Making OSHW design a viable occupation: The OSHW product "Introducer"

[rhb]

For the creator of a complex OSHW product such as a DSO, the major barrier is getting from a working   prototype to a shipping product.  The engineering challenges alone are formidable.  Crowd funding can solve some of the basic economic challenges.  But the business aspects of forming a company are a huge distraction.  In addition, the lack of a track record for a new venture makes selling an expensive product such as a DSO difficult.

What is needed is a company which is set up to handle the business aspects of getting from a working prototype to a shipping crowd funded product, an "Introducer" with a proven track record of delivering.

By their nature OSHW products can only support the Creator and the Introducer for a brief period.  Once a design is successful the bulk of the profits go to the manufacturers and sellers.  The Creators  must make their money by creating and supporting products.  The Introducers by  introducing and providing manufacturing support in conjunction with the Creator.

This is actually how traditional businesses operate.  The design engineers create products which the manufacturing division produces for sale.  The engineering work is paid for by the manufacturing and sales divisions. Each division must make enough money to pay for staff and expenses.

OSHW splits the corporate monolith into design, introduction and manufacturing support, manufacturing and sales as separate business entities.

What is currently missing is a name brand introduction and manufacturing support company focused on OSHW products that bridges the gap between the creator and the manufacturer.  The Creator is freed to focus on the engineering  and the manufacturer gets brand recognition for their product.  While the design is OSHW and anyone can make it, only the manufacturers who pay the Introducer for manufacturing support can legally use the brand name.

Have Fun!
Reg

[OwO]

Usually the result of that is the developers make a pittance compared to what they could have made if they also manufactured and sold products themselves. I think what is needed instead is the infrastructure to make delivering a product doable in the first place. i.e. access to prototyping and manufacturing services. Decent logistics so that you can easily run an ecommerce operation.
Open source hardware in itself already means you make a pittance compared to what you could have made if you kept everything either secret or patent protected, so to make it work at all the developers really need to own the whole operation and take in all the profit, or have other sources of income. OwOComm exists only as a hobby operation, it's not a profitable scheme.

[EEEnthusiast]

I am wondering if money could be made by making the hardware open but keeping the application software, firmware or FPGA code as closed. The manufacturers pay the creator a fee for using the software/firmware etc. Is this a workable model?

I am planning to open source the hardware design files for my power profiler tool ZS-2102-A and see where it goes. It takes a lot of time and money to get things into production and sales. I guess keeping the software as closed may be a way for hardware designers to make some money and a living out of their designs.

[rhb]

It's a difficult conundrum.  Making the HW open but all the SW closed sort of defeats the purpose. 

I think making everything except FPGA bit streams open is probably a good compromise.  That gives users the ability to fix bugs and add features.

The model I suggested is based on the hit that Hugen has with his revisions of the nanoVNA.  He also sells consulting services to other manufacturers.

The fundamental problem is that putting a device on the market involves significant financial risk.  And if your interest is HW design, running a business is a distraction at best and a huge PITA at worst.

Have Fun!
Reg

[EEEnthusiast]

I agree with you fully. Running a hardware company is a lot of cost upfront and pain.
Especially in these tough covid times, the supply chain is stuck and there are lot of uncertainties. By making the hardware open, my assumption is that a lot of small and medium manufacturers may take it up and manufacture it. This can save a lot of costs in shipment across countries, customs duties etc since the product is made very close to the customers. The only hope for the creator is that some one is willing to pay for the FPGA code or for the PC software. May be someone in China who can build the hardware for much cheaper would bundle my software along with it and sell. Don't know if this model will fly...

[T3sl4co1l]

I mean, you could hire the services of a turnkey design house, and open source the design files they produce (given they don't use anything not licensed for such).  But if you don't have six digits USD kicking around, you'll have a bit of an issue doing that.

To put it another way, you're basically asking some firm to solve your problems for you, with no payment, only the promise of royalties.  On the assumption that your product is worth anything at all.  Which most of the time it isn't.  By some orders of magnitude.

Maybe it's not quite that bad.  But there's a lot of people out there, making a lot of junk, and one-offs.  If I'm that firm, I have to be very, very picky about what I choose, what looks promising to invest our time in.  And then we look mean because we're rejecting all these great ideas the community has.  Heh, true, there's a marketable way to turn that around: have the community, and social media, vote on prospective products -- let cool things percolate up, in the same way that any other recommended image/video/article does these days.  But we still have to sort it ourselves by practicality, because most of those "cool" things are going to be cute art projects with no actual production value, or bullshit scams that are physically impossible.  All of which we've seen play out, time and again, on Kickstarter and the like.

Which, really, is what we've basically become, with a different financial model.  I'd say they've done it better, as people put their money where their vote is, and who cares if anyone actually makes a thing?  (Uh, give or take how and when the funds are actually disbursed, I forget.  And there's other platforms with variations on that, so take your pick.)

Tim

[OwO]

Maybe it's not quite that bad.  But there's a lot of people out there, making a lot of junk, and one-offs.  If I'm that firm, I have to be very, very picky about what I choose, what looks promising to invest our time in. 

Very much true. We are a design contractor that will design open source hardware, but we are highly selective about what projects we take on even if paid. If unpaid, the project proposal is just a suggestion.

Really we are just doing our own thing, but we accept "project proposals" as a way to get other people's ideas. I might never do your projects, but I'll take elements from it into our own projects if I think you have a good idea. I do not believe ideas deserve any kind of protection, but implementations maybe. You might tell me to design X & Y & Z, but we'll actually design Z & W, either for another client or to further our own development. If that client actually takes it to production and starts selling it, we may get a small payment, the client will get most of the profit, and you will get nothing (sorry to be frank). Indeed, it's he who actually does the execution, brings the product to market, and bears the risks who deserve most of the profit. If all you did was put together a shitty prototype with arduinos and crap, and the "introducer" had to get that into shape for production, actually manufacture and sell it, and bear the risks if it doesn't sell, then honestly IMO you don't deserve any money.

[OwO]

If all you want to do is hardware design and not worry about production or sales, there are already avenues for that which is to either work at a company or be a design contractor. If you want designs open sourced, you can do that. Provided you are very good at what you do, you can make demands of your employer or clients, such as open sourcing designs or retaining some IP rights. If you think an average quality open source project is worth enough that the person's living expenses should be paid, go lobby your government for a universal basic income scheme or some funding scheme for open source work (like NLnet). Private businesses can not be expected to be philanthropic; their only goal is to maximize shareholder value which conflicts with open sourcing hardware.

[ogden]

Isn't SeedStudio exactly what OP is looking for? https://www.seeedstudio.com/

[JohnG]

Private businesses can not be expected to be philanthropic; their only goal is to maximize shareholder value which conflicts with open sourcing hardware.

This is generally a main goal, but not the only goal. For technology companies, if it is or has become the only goal, the death or at least bankruptcy of the company is coming, because there are much more efficient ways to maximize shareholder value. Over time, it tends to attract more and more parasites.

John

[OwO]

Well, the reality is that any product that sells well gets cloned very quickly if there is no copy protection. Brand name and trademark is not a good enough barrier, they will just sell under a different name and still steal most of your sales. I do think there are ways to make it work though, such as a wide portfolio of products that each individually don't sell enough to be worth cloning, but combined makes you a decent profit. I'm a big proponent of open hardware and am constantly researching new ways to make it work.

[floobydust]

OSHW the "Creator" earns nothing but some dopamine from the fame and glory, and their design will get ripped off by others, uh it is open-source right? So what's the point?

It's just that the creatives get all the energy sucked out of them. This is OK but guarantees the projects get orphaned afterwards, when they are exhausted and burnt out and broke. Nobody remains as the parent or lead and the crowd generally just rips off the design, changes one thing then loads up their blog as if they did all the work, to get some dopamine, or it gets copied and sold for cheap. This is what I see happening amongst the OSHW dreams and myths out there.

I don't think manufacturing/marketing load is the problem, it's just that OSHW doesn't work in a viable way. All I see are one-man bands making something, isolated and on their own and then showing the golden egg to the world. Then it ends there because most of us have jobs and catering to zillions of people on the Internet wanting features, bug fixes, instructions etc. is just an impossible load.

I have worked for a design house and they had (contract) royalties and/or manufacturing control for all designs. They would lock-in inventors for the IP and production. Can't say I liked that approach at all, imagine paying straight engineering hours to get something designed and you didn't own the IP and had to buy the boards/enclosures etc. exclusively from the design house afterwards, including draw-down commitments i.e. 100 boards/month.

[rhb]

If you work as a design engineer for a salary, you get paid for doing a design.  Then you get paid for the next design, not the last one. 

If someone can make salary for the next design off the last design, except for the initial investment in the first design it comes out the same. 

It's a difficult problem.

Reg

[Bassman59]

OSHW the "Creator" earns nothing but some dopamine from the fame and glory, and their design will get ripped off by others, uh it is open-source right? So what's the point?

<snip>

I don't think manufacturing/marketing load is the problem, it's just that OSHW doesn't work in a viable way. All I see are one-man bands making something, isolated and on their own and then showing the golden egg to the world. Then it ends there because most of us have jobs and catering to zillions of people on the Internet wanting features, bug fixes, instructions etc. is just an impossible load.

I agree.

What is the point of open-source hardware?

[rhb]

The Arduino, nanoVNA and $20 LCR/transistor tester are the poster children for OSHW.

Before those appeared that functionality was prohibitively expensive for most hobbyists. 

OEM dev kits used to be several hundred dollars.  And then a cohort of kids started EE classes and wanted to use an Arduino for everything.  Scared the hell out of the other vendors.  TI introduced the MSP430 LaunchPad for $4.30.  Now every vendor has dev kits for dirt cheap or free.

The cheapest VNA was the VNWA at $800.  The cheapest LCR and transistor test was Peak Atlas kit for around $300 for a DCA75 and LCR45.  Now a nanoVNA H4 is $60 and a BSIDE ESR02 Pro is $30 or less. Not quite as good, but at 10% the cost, good enough.

Have Fun!
Reg

[profdc9]

I have to say that the open source designs I have produced I do for several reasons, but these are not related to making a business or a profit:

1)  To educate myself in the technical expertise needed to design and build the device and bring it to a (hopefully) useful state for others. 

2)  To make working examples available that can be examined or modified to be learned from.  Because I expect others may try to reproduce the design, I strive to use parts that are widely available or have equivalent sources from many manufacturers (e.g. jellybean parts) rather than specific parts so that the design will be obsolete once the specific part I rely on reaches end-of-life.

3)  To have examples of my work available so that others may evaluate my qualifications as an engineer.

4)  To generally support my goals of promoting technological literacy and have a technologically informed public that can hold governments and corporations accountable to how technology is used on the public's behalf.

In this sense, I hope that my designs are copied by Chinese and other vendors, and hopefully improved with the improvements contributed back to my project.  But if I was going to start a business around selling devices based on my designs, I would not be able to depend on simply selling the hardware.  I would have to establish a brand around providing the services needed to learn from the experience of building or operating the hardware, kind of like how Heathkit or other hobby electronics manufacturers used to be.  Some Chinese companies have been able to establish good credibility with the hobbyist community, for example, I just bought a Creality Ender 3 Pro printer which is open-source and it is quite good and they seem to have a solid relationship with the maker community and are promoting education about DIY manufacturing.

Dan

[floobydust]

I see you did some great work there and I think it's very rare a project having that much scope makes it out in the wild, works well and has decent documentation. How many hours and dollars do you think you spent?

If I put in hundreds of hours sweat and then people rip off the design, change one part, call it their own, cheapen it so it works poorly - that would piss me off. I'm not sure if I think it's my baby but a project still needs a parent, someone looking after it. Instead, all I see are orphaned projects and repo's of OSHW projects years old. Tossing it
(baby) out into the wild... I can't get comfortable with the notion.

OSHW is largely a "one man band" production thus limited in the project's scope and complexity. The poster children are the initial work of one or a few people, as a one time effort. To go beyond this, be bigger, it needs MONEY. Licensing fees, certification and trademark fees- have to be paid somehow. OP even uses the word "occupation".

Raspberry Pi Foundation (charity) is the sole provider of the hardware, they make money off it. Contrast that with Arduino.
The Arduino lawsuits over their trademark, it was just a massive turd that set them back many years. Who pays for the legal fees, where's that revenue? Knock-off Arduino boards are very cheap 1/10 the cost of genuine boards.

2020 Report State of Open Hardware OSHdata
"The most valuable areas that we are interested in exploring further are revenues for certified products, company revenue, and overall OSH market size."

[rhb]

What I have tried to suggest is a business, run by technically savvy business people, which makes its money by getting OSHW designs from proof of concept to product.

The idea isn't "get rich".  It's make a normal salary as an independent designer of OSHW and comparable compensation as an "Introducer".  The designer has to have the drive to actually produce a working prototype on their own.  But they should be able to stick to the engineering and not have to learn to run a business on top of that.

I know a person who has designed a 100 MHz,  1 GSa/s  Zynq based DSO single handed and is grappling with how to bring this to market without having to become a businessman.  I am on record as saying this could not be done.  I was wrong.

It's by no means complete, but it currently acquires and displays data at Rigol DS1000Z waveform update rates.  It's another year's effort to complete the design.  But he has to support himself and you can only work 80 hrs a week for so long.  He wants it to be OSSW/OSHW but lacks the capital and no one seems to have  a clear idea of how to suppress piracy and make service data and user FW control available.

Reg

[robca]

It seems to me that this would have a massive bootstrap problem.

Nobody with business sense could create a business helping designers with no business sense unless there was a big pipeline of designers with ready design. And no technically proficient designer will spend time designing a full product without knowing that there is somebody to take it to market. You need both sides at once, fully ramped up, in order to be attractive to each other

Also, the "technically competent people with business sense" would need to have above average skills in order to be able to properly evaluate the technical and business feasibility for a variety of completely different projects, from oscilloscopes, to high end musical synthesisers, from audio headphone amplifiers to flight controllers for drones. Those people (if they even exist) can make well above average salaries, and they require businesses with enough turnaround and profit to justify their salaries. Butt by definition OSHW does not generate enough margins, otherwise there is too much room for cloners to use the open design and undercut the cost.

I evaluated a lot of startup business models and I advise private equity on acquisitions/technology investments (software, not hardware). The worst possible ones to invest money in, are the proposals with a "bootstrap problem", since you need to bring both parts to fruition at the same time, both working perfectly

So, nice idea, but anyone with enough business sense to help also has enough business sense to know it's hard to pull off 

[rhb]

It's *very* difficult to pull off.  And that's just on the basis of comparable remuneration.

It's also not viable for anyone other than the top few percent of designers.  But there are people who have the skills and the general social benefit in terms of increasing the pool of people with technical skills because they can afford decent T&M kit  I think makes it worth trying to find a way.

Have Fun!
Reg

[OwO]

The Arduino, nanoVNA and $20 LCR/transistor tester are the poster children for OSHW.

No they are not. I am involved in NanoVNA development and can tell you it's a cesspool of GPL violations (see [redacted manufacturer]), open designs getting ripped off and turned proprietary, and anticompetitive behaviour from the biggest vendors (fake orders traffic and false claims of selling original devices). The vendors doing the most amount of shady practices who turned others' open source designs proprietary are making the most income, while original developers and original vendors who first brought the NanoVNA V2 to market are making a pittance.

I'm afraid open source VNA development may be stalled for some years as all the actual hardware engineers get turned off from it, and you may never see an open source 6GHz design in the near future, all due to endless greed of a few. Tragedy of the commons.

[rhb]

They are successful because they work and are available at a price most people can afford.   The fact that some abuse the situation and don't play by the rules is an indictment of the people who do that, not the concept.

Banks are not responsible for causing bank robberies.

It really doesn't matter if there is ever a 6 GHz nanoVNA.  A lot of people have had an opportunity to learn about vector network analysis that would not have done that otherwise.  That is a net good to society as a whole.

If you do not own the tools of your trade because they are too costly, you are to some extent a salaried slave to those who own the tools.  Read "Max Wien, Mr. Hewlett and a Rainy Sunday Afternoon" by Jim Williams.  A proper home has a library and a lab.

In some places broken test gear is readily available cheaply.  In other places it's not.  Everything in my lab traveled thousands of miles to get to me and 25 year old Tek and HP gear is heavy and expensive to ship.

If you know the story of the development of the nanoVNA as well as you suggest, please write a history.  I'd be very interested to know who the players are besides hugen79 and edy555.  Keeping up with the groups.io list is impossible.

Have Fun!
Reg

[OwO]

Story is very simple:

- 2016 - edy555 (aka ttrftech, Takahashi Tomohiro) designs the NanoVNA from scratch, with his own schematic, PCB layout, and firmware. The schematic and firmware are public and GPL licensed. The PCB layout is unreleased. edy555 sold NanoVNA kits for a short while.

- 2019 - hugen uses edy555's schematic, makes small changes (adding a battery and enclosure), designs a PCB from it, and begins selling it. The hardware is virtually unchanged and edy555's firmware runs unmodified on it. He did not release PCB layouts. At this point I'd say edy555 still deserves at least 90% of the credit for the NanoVNA hardware design and firmware, and hugen deserves some credit for the layout design.

- 2019 - flyoob (deepelec, former employee of sysjoint) designs the NanoVNA-F, also based on edy555's schematic, and without releasing PCB layouts. It runs a modified version of edy555's firmware that has been ported to FreeRTOS.

- late 2019 - HCXQS and OwOComm forms a team of engineers that would design the NanoVNA V2 from scratch. The hardware is entirely a new design and has nothing to do with the original edy555 NanoVNA. The firmware is also half from scratch, with only the UI and menu system copied from edy555's NanoVNA firmware. The hardware design files are all released and we are also fully GPL compliant on the firmware side.


So far so good, but in 2020 is when it all turned to shit. [redacted manufacturer] would start only releasing firmware binaries with no corresponding source (his firmware is almost entirely derived from edy555's, so it's bound by the GPL license). [redacted manufacturer]'s associated companies clone my team's design, with zero changes to the PCB, and flood the market by selling at or below cost to distributors (remember, he has a LOT of capital built up from his NanoVNA sales, so he can afford to do that). As a result all Aliexpress stores that used to carry original NanoVNA V2 devices now sell clones. (Note: do not buy a NanoVNA V2 from Aliexpress. They have little to no QC and a high DOA rate.) [redacted] also uses fake order traffic on taobao to get his new listing to the #1 position and to look like the official listing.

The NanoVNA community (mainly on nanovna-users on groups.io) are kind of naive, they are completely oblivious to the GPL violations by manufacturers but they berated a volunteer contributor for forgetting to keep sources and binaries in sync, they look at my client's product being stolen as if it's nothing and even promote the clone, saying he "improved" the design when he made zero changes to the PCB, did not bother to figure out how to do proper QC, and played foul business practices to attempt to shut down my client.

With that said, I don't blame my client if they are completely turned off from open source hardware in the future. OwOComm is in no danger but I have very little motivation to do VNA projects at this point (I want to focus more on my SDR projects). Some of our team members may still be hired by our client to work on their future VNA projects, but there's no promise it will be open source since it's out of the control of OwOComm.

The current state is edy555 and all the other true NanoVNA developers make a pittance or even nothing (there are countless contributors to the open source firmware like DiSlord, they aren't getting any money out of this). I understand that most contributors aren't expecting to get paid for their work, but if you take their work and profit from it, the least you can do is make your contributions public too. Schematics and PCB layouts are not protected by copyright, so something like GPL doesn't have any legal force. GPL formalizes this for firmware/software but even then people still ignore it (see: blatant GPL violations).

[OwO]

EDIT: removed semi off-topic rant

[profdc9]

Actually, PCB layouts can be copyrighted.  I use the CC-BY-SA 4.0 to copyright my layouts so that modifications to the layout are required to be licensed also under CC-BY-SA 4.0.  It is difficult to enforce, but at least if you can show that their layout was derived from yours, they have violated the terms of the copyright.

For my part, I have tried to make the VNA I designed work with the parts in spec and so have an undistorted signal on the higher harmonics.  I do not drive the SA612 mixers and SI5351 clock generator out of spec so to compensate I have a way to decrease the IF bandwidth to get more signal and use a lock-in type approach to integrate the signal, which slows down acquisition, but at least you can integrate a decent signal at the upper frequencies if you are patient.  My design, as always is at

http://www.github.com/profdc9/VNA

You can now have one of these made with most of the parts populated by JLCPCB.  I myself have been exploring was of getting a 6 GHz VNA design going, which would be using two MAX2871's in conjunction with two SI5351's as well as designing my own downconverting mixer based on the BFP740 transistor.  That is one reason why I have been prototyping mixer and amplifier circuits based on the BFP740 transistor.

Dan