Sophi KravitzWelcome @Mathieu Stephan ! Let's start with an intro
hey everyone! so as you may have seen i'm an open source enthusiast who specializes in designing devices from the ground up. i've worked in many many different sectors, alternated between full time positions and small contracting jobs, to quickly gained xperience
Mathieu Stephan: I've been many times in China, have commercialized around a dozen devices for companies or myselfand I was a writer for hackaday for a year and a half! the mooltipass device started there.
should I start answering the google sheet questions? :)
Sophi: Tell us the story of how the mooltpass happened first!
https://hackaday.io/project/86-mooltipass-offline-password-keeper
HACKADAY
Mooltipass Offline Password Keeper
With time, logins and passwords have become critical elements we need to remember to access the different websites and services we use daily. If we want to achieve good security, each of these credential sets should be unique. We therefore created the Mooltipass, a physical password keeper that remembers and encrypts your credentials so you don't have to
he construction looks complicated with a card and an USB device, why not just an USB device?
he idea was to make something that anyone could work on, remotely. I had this idea in the back of my head to create an offline password keeper. Luckily no one else had this idea... I made a call for contributors and around 20 people started working on it.
he first 20 prototypes were therefore soldered by myself, shipped to the contributors so they could start playing with the first draft of the hardware
hen once we had settled on the final hw and GUI layout, we produced a small batch by someone in china, shipped it to selected testers
e had fw devs, sw devs, designers, advisors... lots of interesting contributors and testers!
@Frank Buss You'll find standard questions answered on our main website here: https://www.themooltipass.com/ but the general idea was to use several devices without carrying them around
, this makes sense, you need just the card, and a device at home, at work etc.
Mathieu Stephan: yup! you may also find all "developed on hackaday" articles here: http://hackaday.com/tag/developed-on-hackaday/
At the time we were publishing updates every month or so, showing hackaday readers our progress and asking for input... heck, I didn't even choose the name mooltipass.
If you read all the articles you'll see what you can typically discover when you start making a device... all the good and bad surprise
Frank Buss: yeah, i've done products for clients and open source, too, I know this :-
Mathieu Stephan: ... then you realize that you can never imagine what can go wrong. I think seeedstudio or similar will be more than happy to do it for you, provided you accept their quote and fund everything yourself. However you'll indeed need to show a functional prototype and I'm fairly sure they won't risk promising anything more than what you show your product can do. I also think they'll prefer that you deal with the certification or they'll simply forward you the certification center bill.
I am an embedded security guy. While building any hardware product what's your level of priority for security?
Benchoff: Another question? We have a lot on manufacturing in China.
certification, and WEEE in Europe, can be expensive
I guess you meant why. Why not? We believe that security through open source is something possible. Level priority of security is a tricky question: anything can be broken provided enough funds. We tried to have the best security/device cost compromise
For reference, the spreadsheet we're working off: https://docs.google.com/spreadsheets/d/1N9VGfSaJrl8cnaMnbAPJAvyg2Or9VmNJTHanQojRsGI/edit?usp=sharing
Mathieu Stephan: @Frank Buss extremely
Mathieu Stephan: I'm trying to find the bill from the fcc/ce center we used in shenzhen. fcc & ce testing for the mooltipass mini was around $2k5
Frank Buss: @Mathieu Stephan that's not too much, so you managed to pass the first EMC/ESD test?
Well, I'm pulling them from the spreadsheet, so knock yourself out either way
Mathieu Stephan: @Frank Buss a small modification was needed for the ESD test, but yes, we passed it on the first try
Frank Buss: nice
Mathieu Stephan: in details, the PCB GND was not connected to the aluminum case. The device would freeze when shocked!
Benchoff: Alright we did the ESD, so here's two."With respect to manufacturing in China, what concerns have you seen with regards to IP\trade secret? I know the topic is open hardware, but it can be discouraging to see a knock off as you're readying your product for launch." and "Related to the above, in the manufacturing world, we've seen counterfiting be an issue (FTDI chips in particular), outside of micromanaging, how can one make sure that authentic chips are being used?" So Chinese IP and counterfeits.
First one: The several factories I visited were actually quite careful about not showing some of their clients products! In the videos you'll find here: http://www.limpkin.fr/index.php?post/2014/07/31/Video-Series-Touring-Factories-in-Shenzhen I had to wait around half an hour to get permission to shoot.
All in all I'd say that as long as the company looks professional there isn't much to worry about.
Mathieu Stephan: Counterfeits
For the critical components, we actually purchase them from their manufacturers, program them, and have them delivered to our product assembler. However, as long as the assembler production volume isn't too big, it's not in his interest to switch some components as he'd lose possible future orders from us
Benchoff: Next one? "I want to make the step from "circuit board prototype" to "consumer-friendly device". How do you get a gadget into a nice case? (Industrial Designers? Are there standard-ish designs that are readily adaptable?)"
ceneblock: So for that two million production run for Kickstarter would probably not be best to let them source the components.
Mathieu Stephan @ceneblock: I wouldn't generalize but i'd definitely be careful if you have never worked with this assembler before
Mathieu Stephan: In my experience it is quite hard to find industrial designers who 1) understand the constraints that results from your component choices & placements 2) have a firm grasp on how their designs would be manufactured 3) know all the available manufacturing processes and their constraints. I'd therefore recommend documenting yourself about standard manufacturing techniques such as CNC machining, injection moulding, and having a go at freecad
Frank Buss: wasn't there some free Autodesk program?
Mathieu Stephan: Ideally case and PCB design should be done at the same time.... but that requires getting familiar with the different case manufacturing techniques... so it may be a lot to assimilate at once
Mathieu Stephan: @Frank Buss I think fusion 360, but am not sure
Mathieu Stephan: it is relatively new if i'm not mistaken
Fusion 360 is awesome for that :)
Mathieu Stephan: I've heard lots of good feedback on it
Maksim Surguy: You can also export DXF from your design and import into KiCad
especially for design to CNC machine toolpath
Maksim Surguy: so that your plastic part + PCB match perfectly
Brian McEvoy: TinkerCAD?
Mathieu Stephan: never heard of it
haydn jones: openScad
I like ViaCAD, but there are a many good 3D CAD programs
BUT HERE'S A GOOD ONE "What do you think prevent the open hardware movement from blowing up into public awareness ? Lack of money, support, publicity or something else?
ceneblock: nice
I think that "mainstream" products aren't open hardware because it's quite hard to get funding when all your sources are out there... hence there isn't a lot of project that can grab the public attention, but that may be a too simple explanation.
Investors / banks prefer funding projects that keep some of their IP private..
people often underestimate how much it takes to go the "extra mile', which will allow a given product to gain lots of customers
Frank Buss: Arduino is open hardware and has public awareness
Mathieu Stephan: ... and its success. which is why open hardware products are generally well known by tinkerers who want to modify/tweak them
@Frank Buss yep! hence my last sentence.
Martin Esser: I asked that question because i have some money and time available and wanted to start a hobby company that finds OS projects and tries to sell them. Do you think thats a possible model ?
Mathieu Stephan: but students / teachers who use them definitely help spread awareness!
Ytitne the Great: Most "OSHW Companies" appear to be service or software companies that just happen to sell hardware or exist only in the Open Source Hardware community. Have you been a part or seen any OSHW companies outside of this arena in competitive spaces succeed?
Mathieu Stephan: @Martin Esser i don't see why not! as i said, it's hard to get funding for open source _hardware_
Martin Esser: ok i will try thank you !
Mathieu Stephan: @Ytitne the Great not really no... but I haven't extensively traveled in the US... just europe and shenzhen area so I wouldn't pretend giving a good answer to that question
Mathieu Stephan: I'm quite sure @Benchoff could give a good answer to that question
Why do you think that is? That is, why don't OSHW companies in competitive industries succeed?
Well, OSHW is relatively new after all... a success isn't made in 2 or 3 years
and a related question: did you make some good profit with the mooltipass?
/rambling because people have conflated free with libre, you can give away software and sell it as a service, but you can't give away hardware and make money /rambling
@Frank Buss none. We use all the funds to fund new mass production batches
Benchoff: BUT HEY NEXT QUESTION
Benchoff: Where does one go to start looking into injection moulding, what are the costs, and how many prototypes can be expected?
:-(
Mathieu Stephan: @Frank Buss that's alright, I prefer using the money to spread the mooltipass word!
@Benchoff That's a tricky question. There are several youtube videos that explain the process so I'd recommend starting there. A typical mould in china for a device like the mooltipass mini will cost you around $2k... which will allow you to produce less than 10k devices. Unfortunately once the mould is made you can't make any more (major) changes to it. I'd recommend 3d printing your case, making sure you added the injection moulding angles to it and then contacting the manufacturer that will double check your design.
Ytitne the Great: Releasing detailed documents that make it easy for someone to replicate your hardware isn't new. I wonder why you feel the trend started 30 years ago.
Mathieu Stephan: @Ytitne the Great is it? I'm relatively young after all... but the few things I remember that you could make yourself would be crystal radios or similar
anyway, 3d printing allows lots of possibilities to reproduce a design now
Ytitne the Great: That is, the trend of not releasing information
right.
Mathieu Stephan: anyway... someone here asked "sorrry, was late getting in. I made a password vault, but it was not really intended as a product, just a cool thing to do. It made me want to get into wearables again, as I once made a wearable WiFi scanner based on the ESP8266. Now I want to build a better one. How would you advise monetizing a product buil from off the shelf parts rather than ground up?"
Mathieu Stephan: my answer: First, see if the parts you chose are not too overkill for the job (I guess with the ESP8266 the answer will be no). Then design the pcb together with the case... check your bom cost and see if you'd be competitive selling it.
it all depends if you want a nice looking device, which can usually only be small because you have designed it completely
Frank Buss: and if you have the BOM cost, PCB, case etc., multiply it by 3 for the end customer price :-)
Mathieu Stephan: there are many many projects on hackaday io that can be quickly made by soldering a few modules together. The project will be functional but may not be pretty / easy to carry / have a good autonomy... so you should ask yourself the question to whom you'd like to sell it to and how many devices you'd like your first production batch to be made of
3 is a good rule, 2.5x is the general one...
you can have a smaller number if you sell it directly through tindie
which takes around 5% in total if I'm not mistaken
SELL YOUR CRAP ON TINDIE EVERYONE
+1 for Tendie. They seem to love us and we love them.
SUPERLIMINAL ADVERTISING
kickstarter has more public awareness
maybe on tindie as well
oops i was getting into my wheelchair.
yah but tindie has a cute robot dog
ok, that's an argument :-)
@Frank Buss getting to kickstarter requires LOTS of work
what i meant was, can i like make kit project videos or something
like... an insane amount of work
i'm not someone who wants to design circuit boards for a living, but i will if i have to
Mathieu Stephan: @Les Hall of course, but keep in mind that we (as humans) are lazy... and busy
yeah, true, so i gotta make it easy for the customer to procure and assemble the parts
If it wasn't for Arduino folks like me wouldn't have known about or felt confident enough to learn hardware. It will definitely help folks become more aware, curious and keep learning more about hardware.
Les Hall: maybe offer them pre-programmed
but right, kickstarter needs a polished video etc., lots of work, Tindie is easier for small DIY projects
if anyone plans to go on KS... it takes at least 2 months of work... and you should have a community behind you before that
great word - community
Mathieu Stephan: as a general rule, each picture takes at least an hour
Frank Buss: this sums at 25 fps :-)
one minute of footage is 3 hours: 1 hour preparing the shot, 1 hour having the right shot, 1 hour editing
Mathieu Stephan: not counting the script that takes days to refine
we shot scenes that were not used because they simply didn't convey the point- a complete day of filming for nothing.
Thats... actually a good ratio.
.... and one should keep in mind that when you are 100% in a project, you completely forget how to sell it to "outside people"
Mathieu Stephan: ... that's when you start to harass your friends for feedback :
IIRC there are companies which does the Kickstarter campaign for you, when you have the product
yes... it won't be cheap... at all
asked for some offers, average price was around $7k
Frank : just like during the gold rush, those who profit are the ones selling the shovels :-)
right :-)
@Mathieu Stephan Hey Matieu
Mathieu Stephan: I'd advise shooting the scenes yourself, and asking a professional to do the final editing and CGI
@Bil Herd hey bil! long time no speak :)
Bil Herd: Yup, once again I am late...
Frank Buss: Bil knows all about realiable mass production
Bil Herd Hi Sophi!
HERE'S A GOOD ONE In regard to the ten devices you've commercialized; Did any require restrictive or costly PCB specs such as HDI or <8mil trace/space?
Bil Herd: Except Bil's brain has become unreliable... >:)
@Benchoff: luckily no! :)
Mathieu Stephan: Another question that was there: "Say I've built a working prototype around a raspberry pi. How does one take this and make it into a product? Reimplementing the rpi seems like a daunting challenge and using the rpi in a finished product seems wrong."
Mathieu Stephan: my answer: "It all depends on the price you want to sell it at and the number of units you want to make. Are you 100% sure it needs a rpi or could you make your product with a standard MCU? Including a rpi in your product is alright, but making a case around it may be not as pretty as a custom solution."
Bil Herd: Also can you get access to the component that represent someone's IP?
Mathieu Stephan: I doubt it
a second question: "What are some things to consider when dealing with different manufacturing vendors? For example why would someone use OSH over Foxconn (for example)"
My answer: It's all about manufacturing volumes. In my experience, after standard vetting, cost is usually inversely linear with the quantity of mistakes that your manufacturer can/will do. If you're starting with product manufacturing, I'd recommend someone local and a bit more expensive so you can learn what you need to take care of during your product assembly. Once you're familiar with all the mistakes an assembler may do, you can start getting your device manufactured abroad. This will however require you to think of everything that could go wrong..
fest: Is there a reason to avoid using Pi Compute module?
I've seen products with just a beagle bone in it, connected to a custom base board
you depend on that module... but as it's something that is likely here to stay... I wouldn't worry about it
you can't buy vertical SODIMM connectors.
ceneblock Thanks @Mathieu Stephan. Getting all my questions answered!
my pleasure @ceneblock , don't hesitate to contact me directly if you have others
Question: "How does one ensure Quality Control of the entire product when dealing with large volumes?"
My answer: For the Mooltipass Mini we're actually doing a quality control process twice: once at the device assembler's and once at the logistics center. The payment terms we enforce are 50% upfront and 50% once the device is assembled and tested. Moreover, we designed a functional test that prevents the device from booting if not all the hardware is correctly working.
Bil Herd: Did you supply testvectors?
I'm still along way from assembling anything, but it is very good to have this kind of background knowledge
testvectors?
test vectors.... sorry missing finger
the fw is self testing, and the device gets sent a few USB commands to check for reliability
I didn't make an ECU so luckily it doesn't need to be more complex than that :)
a new question: "How easy is it to get net terms in China? e.g. Net30 where you don't have to pay the CM until 30 day after delivery?"
Mathieu Stephan: in my experience... very hard... they'll always ask 50% upfront in any case.ften they'll tell you "yes we can do that" but once the funds are sent... well... things are different.ule #1 when asking questions with Chinese fabs: always use open questions, never yes/no questions
oh... an interesting question! "How easy is to handle production without phisically going to China?"
Are you going to make my money disappear? yes/no
@Bil Herd it'll go somewhere... ;)
@Mathieu Stephan Some examples of good questions that we should be asking? ;-)
so... handling productions without going there
it's doable... but tricky
Mathieu Stephan: my recommendation: documentation with very little text but LOTS of pictures
Wow. Considering Net terms make a massive impact on initial outlay and return. https://youtu.be/GJAZtfoVA28?t=528
if your device requires assembly, shoot LOTS of videos
Mathieu Stephan: obviously you won't be able to upload them on youtube... so youku will be your friend
(but is extremely slow)
Okay , we have any more questions?
using videos is convenient for you and your assembler, as he'll be able to watch it again and again over the course of the mass production
Do CMs accept responsibility about incorrectly assembled boards?
OH HERE WE GO: "How easy is to handle production without phsically going to China?"
@fest I am lucky to have chosen assembler that will produce the amount of board i have asked for
that usually means they have to produce a bit more
@Benchoff already answered just above ;
a new question: "Everyone is always talking about manufacturing in China. Is is still possible for a small company to sell some products (lets say, a few hundreds units) while manufacturing partly in europe (pcb assembly) and partly in house (3D printing) ?"
I don't see why not! It may be a little bit more expensive but is a very reasonable option for medium-small quantities
Mathieu Stephan: I forgot to mention that if you plan on producing more, it's better to keep the same assembler
as he'll have made all possible mistakes during that first run
Mathieu Stephan: sometimes they do take shortcuts that you don't know of
fest: Yeah, I know a lot of small EU companies who are using local assembly house for their low volume products
(hopefully)
"Do you have any opinions on using one contract manufacturer or multiple for things like bare PCB, PCB assembly, Final assembly, packaging, fulfillment, etc?"
thanks for the advice :)
t's _much_ better to have only one: in the case of the mooltipass mini we have one guy for the pcb & device assembly and one guy for the case machining
well, some cases were a bit too tight for our device assembler.... so we had to be between them to find a solution
it's doable, but takes time
unfortunately we couldn't have predicted it because the manufacturing process becomes different once you start producing thousands of device rather than just a hundred.obviously having only one guy will lead to a higher final price... as he usually sub contracts several things
So you did a pre-production run and a production run ?
so either you take the responsibility of having "all the pieces fit together" or you prefer paying a bit more to just the one guy
@Bil Herd we did 2 actually
we had 2 "big" orders before we launched on kickstarter. 3 big companies.
Mathieu Stephan: is why you see all these units at the end of the mooltipass mini kickstarter video
"Can you explain the process of hiring the contract software person required for the Mooltipass? Where they were found, rates, terms etc?" > This is by far the most tricky part... we found some of them through upwork, reddit or through friends of friends. Rates were around $50 an hour.... but we were quite disappointed. I'm therefore very familiar with most of the mooltipass codebase as a result.
...the software side that is, i know all the firmware very well
Expect to pay much more than that for quality firmware. :)
contributors are welcome by the way! especially JS / C++ devs :p
@Patrick Van Oosterwijck I'm too paranoid to hire someone to make our fw code :)
contributors on teh firmware were doing it for the love of open source
Frank Buss: some experience from producing in Europe: my Kerberos cartridge was produced by a German company, but only the PCB and soldering the SMD components, for 250 PCBs, because it was cheaper to solder the through hole connectors myself. My advice: don't do this, let the assembler do anything, it is more work than you expect, and possible if it is a medium or high priced product
hence the code quality was _much_ better
"Could you explain the typical shipping situation with a Chineese supplier if you're not there to pick up every order in person? FOB?" we could manage shipments remotely, that was actually never a problem
So they shipped bulk via air?
we chose all our manufacturers in shenzhen area :)
did they ship to the end customer? shipping from China seems to be incredibly cheap
@Frank Buss we found a logistics center in honk kong yes... so that's another person to interact with.
it is way cheaper than if i was to ship orders from switzerland
for 2 mooltipass mini, 3 times cheaper
and honestly... making the packages yourself takes time :/
So you give him a list of orders with qty and addresses and the center takes care of packaging and sending?
yup
I have _many_ python scripts
i have many monty python scripts
the one triggered daily fetches all the orders from tindie, generates an xlsx with all the data, generates all the pdfs for each order and sends an email to my fulfilment center
.... the fluff gets up your nose...
https://github.com/limpkin/tindieorderprintout
limpkin/tindieorderprintout
tindieorderprintout - A python script to automatically generate tindie orders printouts
this sound complicated, doesn't the fulfilment center have a digital interface?
... if anyone wants to reuse it
@Frank Buss I've heard they have.... but a chinese one that isn't easily interface with. An excel file is easy for them, as they just import it in their system
this may be a little off topic but does anyone know of good open source inventory management software that i could use with a cheap USB gun and my laptop?
ok, was wondering about the PDFs, because they are very much write-only for computers
but every logistics center has its own way of doing thing
@TheThriftstoreHacker I don't... but if you find one I'd definitely be interested
@Frank Buss these are for my accounting
and custom
Mathieu Stephan: I had no idea how long it was supposed to last
like an hour but whatever
And he has all his fingers
we're already at two whatevers but that's fine
the kickstarter csv parsing script was the most annoyign one.... I managed to find some errors on kickstarter side
@Bil Herd you need to build your Terminator finger prothese :-)
Mathieu Stephan: @Bil Herd aha :D
I did, it became sentiant, now it mistypes....
lol
@Frank Buss did someone here do that?
Les Hall: @Bil Herd you did the terminator hand?
Mathieu Stephan: Anyone else got any more questions? i think i've covered all the google sheet ones :)
Thank you @Mathieu Stephan ! :-)
you're welcome @Boian Mitov , thanks for the great questions
ceneblock: Yes and thank you very much @Mathieu Stephan!
thanks @ceneblock :)
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.
Odd. Three questions he answered are not here.
Some Bil Herd "look at me" nonsense.
...and what's this strawman doing here?
"Benchoff: /rambling because people have conflated free with libre, you can give away software and sell it as a service, but you can't give away hardware and make money /rambling"
Are you sure? yes | no