your range for little trade-off.
I am assuming here that a DE1+ purchaser is after a lot more than just scale integration.
Anyway just a thought ...... Happy for a simple "Yes good idea" or "No, just too ^&*% busy" answer

Two thoughts:
- "gravimetric dosing" (the espresso machine stops at at a given weight) is only available on two top end pro machines, the Black Eagle and the Linea PB
- we're the only machine, at any price, that shows you "gravimetric flow rate"

Yes, the software is written, but it wasn't free to write, and that cost has to be recouped somehow. And there's more programming planned, with channel detection, inflection marking (more on that in the future), Delta curves, and more. Software doesn't write itself.

The USD$100 cost to buy our BLE scale isn't what I'm seeing as the cost of this set of features. The increased cost of the DE1+ is where people pay for the advanced feature set. If I include the advanced features in the DE1, I remove a key motivation to get people to spend more.

My feeling is that the DE1+ is already hugely underpriced compared to the competition. If a big Italian espresso machine company bought Decent, I wouldn't be surprised if they 10xed the price.

[>decentespresso]

Australian Italian espresso



Australian coffee roaster Ben Johnson explained to me that in Australia, the Italian baristas make a different style of espresso than they do back in Italy.

In Italy, the boiler is typically set to 102ºC, and by the time the water hits the coffee puck, it has cooled to about 95ºC/96ºC.

In Australia, the Italians use a lower brew temperature of around 90ºC. They still use their dark roasted beans, but the lower temperature helps avoid over extracting.

When I was in Italy this summer, I was brewing their beans down at an 88ºC puck temperature for the Italian roasters I met.

I would ask for their lightest roasts (typically medium roast, rather than dark) and this technique would cause the dark-chocolate flavors typical of Italian coffee to disappear, to be replaced with a layered chocolate set of flavors, more reminiscent of blended artisan chocolates such as Lindt.

Thanks to Ben, I now know that this approach is common in Australia's Italian barista community, and I've added "Italian Australian espresso" as a standard profile on the DE1/DE1+.

[Downunder55]

Two thoughts:
- "gravimetric dosing" (the espresso machine stops at at a given weight) is only available on two top end pro machines, the Black Eagle and the Linea PB
- we're the only machine, at any price, that shows you "gravimetric flow rate"

Yes, the software is written, but it wasn't free to write, and that cost has to be recouped somehow. And there's more programming planned, with channel detection, inflection marking (more on that in the future), Delta curves, and more. Software doesn't write itself.

The USD$100 cost to buy our BLE scale isn't what I'm seeing as the cost of this set of features. The increased cost of the DE1+ is where people pay for the advanced feature set. If I include the advanced features in the DE1, I remove a key motivation to get people to spend more.

My feeling is that the DE1+ is already hugely underpriced compared to the competition. If a big Italian espresso machine company bought Decent, I wouldn't be surprised if they 10xed the price.

Fair call, I don't disagree