Anyone used the Bosch BMP085 I2C baro sensor yet ?

[mr.sneezy]

Hmm...
I've written all the 32Bit calculation code now using PBPL and Longs, where they are needed (well nearly everywhere).
Right away I got a list of very strange compiler errors like this
c:/pbp/pbppi18l.lib 695 : Argument out of range ( 4168 not between 0 and 4095)

After having a look at line 695 of that file I still had no clue. So I took a punt that I might have run out of PIC code space, and yes that seems to be right. If I select a PIC in the same sub family with more code space the code complies error free....
Longs sure burn code space like it was endless resource.
Off to find a another bigger PIC again. This will be the 4th PIC tried for this project.


Onward and upwards !!

PS. Is there a simple to use utility, MCS add-on or method that will show me the amount of RAM space my code is using when compiled (as well as program space) ?

[Acetronics2]

Hmm...
I've written all the 32Bit calculation code now using PBPL and Longs, where they are needed (well nearly everywhere).
Right away I got a list of very strange compiler errors like this
c:/pbp/pbppi18l.lib 695 : Argument out of range ( 4168 not between 0 and 4095)

After having a look at line 695 of that file I still had no clue. So I took a punt that I might have run out of PIC code space, and yes that seems to be right. If I select a PIC in the same sub family with more code space the code complies error free....
Longs sure burn code space like it was endless resource.
Off to find a another bigger PIC again. This will be the 4th PIC tried for this project.


Onward and upwards !!

Hi,

Looks linked to too small a program memory ... in 18 pins ... the 1320 exists in 8k ...

Alain

[mr.sneezy]

Hi Alain,

Yes, it compiles OK if I select that one, at about 5200 bytes.

As I have to buy the next size PIC I might as well jump to the 18F2520 with huge code space and 28pins, although I really am only using a few I/O at the moment. Going to be lots of unused I/O's.

Martin

[ScaleRobotics]

PS. Is there a simple to use utility, MCS add-on or method that will show me the amount of RAM space my code is using when compiled (as well as program space) ?

Nice work on the sensor code!

MCS shows you how much code is being used at the bottom of the page after compliling. (Note it does need a successful compile to be able to tell you this.)



You can also peek in the .lst file, and find code used vs left.

MEMORY USAGE MAP ('X' = Used,  '-' = Unused)


0000 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
0040 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
0080 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
00C0 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
0100 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
0140 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
0180 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXX-------
0200 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
0240 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
0280 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX
02C0 : XXXXXXXXXXXXXXXX XXXXXXXXXXXXXXXX ---------------- ----------------
2000 : -------X-------- ---------------- ---------------- ----------------

All other memory blocks unused.

Program Memory Words Used:   665
Program Memory Words Free:   359

Then there is also Darrel's CodeSpace utility here:
http://www.picbasic.co.uk/forum/showthread.php?t=2418

[mr.sneezy]

Nice work on the sensor code!

Thanks, although it's not complete just yet. I'll show you the main sensor code snip in case you are interested. It complies now (if selecting a bigger PIC) and I know the result of the lCTemp section is 100% correct (+ & - Deg C ). The UPres calculations are as of right now untested.
If you see any blindingly obvious code bloat in the calculations let me know...
I will post the complete code as soon as I know it works fine.

I did know what the code size used was telling me in MCS, I was more curious about RAM (variables etc) as I have been caught before going to a PIC with more code space, only to be quickly stumped by running out of variables...

All variables are signed Longs, except MD, AC5 and AC6 which are Words.

'Calculate temperature in 1/10ths of Deg C  from lUTemp     ' Note 2^15 = 32768 Dec or $8000
        X1 = ((lUtemp - AC6) * AC5) / $8000     'find X1 
        X2 = (lMC << 11) / (X1 + MD)            'Find X2
        B5 =  X1 + X2                           'Find B5 from X1 and X2
        lCTemp = (B5 + 8) / 16                  'Hey presto, lCTemp appears... 
          
'Calculate pressure in Pa from lUpres (1/100th's of hPa's)        
        B6 = b5 - 4000
        x1 = (b2 * (B6 * B6 / $1000)) / $800 
        x2 = (lac2 * B6) / $800
        x3 = x1 + x2
        B3 = ((lac1 * 4 + x3) << 5) / 4         'My OSS = 3 (might need to * here not <<)
        x1 = (lac3 * b6) / $2000
        x2 = (b1 * (b6 * b6 / $1000)) / $10000
        x3 = ((x1 + x2) + 2) / 4
        B4 = (lac4 * (x3 + 32768)) / $8000
        B7 = (lUPres - B3) * 50000
        If B7 < $80000000 then 
            lPres = (B7 * 2) / B4
        Else
            lPres = (B7 / B4) * 2
        Endif
        X1 = (lPres / 256) * (lPres / 256)
        X1 = (X1 * 3038) / $10000
        X2 = (-7357 * lPres) / $10000
        lPres = lPres + ((X1 + X2 + 3791) / 16)

        Serout2 SO,16780,["T=",DEC lCtemp,"    "]       'Send Word size number to LCD
        Serout2 SO,16780,["P=",DEC lPres,"     "]      'Send Word size number to LCD