Hi Mike, Yes, the BoilerPlate magnetometer data is fine before adding the extra byte. I assume the magnetometer just stays on its default settings (10Hz, 1Gauss etc), but i don't see the writeDone event corresponding to the runContinuousConversion call. I'm using BoilerPlate here as the exemplar, but in my own apps, I need to be able to write to the magnetometer registers (and receive the writeDone events) in order to change it's settings (gain, update rate etc), and I appear to have lost this ability after upgrading the toolchain and TinyOS tree. Once I start to delve into the platform chip code, I realise I'm quickly getting to the limits of my abilities with TinyOS programming, although I seem to learn more each time I do! I'm using msp430-gcc version 4.6.3. Chris
Date: Tue, 30 Oct 2012 09:12:57 +0000 Subject: Re: [Shimmer-users] yet another toolchain update From: [email protected] To: [email protected] CC: [email protected]; [email protected] Hi Chris, Does the data from the magnetometer look OK to you before you stuff the extra byte in the write packet (despite the fact that you are not seeing the writeDone event)? I have been building BoilerPlate with msp430-gcc version 4.6.3 for months now and I haven't noticed any problem with the mag data. Mike On Mon, Oct 29, 2012 at 4:47 PM, Chris B <[email protected]> wrote: hi after shelving this for a while, i decided to go back and try to solve it. To recap, after updating my toolchain, TinyOS and contrib trees (on the Shimmer-supplied Ubuntu VMWare image) i can no longer achieve Magnetometer.writeDone events. E.g. I've taken the unadulterated BoilerPlate app, stuck an Leds.led0On() before the Magnetometer.runContinuousConversion() call, and an Leds.led0Off() in the Magnetometer.writeDone event. For some reason, the led off call is not reached. Looking at the Log message for SVN revision 5959, it seems the update to msp430-gcc-4.x caused anomaly with the bmp085 pressure sensor i2c writes in the writeReg function, which was solved by stuffing an extra byte into the write packet. I tried doing the same with the corresponding writeRegValue function for the hmc5843 magnetometer, and the Magnetometer.writeDone events are now achieved, but the output data looks dodgy (viewed in the ShimmerConnect v0.2 software the data looks noisy, and goes out-of-scope on the plots). this is where i've gotten stuck again. Any help would be appreciated. regards Chris > Date: Fri, 27 Jul 2012 12:52:30 -0400 > From: [email protected] > To: [email protected] > Subject: Re: [Shimmer-users] yet another toolchain update > > hi chris, > > sorry, i have no idea. > > best baseline sanity check: i have a simple test app, which i will > release later on this afternoon (must run right now), that may help. > run justmag from the -contrib tree. if it works and your code doesn't, > then you have an app bug. > > i'll mail back when i get it out there. > > wish i had better news, > > steve > > On 07/27/2012 12:06 PM, Chris B wrote: > > steve, > > > > thanks for the updates. > > > > Possible breakage... > > > > I don't seem to get Magnetometer.writeDone events anymore, which i use > > in some of my own apps to figure out when it's safe to write to the > > register again e.g. when calling runContinuousConversion, setOutputRate > > or setGain in succession. Checked in contrib BoilerPlate by sticking an > > Led On call there. Any ideas? > > > > regards > > > > Chris > > > > > > > > > Date: Tue, 17 Jul 2012 10:31:57 -0400 > > > From: [email protected] > > > To: [email protected] > > > Subject: [Shimmer-users] yet another toolchain update > > > > > > hi folks, > > > > > > during testing i found that the new toolchain broke some i2c drivers > > > (even the msp430 i2c driver itself). while i did push out fixes > > > previously, i missed changes that i made to the magnetometer driver for > > > the hmc5843; this chipset is on the shimmer "9dof" board, so these > > > changes will interest those users. > > > > > > please pick up these updates whether or not you have upgraded your > > > compiler toolchain: > > > > > > cd ~/tinyos-main/tos/platforms > > > > > > svn update > > > > > > you'll see (if you have followed maintenance suggestions and all else is > > > current) new files pulled into > > > > > > shimmer2/chips/hmc5843 > > > > > > please note that this affects two apps in -contrib as well, since the > > > magnetometer event prototypes are now async. i made these changes in > > > the released versions, so you can update these as well: > > > > > > cd ~/tinyos-2.x-contrib/shimmer/apps > > > > > > cvs update > > > > > > you'll see updates to boilerplate and gyromag. > > > > > > thanks to mike healy for finding this oversight. please keep posting > > > when you find broken things! > > > > > > -steve > > > _______________________________________________ > > > Shimmer-users mailing list > > > [email protected] > > > https://lists.eecs.harvard.edu/mailman/listinfo/shimmer-users _______________________________________________ Shimmer-users mailing list [email protected] https://lists.eecs.harvard.edu/mailman/listinfo/shimmer-users
_______________________________________________ Shimmer-users mailing list [email protected] https://lists.eecs.harvard.edu/mailman/listinfo/shimmer-users
