Starting a Hardware Company with $60K | Hacker News Reader

Starting a Hardware Company with $60K

cobattery.com

189 pointsgwintrob76 comments

The Da Vinci is _not_ a Makerbot knockoff. Its a printer based on open source designs of Reprap.

Makerbot should never get credit for it after they chose to leave the community. Markerbot is actually a Reprap ripoff!

That said, I will be waiting for the next set of articles. This is good! :)

Ha you are clearly more knowledgeable about 3D printers than I am :).

Reprap for anyone reading this is totally awesome. I didn't include it in my post but Reprap is an open source 3D-printer that can make itself...

It's basically the hardware version of the singularity: http://reprap.org/

Well, it can kinda "make itself".

Electronics and strong materials (steel mostly) are not included, and it takes quite a bit of assembly.

But yes, it awesome.

Lathes had been making themselves for ages before 3D printers.

That's a good point! I guess the iteration cycle has gotten an order of magnitude quicker and cheaper with 3D printers.

The most interesting thing about reprap is that it is a long lasting community who does (real) open innovation. There open innovation is not a buzzword. The reprap community didn't invent 3D printing, but they lowered the price and access and are the single (ahem, multiple) cause of the present flood of low cost, reliable 3D printers. They experimented, took many false paths, and continue to push in new direction. They also are the origin of many growing companies : Aleph, Prusa, even the French Dagoma…

Not to be overly pedantic but the Singularity is defined as the point at which AI has reached runaway growth during which it would improve itself beyond anything imaginable by humans.

I think you're just talking about a self-replicating machine like a Von Neumann Probe. (check out the show Lexx if you want to see the hilarious end result of this)

Ha Ha! I have spent way too much time staring at 3D printers to get a good print! :)

If I don't have the time to build a reprap myself, what would be a good version to buy?

If I were replacing my Prusa with something off-the-shelf, I'd buy Ultimaker. My previous maker space had four of them, and they ran those things constantly with minimal fuss.

They are a bit pricey, but there's the option of buying used - lots of people buy 3d printers, and then realize they don't use 'em and so sell them second hand.

Flashforge, about 700 to 800 on Amazon. qidi clone is 400. Monoprice.com has a basic one for 349. Parts cost about 200 if you diy, I am told.

If you can afford one, the Ultimaker 3 is fantastic and extremely easy to use! I've been using one for two weeks. It was pre-built and after unboxing I began a complex print with water-soluble PVA support material in about 20 minutes.

Ultimaker is a supporter of open source hardware: https://ultimaker.com/en/blog/17799-releasing-our-source-fil...

Let me first apologise for the late response. I am sorry I had forgotten about this thread.

The best 3D printer IMO is the Original Prusa sold by Josef Prusa himself. shop.prusa3d.com

3D Hubs also has a great section on 3D Printer reviews and ratings.

The problem with relying on "security through obscurity" for developing your initial CAD models is that you won't be able to patent these designs. You're making a public disclosure about your concept before you file, so if you intend to protect your hardware concepts through patent protection you'd be better off using something like Autodesk Inventor Fusion 360 (which is free for hobbyists, students, and even startups making less than $100k/year, if I remember correctly). I think OnShape made a big mistake by removing the ability to keep any of your designs private without paying.

That being said, if you're working in the open source hardware space or just don't plan on patenting your work, OnShape could be a good option.

Wow, that is an important point that I did not know about. Luckily we do have the resources to pay for OnShape, so all of our files are private and have been private.

I'm using Fusion 360 and highly recommend it -- it's the first affordable/free CAD system I've actually enjoyed using. It runs on Mac and Windows.

Disclosing before filing would indeed, I believe, preclude obtaining a patent in most of the world. In the US, though, there is a one year grace period after the inventor publicly discloses the invention during which that disclosure does not automatically make the invention ineligible for a patent.

Of course it would still be better to keep it private until you file, so as to not lose the possibility of patents outside the US, and so as to not subject yourself to the risk that it might take more than a year to get to the point that you are ready to file.

As a hardware developer, I'm surprised you actually bought a 3d printer. I found that Shapeways and Materialise work really well, and the quality of their SLS prints is basically impossible to match using a desktop FDM printer. Plus, there is no fuss: upload a model, click "order" and get your print in a couple of days.

I'm really interested in the next chapters, especially about dealing with manufacturers, working on injection molding design, choosing your materials, etc.

Also, thanks for writing this! There is too little information on building hardware out there. Nothing like the software world, and each beginning hardware developer has to start on his/her own.

Hey jwr!

We found that buying a 3D printer was essential. They only cost a few hundred dollars and instead of getting a piece for $50 in a couple days, we can print something for a couple dollars in a few hours.

For testing the final design, we still use 3D print services to get really nice prints. But for the initial iteration it's nice to do it as quickly and as cheaply as possible.

Sidenote: We also rigged up a system so that we can run the 3D printer over night and print multiple parts. We will probably write a post on that.

Interesting — I found that given the amount of stuff that I actually needed printed, I am still below the cost of a decent-quality 3D printer. Also, I would not be able to print what I needed with a desktop FDM anyway, and I figured I will save time by not having to fiddle with 3D printers myself (usually the cheaper the printer, the more fiddling is required).

So, that's interesting that you found otherwise! Please definitely write more about that.

Yea, so I think scale, both in volume and in number, is a huge variable.

Our cases are relatively small so they can be printed in a couple hours. And during the development phase, we are going through hundreds of prototypes. (The fit has to be just right, so if a part is 5% off, we scrap it and reprint.)

Also, I love that you mention the fiddling as an annoyance. My co-founder loves fiddling w/ 3D printers so outsourcing all of our printing would take away that hobby :).

Another perspective here. At work we use a mixture of desktop (Form 1) and SLA printing (outsourcced). They have pros and cons. The resin printer is probably better resolution, but the SLA printers can do things like support-free printing which is really really useful for clean designs.

One problem with using printing companies is that they charge by volume, whereas most makers are concerned with the amount of filament they use up. This is cool if you can pack your volume with lots of designs, but it means that even a slight height increment on a big model adds a lot of the price. Then you have order minimums: 3DPrintUK, for instance, has a starting order of £40 which adds up fast.

Time isn't immediately obvious. We can rush an order online to around 3-4 days regardless of size. For small parts it's a no-brainer doing it in the lab, but for big prints there are print reliability issues, it still takes ages at high resolution and requires a ton of cleanup. And yes, fiddling is a pain, I don't mind doing it, but it can take hours to get repeatable prints.

Edit: meant to say we outsource SLS (selective laser sintering) rather than SLA (stereolithography - i.e. the Form) :)

Are you designing things or just printing other people's designs. For me that's the major difference. I almost never print others designs except when testing the machine.

And when I'm done with the iteration I'm likely to machine it.

FDM printers can be extremely limiting. Due to the fact they need scaffolding/support to print anything with overhanging features. SLS prototypes allow for complete freedom in part shape.

I find I design within the constraints of the printer which is not ideal when the final component production method might be an injection moulding.

But having a 3d printer on the desk next to you _is_ invaluable and can help solve issues in a matter of hours rather than days or weeks.

Shapeways or Materialise get quite expensive if you have many models to print. I 3D-printed a hardware prototype on my own recently. The cost was $5 in materials versus $40 for Shapeways.

I would still go the latter route if you have only a few prototypes to print as there is a learning curve in finding the right print temperature, time needed to sand down and refine the parts, etc.

And of course, there's nothing more exciting than, "Right here, right now."

I was surprised your step 1 wasn't "build a working (but ugly) prototype out of pre-existing breakout boards and components."

I might have a pencil sketch or two, but I wouldn't be doing anything in CAD until I knew my idea was feasible and had some grasp of exactly how it might fit into a finished design.

What stages did you go through for the conceptualisation and prototyping of the idea? As someone in the hardware business, I know the circuitry involved is common and simple enough that you probably had little difficulty. But I'd still love to hear how you discovered what you needed where to source it.

Good effort anyway, we need more guides like this on the web!

Yea, that's actually a great point!

I glossed over that pre-CAD prototyping in the post since we built those first "duct-tape prototypes" a long, long time ago. So we don't have pictures readily accessible and it was harder to remember the exact process.

To start with, I think we modified another battery case and basically stuck batteries on the back to try out swapping.

I'll think about how to incorporate that step into the post and see if I can dig up some old photos.

Would love to see that, thank you.

You need to put this through UL, FCC, CE and TUV testing before selling it. Depending on where you intend to sell it you might need to pay for acceptance of the above certificates. For example, Brazil has it's own certifications but they will accept CE for a fee, etc.

Have you gone through compliance testing?

You don't need to do FCC testing if you don't have an oscillator.

This device should have a switching regulator in order to manage the variable voltage coming out of the battery to then power the phone, which means FCC testing is required.

Generally speaking testing labs will run a series of common tests that satisfy the various regulatory agencies. In other words, you don't have to run four separate sets of tests to obtain UL, FCC, CE and TUV. I've done this dozens of times over the years.

It should be noted that if they don't themselves design any of the modules that would require certification, but instead just use already certified third party modules, that greatly reduces the amount of testing their product needs.

I've read that can get the cost of FCC certification down $1-2k and the certification time us under 5 days.

If you design your own module, then it is $7-20k, and will take 4-6 weeks. (Once you've certified that, though, then if you reuse that same module in other products it is like using a third party certified module, so $1-2k for each subsequent product).

Our batteries come with some compliance testing. I'm not sure what other testing we have done for other components. But the batteries are the important one.

Thanks for writing this up, and I look forward to more!

My first company did hardware, and it was kind of an awful experience, and I swore to never do another hardware company. I got burned out from sourcing things, dealing with revisions, and working with manufacturers through the language barrier. But, things look so much nicer ~12 years later, and there are several products I wish existed that I think would have a solid niche market. But, it's still pretty intimidating.

Also, I love the connector. The case is a bit weird, though, with seemingly exposed battery? But, I'm not the target market, anyway (I have an Android device).

Thanks!

Yea, as I said in the post, I think it's drastically easier to start a hardware company now vs 10 years ago. And it's still pretty hard, ha! So I can just imagine what you went through.

Yep, the battery is "exposed" in that we don't have a cover over the battery. But each battery does have protection.

And you aren't the target market, yet! Making cases for just iPhone made things easier, but we are hoping to expand to the most popular Android phones.

You really need to add an email signup for people interested in CoBattery for Android. At least by the time you write the next part of your blog post.

That's a great idea!

Hmm, the post shows that the case was designed in OnShape but the screenshot shows that's its an imported STEP file from Solidworks(?). Admittedly though, I haven't used OnShape since they removed the private project options so I might be wrong.

Understandably, OnShape and Fusion360 (the HSMWorks/CAM integration in it is awesome.) is great for startups and hobbyist. But if you're trying to design some large assemblies or things with complex organic shapes, Solidworks is significantly superior in every way (I'm extremely bias since Solidworks is my favorite CAD software). I know the price is a lot for a startup but I highly recommend any company trying to get into the hardware side of things, to stick with the "foundational" CAD software because in the current CAD software landscape, the investment into these software would pay off in less than a year.

Wow, you paid attention to the details! Yea, the screenshot was just for the purpose of the post.

We did some of the original design in OnShape and some of it in SolidWorks. I didn't want to recommend SolidWorks for cheap work because at the time we only got it free because we were students.

We've found that it isn't that hard to start off a design in OnShape and then move it into SolidWorks as things get more complicated. It isn't ideal, but it's definitely cheaper.

Financially cheaper yes. I've tried Onshape for 6 months and I didn't find it as immediate or as fast as Solidworks. Particularly when modelling multi-component assemblies.

Onshape monthly price for a pro licence is about 1/3 the price of Solidworks monthly (on finance), for me Solidworks is a clear winner here even though it costs more.

But, Onshape is a revolution in the free CAD space. Perfect for simple products like this ;)

Edit: added last sentence

Hey everyone, I wrote this post :). I'm happy to see that it's getting some traffic here.

Let me know if y'all have any questions. I would love to see this post help other people start hacking on hardware, so I'll answer anything I can.

Is that a certified lightning connector? I'd be curious about that process or, if it isn't, the decision to forgo it and whether or not you did any user testing around it.

It's not!

We have some info on the connector here: https://cobattery.com/connector

Apple doesn't actually let you buy a chip that will allow for the right angle connector that we have. So we decided to design our own. It was a pretty time intensive process and we had to search for a while for a good factory. But we are happy with the way it turned out.

Well the idea is obviously OK but it looks really bad in my opinion. Who buys an iPhone and uses this gigantic, bad looking case for day-to-day usage?

How many of those have you sold already?

Ha, well I'll take the first part as a compliment!

We are actually proud that our case is slimmer than most other battery cases on the market. We didn't want our case to be bulky so we designed our own custom lightning connector that snugly fits the phone: https://cobattery.com/connector

What don't you like about the case's look?

Everything. The case looks like a giraffe's tongue, and giraffes suck.

Just kidding. The case looks fine. I'd briefly wonder whether dropping it on the metal connector at the bottom would damage it, and I'd prefer the battery color match the case, but those are just nitpicks.

What!? Giraffes are awesome...

We have found the metal connector to be pretty resistant. But we have one year full warranty, so if anything breaks for any reason we will send you a replacement immediately. :)

Can you write more about how you provide a warranty? Is it simply guessing x% of the products will fail within the warranty period and set aside some money to cover those failures. Or is it more complicated?

Hmmm, it's actually much more simple than that. We want to provide a good product that makes people's lives better. The warranty is a forcing function.

In my opinion, if we can't make money while ensuring people's CoBattery cases works, we shouldn't be in business.

Hey, I just wanted to say it's cool to see ideology in business. It's relatively rare. Best of luck to you!

http://www.apple.com/shop/product/MGQM2LL/A/iphone-6-6s-smar...

Cheers and best luck to you!

Once I upgrade to the 7, I will be buying one of your cases. You've solved all the issues I hated about previous battery cases, which actually had me go back just a "normal" slim leather case.

I.e. no lip, and that I don't have to tether my phone back to the wall when I'm charging it.

Additionally, I believe this will solve the issues I've had with lighting connector wear.

I'm very excited to read about your alibaba connections and sourcing however.

Thanks for the kind words yial!

We've found that a big reason people upgrade is due to a failing battery. So if that's why you are upgrading, I would just save $$ and get CoBattery now ;).

Your product looks interesting. I like that you can swap out batteries without removing the case. Also that you aren't blocking the headphone port. (You should list that as a feature ). One thing that I feel that is missing from your site is how many maH the battery is. I wouldn't know this is a good deal unless you listed this.

Yep, that's something that accidentally got removed on the most recent re-design...

The batteries are each 3,000 mAh.

Good effort, but the "battery" part seems like it is still in need of some attention. Developing a good relationship with a reliable cell manufacturer, and staying on top of quality control issues, is going to consume your life.

Also, get UL certified, or you're wasting your time.

Thanks jpm_sd!

So for the battery, we are actually using a modified version of an older Samsung battery (not the Note 7 haha).

That's something that we didn't want to reinvent. So we are working with a factory who has made these batteries in the past.

I didn't include more about the battery in this post. Is that something you'd enjoy reading more about?

I would!

Sweet! I have some blog posts planned about how we are designing a new charger and our 3D print iteration set up. So I'll think if we have anything else on the battery.

I also wrote up a basic, but goofy, post on "What is a battery": https://cobattery.com/blog/what-is-a-battery

If you have any specific questions, feel free to email me at victor@cobattery.com

Definitely want to hear about all aspects of battery sourcing.

They aren't building battery technology, they are a phone case company.

They simply need to design around a battery type - preferably one that is cheap and accessible - and let the user go from there. Maybe have the first one included, in which case the deal is just in volume discounts.

We actually include two batteries!

I think that's actually a key part of what makes CoBattery special. The case is designed for battery swapping and when you buy CoBattery you have everything you need to start swapping.

Thanks for writing this, the post is super helpful! I didn't see anything about the circuit design/iteration process. Would be great to hear what software was used to design the circuit (if any) and how that eventually translates to mass production.

The circuit is actually really simple in our case. It is just a voltage converter and then an adapter for the lightning connector.

Getting the custom lightning connector and chip was more of an ordeal though! I'll see what I can dredge up on that process.

This is great. It's been a long term goal if mine to do a successful kick-starter robot. This give me confidence that a hardware noob has some chance of making this work.

Awesome! If you end up making something, send it over :). I'd love to check it out.

No love for FreeCad? It's certainly comparable to other desktop software for simple designs. The constraint solver is actually better than several I've used.

I haven't tried it, but it definitely looks cool!

I played with an older version of it a few years back - it has a bit of a learning curve, but well worth the effort.

If you are into coding - there is an interesting tool out there for creating 3D CAD files using a software-based paradigm (that is, you "code" your object):

http://www.openscad.org/

If you prefer something browser based:

https://openjscad.org/

Also:

http://www.implicitcad.org/

Finally - if you want something closer to freecad (mouse/gui) but browser based:

http://shapesmith.net/

There are a few other solutions out there as well that are open source (and note - not all of the above can really be considered "CAD" in the strictest sense of the work - freecad is really the only one, honestly) - but these are the ones I can think of off the top of my head.

Is the battery held in by friction and the connector? If so does it ever fall out?

How do you deal with the lightning port being used the headphone jack?

> Is the battery held in by friction and the connector? If so does it ever fall out?

Yea, it's just friction and the pogo pins at the bottom! It took us a long time to get that fit quite right but now it never falls out. I'll give you a refund and let you keep CoBattery if you buy it and can shake out the battery :)

> How do you deal with the lightning port being used the headphone jack?

I actually have the iPhone 7 and I've totally switched to Bluetooth earbuds. And you can actually pull out the connector if you do want to plug in, but it's definitely a better experience to just move to Bluetooth.

We are working on a version of CoBattery that will have an old-style audio jack. But it's tough to source the audio-lightning components!

We actually started smallworks.com (LEGO compatible iPhone cases) for less, but went through a similar process.

We would all appreciate if you had anything to share about the following:

> 3. Find manufacturers using Alibaba (coming soon)

> 4. Picking the right partner factory (coming soon)

> 5. Scale up production at a Chinese factory (coming soon)

How much time did you spend in China to get through steps 4 and 5 ? Thanks in advance.

I'd love to hear about how you developed the connector.

Ooo, yea the connector, that was tough! We have a proprietary design there so we can't share everything but I will include some info on the connector in the next posts of this blog series :).

Reply on news.ycombinator.com