himanshubdave1426459935

Tested helloArdu.c with SPI_MODE_3 as suggested by Tripwire, works! Thanks. However, the helloRaspi.ino code does not work. I have two options: (i) use two versions of helloArdu.c (one with SPI_MODE_0 for Arduino as Slave and second with SPI_MODE_3 for MSP430 as Slave), or, (ii) Rewrite Arduino code using its SPI library where mode control is available. Both are acceptable as far as teaching is concerned. Thanks all for the help. In passing, I feel that with great popularity of Raspberry Pi -2 (and now -3), which unfortunately does not have ADC's, low-cost micro-controllers having ADC

Thanks a lot Spirilis and Tripwire for pointing out the Errata and other suggestions about ioctl(). I shall try out and report back. Yes, I would like to disturb the Arduino code as little as possible.

As a part of my teaching work, I wanted to demonstrate connecting Raspberry Pi-2 (as SPI Master) to MSP430 LP with MSP430G2553 (as SPI Slave), as an alternative to Arduino-Uno as a SPI Slave. It was aimed at utilizing the ADC's of either of these micro-controller boards, as Raspberry Pi-2 does not have ADC modules. The working of Arduino Uno as the SPI Slave device was smooth and painless, but I spent many hours getting the MSP430G2553 working as a SPI SLave. Ultimately, I hit upon a hack (if there was one) and MSP430G2553 started working as SPI Slave! However, I do not understand why this hac

Thanks a lot. In fact I did tried with -mfloat-abi=hard -mfpu=fpv4-sp-d16 , but did not realize that a different set of libraries should be used. I have found the suggested libraries in energia distribution.

In TivaC Workshop Lab9 on the topic Floating Point, it has been shown that the compuatation time for each iteration loop which contains a sine calculation is approxly 7 microseconds. I ran essentially the same program with Timer1 running in ONE-SHOT-UP mode at 50 MHz, which was started just before the sine calculation loop and the count was read out just after the loop was over. The value recorded was consistantly 171109 which gives total loop time of 342.2 microsec and each loop of 34 microsec. The make file used -mfloat-abi=softfp -mfpu=fpv4-sp-d16 compiler flags, Sourcery CB ar

The TM4C123GXL LaunchPad Workshop only contains Labs, suitable to run on CCSv5, as described in TM4C123GXL_LaunchPad_Workshop_ Workbook.pdf . For Linux users, the problem is that, tho Yough CCSv5 is avialable for Linux, but the flash-programmer does not work reliably for Stellaris/TivaC LaunchPads. I found it much more comfortable to use Sourcery CodeBench Lite (i.e. the Command Line Interface version) along with arm-gdb and OpenOCD, to build, run and debug various examples and Lab's. You can use lmicdi in place of openocd, but it will only allow flash-writing, not full use of debug

TM4C123GXL LaunchPad Workshop Lab7 (bulk transfer-mode USB) requires a client program to exchange text data with the LaunchPad loaded with the project code. TI has provided a client working on WINDOWS. Those working on Linux platforms require an implementation. I have written a "first-cut" version, source attached. It uses libusb. It does not implement the speed statistics calculations, though the required counters are there.usb_bulk_example.c

I have run and debgged all (expect one - where a WINDOWS client was needed) Labs for TivaC LaunchPad Workshop using command-line Sourcery CodeBench and OpenOcd, on Slackware Linux. The details can be given in this list if there is enough interest.