|
Colleagues, My PIC has to talk to the GPS rerceiver, and GPS transmitts data as IEEE 754 floats. I'm trying to write conversion routine in C (CCS compiler), but it does not work. Could you post a working conversion routine? Thanks! Sincerely, Nick |
|
|
Floating point IEEE 754 to Microchip conversion. SOS!!!
Started by ●September 1, 2003
Reply by ●September 2, 20032003-09-02
|
try to use union, like this: union { float f; unsigned char c[4]; } UN; UN un; un.c[0] = 0x??; // LSB un.c[1] = 0x??; un.c[2] = 0x??; un.c[3] = 0x??; // MSB printf ("Float = %f ", un.f); ----- Original Message ----- From: "Nick Alexeev" <> To: <> Sent: Monday, September 01, 2003 9:17 PM Subject: [piclist] Floating point IEEE 754 to Microchip conversion. SOS!!! > Colleagues, > > My PIC has to talk to the GPS rerceiver, and GPS transmitts data as > IEEE 754 floats. I'm trying to write conversion routine in C (CCS > compiler), but it does not work. Could you post a working conversion > routine? > > Thanks! > > Sincerely, > Nick > > to unsubscribe, go to http://www.yahoogroups.com and follow the instructions > > ">http://docs.yahoo.com/info/terms/ > > Esta mensagem foi verificada pelo E-mail Protegido Terra. > Scan engine: VirusScan / Atualizado em 28/08/2003 / Vers: 1.3.13 > Proteja o seu e-mail Terra: http://www.emailprotegido.terra.com.br/ > |
|
|
Reply by ●September 2, 20032003-09-02
|
--- In , "Nick Alexeev" <kender_a@y...> wrote: > Colleagues, > > My PIC has to talk to the GPS rerceiver, and GPS transmitts data as > IEEE 754 floats. I'm trying to write conversion routine in C (CCS > compiler), but it does not work. Could you post a working conversion > routine? > > Thanks! > > Sincerely, > Nick If you find out what the Microchip FP format is and post it here I will try to point you in the right direction regarding converting from IEEE 754 to Microchip format. XCSB deals with IEEE 754 format floating point numbers as standard. Regards Sergio Masci http://www.xcprod.com/titan/XCSB - optimising structured PIC BASIC compiler |
|
|
Reply by ●September 2, 20032003-09-02
|
Hi Marcio, Thank you for your input. Unfortunately the problem is more complex. IEEE 754 float and Microchip float use different formats to store the fraction and exponent. Micorchip's app note AN575 explains the difference (page 3), but I was not smart enough to write a working conversion routine. Thanks again! Nick P.S. App note can be found at http://www.microchip.com/download/appnote/library/00575.pdf ** END --- In , "Marcio" <mzaquela@t...> wrote: > try to use union, like this: > > union > { float f; > unsigned char c[4]; > } UN; > > UN un; > > un.c[0] = 0x??; // LSB > un.c[1] = 0x??; > un.c[2] = 0x??; > un.c[3] = 0x??; // MSB > > printf ("Float = %f ", un.f); > ----- Original Message ----- > From: "Nick Alexeev" <kender_a@y...> > To: <> > Sent: Monday, September 01, 2003 9:17 PM > Subject: [piclist] Floating point IEEE 754 to Microchip conversion. SOS!!! > > > Colleagues, > > > > My PIC has to talk to the GPS rerceiver, and GPS transmitts data as IEEE 754 floats. I'm trying to write conversion routine in C (CCS > > compiler), but it does not work. Could you post a working conversion routine? > > > > Thanks! > > > > Sincerely, > > Nick > > > > > > |
Reply by ●September 2, 20032003-09-02
|
On Tuesday 02 Sep 2003 1:17 am, Nick Alexeev wrote: > Colleagues, > > My PIC has to talk to the GPS rerceiver, and GPS transmitts data as > IEEE 754 floats. I'm trying to write conversion routine in C (CCS > compiler), but it does not work. What are you trying to convert them TO? Ian |
Reply by ●September 2, 20032003-09-02
|
I'll e-mail you the format shortly - as my compiler help describes
it. byte1 EEEEEEEE byte2 SFFFFFFF byte3 FFFFFFFF byte4 FFFFFFFF Where E is exponent, S is the sign bit, F is fraction. Nick > > If you find out what the Microchip FP format is and post it here > I will try to point you in the right direction regarding converting > from IEEE 754 to Microchip format. > > XCSB deals with IEEE 754 format floating point numbers as standard. > > Regards > Sergio Masci > > http://www.xcprod.com/titan/XCSB - optimising structured PIC BASIC > compiler |
Reply by ●September 2, 20032003-09-02
|
According to Microchip App Note AN575 page 3, the Microchip FP format is: (big endian destination) e7 e6 e5 e4 e3 e2 e1 e0 s0 f22 f21 f20 f19 f18 f17 f16 f15 f14 f13 f12 f11 f10 f9 f8 f7 f6 f5 f4 f3 f2 f1 f0 (little endian destination) f7 f6 f5 f4 f3 f2 f1 f0 f15 f14 f13 f12 f11 f10 f9 f8 s0 f22 f21 f20 f19 f18 f17 f16 e7 e6 e5 e4 e3 e2 e1 e0 IEEE 754 single format (32 bit) is defined as: (big endian source) s0 e7 e6 e5 e4 e3 e2 e1 e0 f22 f21 f20 f19 f18 f17 f16 f15 f14 f13 f12 f11 f10 f9 f8 f7 f6 f5 f4 f3 f2 f1 f0 (little endian source) f7 f6 f5 f4 f3 f2 f1 f0 f15 f14 f13 f12 f11 f10 f9 f8 e0 f22 f21 f20 f19 f18 f17 f16 s0 e7 e6 e5 e4 e3 e2 e1 The thing that might faul you up is the order in which the source is retrived from the GPS and stored in memory. f7..f0 could be stored in byte 0 (little endian) or in byte 3 (big endian) to simplify things store it using the same convention as MC uses to store FP So to convert IEEE 754 to MC big endian conversion * take bytes 1, 2 and 3 of the source and store them in the destination. * take bit 7 of byte 0 (s0) of the source and store it in bit 7 of byte 1 of the destination. * take byte 0 of the source, shift it one place left and store it in byte 0 of the destination. * take bit 7 of byte 1 (e0) of the source and store it in bit 0 of byte 0 of the destination. little endian conversion * take bytes 0, 1 and 2 of the source and store them in the destination. * take bit 7 of byte 3 (s0) of the source and store it in bit 7 of byte 2 of the destination. * take byte 3 of the source, shift it one place left and store it in byte 3 of the destination. * take bit 7 of byte 2 (e0) of the source and store it in bit 0 of byte 3 of the destination. Hope this helps Regards Sergio Masci http://www.xcprod.com/titan/XCSB - optimising structured PIC BASIC compiler ----- Original Message ----- From: Nick Alexeev <> To: <> Sent: Wednesday, September 03, 2003 2:15 AM Subject: [piclist] Re: Floating point IEEE 754 to Microchip conversion. SOS!!! > I'll e-mail you the format shortly - as my compiler help describes it. > > byte1 EEEEEEEE > byte2 SFFFFFFF > byte3 FFFFFFFF > byte4 FFFFFFFF > > Where E is exponent, S is the sign bit, F is fraction. > > Nick > > > > > If you find out what the Microchip FP format is and post it here > > I will try to point you in the right direction regarding converting > > from IEEE 754 to Microchip format. > > > > XCSB deals with IEEE 754 format floating point numbers as standard. > > > > Regards > > Sergio Masci > > > > http://www.xcprod.com/titan/XCSB - optimising structured PIC BASIC > > compiler > > to unsubscribe, go to http://www.yahoogroups.com and follow the instructions > > ">http://docs.yahoo.com/info/terms/ > |
