Who can help me SPI EEPROM on PIC24 series

[sindyme]

Hello everybody, I have question about SPI EEPROM on PIC2416KA101.
My PIC2416KA101 can't work on SPI EEPROM.
This is my C code. Who can help me solve this question ? I'm very thankful for any help.

[FvM]

CCS built-in functions have several issues with PCD/PIC24. They are generally working, but often not as
expected. The topic has been addressed in the forum before. I suggest to read also this older thread: (It's a
rather long story, and I used a wrong SPI mode coding in part of the contributions, be aware).
http://www.ccsinfo.com/forum/viewtopic.php?t=38491
There'a a short snippet in the post for reading the SPI-Flash ID-Byte. If it works, you're nearly done.

I have several problems with your post. You don't tell about the SPI flash device, and you don't tell an explicite
reference for the used flash routines. If you are using PCM programmer's "hardware SPI" driver from the second thread,
it doesn't work with PCD, I think. As discussed in the above thread, the combination of spi_write() and spi_read()
doesn't work correctly, you should use spi_read() exclusively (that's my simple solution for the time being, there
are possibly others).

With industry standard SPI-Flash (I used e.g. ST/Numonyx M25P04), mode 0 or mode 3 are correct, mode 1
or 2 can't work.

P.S.: I realized, that you refered to SPI EEPROM rather than SPI flash. They are similar in most aspects, e.g.
SPI modes. But the said read ID command doesn't exist.

[FvM]

You may want to read the above post, particularly regarding mixing of spi_write() and spi_read().

I didn't yet use, spi_xfer(), so I can't tell about it. Also I didn't use V4.104, because the compiler temporarily
lost it's knowledge of some basic C syntax elements in this version. But I'm not aware of changes to the SPI
handling in V4.104.

[sindyme]

Hello Fvm

I use Logic Analyzer to check my SPI signal.
I find the "PIC24F16KA101" only send "SOD signal out" , it no have run spi_read function.
How can i do ...

There is my SPI logic image URL :
http://sindyme.sg1010.myweb.hinet.net/logic_spi.bmp

[FvM]

With a LA, you should be able to clarify the issue in a short. Can you please also tell, which code lines are
generating the shown 4 Byte transfers and what's the assignment of LA channels. SCK obvious, nCS too, the
others aren't clear to me.

[sindyme]

I update my image URL :
http://sindyme.sg1010.myweb.hinet.net/logic_spi.bmp

PS : With Logic Analyzer that I find CCS no run spi_read function.
When I only put "write_ext_eeprom" function at main(), my L.A.
will run logic signal.
If I only put "read_ext_eeprom" function at main(), my L.A
will to wait. ( PIC no run spi_read function .... )


I use these code to test SPI EEPROM

[FvM]

I told you, that a combination of spi_write() and spi_read() doesn't work with PCD. I only forgot, that it can cause
spi_read() to never return.

I also told you, that my SPI mode coding from the begin of the said thread was wrong and has been corrected
at the end.

As your waveform reveal, you actually selected mode 0 instead of intended mode 3. Fortunately, the SPI EEPROM supports both modes. For clarity, I inserted the correct mode moding below.

To avoid the spi_write()/spi_read() problem, simply replace all occurences of sp_write() with spi_read(). (It writes as well, but waits for the previous SPI transaction to finish before.)

[sindyme]

Hello Fvm, I'm very thanks for you.
My SPI can work now.
I use this code control 25LC series EEPROM & "16M" SPI Flash.
This is "16M SPI Flash" URL:
http://www.spansion.com/Products/Pages/ProductDetails.aspx?ProdID=S25FL128P
The size is very very very big for me.

There is my working code. Anyone can use it to test SPI EEPROM.

[sindyme]

Hello Fvm and everybody

I still had the question, my "PIC24F16KA101" already work on SPI.

Now I change "PIC24F16KA101" to "PIC24FJ256GB106" but same code can't work on "PIC24FJ256GB106".

I have note "RPI Select" that I use #pin_select to set my "SDO, SDI and SCK". But "256GB106" still can't work.

And I have use logic analyzer to check SDO, SDI signal.
I find the signal clock have wrong.
My code send 0x05 -> 0x06 -> 0x02 -> 16bit address and data. but the clock have wrong.
There is logic analyzer image for SPI URL :
http://sindyme.myweb.hinet.net/logic_spi2.bmp

and this is my code for "PIC24FJ256GB106"

[FvM]

You are using spi_write() again. It doesn't work.

[bkamen]

And you can't use SPI_Write or XFER and immediately raise nCS.

Think about it:

You write the register:

SPIBUF = some value

Then raise the nCS line:
bset reg.bit

How long do you think it takes at 8 SPI clocks vs 1 instruction? Who gets there first.

Probably your PIC's PC counter wins and you raise nCS before SPIBUF is fully sent.

You need some sort of delay before raising nCS to idle.

-Ben
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D

[FvM]

[sindyme]

[FvM]

[bkamen]

It *is* pretty funny that a "read" function writes data.

Hahaha..

-Ben

EDIT: ok. Maybe not so funny.
_________________
Dazed and confused? I don't think so. Just "plain lost" will do. :D