Decent Espresso news - Page 8

Read about current and future offerings from the site sponsors.
User avatar
decent_espresso
Sponsor

#71: Post by decent_espresso »

It's not easy to speed up manufacturing, yet keep quality high

In the past two weeks we've configured the factory, added 30% more staff, with the intent of going from ~90 espresso machines made per month, to a goal of 120 per month, and then eventually 160 per month. Our https://decentespresso.com/queue started to come down briefly, and then September starts, and this happens:



I had actually hoped that July's peak sales was an anomaly, so that my near-term goal of 120 machines/month would see us reducing the backlog.

In this saturday's zoom, I was asked when the Decent factory was going get faster. The last time we quickly sped up production, we started to see quality problems right away. This time, we've introduced new processes, such as Fabrice, one of my senior R&D engineers, moving to the factory and doing a full review of every machine, for about ten minutes. He pulls every wire, checks every little thing. That's kept our quality up, but it's actually slowed things down in the short term. We've added 50% more staff in the past 4 weeks, and that'll help, but it's not immediate.

Parts shortages are a constant problem for us, as we've moved off a number of our previously-trusted suppliers, because their quality got shoddy, to a smaller number of good (but more expensive) suppliers, and those few are overloaded now with our big orders. We managed to get 10% of the v1.4 parts delivered last week, so we can start on that next week, with the hope that our new main metal supplier (who we love working with) can keep up with regular deliveries of v1.4 parts slightly faster than we need them.

I got bad news last week from our Android tablet vendor, as I'm extremely picky about screen quality, and specify brightness and color balance with explicit measurements they have to pass. The screen they ordered failed. They need another 45 days to deliver our next round of Android tablets, as they need to place a new order with a different LED screen manufacturer. The good news is that I've been working with their Android build manager, and we now have a really slimmed down, custom version of Android, that requires virtually no customizing to be ready for use. We're going to build White DE1XL models for a few weeks, as we have plenty of stock of white Android tablets, but only about 80 black tablets.

I give a much more extensive answer to what's changing at Decent in terms of making machines, in this video:

User avatar
decent_espresso
Sponsor

#72: Post by decent_espresso »

The Impact of COVID on Manufacturing

I was asked yesterday what the impact was of COVID, on manufacturing.

As it happens, I had just received news that the black powder-coating of our almost-finished knockboxes, had been pushed out to the end of September, and was having that very conversation with the company that makes our knockbox for us. I'm attaching below the two comments she made.

From my perspective, what I've seen is previously-reliable suppliers, now turning out low quality work. And, they're skipping steps.

This past weekend we had to reject 100% of a batch of 2000pcs of brushed steel front panels, because they had laser etched our logo at too low power, and it looked terrible. Most suppliers I've worked with make 1 sample with the logo, and then sends me a photo via chat for approval. Or, also common, is for no logo to be etched until after we approve the work in person, and then they put the logo on while my staff member is witnessing it. This is because they're legally allowed to sell product we reject, as long as it doesn't have a company logo on it. ps: I don't mind that law, as I think it's eco-friendly.

So.... low quality is the first thing we're seeing. Secondly, everything is taking twice to three times as long as as they estimated. A 6 week delivery date turns into 12 to 18 weeks. Partially, this is because suppliers were shut down for 6 months, and now there's a huge rush to get things done before Christmas. And also it's because so many suppliers fired staff during the Covid shutdown times, and those people have gone and found other work. Suppliers have tons of new employees, and they're not as good. So, quality and speed suffers.

A glimmer of hope? Yes: a few small companies were well run, and did not fire people during the covid times, are hungry for work and want to take on more. But they can only scale up so fast (just like us) and maintain quality.

So... making stuff is a bit of a challenge at the moment, if you want to make good stuff.

Below is a photo of the chat with my supplier from yesterday.

-john


User avatar
decent_espresso
Sponsor

#73: Post by decent_espresso »

Initial DE1CAFE steam research results

We've built ten prototype DE1CAFE machines and I can now report our initial results from making steam with them using newly crafted firmware.
  • I'm able to make a 200ml (150ml/6oz foamed to 200ml/8oz) latte from 8ºC to 65ºC in 17 seconds. Two drinks at once (300ml foamed to 400ml) took 31 seconds.
  • These DE1CAFE models require a 230V line, and have 2x our-design 2200W heaters in them, both with 2 meter long water tubes inside. There will be no 110V DE1CAFE models. However, this model will work on typical 220V/240V home power plugs worldwide. Commercial electricity is not needed.
  • By picking the right steam wand tip to create 5 bar of pressure, I'm able to raise the boiling point to 152ºC, which (when combined with the 4 meters of the length) allows much greater heat transfer.
  • Once the water flashes to steam, it expands greatly, moving out of the tube quickly, and it's not really possible to heat it up more.
  • Delaying the boiling point is the key insight to greater that transfer, and thus better/hotter/drier steam.
  • This has sped up our steam by 35% from our previous best (not released publicly) efforts, and 50% faster than currently shipping DE1 models.
  • This model will become available next year, and will be priced at USD$1000 more than the DE1XL model, and will be in the same chassis. This model requires a 220V line, and will work in the USA if you have your electrician install that for you, or worldwide on a normal 220V/240V power plug.
  • While home users are welcome to buy this machine, the greater power is really meant for commercial use, where time spent steaming is often the limiting factor in quickly getting drinks to people.
I'm using a prototype steam wand tip made for me by Zubing (who makes the Force Tamper) with three 0.75ml holes in it.
  • Note that the pressure curve below is our intentional "steam profiling" curve, to give you time at lower power to create microfoam with less powerful steam, then plunge the steam wand tip in and get maximum speed heating of the milk.
  • The "steam profiling" also means that steaming larger amounts of milk is faster than steaming small amounts, as the steam energy peaks and then stabilizes at around 50 seconds.

User avatar
decent_espresso
Sponsor

#74: Post by decent_espresso »

Decent Espresso white DE1XL models are currently in stock

I have about thirty of the white v1.3 DE1XL model that I can ship right away. Why? Because we're running low on black Android tablets.

The new batch of black Android tablets have been delayed to a shortage of high quality LCD panels. Well... the tablets would have delivered in time, but the sample I received for approval had unacceptable color balance, so I rejected the entire run, and the supplier had to find another screen supplier. I also think that these delays are there to push us to accept lower-quality product, since it's painful to go out of stock of a key component, like the the tablet. But... we have plenty of white tablets, so that's what we're doing, and I'm refusing the accept lower quality components, even though it's painful.



The good news is that we have plenty of white tablets with great LCD screens, and so I decided to make our white model for a few weeks. New Android black tablets should be arriving in about 4 weeks.

We have 34 existing orders for the v1.4 model in white, which we'll have fully satisfied in about 10 days. We should be able to make 100 more white DE1XL just as the black tablets arrive, so that means we'll be fully in stock of white DE1XL. So... if want a Decent FAST and you order soon https://decentespresso.com/model?de1xlwhite there's a good chance you'll have a Decent Espresso machine by the end of September.

I also have six of the v1.3 white DE1XL in stock, with perfect mirror panels, and they're USD$200 cheaper than the v1.4 model. The only difference is the larger DC power supply, which will enable more frequent background refilling, if you're using our refill kit. Otherwise, they're identical to the v1.4 models. Here's the order link if you want one: https://decentespresso.com/c?s=219+1+220+1

These last few white v1.3 DE1XL (and the early batch of lucky v1.4 white pre-orders) have fully mirrored panels, which were the few perfect ones we received from the 10 different suppliers we bought samples from. Each sample cost us about $100 each, and then we discarded 80% of them as too imperfect, so you can sort of understand why we're not offering mirror panels for now. Once these white panels are used up, all our DE1PRO/DE1XL machines will sport a brushed panel, with a mirrored lip.

I also have three v1.1 white DE1XL models, brand new, all with fully mirrored panels that are slightly imperfect, sold at a 21% discount. Next week, there will be v1.3/v1.4 white and black DE1XL models going for sale on the "Cheap Decent" forum. Direct-Message me if you'd like to be on the "Cheap Decent" forum.

-john


User avatar
decent_espresso
Sponsor

#75: Post by decent_espresso »



50 at once....

6 months ago, we were shipping 50 machines a month, now we need to be making 200 machines a month, if we're going to keep up with demand and shorten the queue https://decentespresso.com/queue

Last week, we made 20 v1.4 machines in 3 days, and then another 20 machines in 4 days. There were 4 small teams responsible for groups of 5 machines each, and I noticed a bit of competition among the groups, that I wasn't sure was entirely healthy. Fabrice and I have been wanting to move to a large-batch, specialist approach, for a while, and so...

This past monday, we did that.
  • we're now building 50 machines at once.
  • each sub-assembly has a "Master" assigned.
  • The Master is responsible for building the assembly, testing, installation, and feedback to engineering.
  • This also means that if something goes wrong, we know who to talk to.
  • It does mean for a less interesting job for each person (less variety) but it should yield higher quality and speed.
  • For now, we're going to aim for 8 days to build 50 machines, which would give us 150 machines per month.
  • That would still be a massive speedup over our previous best of 111 machines in a month.
  • And with time, we'll try to shorten the 50 machine cycle to a 6 day week, and eventually to a 5 day week, so that my staff can get their saturdays back.
As it happens, with this first batch of 20 v1.4 machines, Desmond, the "Master" of testing, though that some of the machines were not regulating temperature at pressures, as well as he thought it should (off by 1.5ºC). He's been doing all the testing/calibrating for a month (when he was hired), so he's become adept at this now. He brought the issue to me, for commentary and validation of his hunch.

Last night, a master repairman (SK) and new hire Keith worked late, moving single parts from one "good" machine to one "bad" machine until the problem went away. We're pretty sure that we've isolated the problem of two different value flow constrictors being installed backwards, which only creates a problem at low pressures, when the precise flow constrictor resistance values matter the most. We're testing that right now, and if so, just need to swap two flow constrictors to fix the 20. And then develop a process to make sure this doesn't happen again, as the flow constrictors are so small that you can't tell which is which, so this is an easy problem to cause.

Meanwhile, we're at step #6 out of 20 steps today, in our assembly of 50 machines. They'll hopefully be finished next week, but it might take a few days longer, as this is our first time changing the process so much.

-john

User avatar
decent_espresso
Sponsor

#76: Post by decent_espresso »




Which is the real 3M?

Decent Engineer Ben Champion noticed today that the tablet on his prototype DE1CAFE wasn't sticking all that well to his tablet stand.

He investigated, and found that the color, material, and stickiness of this "premium product" (3M makes good stuff) does not seem right.

We suspect that we've just been handed a batch of counterfeit 3M adhesive tissue. We're contacting our supplier, who is the same we've used for years.

We're also having 3M officials urgently come to our factory, as they've visited us before, and they've been exceedingly helpful. We're paying for the real stuff, so we want to receive it, thank you very much. And I'm sure 3M doesn't want their reputation tarnished.

If you received a DE1XL from us recently, it's possible that you have this inferior 3M adhesive on the back of your tablet. Naturally, if you have any problems, we will send you a free replacement, just as soon as we sort out with 3M, what's going on.

This might also temporarily delay the shipping of our new v1.4 machines, for a few days, as all v1.4 machines from us now use the same tablet stand as the DE1XL model does.

-john

User avatar
decent_espresso
Sponsor

#77: Post by decent_espresso »

We're just about to ship our new model v1.4 Decent Espresso machine. A little peek inside....

User avatar
decent_espresso
Sponsor

#78: Post by decent_espresso »

Just published yesterday a video showing how to unbox and setup the DE1XL.
Saturday night, I was on chat with new customer Edmond in Singapore, because he new DE1XL wasn't getting past the startup tests, and his pump was making an awful noise. Just as if it had no water. I hadn't yet uploaded this video, so Edmond had set the machine up without any instruction. Well, after 30 minutes of trouble shooting, it occured to me that he wasn't taking the back panel off and lifting the water intake tube out of the way. So... yeah, the machine really was "sucking air" and had no water.

I wish I could have finished this 24 hours earlier, and I could have saved Edmond and I a bit of stress! Hopefully this video will also help others from making this simple (but important) mistake, and they can get making espresso a bit faster.

-john
★ Helpful