I read that article when it came out last september, and added subtle drop shadows to all the "tappable controls", as well as doing the same on the decentespresso.com web site.
I've been a fan of Nielsen for 20 years and we've corresponded. He's used some of my software (Lyris) and written about its useability:
screen 2018-01-10 at 12.17.01 PM.jpg
However, this skin was also meant to be useful if you were blind, because the tap zones are 1/3rd of the screen each, so you can find them by placing your hand on the tablet and using your thumb. The power off and settings buttons are each in a corner. The Settings page is not currently available in a sight-impaired version -- my feeling is that someone with better sight can perhaps set the machine up for the sight impaired user.
To avoid spamming everyone, I'll send you a PM with a better description of the incongruent behaviour as I see it, a set of steps to replicate, and any obvious options I see to address that are in keeping with your description. You can then determine if I have a point, or I'm just crazy!
Appreciate the follow up - thanks John.
The function in the DE1+ with the DE scales will be great..... any progress on scales MKII ?
Will the DE1 have a STOP function available in the API, appreciate it will not be in the provided skin ?
My favourite skin is the Three Cups ... just simple, nice and clean.
Here are a few photos from our new 3000 sq ft (about 300 meters squared) espresso machine factory, where we are starting to build the first 10 machines.
We have most, but not all of the parts, but everything we ordered should be here in the next two weeks.
However, as these are our first production machines, we also expect a few "whoops" where we realize that we forgot to buy some small part. This is part of the transition from R&D to production, and now everything has to be done "according to the manual", where we've been writing out and documenting everything little thing.
As such, the first 10 machines will likely take a few weeks to make, but then things should speed up considerably.
1) as discussed here a year ago, we're not gluing the tablet stand to the tablet ahead of time. Instead, you'll peel off the wax paper and place the stand onto the tablet, as appropriate for your body height and tablet height. We're designing a die-cut plastified piece of paper to include, that will serve as a guide to help you center it (photo: bottom left).
2) we've designed a plastic guide for our use, for gluing the hidden thin iron plate inside the chassis, which keeps the tablet invisibly in place with the magnetized tablet stand. The plate is a large enough magnet that adjusting the tablet (see above #1) still finds its place. (photo: bottom right)
3) we're finalizing the part to assemble the "catering kit" for those who bought the plumbable DE1PRO+ machines. This is built from an aquarium pump, an LED, an RJ45 connector and cable, and a metal enclosure. Middle-left photo shows the in-house prototype, while middle-right photo shows the planned enclosure.
4) the PC boards have not yet been delivered. They are finished and sitting in Shenzhen, but the manufacturer accidentally forgot to order one part, which happens to be the main AC power connector (photo: top right). Whoops. Next week, we should receive 10 finalized sets from them and give our approval for full delivery the following week.
5) we're settling on what colors to assign to which thermometer cables (there are 6 in all) and we'll be receiving the final color-coded fiber tubing next week. (photo: top left)
6) Two new people started this week (EE/technician Parry, and former seamstress Jennifer), while the two people who started last week didn't work out and are now gone. I had hoped that Jennifer's fine sewing skills would mean that her manual speed and dexterity was a transferrable skill, and that's been extremely the case. She's lightning fast and lead mechanical engineer has slotted her for more advanced tasks. We might be hiring more seamstresses in the future!
As we're focussing on building espresso machines faster, we're now paying attention to the steps that take more time, or that take a bit of skill (better if it doesn't) to get right.
"Crimping" in our case means bending a small piece of metal around a wire. We have hand-squeezed tools to do this, but we're finding that our hands are getting tired.
"thermal fuses" are a fire safety precaution, and they are devices which interrupt the flow of electricity if they reach a certain temperature. The tricky thing is that if the connection to the thermal fuse is not reliable, then your device will be unreliable too. And what's more, the wires have to be crimped, not soldered, onto a thermal fuse, because soldering is hot and can damage the fuse (whoops!) in the process.
Today we're ordering two different crimping tools so that our connections are more solid and less fatiguing.
For the thermal fuses, there's the big green machine in the photos. It takes a spool of copper wire, and (kind of like a sewing machine) wraps it around very lightly whatever you've put inside the loop.
For the grounding tabs and the high voltage electrical connections, a separate machine is required. This one runs on an air compressor, is actuated with a foot pedal, and has various different bits you fit into it, for different kinds of crimps.
These machines will arrive next week and should speed up this part of the process.
FYI for machines from v1.2 and later (from August onwards) we plan on having about half of these done ahead of time for us. For example, the thermal fuses will be installed, with a thermostat, onto each water heater, by the new company that will be making those heaters for us. We will still have some crimping to do (so we'll be happy to have these machines) but about half as much.
Thanks for these continuing updates. All very interesting and something I can look back on when telling stories about my new machine.
The past 7 or so days for me have been about setting the included tablet to be as friendly and professional as I can make it. I've intentionally set things up different from a normal Android tablet normally arrives, so as to make the tablet experience a bit easier.
Here are a bunch of choices I've made, that I'm open to changing my mind on:
- the screen doesn't power off if the tablet is plugged in. That way, you can tap the tablet at any time to warm up the espresso machine
- all Google related icons have been removed from the main screen. If you want to run other apps, tap the ... icon on the center bottom
- a custom wallpaper is installed, with only one important icon, in the center.
- you push the power button to wake the tablet. No gesture needed. This is just to simplify usage.
- there is no requirement to have a Google Play Store account, nor to connect to wifi (though wifi is needed when/if you want to software update). Software updates are not automatic because you might feel that your machine is working perfectly well and that you don't want to change anything.
- the tablet will arrive to you pre-Bluetooth-paired to your machine.
- the tablet will arrive preconfigured (just power it on). At the moment, the tablets are set up to be in English, but eventually, I'd like to pre-set up the tablet to whatever language and time zone you used when ordering your machine from us.
- there is a self-standing "Decent Update" program icon, on the bottom left of the screen. This will update your software. This same feature is available in the app. However, I realized that if I accidentally sent out a totally broken version of my software, you would have no way to update to a non-broken one. So, I've made this alternate "Decent Update" for that (hopefully to never occur) eventuality.
- "show touch events on screen" is enabled, as I find it helpful to have little "ghost clouds" appear under my finger as I tap. If you disagree, you can change it.
- the tablet is pre-rooted with a recent free version of supersu, and the kernel is not locked down ("adb" runs in root mode). The tablet is "hacker friendly".
- the DE1 and DE1+ software is installed in /sdcard/de1 (or /sdcard/de1+) and not buried inside Android in a read-only location, in order to make it easy for you to find files and change things as you wish. Each tablet contains the full source code to the GUI.
- I decided to not have the DE1 app automatically launch when the tablet powers up, because it's not how people except a touch device to work. You do need to tap the espresso machine icon.
Things I'd like to improve:
- unfortunately, there is no way for me to remove the Google Search Bar from the main menu. Believe me, I've tried. If someone figures out how to do this, I will be indebted to them!
- I'd like a larger icon, but Android positively fights you over this. There is a way to do it that I know of, by writing self-standing Desktop Widget, and I'd love it if someone figure this out for me.
Abs agree with having the App launch. People are buying a coffee machine not a tablet.
Bear in mind that I do expect you to have some interaction with the tablet OS, for example to set up wifi so that we can send you new DE1 firmware and software.
Also note that auto-launchers (unless they use a rooted method) display the main menu screen for about 5 seconds while boot finishes, and then they run the desired app. That was super confusing to people. The option was to use a rooted launcher, but I wasn't that comfortable with doing that for many reasons. Also, for people who didn't like that behavior, they'd now have to figure out how to undo my hack.
Icons - Materialize
As far as I know, all you need to remove the Google search bar from the home screen is a custom theming launcher which gives far greater control over the user interface - Nova launcher is only one of that comes to mind. But there are others. And since the tab is rooted you have lots of options.
While I am a reasonably advanced android user [been flashing roms for years on different devices], like the other poster's above, I still think the DE app opening with the tab at boot is a good idea. It is a coffee machine first. As long as there is a one click/swipe from start screen to an android home screen - easily achievable with a custom launcher.
Very pleased it is delivered rooted.
When you say the screen doesn't power off if plugged in I assume you mean charging. If the DE is powered down, hence no power to the USB cable, does the screen go off then?
As I've mentioned before, If you got the source code for the ROM from your supplier you would be able to modify and compile whatever you wanted into the ROM and then flash, including updates to Android itself. Any young software engineering student with an interest could do this easily.
Finally, the forums at XDA developers contain an enormous wealth of solutions and information from a diverse and very talented developer community who are wiling to assist with the technical questions. The Alba tablet you use is an Archos IIRC? Maybe this area might help?: https://forum.xda-developers.com/archos-gen
My reticence in auto-launching the Decent app was:
1) unless I use a root-hack, there is a 5 second pause on the home screen before the DE1 app is launched
2) it gives a microsoft-windows-like advantage to the "official app" as it is auto launched, and dissuades people from making their own competing apps. I particularly want to see competing DE1-control apps, for example with a pour-over focus, or a non-espresso extraction focus (other aromatics), and I can think of others.
3) but this is a topic I'm still ambivalent on and might be convinced to change in the future (ie, enabling app auto-start as a checkbox inside the DE1 app, easily disabled).
If the DE1 is turned off, then it is no longer charging the tablet, and then the tablet will turn itself off. At the moment, this is set to a 30 minute time, but I could make it shorter. I occasionally unplug the tablet to show people at demonstrations what is on the screen, and don't want the tablet powering off. Maybe 10 minutes would be a more appropriate timeout? Don't know : this is not that common a use case.
This situation is why Android 8 is appealing to me. It is also why I'm hedging my bets and have made the app cross-platform. I think Google's Android is the best choice at the moment, but I like options.
Also note that this tablet has an unlocked Android, so that "adb root" returns "true". The tablet I bought from Argos (same exact hardware, almost identical OS setup) returns "not available in production OS" to my request for "adb root". I actually got lucky on this, because I didn't know at the time (last year) to specify this non-locked down option of Android, and it's made my life so much easier. The automatic-configuring of the Android tablets that I'm now doing for customers wouldn't have been possible on the locked down Archos version.
Thanks for the XDA referral, I'm sure you're right that there's good stuff there, and I'll look at it in the future when I have the time. Today I was in the factory all day working with Bugs on heater assemblies. It takes time to hire good people, so at the moment we're short staffed and finding all the time we can to put into assembly, with "making the product" better" taking a second place to "shipping to customers who have already paid"
Appreciate your time constraints at this stage John, and the time you gave for a detailed response. With root access and an unlocked boot loader, customisation by end users should be much easier. You might even consider installing a mini google apps install that doesn't include Google Now, or Google search bar at some stage in the future. But for now I'll just watch as delivery gets closer. All speed to you.
Things are a busy here at Decent HQ, as we're all-hands-on-deck to build machines. You can't overnight hire 6 qualified people, so a lot of us are doing two jobs. I did hire 4 people two weeks ago, and two of them have worked out. We're just about to get into the Chinese New Year holiday here, so it's not easy to hire people. Immediately after the holidays, lots of people look for jobs.
I took this photo montage two days ago but haven't had a chance to write this post. So, there'll be another post in a few minutes with more recent news too.
- the cast aluminum drip tray covers saga is at last over! (center left and right photos) When I last reported about 6 weeks ago, they were nicely cast, but the finish was really ugly. Later, the company making them for us seemed to understand the concepts of "deburred" and also "polished" but both at the same time was never happening. We got two sets of samples over several weeks that were either or. Groan. This week we finally received 10 samples of polished, deburred drip tray covers, and boy they look great. Only negative: about 80% are slightly twisted by the drying process. Sigh. Good news: a little hand twisting for a few seconds and a check against a drip tray, and they're as close to perfect as we could hope for. And of course, the price has risen from $2.90 each to $5.50 each, because it's so much work for them to make them look this good. What can you do?
- ten samples of all the PC boards arrived (bottom left). A few real-world problems were found. For example, the thickness of the wire used on the flow sensor is slightly too thick for the small connector on the PCB. This is solvable with a custom-made cable. We don't like the white glob of insulation paint on one board: it's safety compliant but it makes repairing the board next to impossible. They'll change that in production to insulating tubes.
- Our biggest challenge is that we're having trouble finding someone to make the DC power cable quickly, as it's apparently not that common here (it's a Molex Nanofit http://www.molex.com/dpb/nanofit.html which we didn't think was unusual at all) and the cable makers want a lot of time to special order the parts and learn about it. We'll need to get the parts and make them ourselves, or else this'll delay things. In the past, we've just handmade the cables, in small quantities.
- Jennifer has graduated to making pump assemblies (top left). Because our pumps are "vibratory" (no giggling, please) they need to be suspended in air and vibration isolated or they'll make a lot of noise. We use a combination of springs and soft rubber parts to accomplish this.
- Bugs has emerged from her first government-required audit (it took 4 months!) and is pleased as punch to be away from Excel, now wielding power tools and making heater box assemblies. (bottom right photo)
On Friday, we met with Intertek for a few hours, to discuss their findings concerning their testing of our espresso machines for UL safety compliance (for USA and Canada).
There are 8 pages of "fails" which break up into 3 categories:
1) 2 pages of small things we need to change, such as (a) add an internal drainage hole (b) grounding wires need to be colored green/blue (c) our on-off rocker switches follow the wrong UL code and should follow a different one. A dozen other things, all doable within a few weeks.
2) 4 pages of disclaimer and warning text that we need to include in a printed manual with the machine. "For home use only", "Service only by a qualified technician", that sort of thing. No problem.
3) 2 pages of many new, quite annoying requirements because the reviewing engineering sees our tablet as a remote control. This is is our biggest stumbling block, and it is an interesting issue, so I will write more about it below.
The worry, from a safety perspective, is that you might turn steam on by using the tablet while far away from the machine. It would turn on and could potentially burn someone.
This issue was raised 2 years ago in a review we had with Intertek's consulting engineers, and they agreed that the ability to turn coffee/steam via wifi would not be allowed. However, the two Intertek engineers felt that Bluetooth, because it has a very short range, would be fine, and would not cause the tablet to be seen as a remote control. We could even decrease the Bluetooth signal strength, if needed, to limit the range to 6 feet (2 meters).
However, if the tablet is seen as a remote control, boy does this cause complications. Most significantly, there would have a to be an "I accept this remote command" button tap for every single operation "requested" on the tablet, even if the tablet were sitting on the espresso machine.
For other devices, such as (say) a smart oven, a "click to start" after uploading via a tablet makes a lot of sense. But for a coffee machine, where you do several things to make a coffee (make coffee, steam, purge) requiring an extra tap would not good. Imaging tapping "steam" on the tablet, and then tapping again on the machine "really" do that command.
It's a question of interpretation:
a) is a Bluetooth connection more like a short wire (then: ok, no problem)
b) or is a Bluetooth connection more like wifi (big problem)
Intertek's consulting engineers know that they have been telling us for two years that our design was fine. But now, the independent reviewer at Intertek disagrees.
Intertek also knows that if they reject our application, other companies will hear about it, and avoid using Intertek if they have Bluetooth. Other certifying companies (likely Intertek in other countries) have already approved other Bluetooth devices that act like ours.
I do respect that Intertek is in a difficult position. They don't want to be known as "pay us and you'll certainly get approved". At the same time, if they're overly strict, people won't use them.
Before you think this looks like a really terrible situation for us (well, it's not great), let me tell you what the next steps are:
1) Intertek management really wants this Bluetooth issue to not be a problem, and they're currently researching what other labs have done, and what Intertek have done elsewhere. Their goal is to present a technical argument to the reviewer that will convince him that Bluetooth is more like a wire. If this happens, there are no major UL problems left for us.
2) However, the reviewer might counter that Intertek has been inconsistent on this optic, and that a study and a corporate-wide decision needs to be made. That could take a long time. Until then, "Bluetooth is like wifi" to him.
Once we hear back from Intertek more-or-less definitively on this, we'll make a choice:
a) if Bluetooth is fine, great, we're on a clear path to getting approved
b) if Bluetooth is a problem at the moment, then we will glue some sort of proximity detector to the tablet (like a window alarm sensor) and require the tablet to be sitting on our espresso machine, in order to work. This is "suboptimal" but it does allow us to ship now. You would not be allowed to control the DE1 with your phone. If in the future Intertek changes their mind, we can remove the need for the proximity detector. Intertek management really hate this option, because it means (a) they gave bad advice (b) they are against Bluetooth.
Interesting issues, hopefully to be wrapped up in the next few weeks.
Can the 'proximity detector' simply be disabled by the end user (me) so the machine just obeys orders as you intended?
John, is Bluetooth proximity detection any use to you? I know you guys are technically across it all, but make the suggestion in the unlikely event you haven't already brainstormed that one. eg: https://ewenchou.github.io/blog/2016...ity-detection/
All a bit of a letdown at this stage though. Hope a door opens soon.
I did point out in my text above that this is all subject to change (improve) with time, and so it's possible that the 'proximity detector' would be in place for a few months and then a software update could be sent out that removes the need for it, if/when Intertek decided to go down the "all global Interteks need to agree on this" route.
The issue with signal strength as a safety, is if Intertek sees it as residing in software and not hardware, then the safety feature isn't worth anything for passing-UL purposes. We'd need to find a BLE-proximity detector on/off relay, that is UL approved. Not likely easy to find.
The problem for me with a simple proximity sensor/switch glued to tablet, is that if the tablet dies for any reason at all - not likely I know, but in a small percentage of cases entirely possible, the DE1 won't work unless we have a spare sensor [hope they are cheap! ]
It's alright. I figured it - the sensor on tab is passive I assume, so just leave it in place and load DE software onto another android device and use that? Would BT pairing be an issue in that case or just the usual pairing process?
Last edited by gc; 22nd January 2018 at 01:21 PM. Reason: Further consideration
If tablet charging is via USB that is connected to the espresso machine, can it be used as a connection to control the espresso machines function as well?
As removing bluetooth pairing would be a "software safety feature" and thus of no additional UL approval value, there's no reason for me to lock down pairing. In fact, if I do, it makes UL approval worse, because now the included tablet has to be certified as part of the machine, as opposed to it just being a self-standing device. It's easier for us to get certified if we allow you to use your own Android device, because we don't need to certify it.
Yesterday we standardized and documented the way the group head should be built (left photo), and we test that procedure out by having a totally mechanically untrained person (Jennifer) follow the directions.
We’re moving through the stages of reducing competence needed to build something, but also then nailing down any variability so that they’re all built the same way.
I was looking at how competitors wire up their group heads and thought you might be interested to see how they’re different. The photo on the right is of a competing home-espresso machine.
What’s interesting to me is that they’ve been allowed to get away with two thermostats (those are self-resetting overheating safety protectors) and only one thermal fuse (not auto-resetting). We were told we had to have two non-resetting safeties (thermal fuses) and that the thermostat, being resettable, was not acceptable. We put one in any way, so that if there’s an overheating, the thermostat safety occurs before the fuses, because if a thermal fuse blows, you have to send your machine in for repair :-( — it’s interesting to me to see the variability in what’s allowed to pass through different safety certification authorities.
Also, the competitor uses black shrink tubing on the electrical connections, which is quite safe, but hard to repair. Because those are self-locking connectors, you have to cut off the shrink tubing to unplug the connection. We’ve preferred to go with sliding insulating covers to improve repairability.
Correction: Jeffrey pointed out to me that this other group head is probably using the thermostat to stop pre-heating the group at a set temperature, so that no PID is needed. Like an oven.
Those are thermostats for controlling the temperatures for brew and steam, not safety fuses. Nothing wrong with heat shrink for additional insulation. Did your Intertek say anything about your safety fuse not being on the brass group itself?
The boiler on the right is presumbly from a generic entry level chinese made espresso machine such as a Hamilton Beach so surprising that you consider them a competitor, but the next level up, for example Brevilles, do have stainless steel boilers.
Well noticed on the fuse not being on the brass, and the answer is that Intertek's reviewer didn't take apart our group head to check this, and last friday in the meeting asked us to send them a sample group head not assembled. So, we don't know yet, and this week we're sending them the group head apart from the body.
Intertek's consultants didn't flag this previously as a problem (they've reviewed this) but, in previous versions, the brass and aluminium parts were in direct contact. About 4 months ago, we added a fiberglass pcb insulation layer between the two, so that the hot group head didn't bleed heat into the chassis. I'm not sure that Intertek noticed this change and opined on how that may effect the thermal fuse as a safety.
This brass part is CNCed, so we can make a new feature to put the thermal fuses on the brass if we find out next week that we need to, but for the 300 we already made, we might have to weld a mounting feature on the brass by hand for these 300. This brass part is one of the most expensive on the machine, at USD$ 85 each, so I'm pretty motivated to repair rather than replace.
Last edited by decentespresso; 24th January 2018 at 07:31 PM.
As we're building our first batch espresso machines (rather than one prototype at a time), we're paying close attention to tasks that are quite difficult, slow or error-prone. We'll then try different approaches, or sometimes we work on making tools to simplify or dummy-proof.
- the way electricity comes into the espresso machine is fairly complicated. Why? Because instead of the power switch directly turning electricity on, the way it works is that the power switch supplies voltage to the 24V converter, which then boots up the logic board, and then if the firmware is working correctly, it turns on a relay which then connects mains power. The idea behind this complicated wire up is that if the firmware "crashes" (it is software, after all) then the mains power is immediately cut off by the relay. One consequence of this wire-up (top left photo) is that in two places, we have two wires that combine into one spade connector. We tried 3 different ways to do this, and at the moment, we think that using our newly-purchased copper wire-wrap-crimper (the "green machine") is the best way (bottom left photo). If you've got an electrical background and know of a better way, please chime in.
- the insulating box around our heater is assembled by us by soldering several fiberglass boards together. It's been very time-consuming, about 30 minutes each, to make these. The biggest problem is that the box tumbles apart as you try to solder it. You're constantly using a lot of tape to temporarily hold it together. Clumsy. Johnny has been working on a routed plastic guide to hold the box together, and our EE student intern Stanley has been at it for two days (bottom right photo). We've cut the soldering time down to 10 minutes each. That's still 100 hours total to make the 600 boxes needed to ship 300 espresso machines, but a lot better than 300 hours previously. There's still about 10 minutes of wiring to do afterward on each heater, but that process lends itself to repeating small tasks, and thus speeds up with expertise gained. I've attached a not-yet-closed photo of the heater box and wiring, in case you're curious what's going on inside.
- there is a latch on the back that you lift up, to lift the water intake tube out of the water. Johnny had designed a cut-rubber piece that went around that latch, which makes it feel slightly smoother as you lift. However, he designed it a day after I declared "feature cut off" some months ago, so it's not making it into the first 100 machines. It may seem like a small thing, but I'm worried that (a) the glue might not stick permanently or (b) the rubber guide might not be mounted perfectly aligned, and thus be worse than nothing as it "gets in the way". You can see Johnny with his mounting-guide tool (also from routed plastic) to put the rubber piece in. You can also see the rubber piece on the espresso machine behind him. The glue is 3M supplied and appears sufficient, but the perfect-alignment issue is still to be solved. People often email me asking "what's going to be different between v1.0 and v1.1 of your machine" and this is the sort of small improvement that'll come over time.
- speaking of routed plastic, some months ago we bought an inexpensive open-source CNC machine, hoping we could use it to etch the serial number, model number, etc on the back panel. It turned out not to work for that purpose, because (a) for a drill bit to carve text, the thing being carved needs to be perfectly level, and that turned out to be really difficult and (b) the optional laser cannot "write" on the rear plastic panel. Happily, though, this low-quality CNC machine is perfect for routing out simple plastic shapes, and so it's been on nearly continuously for two weeks, creating tools for assembly. Photo attached of the little CNC machine that could.
Industrial electrician here. Those connectors are the best I'm aware of for your application, however I have two concerns: are the crimps designed for crimping over the small brass crimps as shown in the bottom left photo? I believe that the upper part of the crimp (directly below spade part) is to crimp onto the bare conductor and the lower part of the crimp is to hold onto the insulation so that the insulation cannot pull away from the crimped part of the conductor and thus out from under the outer insulation if used. I'm hoping the bottom left photo is just a stock photo to show the crimp, and after having a second look it appears to be correctly crimped (at least on the one blue wire it's visible on) in the photo above. I'm curious about the crimped look of the insulation over the crimped portion of the connectors. I'd be concerned they are no longer insulating properly all the way around unless the insulation is designed to be done like that.
Also the lack of insulation around the spade portion of the blue connectors in my experience will lead to corrosion ruining the connection significantly faster than with the insulation (like the red connectors bottom right of photo), also significantly increases likelihood of shorting between contacts should any moisture or stray conductive parts avail themselves. I know that both those situations are outside planned operating conditions but as far as best practice goes that would be my advice.
7083768_orig.jpg It would be easier to fit onto the slot sides without templates and a lot of mucking around, and depending on the material, may even be stiff enough to hold in place without adhesive.
I agree with you that a spade connector is designed with a double-crimp surface (1) for the wire (2) for the insulation, as strain relief. However, what I haven’t shown is that our plan is to use heat shrink tubing over this entire part, for about 3cm over the insulation as well. Something like this:
That should provide reasonable strain resistance. However, strain resistance isn’t too much of a concern for me in this use, because the cables are about 10cm long and connects to/from a PCB or a relay. Nothing connecting here moves, so I can’t imagine there being any strain on the wires. What I do worry about is vibration, which is why the spades are all “positive lock” with a tab/spring, and they don’t come off once inserted.
So yes, I agree with you that the way we’ve done this is not typical: we’ve used the copper wrap as if it were the insulator, and crimped twice: (1) on the stranded wire, and (2) on the copper wrap. Both crimps appear to be very solid, and we’re using a pneumatic crimping machine. I would like to think that the shrink tubing will prevent the wire from moving with time.
The way the two-in-one is usually done is with this sort of connector:
However there isn’t space on the rocker switch for two of these. We used one of these on the top left connector of the rocker switch, but the bottom left connector is recessed and there is no space. Also, the Y connector on the top left does crowd things more, thus increasing the likelihood of arcing.
The two-wires-in-one-spade technique we’re using isn’t unheard of. Googling finds things like this (from the Nissan Leaf car), which looks to me like it’s just stuffing two wires into one crimp.
It’d be helpful to hear from you if you think there’s a better way for us to do this, or perhaps if you think the heat shrink tubing will do the job.
One way would be to have a short cable (say, 5cm) with a single spade connector to one of these Y connector spades. However, space is tight in the PCB area, and I’m not sure that adding more cables, with more insulation, doesn’t create more problems than it solves. This sort of thing:
Those blue connectors are only used in the places where the switch (or relay) uses the smaller spade connector standard, and we thus have to use a different kind of crimp. On the larger spade connector standard, we can cover it.
Here are the two sizes side-by-side.
Nonetheless, I like your suggestion, and in an “abundance of caution” I don’t see a problem with putting a bit of heat shrink tubing over that exposed metal. It will mean that a repairman has to take a blade to that shrink tubing to get the connector off, but that’s not too big a deal.
At any rate, I like the suggestion and will add shrink tubing to our assembly to-do-list on these cables.
More to the point, I myself never actually had a problem with the feel of the lever moving through metal. It doesn’t grate as the fit isn’t tight. This was a “problem” that my engineer decided needed solving, but as (a) nobody including me thought there was a problem and (b) the solution might cause problems, I’m not going ahead with this anyway, at least not now. If my engineer solves how-to-position-it problem and it performs well under testing, we'll add it in a month or two, but I regard it as a tiny thing.
Here's a large photo montage that shows all our progress. Busy busy!
Going from left-to-right, then top to bottom:
- 10 "espresso flush assemblies" (with the "slow down the water box"
- fully wired up 10 of our logic boards, with all probes, valves, sensors and connectors. Wrote a C++ firmware program to check everything. Fully tested 10 boards in this way. Documented the wiring and test process.
- received 300 coated-paper guides to help you put the tablet stand on centered, at the position you want
- wrote a manual on assembling the steam wand
- a manual on how to assemble our two mixing/sensor chambers.
- built 10 water-update assemblies
- assembled and quality-controlled 100 steam wands with a real water pressure test
- built 10 group heads
- second revision of wiring up the power switch, relay and PC boards
- a jig was successfully made and documented, to speed up soldering together our water heater insulation boxes
- 10 pump assemblies assembled (still need to write a manual for this)
- a jig for placing the under-the-case magnet arrived, and it was modified by a few millimeters as the magnet position was not perfect.
- an 8mm hole was hand-drilled into 10 chassis, to comply with Intertek's drainage-hole safety requirement.
- a color coding scheme for all the temperature sensors was decided upon
- 20 heater boxes almost finished (enough for 10 machines)
- detailed instructions for inserted a vital flow restrictor
To do this week:
- build the main mixing chambers, and write a manual about how to do this
- decide how we want to make the high voltage cables to the water heaters and group heads. We've made these by hand in the past and now need to really formalize the process.
- assemble one machine, see if it actually works (!) and then build the other 9 machines.
Still to do before we ship to customers:
- ship 6 machines to the "insiders" early beta users. These are my advisors for the past 3 years, who I first met when we were thinking about taking on the ZPM espresso machine project. They'll honestly tell me if we messed up somewhere.
- finish programming/testing some firmware/functionality. Firmware updating is coded, but not torture tested. Group head flushing and a cleaning cycle need to be coded. Final calibration (of flow and pressure sensors) numbers need to be moved to bluetooth and not hard-coded.
- get a final list from Intertek of their safety concerns regarding our construction. They gave us their report last week, but did not opine on our water heaters or group head, because they hadn't disassembled them. We sent them disassembled parts this week for their review.
- the Chinese New Year holiday season is coming up in February, with Mainland China closed for two weeks, and Hong Kong closed for a week. We're trying to get our machines to our beta testers before then so we can hit the ground running upon the end of the short holiday.
(a) hunt down crimps for larger gauge wire so we can more comfortably double-wire (this looks like what Nissan is doing on those Leaf car cables)
(b) find over-the-spade insulator sleeves for the smaller spades.
Thanks a million for taking the time to give your expert advice, and it comes just when I needed it.
We've sent Intertek all the supplemental documentation and disassembled parts that asked for, and we're now waiting for the final results of their "construction review". Once received, we'll make two new espresso machines with the changes they requested, and then will move toward full review.
The "prerelease" buyers of our espresso machine will receive machines with all the changes Intertek has thus requested, but the machine will not yet be certified.
In other news, Intertek is still discussing internally whether our Bluetooth tablet is an external remote control or more like a short wire. The difference in interpretation is crucial to our moving forward.
However, in order not to further delay shipping, I have hand made a "tether" in the photo above, and proposed it to Intertek as a solution if they are adamant about the tablet being a remote. I believe that this tether should solve the problem, and I'm discussing it with them this week.
The end of this tether would have a torx screw, because Intertek considers flat-head and phillips-head screws to be "user removable" whereas removing a torx screw requires specialized tools. The other end is secured with a cable tie. If the user removed this cable, the machine would no longer be considered UL certified.
This is a question for you guys: how offensive do you find this solution?
Here is a more professional version of the same idea, sold for stores that have tablets and other devices on display.
screen 2018-01-29 at 3.46.12 PM.jpg
I do think it would be better to tether the tether at the back of the machine if possible, I really would prefer not to see that on the side of the machine ruining the nice clean lines, but ......
I must say I do struggle, with them considering the tablet as a remote control ....it is the primary control, not a secondary remote control.
I do fear by giving Intertek the option of a tether, they may just see it as an easy out for them.
It is odd they wouldn't accept the good pragmatic solution as you suggested of "We could even decrease the Bluetooth signal strength, if needed, to limit the range to 6 feet (2 meters)" that is clearly a good solution.
Thanks for your tenacity
I think it looks like crap to be honest, like its going to run off on its own and elope with the grinder if it's not tethered. If it's a last resort and the only way to get the machines out then I agree a more discreet fixing on the back.
And plus one on the tenacity!
John, that solution in the first image would drive me nuts. I'm not sure what you are saying about the "more professional version" - is that an option you are considering instead of the first photo?
I too would be concerned about giving Intertek an easy way out with that offering, and truly hope the bluetooth option is ultimately approved. Although I have opted for one of the CE/UL approved DE1+'s, if all other rational requirements for CE/UL had been met, I would rather the cable was omitted from mine altogether, or was easily removed on delivery without defacing the espresso machine or tablet. My heart really sank when you revealed that Intertek were treating B'tooth as a remote.
Can you not run this tether (if it is needed) through the mount where the torx is hidden under the mount? ie so it is all invisible?
Or have some other mounting mechanism where once the tablet is clicked in, it needs a torx to be removed (if it cant be shipped in a mounted state for example).
Pardon my pseudo latin John, but noli illegitimi carborundum!
Like others have said, I don't like it and would certainly recommend not putting it out there to give them an easy out. Keep it as a hip pocket solution, only to be revealed if you have no other way out. If it must be, it should connect on the back panel only.
Keep your powder dry!
Last edited by HBDecentRefugee; 30th January 2018 at 04:06 PM. Reason: spelling
Many thanks for all the feedback about the tether idea. I'm still waiting to hear back from Intertek as to whether we'll need to go that way.
My current understanding is that if we do need a tether, we will be able to supply it as a "some assembly required" step for the customer to do and an instruction sheet.
I've ordered a sample of a very nice "anti-theft cord" that plugs into the audio jack of the tablet, which would probably be my choice. It looks identical to what you find in an Apple Store to prevent theft. The other end is a small metal disc with 3M glue pre-applied behind some wax paper.
Some users, such as cafés, might be especially interested in the anti-theft tether so that nobody walks off with their DE1 tablet.
For the past 3 years, we've mostly lived in one 3000 sq ft space, where all our espresso machine R&D has taken place.
Now that we're moving into proper manufacturing, we're taking away two work tables to move them into the newly-rented factory space. The reclaimed space will now house ceramics, touch tablets and give us more warehouse space. It's amazing how much space the "components" that go into an espresso machine take, because of all the protective packaging around each part.
The custom-made cardboard shipping boxes arrived today.
The knockbox that Joao and I designed a year ago, is now finally shipping. Boxes are queued up today from the preorders, and we have another 370 in stock. https://decentespresso.com/knockbox
As I've documented here, there have been a number of manufacturing challenges because of the unusual shape and my choice to make it out of 3mm thick aluminum rather than molded plastic. It holds a fairly large capacity (about 25 pucks) because we make a lot of coffee in the office and nobody likes to empty the knockbox "compost" out.
Because the knockbox is black powder coated over a large surface, about half of our stock is cosmetically perfect. I've decided that if there is 1 imperfection, I'll discount it 30%. 2 or more imperfections means 50% off.
I've attached two photos showing the kind of imperfections we're seeing. Typically it's a scratch or a variation on the black color (a "blotch"). Sometimes there are some scratches on the knockbar because the hole was slightly too tight.
I know that some people advise only selling "absolutely perfect" products for "branding reasons", but (a) I dislike the waste and (b) some folks are happy for the discount and plan to bang the hell out of our products anyway, so they'll soon be imperfect in real use.
If you like buying things on sale, look here:
I've been wondering how the bluetooth scale development is coming along too. I've been hoping it might be ready by the time my certified DE1+ is ready for delivery so I can get them together.
but our Decent Scale is on an indefinite "6 months away", as we won't devote significant focus to it until espresso machines are regularly rolling out of here.