Representation of 24-bit and 40 bit values

RE: Representation of 24-bit and 40 bit values

awneil — Tue, 28 May 2019 22:27:14 GMT

[quote userid="74350" url="~/f/nordic-q-a/47739/representation-of-24-bit-and-40-bit-values/189720"]do you mean to define it as a structure[/quote]

Or an array.

RE: Representation of 24-bit and 40 bit values

DeveloperZ — Tue, 28 May 2019 21:46:27 GMT

[quote userid="6462" url="~/f/nordic-q-a/47739/representation-of-24-bit-and-40-bit-values/189717"]but you could use 3 bytes, if you like.[/quote]

By this, do you mean to define it as a structure as he said above or is there any other way to define a variable of size 24-bits?

RE: Representation of 24-bit and 40 bit values

awneil — Tue, 28 May 2019 21:25:06 GMT

[quote userid="73680" url="~/f/nordic-q-a/47739/representation-of-24-bit-and-40-bit-values/189710"]15-bit; should I use 16-bit[/quote]

Well, you can't use less than 16.

Being a 32-bit architecture, using 16 against 32 may not gain you much.

[quote userid="73680" url="~/f/nordic-q-a/47739/representation-of-24-bit-and-40-bit-values/189710"]for 24-bit; should I use 32-bit?[/quote]

That'll be easiest - but you could use 3 bytes, if you like.

[quote userid="73680" url="~/f/nordic-q-a/47739/representation-of-24-bit-and-40-bit-values/189710"]40-bit; should I use 64-bit[/quote]

Again, 64 will be easiest, but 5 bytes would do it.

[quote userid="73680" url="~/f/nordic-q-a/47739/representation-of-24-bit-and-40-bit-values/189710"]and specify the limits[/quote]

Not sure what you mean by that?

[quote userid="73680" url="~/f/nordic-q-a/47739/representation-of-24-bit-and-40-bit-values"]should I define the limits in the stdint.h [/quote]

You should not fiddle with stdint.h.

If you want custom types, then create a custom header...

RE: Representation of 24-bit and 40 bit values

Aryan — Tue, 28 May 2019 18:20:41 GMT

Thanks for the detailed explanation.

For the other sizes like 15-bit; should I use 16-bit?

for 24-bit; should I use 32-bit?

for 40-bit; should I use 64-bit?

and specify the limits?

RE: Representation of 24-bit and 40 bit values

ovrebekk — Mon, 27 May 2019 11:00:58 GMT

Hi Aryan

You can't create variables with a different size than the basic types (8, 16, 32 or 64), but you can create structs that contain arbitrarily sized fields (up to 64 bits maximum) by using bit fields.

For the 76-bit value you mention this would look something like this:

typedef struct __PACKED
{
uint32_t year : 7;
uint32_t month : 12;
uint32_t day : 5;
uint32_t hour : 5;
uint32_t minute : 6;
uint32_t second : 6;
uint32_t day_of_week : 7;
uint32_t action : 4;
uint32_t transition_time : 8;
uint32_t scene_number : 16;
} schedule_register_t;

The __PACKED tag ensures that the struct is byte aligned, rather than 16 or 32 bit aligned, to keep the size in memory as small as possible.

Please be aware that the size of the struct will be 80 bits (10 bytes), and not 76, since the minimum alignment for any struct or variable is 1 byte.

Best regards
Torbjørn