As many of y'all have noticed, the Embedded Systems Show was last week. Now that I'm back home and again have decent Internet access (I was using the AltaVista free access and I can report that it was just awful), I thought that I'd report a few observations and rumors. I know that some others on this list were there as well and I'd encourage them to do the same. The main on-topic things I can report are the following rumors, which I will leave unattributed and unauthenticated: * Microchip's sudden shortage in the 16C8xx range may in part be related to shake-out in a new assembly plant in Thailand. Microchip makes the wafers here in the US, cuts them into dice and ships the dice to Thailand for assembly into the plastic packages. They have been starting a new plant there, and it has taken some time to get everything up to speed. Supposedly the dice exist, it is just an assembly issue. One of uChip's booth staff did give me a sample of the PIC16C877, and they also had samples of their new op-amps and standalone SPI A/D devices; I got two MCP601s and one MCP3208. * The 18Cxxx parts are sampling and "should be available" from local FAEs (I should get me one of them FAEs). Supposedly there have been no major issues and production quantities should be available "soon", I had the impression that this meant probably by year-end, FWIW. * The main bad news had to do with in-circuit emulation. The Tech-Tools guys said that the current Clear View Mathias didn't have the guts to handle the 18Cxxx, at least in 16-bit mode. Other than vaguely admitting that they've always got something in development, I wasn't able to find out anything about any new Clear View products. I also asked at the uChip booth about whether I should be upgrading to the ICE2000 for the 18Cxxx, and I was warned off; again it seems that the ICE2000 will likely support the 18Cxxx in 14-bit mode, but not in 16-bit mode. Thus, I guess that we can expect some new ICE models coming from uChip soon. (That's your cue, Tjaart!) It didn't seem that all these details were settled yet, at least as far as they were able to tell me. In the slightly off-topic range, * An Atmel rep told me that the bond-outs for the ICE200 pretty much sold out, and thus the supply of ICE200 units will be limited to inventory for the next couple of months. * Kanda, the people who make the AVR STK200 and ATmega STK300 development boards, are coming out with an STK100 board for the ATtiny chips. * Scenix had a sizable booth. One nice thing is that they will soon be shipping development boards for low cost. I guess that they've been delayed by the fact that the board's first manufacturer got bought out by someone else. They looked like they'll be nice boards, when they come out. * Zilog also had a pretty big booth, with a sports car in it and everyone wearing racing uniforms. * Philips will soon have out the P51XAG49 flash based chips. These will have 64KB of flash and 2KB of RAM, in a PLCC-44 package with all the normal XA features, such as an 8/16/24-bit memory bus, 16-bit math with 32-bit extensions, separate system and user stacks, paged access to up to 16MB/16MB RAM/ROM storage, two UARTS, etc. * Infineon had a huge booth, and they were handing out a "pre-release" copy of D.A.v.E. 2.0 (Digital Applications virtual Engineer). I attended a presentation on D.A.v.E., and have to say that it is a pretty stunning piece of work. Basically it amounts to an automated databook; instead of pulling up datasheets and digging through them for register bitmaps and such, you pull up this GUI application that will allow you to select a part, and then step through an extensive series of menus and forms. You point-and-click to select things like UART and timer settings, interrupts and interrupt priorities, etc. When you're done, D.A.v.E. will generate C code either for the Tasking or Keil C compilers that has pretty much a full outline for a standard event-loop application, with all the registers configured correctly and such. The skeleton code then has "your code goes here" type comments; as long as you leave all the other stuff untouched and only put your code where it gives you a place, you can go back and change settings in D.A.v.E. even though you've been modifying the generated code. It will do this for the 8051-based C5xx series as well as the 16-bit C16x series. I guess it will also do the 32-bit TriCore, but I didn't pay much attention to that. I fooled around with it on my notebook PC and it seemed to do what they said it would. * I spent a bunch of time looking at the products at the Phytec -- http://www.phytec.com/ -- booth. They have some very nice development boards for a reasonable sum of money. One *really* nice thing is that they have designed many of their boards to use little processor/memory/oscillator modules that have rows of pins to connect them into a main development board. The nicest part about it is that the processor/memory modules are reasonably priced and could easily be used to take a prototype off the development board and onto a custom board for a small pre-production pilot run. Not yet on their website, they are just bringing out a new style of module called the "phyCORE", that is purported to have much better EMC characteristics then the older ones. The first units will use the Philips 8051 and 8051XA chips. The first two XA modules will use the flash chip described in the last item and a new XA version with integrated CAN. The first two 8051 modules will use the P89C51RD+ and, again, one with integrated CAN. All those modules use PLCC sockets, so you can easily use them with an emulator. * emWare -- http://www.emware.com -- was just about everywhere. Besides their own booth, they had a presence in the Philips, Microchip, Mitsubishi, Motorola and Hitachi booths. * Linux penguins weren't the star of the show, but they were pretty common. Montavista Software was at the Ziatech booth selling "Hard Hat Linux", and most X86 and some PPC SBC vendors were claiming Linux compatibility. Sun Microsystems had one of the larger booths, and they were aggressively pushing both their Java hardware and software, as well as the SPARC-based SBCs and motherboards. They did have Linux running on at least a couple of systems. * Something of a surprise to me was the size of Cygnus' booth -- they were also one of the larger on the floor. And, in solidly [OT] territory: I attended the tutorial on "System Architectures and Archetypes", which turned out mostly to be about UML and generally using modeling languages to generate code for embedded systems. It was fascinating but I think they have a long way to go yet. Also, I attended an introductory tutorial on DSP, which I thought did a good job of introducing DSP. The other classes worth mentioning were those given by Jean Labrosse on Real Time Kernels. Although it was pretty much focused on the worldview from uC/OS-II, I think that he did a good job of giving a sense of what RTOSs do and don't do. The trade show was of course a bunch of fun. I got to meet Clyde Smith-Stubbs, Bob Blick and Scott Dattalo, among others. I know that Bob took some pictures, I don't know if he'll have them up on the Web at some point. It was a good show for T-Shirts, I got four and could have gotten maybe a dozen if I'd had the patience to sit through a bunch of inane presentations. The Infineon booth had the best deal -- everyone who was willing to sit down got a cup of cappaccino or latte from a real barrista, and everyone who sat through the *whole* show got a stainless-steel insulated coffee tumbler -- with a one year warranty no less. I got a real hard hat with a Penguin on it from Montavista Software (Hard Hat Linux), Klix toys from Atmel, and an Intel Bunny-Suit doll. (That was another one that required attendance at a presentation, and the lady presenter at the Intel booth looked as if she was powered by a Pentium III; one rarely sees that tight a performance at a show like that. I also recall absolutely nothing of what she said -- I spent most of the time just staring at her in awe...). Virtually everyone (except Analog Devices, dammit) was handing out CDs with data sheets and demo software. I must have gotten 30 or so such freebies (including uChip's latest). The Integrated Chipware booth was particularly vile -- they had a lady that looked old enough to know better doing a "magic" show and trying to lead the audience in an "Integrated CHIPware" cheer; it was somehow necessary to jump up in the air when pronouncing the syllable "chip", and the most enthusiastic members of each audience would win prizes. It was disgusting; even though I didn't mind the extensive use of "booth babes" (to quote Dilbert) generally, I think that she made pretty much everyone cringe. The Hitachi booth was out in the main hall; you had to pass through it to get to the main entrance to the show floor. At strategic corners they had lithe young ladies wearing the shortest shorts practical, holding out bowls of chocolates and trying their damnedest to get each passerby to take some. They were pretty good chocolates, FWIW. That kind of stuff was far from universal, however; the staffs in most booths were reasonably well-qualified sales or technical people, regardless of gender. Sorry for the long report, but I thought y'all might be interested. --Bob -- ============================================================ Bob Drzyzgula It's not a problem bob@drzyzgula.org until something bad happens ============================================================ http://www.drzyzgula.org/bob/electronics/ ============================================================