Decent Espresso Machines (DE1) - Any thoughts?

[JMcCee]

That Christmas tree's looking a bit scrawny..

[>decentespresso]

Heater Stand-off



Intertek has asked for detailed information on exactly how our hot water heaters are built as part of their UL compliance review. Unfortunately, our manufacturer considers this information confidential and will only share some of the details. A bit of a stand-off now exists between the safety compliance reviewer and a desire to keep information secret by the hot water heater manufacturer so that they're not cloned. Both sides have a reasonable position, and we're negotiating in the middle to reach a compromise.

To move things forward, we sliced through one of our heaters in order to see what's really inside, as much as we can. We know that the water tubing is 304 stainless, and the insulating material around the heater coil is magnesium powder. We don't yet know what the heating element is (but it's probably nickel-chrome) and we're pressing to be confirmed that's the case, and at what diameter (though we can dig it out and measure it).

[JMcCee]

It sounds like Intertek don't want you to build the machine.

[>decentespresso]

Mechanical Engineering Student Interns Tommy and Charlotte are today using our new tube-cutting machine to precisely cut lengths of silicone water tubing. This tubing is used to bring water into the flow meter and pumps, and at that point, there is only low pressure, which is why silicone tubing is used.

We just recently decided--down to the millimeter--how long these tubes should be, which is why they're cutting them now, in quantity, for 300 espresso machines.

The tube lengths aren't something you can simulate in software, you have to actually build a machine, try different lengths, and decide what's best. That's why it's being done now, just before building machines in quantity.

The students measured the tubes and found a 2mm error margin on a 200mm long tube: about a 1% tolerance. That's pretty accurate.

[bigdaddy]

Hi John,

A couple of questions about your grinder...

When will it be available to buy? and

In one of your videos you mentioned it has conical burrs...I'm just wondering...Does this machine have conical burrs or flat burrs?

Cheers.

[>decentespresso]

A couple of questions about your grinder... When will it be available to buy? and In one of your videos you mentioned it has conical burrs...I'm just wondering...Does this machine have conical burrs or flat burrs?

I assume you're referring to https://decentespresso.com/pro_grinder which is in this photo below.

These have been available for pre-order for a while, and I have a pallet of 50 grinders in stock. However, these grinders are then hand modified by us to have a X/Y/Z adjustable portafilter stand, as my focus with this grinder is getting the mound centered and also weighed. The scale stands were the last design for us to really nail down. In the photo, if you look carefully, you'll see that the left hand side grinder is missing the thumb-screws that the right hand model has. We previously had height adjustment holes at set intervals, but I decided to move to thumb screws and continuous adjustment.

The sheet metal company suddenly wants an additional USD$680 to deliver these, which wasn't agreed to. I suspect this is a "get it before chinese new year" negotiating tactic (sigh) so we might have to wait until they come back from their 2 week holiday to start assembling our grinders.


BURRS

I also wanted to sort out an upgrade (and replacement) path for the burrs. They are 64mm flat burrs, as the web page mentions. I need to totally redo the video as soon as we start shipping the grinders, which will be immediately after Chinese new year (about 2 weeks). You are correct that in the video I incorrectly state that they're conical.

We've been working Hansung@SSP burrs https://www.instagram.com/SSP_grinding/ to use his burrs as an upgrade/replacement burr. The burr upgrade effectively increases the quality of the grinder from a Mazzer Super-Jolly level to a Mahlkonig K30 level, and this is guaranteed to be the case by SSP. The SSP burrs are much admired also by Socratic https://www.instagram.com/Socraticcoffee/ and their testing has yielded some nice results.

We're also going to be selling the SSP burrs to people who want to upgrade their existing grinders, as I'm not sure there are any sources besides SSP in Korea at the moment, for them.

On Home Barista they've been talking about the SSP burr upgrade recently:
https://www.home-barista.com/grinder...od-t48458.html
with what appears to be good results.

[>decentespresso]

Calibrating to Reality



Our "first production machine" is now up and running, and we're working through the final problems before making a lot more identical machines.

Here's what we've found so far:

FLOW METER DIDN'T WORK
It turns out that which-wire-does-what changed between model numbers and so our flow meter cables need to have their pins re-arranged. This takes about 30s per cable, and Jennifer can be seen doing this on the top left.


CALIBRATING PRESSURE
Photos: left middle, top right.
Next, we got to test the calibration steps that I've written for the Android tablet. Why do we need to do this? Because: the specifications a manufacturer gives us for a part aren't necessarily what you actually get. The pressure sensor comes with a "magic number" to convert resistance to pressure, but physical variation means the "magic number" will be slightly off. To test the pressure, we use a Scace II portafilter, which is a $600 piece of test equipment that independent reports pressure and temperature. We then change the "magic number" in our pressure sensor so that our readings now agree with the Scace. 91PSI=6.27 bar, so we were 4.5% off until calibrating.

Incidentally, on the middle-left photo, you can see the cooling effect of the unheated stainless steel mass of the Scace 2. Our temperature probe is reading the effect of this cooling, and our water mixing technology tries to compensate for it by putting in hotter water. In this case, a -2.5ºC cooling effect was over-compensated for between seconds 10s->25s. Running a hot water shot or two, to preheat the Scace, would remove this temperature fluctuation at the puck.


CALIBRATING FLOW
Photos: bottom right and left.
Using the Bluetooth scale to measure flow rate into the cup, we can obtain an accurate flow measurement, based on weight rather than moving water. At lower speeds to water, the flow meter readings get quite noisy (see on the oscilloscope) and a bit of tweaking on the noise filter was needed to make this wor better. Here, the calibration was needed to correct a 4% to 8% error.

[bigdaddy]

I've just cut and pasted post 966 and put it into the decent espresso grinder thread and added a post so not to clog this thread up with grinder questions.

Cheers.

[>decentespresso]

Down to the wire



We are color coding all our wires and sensors in 11 different colors, using a colored fiber sleeve.

To make the colored sleeve, we've been using our automatic-tube-cutting machine, and then cleaning up the splayed mess (bring right tubes next to Josephine's left elbow) with a hot air gun (bottom left photo).

The resulting tube no longer splays (blue tube in the bottom right photo) but it looks sloppy and takes up precious space next to other connectors on the PC boards. I've not been happy with this, and have looked to improve it.

This week, a temperature controllable "hot wire knife" arrived, and the results (grey tube, bottom right photo) are so much better.

You can see Josephine in the photo redoing the already-cut tubes so that they have a clean, melted cut. I'm not super-happy with the safety aspects of this process, so we'll be soon making a plastic guard with a "cut slot" in it, to prevent an accident.

And speaking of accidents, many thanks to Jeff W on Coffee Forums UK for bringing to my attention the risks of loose hair in the shop, which he noticed in one of my recent photos. A humorous "hair on fire" poster now graces the shop wall, and we'll order a shop helmet after Chinese New Year.

[>decentespresso]

Getting organized



Our espresso machine factory is slowly starting to look a bit civilized, as we prepare to ramp up production (and hire more people) in about 5 weeks from now.

[JMcCee]

Thanks for the update. How's it going with Intertek?

[bigdaddy]

I'm just thinking a bit into the future and if I was to buy. With all the new electronic sensors and technology on this machine (which I think is fantastic)... Will it need any type of regular calibration and continual/extra maintenance to keep it accurate in the future?

I can see one thing which would be no big deal...I can see the need for regular updates to be installed on the tablet, but I was thinking more on the hardware side of things...I don't know, but maybe there will be a requirement for yearly serving where all these things may or may not be adjusted or maybe there is no need or maybe there will be a need to do it at home?

Cheers.

[>decentespresso]

Thanks for the update. How's it going with Intertek?

WhatsApp message from the big boss (a few days ago): "we haven't forgotten you. We're having an internal conversation and need a few more days".

Separately, as I mentioned a few days ago, the examiner wanted specs on our heaters. By cutting them apart, and measuring various parts (resistance can be measured) as well as a few yes/no questions the manufacturer was willing to answer (until they decided we'd asked too much, got nervous, and stopped answering), we were able to make quite nice drawings with most specifications of both the group head cartridge heater, and the water heater. That's now back with Intertek for review. No new "fails" yet.

I'm just thinking a bit into the future and if I was to buy. With all the new electronic sensors and technology on this machine (which I think is fantastic)... Will it need any type of regular calibration and continual/extra maintenance to keep it accurate in the future?

For the flow meter and pressure sensor, our understanding is that mechanical variation at manufacturing is the issue, though at high altitude the pressure sensor will be incorrect at zero bar. For the temperature sensors, they might drift with time (as many thermometers do).

It's really a question you have to ask yourself:
--> if the device consistently gives the same result, but is off by 5% to the official measurement, does that matter to me?

In other words if your coffee tastes best at what you believe to be 8.2 bar, but actually you're at 8.6 bar, do you care?

If you do care, then you'll want to have pressure, temperature and flow recalibrated.

For flow rate, you would buy our $99 bluetooth scale, and make sure you have a DE1+ or better model.

For temperature and pressure, you would either buy a Scace 2 portafilter, or (when we ship it) our alternative to the Scace.

In both cases, the protocol is very simple:
1) run a 6 bar shot (no puck), write down what pressure the Scace sees it at
2) run a 2.5 ml/s flow profile shot (no puck), and write down what the bluetooth "flow rate into the cup" as measured by weight.

There will be a page on the tablet software where you can enter the test you performed, the goal and the measured value.

I can see one thing which would be no big deal...I can see the need for regular updates to be installed on the tablet, but I was thinking more on the hardware side of things...I don't know, but maybe there will be a requirement for yearly serving where all these things may or may not be adjusted or maybe there is no need or maybe there will be a need to do it at home

If you have your DE1 serviced for whatever reason, part of the process will be a no-cost recalibrating of your machine.

As far as regular updates to the tablet software, that is the plan, but I don't see the *need* for it. Most kitchen devices don't get better (more features, more accuracy) with time.

[gc]

WhatsApp message from the big boss (a few days ago): "we haven't forgotten you. We're having an internal conversation and need a few more days".

Separately, as I mentioned a few days ago, the examiner wanted specs on our heaters. By cutting them apart, and measuring various parts (resistance can be measured) as well as a few yes/no questions the manufacturer was willing to answer (until they decided we'd asked too much, got nervous, and stopped answering), we were able to make quite nice drawings with most specifications of both the group head cartridge heater, and the water heater. That's now back with Intertek for review. No new "fails" yet.


For the flow meter and pressure sensor, our understanding is that mechanical variation at manufacturing is the issue, though at high altitude the pressure sensor will be incorrect at zero bar. For the temperature sensors, they might drift with time (as many thermometers do).

It's really a question you have to ask yourself:
--> if the device consistently gives the same result, but is off by 5% to the official measurement, does that matter to me?

In other words if your coffee tastes best at what you believe to be 8.2 bar, but actually you're at 8.6 bar, do you care?

If you do care, then you'll want to have pressure, temperature and flow recalibrated.

For flow rate, you would buy our $99 bluetooth scale, and make sure you have a DE1+ or better model.

For temperature and pressure, you would either buy a Scace 2 portafilter, or (when we ship it) our alternative to the Scace.

In both cases, the protocol is very simple:
1) run a 6 bar shot (no puck), write down what pressure the Scace sees it at
2) run a 2.5 ml/s flow profile shot (no puck), and write down what the bluetooth "flow rate into the cup" as measured by weight.

There will be a page on the tablet software where you can enter the test you performed, the goal and the measured value.


If you have your DE1 serviced for whatever reason, part of the process will be a no-cost recalibrating of your machine.

As far as regular updates to the tablet software, that is the plan, but I don't see the *need* for it. Most kitchen devices don't get better (more features, more accuracy) with time.

John, one consequence of DEx machines' variation from spec will be the impact on one of the main future features - sharing recipes in the cloud. As you say, in the home environment, why should we care about absolute values if whatever settings we choose work for us. But that changes in the sharing environment. This will mean that sharing a recipe is only an approximation that has to be tweaked in unknown ways to replicate someone elses success. I know this is only a practical consequence of component variability and quality control issues which you've discussed before, but it brings home to me that these machines aren't 'reference machines' that are standardised despite the appearance in those graphs slithering across the screen of the tablet.

Although I will get the bluetooth scales I'm very unlikely to stretch to the expense of a Scace for occasional calibration.

Appreciate your updates very much, and at this stage, especially the updates on Intertek, which is obviously defining the timing of the production process and delivery dates going forward.

[bigdaddy]

Perhaps I'm also, or more to the point, asking if there are any expected corrosion or build up etc issues on the sensors or not. It's all fine in the kitchen if everything is equal then it doesn't matter where your datum point is or what it reads so long as it is consistent with the other readings but I'm asking if there is any forseen corrosion or build up issues that could vary the individual readings over time thus causing the need for calibration and maintenance from time to time.

If I could indulge and draw an analogy to cars and I know they are a completely different beast, but the sensors etc are checked regularly and dealt with if any issues occur..Just wondering if regular checking would be needed on the DEx models or not.

Cheers.