This repository has been archived on 2017-04-03. You can view files and clone it, but cannot push or open issues/pull-requests.
blog_post_tests/20120115212201.blog

11 lines
3.0 KiB
Plaintext

Calling all hackerspaces
<p>This is a shout out to all hackerspaces and engineering schools within easy reach of Philadelphia. I&#8217;ve got a nice little design-and-build project that would do the world some good, but I don&#8217;t have the skills or facilities to do it myself.</p>
<p>The problem: build a ruggedized special-purpose test enclosure to be mounted on a roof or utility pole and host a bunch of GPS sensors. The tricky part is that it needs to be outside and not under top cover (for good skyview) and thus weatherproof, but also transparent to the GPS radio frequencies. Another part of the design problem is getting data and power cabling back to my development computer.</p>
<p>UPDATE: I&#8217;m now pursuing a different path &#8211; trying to figure out how to build a GPS repeater on the cheap so I can effectively pipe the RF from a roof antenna to be retransmitted in my office. This has the obvious advantage that the GPS test rack will be able to live inside, near my desk, rather than outside in an enclosure that can only be reached with a ladder. So now I&#8217;m looking for a hackerspace frequented by radio hams.</p>
<p><span id="more-4060"></span></p>
<p>I need this because I&#8217;m the lead programmer of <a href="http://catb.org/gpsd/">GPSD</a>, responsible for the correctness and robustness of software used in hundreds of thousands of deployments in GIS software worldwide. GPSD is used for navigation, fleet tracking, scientific telemetry, autonomous-vehicle guidance systems, and disaster response. Lives may depend on our quality.</p>
<p>Most GPSes can be simulated with canned datafiles run through a test framework. A few cannot be, because the handshaking between host software and device is too complex to be reliably simulated. These problem children need frequent live testing to avoid regressions, and the problems with doing that have made for a troubling weak spot in our test regime. I want to be able to push a button and get test results every time I perform a potentially sensitive commit.</p>
<p>The test enclosure needs to be able to host a minimum of four RS232 devices and four USB devices; more, up to a maximum of 16 each, would be better. Cooling isn&#8217;t an issue as they&#8217;re all milliwatt devices. Swapping in new test devices needs to be easy. There are a bunch of practical issues about where to site the test enclosure for both good sky-view and serviceability, how to power it, and so forth. </p>
<p>I&#8217;m looking for student and hobbyist EEs that would enjoy developing such a design and the real-world experience of building and deploying it. The deliverables for the project would (1) be a set of blueprints, to be published under Creative Commons, (2) two working copies of the test enclosure (one for me, one for my test guy on the West Coast), and (3) an installation expedition to mount one of the copies out here where I live.</p>
<p>This would make a good practical engineering project for a class in power and signals engineering. Anybody interested?</p>