The USB Condom

So a while back, @drakkhen and I were chatting and he suggested a simple device that would let him “airgap” his mobile devices from his computer but still use them for power. Like a “Charge-Only” USB cable but in an “adapter” form that you could use on normal USB Cables (the only previous alternative was to destroy one, or remember to carry your charge-only cable). We were surprised nothing existed already…how could that be?

@drakkhen apparently knew I was working on Tally/Osprey, manufacturing FaceDancers, and working more and more with designing and reversing embedded systems. So I guess he thought I was qualified to check into it. He was wrong (I am woefully unqualified) but eventually I came up with with something crude.

The result was something we called the “USB Condom”. I designed the early versions in Eagle, and then moved to PADS when the factory I was getting quotes from requested it. I am a noob to circuit design so I need every one of my designs (specifically for Tally, a much more complicated RF board) reviewed by someone that knows what they are doing.

The USBCondom may look simple (and it is) but a lot of work went through pouring over USB parts spec sheets to make sure we selected plugs and receptacles that geometrically aligned, so that we could only print one board that worked for USB TypeA, Micro-B, and Mini-B. I poured through the spec sheets for hundreds of receptacles. But after several revisions and a few weeks of working with the engineers at the factory we had some working prototypes so I tweeted about it and went to bed.

The next morning I awoke to an inbox FULL of messages (my email address was on the circuit board schematic I’d tweeted) mostly from media outlets. My Twitter feed was also on fire. By weekend’s end there’d be over a million hits on the website.

I’d only initially planned to do a small manufacturing run and give them to friends and colleagues at conferences and sell as a gag-gift on Int3.cc (which is where we were selling Facedancers). But the response had us reconsidering such a small run….You can read all about that experience (the media storm and whatnot) over on the Int3.cc Blog.

Nonetheless, the huge response on the net had me rushing to put up a web-page  more functional than just the GitHub Pages hosted splash page that was there. It also had me hurrying to respond to journalists and interested customers. We decided to up our initial manufacturing run from a hundred to one thousand. I opened the web-store on Int3.cc after confirming the order with the factory, and orders rushed in.  Within three days we were sold out.

Then the trouble struck. Complications with the manufacturers. Overpriced assembly (we suspected the factory saw the media blitz/demand and upped the prices). And then there was the trouble of figuring out how to fill so many orders (we only have a few people at Xipiter offices working full-time already). Do we use a “drop-shipper” or contact a fullfillment contractor or use Amazon?

The most costly mistake (we were paying very close to what we charged for the rushed creation of the units) was the manufacturing errors. We just didn’t error-check designs before they went off to fabrication. After all, what can go wrong on such a simple board? We ran into a really lame error in design caused by a faulty spec-sheet on one of the USB components (the pinouts were wrong). This resulted in a “power to ground” mis-wiring. When one thousand units arrived I eagerly plucked one from the box and plugged one into my phone and fried my phone!

In the end, not only was I out of all the money  with none of the units usable but I also had hundreds (close to a thousand) people emailing me wondering where their orders were. My phone was fried so I couldn’t even call the factory to yell at them.

In the end I had to eat the expense. It was a solid business lesson learned. “Trust, but verify.” and don’t rush.

Everything has since been fixed and we are happily shipping USBCondoms. We’ve also overcome a few of the early design complaints and when we can carve out the time, we’ll improve the design a bit (maybe shorten the board a bit to make it a little more portable, put it in a case, etc.)

Nonetheless, something that started as an “after hours” thing blew up to consume way too much of our “paid work” time, but it was a good learning experience, and people are pretty happy with it.