One step left to Kickstarter - is there a chance for the project?

For 9 months now we have been developing our project - a miniature Black Swift network computer. It's time to give birth, so to speak.



I would like to share my experience and learn the opinion of the community: is it not in vain that we all started this? We are almost ready to go on Kickstarter and me, as the initiator of the project, is slightly shaking from periodic adrenaline rush. I will not hide, I am waiting for words of support and approval, but healthy criticism will be extremely useful.

Everyone who is interested in our story, please read and comment.

The birth of an idea

Here here , I described how this happened. The customer came with the project, we started looking for technology platform options, found several boards on the Atheros AR9331.



We looked, thought, and decided to make their own implementation. Firstly, it was interesting. Secondly, having calculated the cost of production of similar boards, we realized that it would be possible to do our job somewhere at 30-40% at least cheaper than buying ready-made. Thirdly, several ideas immediately appeared on how to do better than the "competitors". From the very beginning, one of the ideas was to make the project completely Open Source - both from the software and hardware side. On the one hand, there is a lot of information on AR9331 everywhere. However, anywhere so far an accessible working circuit of a device based on this SoC is not observed. Hundreds of projects people have already implemented, but an open, ready-made platform has never appeared.

First pancake

Our ingenious engineer Valera developed the first revision of the board for about three months.



We wanted to make the board tiny and fit in four layers (to make it as cheap as possible). However, this is very difficult to do if you use DDR2 memory. Here's what the board looks like under the memory chip:



After much torment, everything worked out, and we sent the documentation to our production specialist, Alexei. Alexei lives in Hong Kong, a professional electronic engineer, and in the field of electronics production in Shenzhen ate (thank God, not literally!) More than one pack of dogs. They didn’t look for a factory for the manufacture of printed circuit boards, they used the same one that Virtrilki did. But with the search for a factory for installation, we are tormented. Nobody really was eager to do a rather complicated installation (we actively used 0201 components with dimensions of 0.6 × 0.3 mm) for a circulation of 20 boards. Well, yes factories, thank God, in Shenzhen a lot, found, done. And in June, we finally got the first revision of the boards on hand and began to deal with them. Of course, the boards did not work right away (who would doubt it?). In a week, Valera found several of our jambs and about the same number of Chinese (transistor rotated 180 degrees during installation - what?). But his golden hands still allowed us to launch most of the boards. Of course, our joy was crowned by an article on Habrahabr.

Step Two, but Not Last

For a month of working with the board, it became clear what was missing and where to move on. Of course, the bugs were fixed first. They added a button - without the ability to make a hard reset or enter the boot loader with the default settings, it is very difficult to work with raw firmware. We added an FTDI chip so that you can not only power the board from the computer’s USB port, but also immediately see the console. They added the ability to flash the NOR flash chip without evaporating it from the board (yes, and hard reset does not always help with development, alas). We increased the power of the DC-DC converter, which makes stabilized 3.3V from 5V from USB to power the board itself.

Using mrmagerdeveloped a plastic case. Prototypes printed on a 3D printer. The thermal printer is a decent result, alas. But the photolithographic impressed (albeit expensive, of course):



Yes, in between cases they came up with the name of our computer - Black Swift. Black Swift . Swift is a small, but very fast bird, which seems to hint. Well, black - Alex successfully selected a color for the board, we liked it. Our designer, Tatyana, drew a nice logo.



Well, with the software, of course, tormented. The bootloader and firmware (u-boot and OpenWrt 14.07 Barrier Breaker) worked as they should.
And at the end of October, the second revision of the boards came:



This time there were no critical problems, and the boards earned the very next day. But, alas, there were absolutely no jambs ... So there should be a third revision.

And one more our article appeared on Habrahabr . Where almost a thousand people expressed a desire to purchase our computer. Motivates, however!

Our plans

And so, we have come to the current moment.
There is a working prototype of a miniature network computer, made at a very decent level. In our opinion, a very useful thing in the household.

There is an understanding: it is impossible to make money on sales of the board itself. Simple arithmetic: even if you make a profit of 5 USD from one board (20 USD actual operating cost in Moscow, 25 USD retail price), then with a circulation of 1,000 boards we will earn as much as 5 thousand USD! Absolutely crazy money, yeah (sarcasm). And it’s unlikely that we can sell an order of magnitude more in Russia;

Therefore, we decided:

1. The board should not be “sold”, but “distributed”. If you “give” it to people at cost, then the audience will grow much faster, which can make Black Swift (especially considering the openness) another common technological platform, a basis for independent development. If people come to us with their projects, we will be happy to help organize circulation production and (we hope!) We will earn our bread.
2. It is necessary to enter the world market through Kickstarter.

So, in December we are going to do another test revision of our board (in two versions - the regular and the Pro version). I hope this revision will be final, without jambs (spat three times over my left shoulder).
In January we want to make a trial run - at least 500 boards. If we dial pre-orders for this, of course (if there are no people who want to buy a computer, then why?). Here, the main risk factor is the Chinese New Year. Almost all of February, China does not work, so you need to have time to make and deliver the fees in January, otherwise the next deadline is already mid-March.

Well, from April, I hope the board will be available for immediate purchase, without pre-order.

Next week we plan to start raising funds on Kickstarter. Registered legal entity in New Zealand, we are preparing materials with all our might. As usual, everything is not easy when you do it yourself (we decided to refuse the services of intermediaries). Scary, already terrible. Do you think we can do it?

PS I beg people who vote “there is no chance on Kickstarter” - could you please reveal your point of view in the comments? I really need constructive criticism.