Hi Gaurav More,
Thanks for your reply.
According to the AN11508, the clock configuration supports two kinds of modes, and they are select freely the developer.
For the MCB4357 board, it uses the EMC_CLK0 to drive the SDRAM MT48LC4M32B2P when the EMC_CLK0 pin works as the function 0, in addition, the EMC_FBCLK0 feature is enabled too.
I've attached the schematic about the MT48LC4M32B2P connection, and the related demo project, please refer to it for details.
Have a great day,
TIC
-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!
- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------
Hi Gaurav More,
Thanks for your reply.
1) I am not able to identify the code section where EMC_FBCLK0 is enabled in the code that you shared.
-- Please check the below figure.
Have a great day,
TIC
-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!
- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------
Hi Gaurav More
Thanks for your reply.
I've done some testing with your code in the periph_memtest demo, however, in debug, I don't encounter the similar phenomenon you mentioned before,
My board is MCB4357, IDE: LPCXpresso v8.2.2.
Fig 1
Fig 2
Have a great day,
TIC
-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!
- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------
Hi Jeremy,
I check with my code for SDRAM and while running the code I found the following behaviour,
While loading the code and running the code it is observed that it is storing the data on multiple location.When I am running the code mentioned below with stepwise execution,
__SECTION(data,RAM8) uint8_t data_buffer[1024];
Length = DRAM_SIZE;
for (i=0;i<Length; i++)
{
data_buffer[i] = 0x0A;
}
I just run for i=0 and data_buffer[i] = 0x0A;
and then checked the memory. So follwing is the observation.
Instead of storing the data in only data_buffer[0] = 0x0A; sinze i=0 it is also storing the same value 0x0a in other locations as well refer the screen shot for observation.
Also one more observation is when I am stopping the execution stepwise and checking the data in expression and memory as shown below then again while proceeding for the next step then it gives below error
But if without the checking the debug value and running step by step the it works but it is storing the value in all other locations as well as i have mentione above.
Also I would like to suggest you to please check my code with your board by doing the modifications in the code w.r.t your eval board and check the behaviour. mean while i will check for the clock and other issues if exist.
There is one more observation reagrding the memory that we are using.
You can see highlighted part where is it mentioned the memory configuration. But similar memory configuration is not available in the user manual of LPC43XX.
is there a probelma related to this? Since the same behaviour is ther there after doing following method.
for (index = 0; index < 20; index++)
{
*(uint32_t *)(sdram + index) = 0xA5A5A5A5;
}
while(1)
{
Chip_GPIO_SetPinToggle(LPC_GPIO_PORT, 5, 19);
for (index=0; index<0xFFFFF;index++ );
//sdram = (uint32_t *)EMC_ADDRESS_DYCS0;
for (index = 0; index < 20; index++)
{
if (*(uint32_t *)(sdram + index) != 0xA5A5A5A5)
{
Chip_GPIO_SetPinState(LPC_GPIO_PORT, 5, 19, (bool) true);
while(1);
}
}
}
Here also it it writing the whole memory with 0xA5A5A5A5 instead of only 20 locations.
Also refer the attached code as well for your reference where I have made the modifications regarding your inputs as well.
Thanks
Gaurav More
Hi Gaurav More
Thanks for your reply.
Q1) I've not encountered a similar phenomenon as you mentioned when using below code
__SECTION(data,RAM8) uint8_t data_buffer[1024];
Length = DRAM_SIZE;
for (i=0;i<Length; i++)
{
data_buffer[i] = 0x0A;
}
Q2) The highlight configuration is suited for the AS4C4M16SA.
Q3) I've not encountered a similar phenomenon when using the below code.
for (index = 0; index < 20; index++)
{
*(uint32_t *)(sdram + index) = 0xA5A5A5A5;
}
while(1)
{
Chip_GPIO_SetPinToggle(LPC_GPIO_PORT, 5, 19);
for (index=0; index<0xFFFFF;index++ );
//sdram = (uint32_t *)EMC_ADDRESS_DYCS0;
for (index = 0; index < 20; index++)
{
if (*(uint32_t *)(sdram + index) != 0xA5A5A5A5)
{
Chip_GPIO_SetPinState(LPC_GPIO_PORT, 5, 19, (bool) true);
while(1);
}
}
}Note: I've done the above testing base on MCB4357 board, the demo code is based on the periph_memtest demo which is come from the LPCOpen library.
Have a great day,
TIC
-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!
- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------
Hi Jeremy,
As per your input I the modifications in the code and also refered the application note (ANN11508) for SDRAM interface with LPC43XX.
After doing the modification I am able to write but the data written at the alternate address is not proper means not the one which want to write.
Let explian in detail. I am writing
uint16_t Data = 0xA5A5;
uint16_t Data1 = 0x1234;
In alternate location. and I am writting data using following logic,
for (index = 0; index < SDRAM_SIZE; index++)
{
if(Flag == 0 )
{
sdramAddr = Data;
*(sdram + index) = sdramAddr;
Flag = 1;
}
else
{
sdramAddr = Data1;
*(sdram + index) = sdramAddr;
Flag = 0;
}
if(*(sdram + index) != sdramAddr )
{
while(1);
}
}
after cehcking the memory location 0x28000000, below is the observation where it is writing 0x1200 instead of 0x1234 and it is happening only for every alternate address address writing. it is writing 0xA5A5 properly.
It is storing 0xA5A5 at 0x28000000, 0x1200 at 0x28000002.. Likewise..
You can see it is writing 0x1200 and 0xA5A5 at every alternate location instead of writing 0x1234 and 0xA5A5.
Please let me know the reason for the same. Also find attached code for your reference. in order to check if any configuration is missing.
Thanks
Gaurav More
Hi Jeremy,
There is one observation in the functionality of SDRAM interfaced with LPC4367 in my customized board.
The observation is in the below mentioned code while writing the data
uint16_t Data = 0xA5A5;
uint16_t Data1 = 0x1234;
In alternate location. and I am writting data using following logic,
for (index = 0; index < SDRAM_SIZE; index++)
{
if(Flag == 0 )
{
sdramAddr = Data;
*(sdram + index) = sdramAddr;
Flag = 1;
}
else
{
sdramAddr = Data1;
*(sdram + index) = sdramAddr;
Flag = 0;
}
if(*(sdram + index) != sdramAddr )
{
while(1);
}
}
when #define SDRAM_SIZE 128
It works fine and writes proper in all the location. So as per my understanding I am able to write 0x28000000 to 0x28000100 means upto 256 bytes.
Patern mentioned below,
Address Data
0x28000000 0xA5A5
0x28000002 0x1234
0x28000004 0xA5A5
0x28000006 0x1234
like wise till 0x28000100 Which is correct one.
.
when #define SDRAM_SIZE 129
Above 0x28000100 means if address location is 0x28000102, it writes 0x1200 and it also corrupts the previous location as well, means from 0x28000000 to 0x28000100 address location by writing 0x1200 in every alternate location.
Patern mentioned below,
Address Data
0x28000000 0xA5A5
0x28000002 0x1200
0x28000004 0xA5A5
0x28000006 0x1200
like wise till 0x28000102 which is not correct.
Please let me know the reason for this behaviour mentioned above on very higher priority. IS there any thing missing in the code which is not allowing to write above 0x28000100 address location?
Thanks
Gaurav More.
Hi Gaurav More,
Sorry for reply late.
If the SDRAM_SIZE becomes 129, it means that the writing byte will be 258 bytes which consist of 0x2800101 and 0x2800102(will be written: 0xA5A5) and the previous 256 bytes.
So it's correct.
Have a great day,
TIC
-------------------------------------------------------------------------------
Note:
- If this post answers your question, please click the "Mark Correct" button. Thank you!
- We are following threads for 7 weeks after the last post, later replies are ignored
Please open a new thread and refer to the closed one, if you have a related question at a later point in time.
-------------------------------------------------------------------------------
Hi Jeremy,
Thanks for the reply.
As per you input i did some more test to checl the behaviour. I made some modification in the code mentioned below and checked the behaviour
volatile uint16_t *sdram = (uint16_t *)EMC_ADDRESS_DYCS0; /* SDRAM start address. */
uint16_t sdramAddr;
uint16_t Data = 0xA5A5;
SDRAM_SIZE = 130;
for (index = 0; index < (SDRAM_SIZE); index ++)
{
sdramAddr = Data;
*(sdram + index) = sdramAddr;
}
And after running the code below is the result that i got whihc is different. I just wanted to check if i write the same data accross al location what is the output? Some plase it is writing properly and at some places it is writing 0xA500 instead of 0xA5A5.
You can see that it is writing randomly. Now I tried with the EMCCLK Delay 6 and 7 as well but still the same result.
What is the resaons please let me know..
Even i tried
memset((uint16_t *)sdram, 0xA5A5, 130);
but still the same result. It is not writing the data consistently.
I feel I am close to the solution but then again not getting the result.
Thanks
Gaurav More
