Jump to content
43oh

jpnorair

Members
  • Content Count

    718
  • Joined

  • Last visited

  • Days Won

    18

Everything posted by jpnorair

  1. I'm pretty sure Contiki is supporting the MSP430 now. I'm not sure if it supports Jennic's stuff, but Jennic has always been a ZigBee shop. In that case, it surely does not. ZigBee is not a very open standard -- or, really, a set of standards. ZigBee SE 2.0 uses IPv6 IIRC.
  2. If you do these things, the reference design should work as advertised. 1. Do what you can to copy the reference circuit layout, that will simplify things. 1b. Don't put inductors parallel to each other. Put them in a straight axial line, or perpendicular. Otherwise you get transformer behavior. 2. Use a 4 layer board. Just do it. Ground layer right under RF filtering components. 3. vias connecting ground layers... all over the place, but particularly at edges. You do not have nearly enough in that image. How are you connecting to the antenna? That is another list of consideratio
  3. OK, I would use a Johanson 920 MHz antenna on the top part of the 48x10mm section, and have a trace come out of it in a L-shape and go along the the top 70mm edge. This should generate 433 MHz, or maybe a bit less, and it is easy to tune -- for basic tuning, just trim the trace. I would start with an 0.5mm trace, and maybe in design revisions you find a better trace width, but it shouldn't have a huge impact in this design. Incidentally, W
  4. I don't mind minicom, but I need something a bit more special-purpose. I suspect it will want to have a GUI someday, but since the meat of it is POSIX C, all the GUI needs to be is the GUI. That is the only way, I've found, to allow a good degree of independence of platform, independence of library version (I'm looking at you, Qt), and independence from having a damn terminal app that is a 36MB application and eats 100MB RAM.
  5. Any type of dual-band antenna that you can build without a lot of antenna knowledge and experience with numerical modeling is going to be a compound antenna. Therefore, I recommend using two antennas -- easier to design, not really any larger. In fact, with the proposed GPS antenna, you probably won't even notice it is there. 48x70mm is enough room for a 433 MHz antenna and a GPS antenna. With antennas, if you start making them funny shapes you usually lose efficiency. If you add inductive loads you lose bandwidth. Fortunately you have enough room to build a 433 MHz antenna with at le
  6. My current "night time coding projects" are: (1) implementing AES128-CCM cipher for non-aligned frames. (2) Building a binary serial terminal app in pure, POSIX C for good-old command line shell. Why is it that all the binary serial terminals are married to one horrible OO graphical framework or another? I do enough hardware stuff during the day.
  7. Anything above 25 MHz is YMMV, but for these sorts of MCUs, usually the number one limiting factor is the speed of the flash memory. If you run the code from RAM, you usually stand a better chance at reliably overclocking.
  8. jpnorair

    libemb

    "libconio" is not the best thing to name a library What does the word "conio" mean?
  9. What size of components are you happily soldering with those, @@RobG?
  10. I assume you want good introductory books rather than masters-level textbooks. I learned RF & Communications in an academic setting, so I cannot help you much there. My favorite book in that genre is: Principles of Communications: Rodger E. Ziemer, William H. Tranter. It is excellent, but it's very mathy and best suited for coursework. However, I did need to teach myself antennas. I can recommend one for antennas: Antenna and EM Modeling with Matlab: Sergey Makarov. It does not go too deep into the theoretical calculus or the theoretical mathematic models used to describe antennas.
  11. I want to have a USB-UART bridge for the CC430, and it is cheaper, better, and more interesting for the hacker to have it be an MSP430 than a FTDI chip. The thing is, USB-UART bridge firmware does not take a lot of space or resources. So, I'm thinking what else I can put on there. FET is certainly an option. But since it is programmable, if you don't like the FET, you can reprogram the MSP to do something else.
  12. I did an audit of GoodFET about 2 years ago, plus I read the app notes on how the JTAG/SWD works, and it seems that it is definitely possible to port it into a small size. TI does not tend to ship tight code. I would guess that a CC430-only version, or even MSP430F5-only version, could be implemented in less than 16KB. Maybe even 8KB. Soon I will do an audit of TI's code. I will certainly post my results.
  13. An even better part is MSP430F5503, since it is just like a 5510 but without the ADC and some other features that would be irrelevant to a FET. I have an old CC430 + 5503 board design, where 5503 acts as a FET and USB bridge to the CC430. I haven't gotten to doing much of the programming work, but this EZ-FET lite will make it a lot easier. I think I'm going to open source the board design some time in the next few months, since I don't really have the time to do all the firmware myself. It would be nice to port Energia onto it, like with PanStamp.
  14. I would avoid the EM430Fx137RF900 kits, if only because the cheaper, smaller, and nicer PanStamp is around the corner. In the meantime, there is also the Wizzikit, which is easy to buy if you are one of our European friends. If not, just stick with the EXPCC430.
  15. jpnorair

    RF-BSL?

    Do you know OpenChronos? Check google, there are a ton of forks.
  16. Everything is RoHS in USA, also, but you can still buy leaded solder. For reworking and prototyping in general (but especially reworking) leaded solder is so, so much better than lead-free is.
  17. Thingsquare already supports this device, IIRC, so it should be in the fully FOSS Contiki repository sooner or later. Thingsquare will still send you the code, you just need to ask. Contiki was AVR-centric for a long time. Now it appears that TI is supporting it, so you can expect that support for MSP430 and CC parts (not so much CC430, which is even too resource-limited for uIP) will be pretty good. In full disclosure, I sort-of compete with Thingsquare, but not really. It doesn't mean I don't want you to get IP on an MSP430 if that's what you want, and Contiki/Mist is probably the
  18. I once wrote an I2C bridge in VHDL. It's very possible to do, of course, but it will probably take you enough time such that spending $800 will be less than the time cost. If you have already written the I2C slave interfacing code for the CC2541, and it is less than the limit, another option is simply to add another slave (which would be the MSP430) and just pipe whatever comes-in from the slave immediately to the MSP430 slave. That might be a day's work. One thing that sucks about 8051 is that they are all slightly different, and the decent compilers are very expensive. In the future,
  19. Someone told me this last time, so I cut a slot in the insulation (it was a flat cable). I didn't think to test if it would work with the ground line as well, but for single phase AC I bet there's a workaround regardless of the cable type. I'll look into it.
  20. Just thinking about this more: With appropriate circuitry, you could even power a device with the induced EMF. I think I'm going to try that sometime Q1 with one of my solar-powered wireless boards. Of course, I will remove the solar and replace with some rudimentary rectification and voltage limiting circuitry.
  21. If the AC current input correlates to the state of the system, then you can hack a solution by wrapping a helical coil around the power cord, and rectifying (and quite likely dividing as well) the induced voltage. Then you can use an ADC to capture the resultant waveform, and measure the amplitude. I've used this as a way to non-intrusively monitor a washing machine. Each stage of the cycle uses different current, which can be monitored and reported.
  22. The other obvious option is Maple (leaflabs.com), which is an Arduino derivative for STM32F10x. Otherwise, I use STM32's quite a lot. There are a ton of tools and resources for STM32, more even than for MSP430. Unfortunately, there is not a nice user forum for STM32 (maybe an opportunity for someone) despite the fact that there are more STM32s in the wild than all other ARM CM's combined. Anyway, Google is your friend. It won't take long. Just snoop around the web and come back when you have a more focused idea on what options you like. I'll tell you what I think.
  23. One bad assumption I made is that the power should be low. With 1mW, it requires a lot of finesse to get it. Sure, it's possible to find corner-cases, or to use more power. That was my mistake. Are you referring to the Synapse module? That's an 802.15.4 module. I know that technology well -- for the record, it uses MSK with a 4:32 spreading sequence. In most places it is limited to 1mW, hence the bad assumption from above. Repeat test with two nodes, at ground level, in an NLOS environment. In the dry, then in the rain, then with the target node con
  24. Implementing this is a lot more difficult than simply using the SX127x, which has some other built-in features for achieving long range. HopeRF sources Semtech components, or possibly pirates the HW -- I have no way to verify. Either way, SX127x might be used in a HopeRF module fairly soon. The SAW-based transceivers certainly do not supply coherent receivers, PSK, or integrated convolutional code, so you need to use a different device. Implementing RS can be done in firmware, but it is not easy. It is even less easy to do it fast in firmware.
  25. @@zeke You are free to write up a similar response. As you acknowledge, just saying "use this module" is not enough. You seem intent on proving that I am wrong, which is pretty common in discussions between engineers so I'm not terribly bothered. I am making three assumptions: - the user wants to achieve 300m with minimal other concerns. - I know more about RF communication than the user does. - I know more about RF communication than you do Maybe my assumptions are wrong. Don't take it the wrong way. 43oh.com isn't exactly a huge repository for knowledgeable RF communications e
×
×
  • Create New...