View previous topic :: View next topic |
Author |
Message |
micro_debugger
Joined: 08 Oct 2009 Posts: 73
|
|
Posted: Sun Nov 03, 2019 1:26 pm |
|
|
Byte is 500 ns, not bit, sorry my mistake, but with gap of 1us between each byte, and sending of 7 bytes per each write, we get the 10us write time per one byte of data. We have 7x500 ns for data + 6x 1000 ns for gaps = 10 us per each byte written to the FRAM.
The 15 MHz is the limitation of the dsPIC I'm using, can not be faster.
Check also the other functions, to see if they are faster. |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19539
|
|
Posted: Mon Nov 04, 2019 1:47 am |
|
|
You will find spi_read and spi_write are significantly faster.
I think I'll actually have a word with CCS about this. Basically spi_xfer
always has a wrapper round it to support the multi-byte transfers. Really
this should be removed by the optimiser when single byte operations
are used. |
|
|
micro_debugger
Joined: 08 Oct 2009 Posts: 73
|
|
Posted: Mon Nov 04, 2019 3:04 am |
|
|
I'm already at single byte write at 4.9 us, read of single byte 4.1 us, and block of 21 bytes read 19us, so it is acceptable with my project requirements.
More will write you later on and release the final library to community.
However it is still expensive pleasure to use such FRAM as it cost about 25 EURO per IC, so only on dedicated application can be used. |
|
|
Ttelmah
Joined: 11 Mar 2010 Posts: 19539
|
|
Posted: Mon Nov 04, 2019 7:13 am |
|
|
It's a high capacity device, so expensive.
They drop quite a lot once you buy in reasonable quantity though, so
with a bit of searching around you ought to be able to do quite a bit
better. |
|
|
|