Decent Espresso Machines (DE1) - Any thoughts?

[>decentespresso]

new ideas for group head water diffusers

Ben Champion suggested we change the group head design to make experimentation easier, which we've now done. Ben and Ray are now trying several different group head diffuser ideas, some based on calibrated orifices, some on turbulence:



All of these are performing well under simulation, and all are much better than the single-hole approach we had (and that you find in other espresso machines). However, it's unlikely that we'll decide which one is the "bee's knees" of diffusers in the next few weeks. We'll likely choose one of the calibrated approaches, because it's known to work well (turbulence is very hard to understand and model).

We'll keep banging away on these ideas in the months to come, and I'm hoping that once the DE1 is out, other people will join in with their own ideas, and have them CNCed (local CNC cost is about $150 for this part) to see if they work better. With lots of people banging on the problem of even water flow (at low velocity) under varied conditions, I'm hoping that a winning approach will emerge (or maybe several). We'll be posting the CAD files for our own ideas, in case you want to start with that and modify it.

Here's what some of the simulations look like. We're trying to have even water flow, but also quite low velocity, to minimise "drilling into" the coffee puck, at varied pressures and flow rates.

Our next set of simulations, which Ben has started to work on, involve simulating the puck's absorption of water. We're hoping to gain a better understanding of what happens during preinfusion through this work.

[>decentespresso]

espresso tablet artwork in other languages

In the past 48 hours, 23 people have volunteered to translate the DE1 tablet software, and we now have added Swedish, Hungarian, Thai, and Danish to the list of supported languages (in addition to Japanese, Korean, Chinese, Spanish, Italian, Portuguese and German).

There are about 40 artistic tablet styles that will be available on the DE1. Instead of doing real work, this past summer I worked with 40 different artists, from all around the world, to create these. The translations are automatically used on all of them. Below you can see some examples of this, in Korean, Japanese and Swedish.

[MrJack]

Ben Champion suggested we change the group head design to make experimentation easier, which we've now done. Ben and Ray are now trying several different group head diffuser ideas, some based on calibrated orifices, some on turbulence:

[ATTACH=CONFIG]15628[/ATTACH] [ATTACH=CONFIG]15629[/ATTACH] [ATTACH=CONFIG]15630[/ATTACH]

All of these are performing well under simulation, and all are much better than the single-hole approach we had (and that you find in other espresso machines). However, it's unlikely that we'll decide which one is the "bee's knees" of diffusers in the next few weeks. We'll likely choose one of the calibrated approaches, because it's known to work well (turbulence is very hard to understand and model).

We'll keep banging away on these ideas in the months to come, and I'm hoping that once the DE1 is out, other people will join in with their own ideas, and have them CNCed (local CNC cost is about $150 for this part) to see if they work better. With lots of people banging on the problem of even water flow (at low velocity) under varied conditions, I'm hoping that a winning approach will emerge (or maybe several). We'll be posting the CAD files for our own ideas, in case you want to start with that and modify it.

Here's what some of the simulations look like. We're trying to have even water flow, but also quite low velocity, to minimise "drilling into" the coffee puck, at varied pressures and flow rates.

Our next set of simulations, which Ben has started to work on, involve simulating the puck's absorption of water. We're hoping to gain a better understanding of what happens during preinfusion through this work.

[ATTACH=CONFIG]15631[/ATTACH] [ATTACH=CONFIG]15632[/ATTACH]

John, Would be interesting to see whether there would be any impact from changing the orientation (or length) of the inlet piping. From the simulation results, looks like the change in direction in the elbow might be influencing the flow through the distributor (more bends in series typically makes this worse).

What is it being modelled in? There looks to be some high velocities in areas I would not expect.

I would hazard a guess that the way the chamber fills at the beginning of the shot (they real meaning of pre-infusion?) would be more important than pressure drop mid shot.

The easiest way to reduce velocity is to increase the flow area

[>decentespresso]

The easiest way to reduce velocity is to increase the flow area

This is being modelled in solidworks, and there are LOTS more images to post, but this forum software limits the number of graphics per posting. If you've got an engineering background and would like to join in the conversation, I'd be happy add you to the engineer's forum that is discussing this.

At the moment, the best performing design in simulation is Ben's, which uses turbulence to reduce velocity. However, we haven't tested it in the real world, and my self-imposed end-of-R&D deadline is coming up in two weeks.

We're going to need to pick a design which "works well in simulation" and "works well in the real world" on to start with that, with the understanding that coming up with an ideal design might take another 10 months of iterating between design, simulation, CNCing, debating and real world testing. Regardless, this design will perform a lot better than the single-hole design we had before, but there's room for improvement.

[>decentespresso]

A demo of our DE1 Espresso Machine tablet interface.

[gc]

It may just be my old laptop, but the tablet GUI - especially the sliders - doesn't seem to have very strong contrasts. As my vision needs some help these days, will there be a skin available with stronger contrast and colours for visibility?

[MrJack]

If you've got an engineering background and would like to join in the conversation, I'd be happy add you to the engineer's forum that is discussing this.

Sounds interesting. I'll send you a pm.

[>decentespresso]

It may just be my old laptop, but the tablet GUI - especially the sliders - doesn't seem to have very strong contrasts. As my vision needs some help these days, will there be a skin available with stronger contrast and colours for visibility?

Yes, we've made a high contrast skin for folks with poor vision (see below), but the contrast on the sliders had evaded me as a consideration. I will fix that, and thanks for mentioning it.

[>decentespresso]

NEWS: steam wands ordered today Today, we start ordering parts for the 180 DE1/DE1+

Today, we start ordering parts for the 180 DE1/DE1+ machines that have been pre-ordered. We're buying enough parts to build a first run of 500 machines. Depending on the part, it can take up to two months to receive an order, which is why we're starting now, so we can start assembling in June.

We ordered 500 units of the steam wand for the DE1/DE1+ models. This is an off-the-shelf part from our espresso parts supplier. We're also ordering 500 units of our own-design steam wand for the not-yet-announced DE1CAFE model. In the photos below you can see the difference between them. The cafe model steam wand is meant to extend beyond the front of the counter, and to be the right height in a countersunk espresso machine, and to easily accommodate 1 liter jugs. The DE1/DE1+ steam and is more compact and is optimized for the shape of that espresso machine.

As promised, we will be sharing CAD files for every part we've created. Here is the CAD file for the DE1CAFE steam wand: https://tinyurl.com/de1cafe-steam-wand (you'll need a free onshape.com account to see it)

[Gorokabeanz]

I know you guys have a lot on your plate at the moment but am I able to enquire as to an update on your distribution tool?

Im really loving the uber-isation you are bring to the industry and am following with great interest. Keep up the good work!

[>decentespresso]

I know you guys have a lot on your plate at the moment but am I able to enquire as to an update on your distribution tool? Im really loving the uber-isation you are bring to the industry and am following with great interest. Keep up the good work!

Regarding the distribution tool, it was only a concept 10 days ago, drawn in 3D, and the CNCed version of it, as a "blade" that can be inserted into our funnel, arrived on my desk this morning (photo below).

We don't even know if this is a good idea yet. I won't manufacture this until (a) I know it actually works and (b) there's no obvious way to improve it, so it's unfortunately months away still.

[Gorokabeanz]

Thanks for the update. Hope it does come into fruition in one form or another. Good luck on the endeavour .

[>decentespresso]

Burning Man 2013 helped design the Decent Espresso machine

Back in 2013, just getting started on this Decent Espresso project, I was at Burning Man and was floored by the LED "eye candy" display by the Disorient camp Disorient - Disorient wiki

Here is what it looked like:

Being a geek, I scrutinized the circuit board, power supply, and I was really impressed how well engineered it all was. Having something this complicated survive 110ºF and sand storms is not easy. I'm not the only one who was impressed, here's an article about the tech : https://blog.adafruit.com/2014/03/18...eagleboardorg/

I happened to meet Disorient's engineer on this project, at a free coffee stand run by some crazy Italians and I picked his brain for a half hour.

One major take-away: he absolutely loved, loved, loved Meanwell power supplies. They'd never failed him, in all his years at Burning Man. That's quite an endorsement.

Today, we ordered 500 units of this Meanwell power supply for the first manufacturing run of the DE1:
LPV-35-MEAN WELL Switching Power Supply Manufacturer

It's totally sealed (airtight), fanless, and we've never had a problem with any of them. It's a pricey part, when you're running a power supply inside a hot and humid espresso machine, you want something really, really good.


[>decentespresso]

NEWS: bluetooth module ordered

Today we ordered the bluetooth module BLE Nano made by Red Bear Labs to build the 90 pre-release Decent Espresso Machines in June.

This may sound like minutiae, but because all communications between the tablet software and the espresso machine go over bluetooth, we need to pick something that has been around a while and is known to be reliable.

Lots of people have used this module for their devices, and it's known that Android and iOS apps work well with it. That's important, because--despite its ubiquity--bluetooth can be frustrating unreliable, especially if you cut corners and get cheap parts.

Thinking about the future, going with Red Bear also gives us an upgrade path to offer Wifi upgrades, simply by swapping out this chip.

We also like Red Bear labs because they support open source development, and their bluetooth module has a CPU on it that can be programmed using standard Arduino development tools.

This means that "coffee hardware hackers" can write their own Arduino software that they upload to this tiny computer, and it will enjoy access to the Decent Espresso computer and all the sensors.

For our fans who are software developers, we've written a Python program (that we'll give away) which emulates a Decent Espresso machine, but from your computer. You plug the BLE module into the USB port adaptor, run our Python script, and you can now program much faster without having actual hot water being dispensed as you debug.

We only bought 90 of these, because we'll be switching to a less-large, less expensive version of the Nano for our production machines, starting in June. There is some firmware programming we have to do to make the switch, and we like the fact that this version has a socket in case we need to changing anything out during the pre-release rollout.

[gc]

Just a bit more info for John on GST. eBay threatens to block Australians from using offshore sellers