The Ware for October, 2005, is shown below. Click on the image for a much larger view.

This month we’re going old school. I’m thinking this is probably one of the more obscure pieces of production hardware that I have sitting around in my house. To be honest, it took me some time to figure out what this thing does, but once I found the documentation on it, I was reminded that I take a lot of technology for granted these days. I’ll be impressed if anyone can name the exact piece of equipment this comes from, or if they can guess its precise function.

Some side notes–sorry the blog has been pretty quiet lately. Last month was brutal at work, I was putting in 100-hour weeks to get a chip out on time; fortunately, my new girlfriend is very understanding of my situation. I think that the hard work will pay off in the end. I’m now taking a little time now to catch up with friends and relax a bit before jumping back into the groove. I also got notice yesterday that one of my paper submissions to ISSCC was accepted, so I’ll be in San Francisco next February presenting a talk on a 10 Gbps integrated CMOS photonic system technology!

Another side note–you may have noticed the ads on the blog sidebar. The bandwidth overage charges from the comment-links in slashdot are starting to pile up, so I’m using Google’s adsense service to help cover the costs. I think the ads are fairly tasteful and inconspicuous (although poorly targetted due to the dynamic content generation infrastructure of this blog software); if they cause problems for anybody let me know. I don’t want these things interfering with the quality of this website.

I was really impressed at the quality of the responses to this past month’s Name that Ware contest. The analyses of the hardware were very in-depth, and several correct answers were given in the end. The device is in fact a Blackberry 7290 using the Cingular network, and again, Nate gave the correct answer first (using the SHA-1 sum hash to seal his answer): Nate even got the network provider of the phone correct (I’m surprised he didn’t also figure out the IMEI somehow). I think I should also note that Mike Miller also gave a correct answer. Congratulations! I’m hoping I can provide a bit more of a challenge next month.

The phone board you saw is in fact the guts of the phone that I use. As a sidenote, I’m pretty happy with the Blackberry; it does what it does well, and it doesn’t crash like the Treos do (hurray for not missing calls because my phone crashed). Normally, the RF perfromance of the phone is top-notch; I’ll get signals where others don’t. However, recently, my phone started acting up. My phone was having problems finding the network and it was dropping calls after about two minutes of talk time. It was getting pretty intolerable, so I decided to take a look inside to see what’s up. It seems that the antenna connectors on the phone had built up a little bit of oxide on their terminals. I think this happened because I tend to take my phone with me when I go out dancing or jogging. When I take the phone out of my pocket during such activities, the screen would be wet with condensation. Cleaning off the connectors and reseating the patch antenna seemed to do the trick, and the phone is back in fine working order again. Hurray for not having to drop another $300 on a new Blackberry!

Below is a picture of the relevant region of the Blackberry with some of the more interesting features called out. The antenna connectors are press-fit connectors with minimal wiping action. These types of connectors in my experience tend to be fairly unreliable; I used something similar to this once and I had about a 1 in 1,000 contact failure rate, which is pretty bad if you are trying to build a board with thousands of contacts (which I was doing). I also think it’s pretty neat to see how small they’ve gotten the acoustic wave filters that they use for the RF front end on these phones. For those who are not familiar with this technology, many phones use electro-mechanical filters to achieve the stringent spectral shaping requirements for cellular phones (yes, these filters really have parts that mechanically vibrate at frequencies of 1 GHz+!).

The Ware for September, 2005, is shown below. Click on the image for a much larger view.

Yes, we all probably know immediately what kind of board this is by the looks of the prominent SIM card slot. However, the question is: what’s the model and make for this board? This one is actually the insides of my device; I had to do a little cleaning on it tonight to keep its performance up, and I love looking at how they put these things together…so compact, and so clean! And look at those fine-pitch blind vias in the upper right hand corner. They look like they are laser-drilled…

Sorry this month I was late! It’s been a rough month work-wise. Next month should be better, hopefully. I want to catch up on my printer hacking too, and there are some other neat projects that I’d like to play with soon.

This past month’s name that ware was again a breeze for the contestants. I may take Nate’s suggestion for not this month’s, but for the month after’s contest (blocking out major manufacturer labels and part numbers to try and up the ante). For now, I’m just glad I found the time to put up a new ware this month and declare a winner for last month.

DavidR had the first correct response with his clever SHA1 sum hashed response; congratulations to DavidR! The board is indeed an Intel Paragon node board with an I/O daughterboard. Drop me a line for your prize. Nate as always had impressive comments and impressive response time. I’ve gotta go dig around my archives of obscure boards for more challenging stuff.

A little short on time, I’m not going to post the detailed breakdown of the board’s contents, as DavidR and Nate already did a great job of that in their responses. As for the fatal flaw, back at MIT my PhD advisor, Tom Knight, handed me one of these boards and asked me, “what’s wrong with this board?” and my response–which we can debate on its correctness–was “the tiny network connector on the left.” TK did happen to agree with me on this point, as this machine was bandwidth-starved and had trouble getting scalable performance on programs that did not partition “just so” to the machine’s granularity and architecture. In contrast, some of the better balanced machines out there, such as the Cray or IBM supercomputers, seem to be made out of mostly wires and connectors with a few chips interspersed throughout. It’s all about the wires, baby!

I just thought this was such a cool/useful hack, I had to post a link to it here. DC wrote a bitmap viewer/rotator. I can think of dozens of situations in RE work where something like this can be really handy…