Monday, April 03, 2006

Universal Digital Ports

Had an idea the other day that could be a business opportunity: a "universal digital port". In the aviation electronics world, and probably in the electronics world more broadly, there are a lot different digital data protocols, like serial ports in RS-232, RS-422, RS-485, ARINC 429, ByteFlight, Ethernet, USB, parallel ports, game ports, etc. In aircraft cockpits, you have all these different devices using different formats and getting them to talk to each other is a pain. There are companies out there that make conversion devices (we call them "happy boxes" here at the office) but they are often very expensive, and specialized to for a particular device using a specific format.

Why not do the conversion in software? The device itself would be a generic embedded computer with an array of digital inputs and outputs whose voltages can be varied across the range found in the various data formats. The actual protocol, timing, voltages, etc and the conversion from one protocol to the other, is done in software that can be easily written and loaded to the device. You could even write the software in an extensible "plug-in" style and release an API so that people who buy your device can create their own protocols.

Using one common hardware configuration gives you pretty good economy of scale in production. Opening the API offloads a lot of the programming work, so you would just have to get the basic OS and the software framework going to have a viable product.

In a way, this is kind of similar to software radios, such as the open-source GNU Radio and the military's JTRS but applied to digital device-to-device communication instead of audible radio. Any thoughts? Kevin, is there already a VC-backed startup that's been doing this for the last three years?


At 5:06 PM, Blogger Kevin said...

... I don't know of anyone doing this right now, but we don't focus on anything involving hardware.

Couple questions...

1. How big of a market do you think this is? How much could you sell this for, and what are reasonable assumptions for number of units?

2. What are the happy box makers doing? Who are they? Are happy boxes just a small part of their business? (I am wondering if #1 is big, why the happy box makers would be asleep at the switch.)

... Learn to be a VC in five easy steps! To be continued.

At 5:29 PM, Blogger Dan Craig said...

Well, there are a few different markets out there, but their needs are very different, as are the price points. One the cheap side is the experimental aviation and the general hobbyist market that might include people into cars, game systems, or stuff like that. I'll have to do a little research to see how big the market for that would be, but the price would have to be pretty low to hit this market. Probably on the order of $100. Converters that do only fixed formats seem to be in the $50-$70 range.

Then there's the certified aviation market (or anything else that's safety critical). In this case, you have to do some more extensive software development and testing processes and that tends to make just about anything cost at least 10X more than it would otherwise. By that reasoning, the happy box would be priced around $1000. I don't know how that compares to others, since a lot of the companies tend not to list prices. But there was an analog to digital device that I found that one would use to convert analog navigation radio signals to digital that was priced at an absurd $9000.

I'll dig around more for additional happy box makers. The one I'm most familiar with is Shadin, and they do a lot of other business in aviation sensors. There's another one that I saw an ad for last week, but I can't find that magazine now.

At 10:03 PM, Blogger Thomas said...

not knowing anything about the subject (when did that ever stop me from "disagreeing") I am not sure how cheap you could produce this general purpose happy box. usually the more general purpose the more these things typically cost. I don't imagine many people have much use for a general purpose converter. If I was in the market for one I want a specific purpose one (I have a gps that speaks rs-232 and a laptop that only has usb ports) so I go out and buy that converter (it was $10) I find a lot of these converters are really cheap if they are for the really typical applications. When you get into arbitrary conversion you get into much more computation between differnt protocols aside: i mean what does usb to ethernet really mean.. are you turning signals into packets? are you just using ethernet becuase it is a longer cable? not really knowing the translation needing to take place i imagine allowing any arbitrary number of these you will need a pretty good processor and some damn good software (both of which aren't cheap, or small, or non-error prone) So while i guess I can see someone who is tinkering with stuff every day for their job needing one of those. hobbyists/low level professionals will probably want to stick with the all hardware/specific purpose one.

*caveat: I realize that the really cheap ones are usually just adapters that rearrange the wires and plug size/config.


Post a Comment

<< Home