Question About Delay_ms

[jorgeal 0]

Hello, I´m a newie. I started with the proyects posted in section for beginners and I have the following problem:

 

I tested the proyect "binary progression" and all works ok in simulation, but when I wired in a proyect board,

the delays programmed run faster than should be. I measured about 200ms with my oscilloscope and the programed value was 500ms , its a big difference.

 

Later I changed delay_ms(500) for delay_s(2). In this case it works ok, ie programmed values are very similar to measured ones.

 

Is there a range values that could be used inside delays functions like delay_ms ??

 

any help?

 

thanks

Jorge

[Reynard 0]

delay_ms only takes an unsigned char value therefore should be in the range 1 - 255. Page 69 of manual.

 

To get 500ms call the delay_ms(250) twice.

 

Cheers

 

Reynard

[jorgeal 0]

Thanks Reynard, I suspected it!.

So there are in the Forum at least 3 or 4 wrong examples using delay_ms(500).

The corious thing is that they work fine with the simulation software without any warning or error message.

 

bye.

 

Jorge

[Pavel 0]

... So there are in the Forum at least 3 or 4 wrong examples using delay_ms(500).

The corious thing is that they work fine with the simulation software without any warning or error message...

 

You'll get a warning if you use -W2 compiler command line option (all warnings). And for simulator there is nothing to report because by the time code gets executed under sim it already uses a converted value.

 

Regards,

Pavel

[Reynard 0]

Hi Jorge,

 

As Pavel says, if you turn on 'All Warnings' you would have gotten a warning informing of loss of data.

 

You can use the command line setting or in the IDE use 'Settings -> Options...' then click All Warnings.

 

The compiler demoted your value of 500 by truncating to the last 8 bits leaving you with 0xF4 or 244ms as you got.

 

All Warnings is good for beginners as it can pick out simple unintended errors of loss of data etc. It is also useful for experienced programmers to run their code through the All Warnings occasionally. We all have senior moments.

 

Cheers

 

Reynard

[jorgeal 0]

good, I did it and now I can see the warning.

Thanks for your time.

[russ_hensel 0]

Most ( all? ) times you use a variable that is too small for a value the value will "roll over" 255 + 1 = 0, etc. You can use a bigger variable type, but if the function does not

honor it you will still not get what you want. Sticking to unsigned char gives the fastest code, but the most limited.

 

 

 

good, I did it and now I can see the warning.

Thanks for your time.