screen 2017-10-19 at 2.52.57 PM.png
I've collected all the advanced features that can cause info overload, and made a page where you enable/disable them. Now, it's your choice how much data you want to look at.
The charting of pressure and flow rate changes is particularly experimental at the moment. In theory, these lines should give you insight into channeling and current extraction rates.
- For instance, under constant flow, a pressure that no longer declines rapidly probably indicates that your coffee material is mostly extracted.
- Under pressure profiling, short leaps in flow likely mean a channel opened up and healed itself.
- During a pressure rise at the start of a shot, if pressure drops and recovers, this probably means your grinder is not very good.
- It's also possible for the chart data to mean something else or even nothing at all. Maybe a temporary electrical blip from turning the blender on while making espresso?
All these insights are pretty advanced and I also feel that Decent and its user community will need to work hand-in-hand to refine them to make them more helpful and to remove "false positives" cases.
Thus, I've made these features optional.
There are another dozen or so other configurable features that I haven't yet exposed to the GUI, but I might in the future based on discussions with you.
For example, how many degrees +/- the goal temperature, do you want the chart Y axis to be? For the time being, if you want to change that sort of thing, you'll need to edit the settings text file on a computer.
screen 2017-10-19 at 2.53.20 PM.jpg
A lot of my recent work has been deeply incorporating weight-in-cup data now that I have access to a bluetooth scale. For instance, yesterday this showed me that there is a 2.5 second delay between our opening the water-on valve and the water actually hitting the puck (the time for the water to travel). We need to adjust for that. Here's a photo of that: a shot w/o a portafilter mounted, showing water in (blue) vs water out (brown).
I've also learned a lot about puck preinfusion now that I have "flow rate into the cup" data, and I can assert that preinfusion rarely fully saturates a puck, and that increasing flow rates into the cup are the norm in espresso, even if you put a constant amount of water in (ie, flow profiling vs pressure profiling). I'm now of the feeling that much longer (~20s, aka "Slayer" shots) preinfusion times are necessary if you want to fully saturate the puck.
But if my postings are growing tiresome, I can also back off, or post different things, and we can restart thread when some of the people here have their machines from us.
As you guys wish...
https://www.instagram.com/p/BYsVDhDh45R/), but it's good to have some confirmation. It won't always be 2.5s of course (as it is flowrate dependent). There will also be some water required filling the headspace above the coffee.
The past few days, mechanical engineer Fabrice Martin has been building the DE1 from scratch, alone, using nothing but our assembly manual. The goal is to find missing steps, dependencies, very difficult steps, etc, in advance of receiving all the parts to build 300 (and then 1000) machines. He's never done it before, so he's a good guinea pig.
Everything is being built with samples from the actual manufacturers supplying us with parts. Nothing is from a high-end prototyping CNC service.
For instance, we found that soldering wires to the thermal safety fuses was impossible, and it turns out that's for a good reason, as that's not UL allowed. The soldering gun is hot enough to potentially damage the fuse. You have to use special double-sided crimps. And it all has to be insulated with tubing that, if it were to catch on fire, quickly self extinguishes (ie: Flame resistant tubing). We also found that our fancy custom made bracket to hold the thermal fuses didn't hold them tightly enough and that a simple washer actually worked better.
We've had small leaks in previous group head designs, so we're a bit paranoid about that potential problem. Today we held the group (with a blind basket and portafilter) at 19 bar for two minutes with no leak (using a summer-intern built rig for this purposes), so we think we've licked that problem.
Once this machine is built and has no problems, we give the final approval to the existing parts manufacturers, and they deliver the total quantity already paid for.
IMG_7358.jpg IMG_7360.jpg IMG_7361.jpg 9411508473957_.pic.jpg
Depending on the solder and the rating of the fuse the solder can melt first in some situations.
Attached below is a photo of what we used instead, and I've now heard from two people (via other fora) that this is common.
Oh yeah it'd be bad! Crimping is the only way to do it.
I for one, have been happy to read your updates and development issues - both software and hardware - on this and other forums. It has helped me take an interest in, understand a LOT more about coffee making and appreciate the real life case study of issues in product development. Understanding how the machine is made will certainly help me deal with mine when it arrives.
My only real disappointment so far is the design decision which precludes steaming milk and pouring a shot at the same time because of the decision to use an inadequate power supply. Since you revealed that, I've been practising doing both sequentially instead of simultaneously [as I usually do] on my HX machine and to tell the truth, I think its going to annoy me every day of my life with the DE1+. If there was an option to upgrade the power supply to enable both functions at the same time I would certainly have opted for that.
I thought on 240v you could do both at the same time.
However, I want to clear up the misunderstanding that it's simply a question of a bigger power supply that keeps the current machines from making steam during espresso. If that's all it was, I agree, we'd have done it.
Steaming during brew is not available on our current models for 5 reasons:
1) keeping the machine cost down
2) the small size of the chassis
3) heat management
4) our technology
5) shipping the 220V version at the same time as a 110V version
To keep the price down, all our current models have a single water path. Steaming during brew would mean another pump, and a new water tubing route, dedicated to steam. Because we heat water on demand, we can't just tap the top of a boiler to vent steam, like a traditional machine can. The extra pump would also take space we don't currently have.
In order to get proper powered steam from 220V, we also need to add another heater, or 2x the size of the current heater (which is speced at 1350W, appropriate for 110V levels of power). We don't have the space in the current models to put in another heater.
Another heater would also generate more waste heat inside the chassis, and we must keep our insides under 60ºC. We're typically at 51ºC now. Another heater would mean more heat management R&D, which probably means one or more fans (some noise too, then). That does mean (yes) a larger power supply but you don't just plug a fan in and get efficient cooling, not in such a small space as our chassis. We have to study and alter the internal air flow characteristics so that air flows through the entire machine, and that there are no hot spots.
Steam during brew, if you're steaming during brew, is very challenging technically if you're heating on demand, because the amount of power used to make espresso is going to fluctuate constantly. That would cause steam pressure to also fluctuate, which you would find very frustrating. We've calculated that even at 220V, steam-during-brew, using an on-demand heating approach, will not be competitive with dual boiler machines.
In order to be competitive with a dual boiler at 220V, we need two 220V power plugs, with one dedicated to espresso, the other to steam. That's what we're doing in the DE1CAFE.
And finally, we wanted to ship the 220V at the same time (almost) as our 110V machines. You can see from above that steam-during-brew is non-trivial to add, and we didn't want even more reasons for delays.
If steam during brew really is essential to you, then I suggest you wait until we have the DE1CAFE model out. There will be a tabletop as well as builtin version of it.
However, we do give you quite good steam the instant espresso ends, which most people seem to feel is an OK compromise.
The Decent approach, using on-demand heating, yields a different sent of benefits and compromises compared to traditional boiler technology. Since boilers store up heat over time, they can vent steam easily at any time. Our approach can't, but it does offer other benefits in exchange.
It's up to you whether the different mix of benefits is attractive and worthwhile.
I suppose, though, that this question could be answered by offering a UI where you could "pinch in or out" with two fingers, on any of the charts. For now, though, these sorts of concerns are seen by me "for a future version" and that's why they're in a text file and not in the UI.
Good answer. I use a commercial machine in a domestic environment and simultaneous brew/steam is a non issue. In other words, the DE1/DE1+ in the same environment is an even less of an issue.
Regarding the temperature range, I find it hard to visualize what would be a useful graphical presentation without actually seeing the options, but I assume the config text file will be user editable and can be changed without too much fuss. [Will the tablet arrive unlocked and rooted?]
https://www.greenbot.com/article/245...or-tablet.html) then you will be shown the "OEM unlocking: on/off" choice in "Developer options"
Yes to rooted: using Kingo Root (KingoRoot for Android, the best One Click Root Tool/APK for free.), and the tablet auto-updates the rooting program to the latest version, to avoid security problems.
Note that I have tried to expose, via the tablet gui, everything that most people would want to change. The editing-text-file thing is just for adventurous types who want to "unlock secret levels".
If you screw up and edit the settings text file in ways that cause havoc (and didn't keep a backup), you can delete the file on your tablet and it'll be recreated with sensible defaults.
screen 2017-10-24 at 10.50.00 AM.jpg
Regarding this hidden setting we're talking about (and probably overly so <grin>) -- it is what the Y axis range should be on the temperature chart. Since the DE1+ tends to not go over temperature (the worst we've seen is 2ºC, but we think we can keep it under 1ºC for the release for all cases except clogged pucks) the default is for the Y axis to chart up to 2ºC above the goal, and 9º under the goal. These are probably sensible default for most people, which is why I decided to not over complicate the UI with settings to change this inside the tablet software.
The pressure and flow charts automatically set to sane ranges (0-11 bar, 0 to 6 ml/s).
However, you can drive the DE1+ to 13 bar, and if you find that you want to make espresso up in those regions (I don't think many people currently do) you may want to change the Y axis for the pressure chart as well.
A Scace does not behave like coffee because, using the WBC protocol
you are supposed to pull multiple shots on the Scace to warm up the "plastic puck simulator" as well as the stainless steel of the Scace itself. With a normal shot, the coffee is at 20ºC, and not preheated like it is on a Scace (excepting the special case of preheated grounds from a Mythos grinder)
The DE1+ achieves the goal water temperature within half a second. Our Scace shots look good. At Synesso, when I demoed to them last year, the DE1+ was 0.1ºC above the goal temperature, as measured by their home-made Scace equivalent.
However, the point I'm attempting to make is that coffee grounds don't behave like a Scace.
1) coffee grounds are at 20ºC (room temperature):
2) when hot water mixes with 20ºC coffee grounds, the resulting slurry is at a lower temperature than the input water temperature
On the DE1+ we have two temperature probes in the group head:
1) water temperature into the group
2) water temperature 2mm above the coffee puck (directly behind the shower screen)
When most people talk about water temperature, they're actually talking about another temperature:
3) water temperature inside the boiler
We track this number internally, but don't display it to the user (currently) as we think that the two temperature we do report: (1) water-inside-the-group-head and (2) coffee-slurry temperature, are the relevant ones.
Boiler based machines put constant temperature water into the puck, so that the temperature at the puck tends to rise gradually and linearly.
The DE1+ can vary the temperature of the water hitting the puck, so that the puck temperature can be raised faster than a boiler can.
Is that any clearer?
John, I see it is a bottomless portafilter included, is there an option to upgrade to a dual spout option ?
https://decentespresso.com/portafilter or you can use any La Marzocco compatible portafilter, as you wish.
If I can permit myself to mark up a graphic from the article above, the thin line I've drawn below is a rough approximation of what the DE1+'s extraction temperature curve looks like. This is what I meant by "2ºC within 5 seconds, 1ºC within 10 seconds"
Lately, though, we've been doing much better than that, because we have recently deployed a machine learning algorithm, which uses historical shots to predict the future to compensate ahead of the sensors. In other words, your first shot might see a 5 second delay in reaching 2ºC of the goal temperature in the puck, but successful shots will get closer much faster. This is a *very* new feature for us, and I haven't wanted to promise anything from it yet, but here is a photo of the puck temperature of the shot I just pulled: you can see that we're now precompensating for the cooling effect of the puck.
screen 2017-10-24 at 4.02.48 PM.jpg
I'll be releasing a full video of this shot tomorrow (I have other stuff to work on at the moment). I took the video in response to the comment a few days ago about "does it make coffee?"
Here's a preview of this last major feature for the DE1+ espresso machine. I made this movie to get your thoughts before I finish coding it.
[QUOTE=decentespresso;618628]Yes: we sell a double spouted portafilter https://decentespresso.com/portafilter or you can use any La Marzocco compatible portafilter, as you wish.
I was thinking more like pay the difference between the two, .....can't see I would ever use a bottomless one ?
1. 93.5 - C
2. 91.5 - C
3. 89.5 - C
That way there would be one place you can see all the settings at once for a shot. You wouldn't need to edit them on that screen, just not have to move back and forth between tabs to see the numbers.
For a while at least, it might be helpful to have info bubbles that appear as you touch the item on screen which explain each option in the advance menu - pretty much as you did in the voice-over in video. I found your voice-over necessary to make sense at all of the usage of the advanced features you were describing. Maybe have an option to switch this help option off when experience and familiarity advances and the help bubbles become intrusive and unecessary.
It would be possible to "press and hold" to get help, but the downside to that is then every action will need to start after the finger is lifted, in order to count finger-down time. Currently, the UI reacts to a touch when the finger touches it, which feels faster than reacting when the finger lifts.
My intention is to have a ? icon on most pages, which gives help for each page. Possibly even a how-to video. That feature will come in a later version.
Last edited by decentespresso; 25th October 2017 at 01:26 PM.
small 'i' icon in the corner of each box?
Or, just the ? icon as John suggests would work for me.
Wow John, this is such a great milestone ..... congratulations !
It has been a fascinating journey indeed, to follow the progress, the trials and tribulations to finally get to the final "build" machine.
Appreciate your transparency through the process, it been like that addictive TV show, waiting for that next post to see the next twist and turn of the tale ......... Yes, I've been told I need to get out more !
Certification the next 'Episode' ?
CongratulatIons on your project and involving coffee enthusiasts in the development process. It certainly looks like a game changer with a surprisingly low price. The spiel for the DE1PRO+ includes "upgrades the internals for longer life" and I couldnt help thinking that potential buyers would want that sort of quality built into all models. It raises the question about the expected life of the DE1 and DE1+
On the DE website the DE1 and DE1+ has an expected service life of 10,000 shots, that represents quite a few years for general usage I would think, the Pro seems to be cafe oriented and has a premium price tag to cater for the extra service cycles to 100,000.Longevity.JPG
I guess it comes to where does John draw the line for a commercially acceptable price-point for a general user.
He then offers a more robust unit if you want to pay the premium, good product segmentation I think
I wanted to show you how the DE1+ display lets you understand the complicated reality of an espresso shot.
The photos below show the program I set, and what actually happened. This was a "slow to develop" shot, with a long preinfusion, slow drips at the beginning, and then a pressure decline.
The resulting shot was thick, delicious, with a taste of toasted peanuts that I don't get from these beans when I do flow profile shots (which is what I usually do).
I am also showing you these two photos so you can see the difference between the desired program, and what happens when you bring reality (ground espresso) into the picture.
FYI the shot ended before the pressure reached 6 bar, because the in-cup weight of 28g was reached (this was a 14.5g dose).
It’s been awhile since I’ve shown you a shot being made on a current DE1+ machine. So here you go…
A year ago, Jeff came up with the very clever idea of insulating our water heaters by pouring silicone around them. He built a mold out of legos to try the idea out. You can see that experimental heater 3rd from the-right, bottom row.
We then spent a good nine months trying to move the idea into production. The problem was that silicone with no air bubbles didn’t insulate very well (bottom left, middle and right heaters, in the photo).
As we added bubbles to the silicone mix we ran into two problems:
1) the bubbles caused a slow, creeping expansion while drying. This tended to break the mold or for the silicone to ooze out unpredictably, sometimes even uncovering areas (2nd heaters from the left, on both rows)
2) as the silicone dried, the bubbles migrated upwards, leaving the bottom very high density (low insulation) and the top overly bubbly (top left in the photo).
The solution to both these problems is to do this properly, with a metal mold under compression, and having a machine rotate the product constantly as it dries. That means we have to outsource this to someone who actually knows what they’re doing, because we clearly don’t.
So here’s what we decided…
For the 300 machines we are building now, we have a solution that works well, but is very time-consuming (about 30 minutes each). We make a box from fiberglass boards, and stuff the inside by hand with mineral wool. That's the two heaters in the top right of the photo.
For the next 1000 machines, we’ll be working with a new company to make our water heaters, using our custom design (a bit more steam power, even at 110V), and they will also make a two-part mineral-foam insulating box for our heater. That will save us a lot of time.
Alas, sometimes research leads you to a dead end…
Will the future decent scale work with the DE1 interface or only with the DE1+ interface? This has been a great ride watching the development process.
On the DE1+, with the scale you see total shot weight and flow rate into the cup, as well as adding a end-shot-on-weight feature (like the Black Eagle and Strada EP have). As those are advanced features, they're targeted for our more advanced machines.
1) this is a new mold, and will take 90 days for the mold to be made and a first sample produced, then 60 days after that to get quantity. That would be a 5 month delay. The mold itself costs USD$17,000 vs $17 each for the heaters we buy for these 300 machines. For USD$10k (600x$17) we can get all the heaters we need for these 300.
2) and the minimum order quantity is 10,000 water heaters, ie enough for 5000 espresso machines. The per-heater price goes down to $12 each, but we're still looking at writing a cheque for $120,000 + $17,000 = $137,000, just for the heaters.
3) we haven't worked with this company before, and we want to test the heater for at least several months before putting it into production.
The dates I've put above assume (a) no testing on our part and (b) everything we get from these people is perfect and doesn't require any iterations. So, add several more months to compensate for reality.
It's actually quite a challenge for any company to make only 300 machines. When we finally can produce several thousand machines per year, we'll have more options. I expect that every hardware revision of our espresso machine will improve slightly, as is the case for our accessories (we're on revision 6 for our milk jugs).
Just to be clear, the existing box for the heaters insulate just as well as the new box will. We're hand-stuffing mineral wool into each box, and the air temperature currently reaches 60ºC around the outside of the heaters, despite the steam heater being at 180ºC. As far as a power increase goes, we're only able to get 9% more steam power from the new future model, and then we're at the limit of what we can do without potentially blowing a fuse once in a while.
I don't know if you wanted all that information <grin> but now you know why we did what we did.
Last edited by decentespresso; 30th October 2017 at 05:36 PM.
Last edited by decentespresso; 30th October 2017 at 07:16 PM.