Decent Espresso Machines (DE1) - Any thoughts?

[>decentespresso]

Shipping the 300 DE1 espresso machines : current status.

Over on coffeeforums.co.uk I was asked "Is there a fairly solid target date for assembly at this point, and if so what is it? If not, what are the unknowns that prevent a solid target date? "

The short answer is "parts start arriving in November, machines likely to ship in early December, but will take a few weeks to be all built"

Here's what longer answer which I wrote, if you're curious to know more...

The ceramics were ordered a while ago and were supposed to arrive this past week. They're "in the mail", is what we've been told, and should arrive this week.

We haven't ordered the drip tray cover yet, because we found too much variability on the drip tray, and wanted to wait until the real McCoy showed up and we could see what we were really dealing with. Since it's a cast piece, it will take 30 days for the cast, and 35 days then for delivery. This makes it one of the very last pieces we're expecting (end of November) but the good news is that it does not delay the rest of assembly.

Bear in mind that currently, it takes us about 30 man hours to build one DE1. With 8 people here, even if they're all working on it full-time, that would be 10 machines shipping per week, hence 30 weeks. Obviously, we're going to hire people to help, but keep in mind that assembly will take time too. And, totally assembly time will be less than 30h since we'll be making things in batches. A lot of the v4 chassis changes were to simply assembly and reduce disassembly/repair time.

The heaters are being ordered next week, and are to take 30 days. They're the most important part because they go in the center of the machine, and if we don't have them, they prevent the rest of the machine from being built. We were not happy to find a 15% defect rate in the 40 samples we had ordered from our supplier, and so we have switched to a heater manufacturer recommended to us by ODE. As ODE has been exclusively in the espresso parts business for decades, we tend to follow their advice. We tested their heaters for the first time in our v3 machine and they look good.

The all-metal chassis has not yet been ordered, as we are waiting for the v4 chassis to be CNCed and arrive here for final testing. That will be ordered in mid-October, and there is a 30 day lead time on that. There were 38 changes from v3 to v4, none of them "risky" but they do need to be seen & tested before we build 300 machines. For example, we added 5mm to the depth of the DE1, because we found that once it was fully assembled, the fit was too tight to remove the low-voltage PCB for repair purposes (a fan got in the way). You had to remove the pumps to remove the PCBs and that would annoy the heck out of a repair person.

So... to succinctly answer your question, parts will start arriving early November, and we'll start by first quality-control testing each part. Then, we'll put together the subassemblies that we can, as the parts arrive. Ideally, the chassis will arrive mid-November and the sub-assemblies will slide right in.

The biggest worry I have at the moment for schedule slippage is the mixing chamber, which is by far the most sophisticated part of the machine, is CNCed from a somewhat exotic material (Ultem) and requires two suppliers (CNC and custom valves) to work together. That's supposed to arrive between 7 to 9 weeks from now.

And of course, the other concern is that Christmas is arriving, and as a relatively small company, our orders might get bumped in our supplier's schedule, to make space for "more important" clients.

The 110V "early access" machines already sold will go out first, while we wait for the testing lab to certify our "final release" and 220V machines. We expect the testing to take 2 months.

Nonetheless, we're very close to the end of this journey, and espresso machines will soon be shipping.

And the good news is that once we know how to build 300 machines, we'll be in a good position to build the next 1000 machines at a much faster rate.

[>decentespresso]

That is truly fascinating. The time that mass in the cup is first recorded really surprised me; slightly earlier than I would have expected looking at the other data.
Is the flowrate measurement (blue line) in that shot from a physical or "virtual" meter?

In those charts, it's purely virtual. Virtual is what we've had for most of the past two years, and one advantage of that approach is that very short, fast changes in flow can be noticed. With a hardware flow meter, they're not. This quickness is useful because we can spot (and chart) channels opening and closing, as well as quick & short drops in pressure during the initial pressure rise to 9 bar. The 2nd one really helped me improve my puck prep skills and also "neg" some grinders as being poor quality.

Our intended "final" approach is to use the flow meter to give us the overall speed of the water, and the physics model to modify that overall speed with quick changes we notice.

At the moment, the hardware flow meter is disabled as we're focussed on getting the physics model right. In a few weeks, the integrated flow model should be done, and the charts will be using that.

[TampIt]

It's really interesting to be able to see the rate at which espresso pours into the cup. I think it's going to lead to some insight.

In the espresso below, I am using a somewhat fine grind, 18g dose, the coffee is fairly old (4 months) and medium-dark roasted (not so yummy). The two graphics are the same espresso: one is a "zoom" of the other.

[ATTACH=CONFIG]17126[/ATTACH]

What's interesting about this shot is:
- even though preinfusion has definitely finished (pressure is high)
- the puck continues to absorb water
- which you can see because the water out of the puck (brown line) is half the speed of the water into the puck (blue line)
- and as the puck becomes fully saturated (toward the end of the shot), the flow rate in finally equals the flow rate out

I think that a slower, longer preinfusion would fix this, as would a coarser grind (the beans would have less water holding capability).

That would change the flavor, and it'll be interesting, with this new data, to see if it's possible to "make the chart look good" equal "espresso tastes good"

[ATTACH=CONFIG]17125[/ATTACH]

G'day John

I reckon a major set of insights are about to hit. Using a manual lever the best range of shots have a fairly uniform flow rate "into the cup". That is why manual levers can excel in the first place. Clearly your chart shows the lag "post preinfusion" is substantial. Who would have picked that? I now feel that a fair part of the reason why my manual lever Electra always did a better cuppa than my 220V GS3 is on your chart.

FWIW, my take is that you should not grind coarser (I have rarely improved a good espresso setup doing that), it would be to increase the "early post preinfusion" pressure to give a uniform flow rate into the cup. Considering you have the real time data from your scales I also reckon an automated link there to set the flow rate would be a total gamechanger.

TampIt

[>decentespresso]

I reckon a major set of insights are about to hit. Using a manual lever the best range of shots have a fairly uniform flow rate "into the cup". That is why manual levers can excel in the first place. Clearly your chart shows the lag "post preinfusion" is substantial. Who would have picked that? I now feel that a fair part of the reason why my manual lever Electra always did a better cuppa than my 220V GS3 is on your chart.

I agree with you there: lever machines have always had much more subtle espresso control than flat-9-bars pump machines, and I agree, they taste better.

FWIW, my take is that you should not grind coarser (I have rarely improved a good espresso setup doing that), it would be to increase the "early post preinfusion" pressure to give a uniform flow rate into the cup. Considering you have the real time data from your scales I also reckon an automated link there to set the flow rate would be a total gamechanger.TampIt

Good suggestion. I think the way to do this is to slow down the preinfusion rate, to give the puck more time to absorb the water. Pressure is building up during the fast preinfusion, but I've gone too fast now.

I'll try slowing PI down to 3 mL/s and see if that helps. Thanks for the insight.

[>decentespresso]

A "Stop Espresso at This Weight" feature has been added and now works on the DE1+.

If you buy a DE1+ and as well as our Bluetooth "Skale 2", you will be able to do "gravimetric" shots like you can on the Black Eagle and La Marzocco Linea PB.

A new feature will appear in the DE1+ profile editor, letting you set "stop at weight", if you have a Skale. The feature won't appear otherwise.

The DE1+ will automatically stop making espresso when 97% of that weight is reached. That last 3% is to account for the pressure remaining in the coffee basket, which will still cause some coffee to flow into your cup for a few seconds even after the pump has turned off.

In a few weeks, I'll make the feature a bit more pretty in the Tablet GUI. Not to worry: I know it looks a bit cramped in there now.

The "Skale 2" is a good quality scale, but not meant for use in wet environments, which is why we're working on our own "Decent Scale". However, in the interest of not delaying the use of this useful feature, I've implemented it with the Skale so you can get started using it right away. Eventually, we'll have our own Decent Scale and this feature will work with that too.

[MrJack]

In those charts, it's purely virtual. Virtual is what we've had for most of the past two years, and one advantage of that approach is that very short, fast changes in flow can be noticed. With a hardware flow meter, they're not. This quickness is useful because we can spot (and chart) channels opening and closing, as well as quick & short drops in pressure during the initial pressure rise to 9 bar. The 2nd one really helped me improve my puck prep skills and also "neg" some grinders as being poor quality.

Our intended "final" approach is to use the flow meter to give us the overall speed of the water, and the physics model to modify that overall speed with quick changes we notice.

At the moment, the hardware flow meter is disabled as we're focussed on getting the physics model right. In a few weeks, the integrated flow model should be done, and the charts will be using that.

A rising flowrate and falling pressure make perfect sense to me, from a puck resistance point of view.
Ongoing accumulation in the puck, less so.

That begs the question, is the difference between the measured outflow and virtual inflow real, or virtual (or a combination)?

I presume your virtual flowmeter is essentially a calculation of speed x displacement?

Here is a plot of the data from that shot, showing inflow and outflow against the pressure, from 12s onwards. Also plotted is a pump curve for an Ulka vibration pump. I can't help but see the resemblance of the outflow measurement (orange) to a pump curve (green) but operating at say, 50% speed. The blue line would only be possible if the pump speed changed.




Which makes me think that the inflow calculation is inaccurate (perhaps it is not considering 'slip') and that at least some of the difference is virtual. Although there are other factors such as temperature, density and incomplete saturation of the puck, which could also be at play.

The good news is, if that is the case, it should be fairly easy to generate a detailed model of pressure vs flowrate vs speed for the pump, to improve the accuracy of the virtual flowmeter (if indeed my assessment is correct).

[>decentespresso]

Which makes me think that the inflow calculation is inaccurate (perhaps it is not considering 'slip') and that at least some of the difference is virtual. Although there are other factors such as temperature, density and incomplete saturation of the puck, which could also be at play.The good news is, if that is the case, it should be fairly easy to generate a detailed model of pressure vs flowrate vs speed for the pump, to improve the accuracy of the virtual flowmeter (if indeed my assessment is correct).

Correct: we know that our virtual flow rate numbers are not yet correct. I mentioned in a previous post, that this in fact was the reason for integrating a scale into the DE1+ at this time, in order to be able to fix the model.

The main mistake in our model seems to be over-estimating flow when there is pressure, and under-estimating it when there is not. Vibratory pumps move a differing amount of water depending on what they're pressure they're "fighting against".

In a few weeks, we'll have improved our virtual model, but for now we'll only be using to display quick, transient changes in flow (to help notice channeling), but using the hardware flow meter for accuracy of reporting flow.

[>decentespresso]

Flat UI be gone

A few weeks ago, the Nielsen Norman Group published a study indicating: "The popularity of flat design in digital interfaces has coincided with a scarcity of signifiers. Many modern UIs have ripped out the perceptible cues that users rely on to understand what is clickable." https://www.nngroup.com/articles/fla...e-uncertainty/

I found the study very convincing, as I've found my iPhone harder to use since they've made all the UI elements "flat".

In early September, I changed the Decent Espresso web site's "CSS Theme" to use drop shadows on the buttons (previously, buttons were "flat"). I sat back and waited for the complaints to roll in. They didn't, and the "how do I buy something on your website?" tech support messages seemed less frequent.

On our espresso GUI, I've often found it a bit hard for the to track "what should I do next" ?

Today, I've taken the Nielsen Norman Group advice to heart on our espresso machine tablet software, and all "tappable" buttons and tappable icons have a 3D "lift" to them. I'm hoping this makes it a bit easier for baristas to intuit what to tap on.

Below are some animations showing the DE1+ gui and also all the configuration pages. FYI I use photoshop to make the tablet GUI, and its animation feature to test what common commands might look like to the user.

---

And for a walk down history lane...

I'm a big fan of Jakob Nielsen, whose "Green and Blue" usability book as a huge early influence on me in my 20s. My employees grew so annoyed at this book that, in jest, they had a "What Would Jakob Do?" t-shirt and mouse pad custom made as a gift to me.

Jakob had very kind things to say about the Lyris Email List Server software I wrote (he used it, back in the deay) and he even studied how usability increased his newsletter subscriber size: https://www.nngroup.com/articles/mai...ist-usability/ - a very tangible proof of usability making business sense.

[>decentespresso]

No burning your fingers

Steam wands can get hot, and espresso machines usually put a short rubber tube around a bend to help you move the wand without burning yourself. The problem is less acute with "cool touch" steam wands, such as our machines, that have a Teflon inner tube. However, the wand still gets hot simply by being immersed in hot milk, so a rubber tube is welcome.

Two years ago, I bought from several suppliers the protector models they sold. Generally, they were on the harder side (sometimes plastic), with a slight air gap between the tube and the protector. Most tended to slide around instead of staying put. And they are generally fairly small, so you do need to pay attention when reaching toward the hot metal tube.


I wanted the protector to be as long as possible, and to stay put. So, I decided to make my own protector from silicone rubber irrigation tubing. I could make it longer, at 60mm (just far enough to not interfere with a 1 Litre jug). The finish was matte black and looked nice. Heat insulation was not great, however, because of a lack of air gap and the high density of the rubber, so the rubber "protector" itself got hot.

Also, it was really difficult to cut the tube in a straight line, because it deforms as you push down to cut.



This good-but-not-quite-right tubing is what has been on all our machines until now.

This past summer our mechanical engineering intern Tommy studied a variety of tubing materials, to solve this problem. As a heating test, he ran steam continuously for 3 minutes and measured the surface temperature of the rubber. He also overheated the rubber to see if it gave off a smell or deformed. Finally, we wanted 3rd party food-safety certification.



Two days ago, the final tubes arrived, cleanly precut to the right size. The surface is a bit shinier, but the heat insulation is vastly better than my attempt. The fit is tight, so it doesn't move about, and there's a trick to easily sliding it on: soapy water.



The final part costs much less than $1 per machine, but it's such an important part of daily interaction with an espresso machine, that I wanted to be sure to do it right. There's nothing worse than burning yourself in the morning as you make your wakey-wakey espresso drink.



-john

[TampPolice]

The 2nd one really helped me improve my puck prep skills and also "neg" some grinders as being poor quality.

Thats interesting. In the 2017 decent era of espresso we can now evaluate our grind and distribution methods by looking at the charts.
Out of curiosity, in your experience, which grinders have been able to produce the best looking charts? Apart from the decent grinder

[rawill]

I think the way to do this is to slow down the preinfusion rate, to give the puck more time to absorb the water. Pressure is building up during the fast preinfusion, but I've gone too fast now.

I'll try slowing PI down to 3 mL/s and see if that helps. Thanks for the insight.

Now this is interesting, as it has been suggested to me (by Tampit) to preinfuse, stop the machine, do something else, get milk in the jug, make the toast or whatever, then steam the milk and do the pour.

The reason being to let the puck get wet through before starting the pour.
Does this add up?,
I use an EM 6910, among others!

[LeroyC]

Decent Espresso Machines (DE1) - Any thoughts?

Now this is interesting, as it has been suggested to me (by Tampit) to preinfuse, stop the machine, do something else, get milk in the jug, make the toast or whatever, then steam the milk and do the pour.

The reason being to let the puck get wet through before starting the pour.
Does this add up?,
I use an EM 6910, among others!

That would only really work with a machine that allows low pressure preinfusion. So it might work on an EM6910 that has preinfusion of sorts, but wouldn't really work on anything else that lacks preinfusion. I experimented with this on a couple of my old machines (Gaggia Classic and Brugnetti Simona Top) as I saw it suggested by a few people, but as they hit 9bar straight away it didn't really work in my opinion. Haven't tried it on the 6910.

[>decentespresso]

That would only really work with a machine that allows low pressure preinfusion. So it might work on an EM6910 that has preinfusion of sorts, but wouldn't really work on anything else that lacks preinfusion. I experimented with this on a couple of my old machines (Gaggia Classic and Brugnetti Simona Top) as I saw it suggested by a few people, but as they hit 9bar straight away it didn't really work in my opinion. Haven't tried it on the 6910.

I've done something like this, putting in a 60 second pause after preinfusion. The water starts to drip through a little bit, and in the end, you get something that tastes like a mediocre pour over. Others may have better luck, but the DE1+ definitely supports this sort of program.

Thats interesting. In the 2017 decent era of espresso we can now evaluate our grind and distribution methods by looking at the charts.

Absolutely so. With a constant flow shot, the higher pressure you reach with the same dose weight and beans and grind setting, the better your puck prep was.

I was disbelieving for years, but what Rao does (gentle side taps, and nothing else) produces the highest pressure according to this test. Rao also angles the portafilter while he taps, which I don't do, but otherwise I've ended up at the same technique he uses (and that was not my intention).

So far, my results are the same as Socratic Coffee's, namely that more fussing with the grounds yields more variability.

Out of curiosity, in your experience, which grinders have been able to produce the best looking charts? Apart from the decent grinder

The Mythos is fantastic, my only wishes for improving it would be (a) weighed doses and (b) a locked position for the portafilter which results in a centred mound.

Our current grinder is OEMed from a subdivision of Mahlkonig (from HeyCafe, and then we modify it further inhouse) and my goal with it was to have a very acceptable under USD$1000 grinder with dose weighing.

The Mythos is in a different league.

I've had some Peak shots and they were excellent, but reliability in a production setting is at the moment, not great with those grinders. Once they solve that, the Peak will probably be in the same league as the Mythos.

Roburs are to be avoided.

On the lower cost side, I think very highly of ANFIM.

[HBDecentRefugee]

The final part costs much less than $1 per machine, but it's such an important part of daily interaction with an espresso machine, that I wanted to be sure to do it right.

Attention to detail is at the heart of good design. Love to see this in the DE1 design and prototyping process...

[LeroyC]

I was disbelieving for years, but what Rao does (gentle side taps, and nothing else) produces the highest pressure according to this test. Rao also angles the portafilter while he taps, which I don't do, but otherwise I've ended up at the same technique he uses (and that was not my intention).

So far, my results are the same as Socratic Coffee's, namely that more fussing with the grounds yields more variability.



The Mythos is fantastic, my only wishes for improving it would be (a) weighed doses and (b) a locked position for the portafilter which results in a centred mound.

Our current grinder is OEMed from a subdivision of Mahlkonig (from HeyCafe, and then we modify it further inhouse) and my goal with it was to have a very acceptable under USD$1000 grinder with dose weighing.

The Mythos is in a different league.

I've had some Peak shots and they were excellent, but reliability in a production setting is at the moment, not great with those grinders. Once they solve that, the Peak will probably be in the same league as the Mythos.

Roburs are to be avoided.

On the lower cost side, I think very highly of ANFIM.

I know it's a bit premature when your espresso machine isn't even in production yet, but it almost sounds like you'd be keen to look at doing for home grinders what the DE1 is going to do for home espresso machines. I like your upgraded Tontile grinder and it'll be a more than adequate option for anyone buying the standard DE1, but fit the real coffee geeks that are looking at the + and Pro they're likely to be keen on something better.

And yeah I agree that Anfim are great grinders. As the company that is essentially the Italian arm of Mahlkonig you'd expect nothing less. Not sure that they can be thrown fully in the 'lower end' category though. The SP450 is over $4k here in NZ and equal or better than most other modern electronic commercial grinders.