Multi-digit word variable to string

[richard]

Now i'm trying to figure out how to add the remainder in the command.

i hope this is not a medical diagnostic device
your temperature calculation is not accurate and the remainder calculation is mathematically incorrect

temperature  = tempir/50         ' based on the manual we need to divide by 50 or 0.02
temperature  = temperature - 273 ' subtract 273 for Degree in C,actually is 273,15

yields the temp in deg C to about 1 deg of resolution with an error of +- 1 degree

remainder    = (tempir*2)//100   ' multiply x 2 and get the remainder //100

calculated the 100 modulus of the temp in deg kelvin

adding them together is nonsense , trying to display a two decimal point result is from any of these results is just not right

wtf tempir = (temp.highbyte << 8) + temp.lowbyte ;just Nooooo! its the hardest most convoluted way i have ever encountered to go 
tempir = temp which serves no purpose anyway




don't like this either, it pisses away too much resolution hence its not particularly accurate 
temperature  = tempir/50         ' based on the manual we need to divide by 50 or 0.02
temperature  = temperature - 273 ' subtract 273 for Degree in C,actually is 273,15
remainder    = (tempir*2)//100   ' multiply x 2 and get the remainder //100


;this is good  although it could be even more accurate [temperature  = temp*2  - 27315 ; temp x 2 - 273.15*100 ] >==>temperature*100] if the raw value won't exceed 32768 


temperature  = (temp  - 13657)<<1  ; (temp-[273.15*50]) x 2   <==> temperature*100; is good




i make the assumption temperature * 100 is the value required by the slider display


tempir and remainder are not needed

display like this

'------------------------------------------------------------------------------*'                    IR  Display 1 the parameters                              *
'------------------------------------------------------------------------------*
display: ' IR TEMP 
hserout2 [$73,$00,$02,$12,$00,$ff,"IR TEMP:",00]
Hserin2  timeout,error,[wait(6)]
hserout2 [$73,$08,$02,$12,$F8,$00,dec2 temperature/100,".",dec2 temperature//100,"*C",00] 
Hserin2  timeout,error,[wait(6)]
 


gosub checksum2    
hserout [$01,$0F,$00,temperature.highbyte,temperature.lowbyte,IRCS]   
Hserin  timeout,error,[wait(6)]
return


'------------------------------------------------------------------------------|
'                               CHECKSUM IR                                    |
'------------------------------------------------------------------------------|
checksum2:
IRCS = 0
IRCS = $01 ^ $0F ^ $00 ^ temperature.highbyte ^ temperature.lowbyte 
return

then lets have a look at the i2c stuff , its not right either really


ps if negative temps are possible then additional code req

[astanapane]

Richard,

really thanks for your time.

I found this,if i remember correctly somewhere related to this specific sensor.

tempir = (temp.highbyte << 8) + temp.lowbyte

Then in the equation i have added the following

temperature  = tempir/50         ' based on the manual we need to divide by 50 or 0.02
temperature  = temperature - 273 ' subtract 273 for Degree in C,actually is 273,15

Your way is much more efficient and works just as expected. Im very happy.

Now what is wrong with the I2C command?

[richard]

I found this,if i remember correctly somewhere related to this specific sensor.


Code:
tempir = (temp.highbyte << 8) + temp.lowbyte

maybe but in conjunction with
i2cread SDA, SCL,addr,reg,[temp.lowbyte,temp.highbyte] ' reg = $07

its not useful in any way

Now what is wrong with the I2C command?

all this achieves is exactly nothing

reg    = $07             ' based on the datasheet $07 is the object's temp


grave doubts about all of this  it looks like some mishmash of a C code method where no read/write register command exists as it does in pbp 
addr   = %10110100       ' $B4  binary value: %10110100
i2cwrite SDA, scl, addr  ' send write command, shift the address B4, 8 bits where LSB is W = 0
i2cwrite sda, scl, reg   ' register is $07


i2cwrite sda, scl, addr,reg, [some data of some sort]  ' register is $07 ;would be the normal way if you had any data to write 


totally bogus in pbp bit 0 should always be 0
addr   = %10110101       ' $B5  Binary value: %10110101


a form of torture to i2c chips with an incomplete transaction
i2cwrite sda, scl, addr  ' sends read command, shift the address B5, 8 bits where LSB is R = 1, clock idles low

all that's needed
reg = $07
addr = %10110100
i2cread SDA, SCL,addr,reg,[temp.lowbyte,temp.highbyte] ' reg = $07 ;would be the normal pbp way

[astanapane]

Richard,

now i got a bit comfused. If i only add the following, it doent work as expected

all that's needed
reg = $07
addr = %10110100
i2cread SDA, SCL,addr,reg,[temp.lowbyte,temp.highbyte] ' reg = $07 ;would be the normal pbp way

But if i add the write mode and I2Cwrite command it works.

reg = $07 ' this is the register for IR temp mode
addr= %10110100 ; this is the write mode

i2cwrite SDA, SCL, addr 
i2cread SDA, SCL, addr, reg, [TEMP.lowbyte, TEMP.highbyte]

the following also works.

reg = $07 ' this is the register for IR temp mode
addr= %10110100 ; this is the write mode
i2cwrite SDA, SCL, addr 

addr= %10110101 ; this is the read mode
i2cread SDA, SCL, addr, reg, [TEMP.lowbyte, TEMP.highbyte]

if i do not add the following line the code doesnt work.

i2cwrite SDA, SCL, addr

Shouldnt we write first and then read? ok i understand that we have nothing to write. But why it doent work if i do not add the i2cwrite line?

[richard]

Shouldnt we write first and then read? ok i understand that we have nothing to write. But why it doent work if i do not add the i2cwrite line?

which is precisely what the i2cread command does for you automatically. it should not ever be necessary to add a extra write , no i2c chip i have ever encountered has needed such treatment

i2cwrite SDA, SCL, addr

is an incomplete transaction. if your chip needs it then it is not a true i2c device or something else is occurring.
a logic analyzer would show what's really happening. its not normal and should be examined further to have any confidence
that the code is really viable in the long term

addr = %10110100

trying to set bit0 [the r/w bit] of a i2c address with pbp's i2c commands is not possible and will not have any effect
the code will set the r/w bit as needed

[richard]

it could be the chip is not really i2c and behaves differently

try this, at least its a nicer transaction
i2cwrite SDA, SCL, addr,[7]
i2cread SDA, SCL, addr,[TEMP.lowbyte, TEMP.highbyte]



looking at the wiki link you posted, this would do exactly what the smbus timing diagram describes as a proceedure

PEC var byte
i2cread SDA, SCL,addr,reg,[temp.lowbyte,temp.highbyte,PEC] ' reg = $07 ;would be the normal pbp way

worth trying

[astanapane]

Hi Richard,

sure i'll try that, and i will try to figure out what is going on with a logic analyzer.

I have a saleae logic 8.

Will try that later today. Many thanks for your time once again.

[richard]

If there is only one MLX90614 sensor, the 7-bits address is 0x00 by default.

according to wiki addr =0 where did you get B4 from ?

[Ioannis]

Richard,

Then in the equation i have added the following

temperature  = tempir/50         ' based on the manual we need to divide by 50 or 0.02
temperature  = temperature - 273 ' subtract 273 for Degree in C,actually is 273,15

Well, if that compiler could do floating, maybe it could be used. But in PBP integer math? I do not think so. You need what Richard did as a workaround.

Ioannis