ISP over I2C - How?

[kbk1]

Hi

I am working on a host application, which will program a LPC84x MCU via ISP by the I2C interface.

From the user manual I can read that the FAIM register default value is ISP over UART.

So, how do I access the FAIM registers, to change it to use ISP over I2C?

Isn't it possible to program the LPC84x by ISP/I2C without having to access the MCU registers first?

Thanks

Kaare

[brendonslade]

The LPC84x devices will boot from I2C under certain conditions. You really need to refer to the flowchart from the UM to understand this. For a part with invalid (unprogrammed) FAIM, I2C ISP boot should be working but at fixed pin locations (PIO0_11 for SDA and PIO0_10 for SCL); if the FAIM has been programmed (is valid) then FAIM word 0 bit 30/31 will determine if I2C ISP will happen (as opposed to SPI or USART):

Also... just want to let viewers of this thread know that the free Flash Magic tool includes a FAIM programming utility. Under the ISP menu, open FAIM to view the current FAIM settings and write new values. .

This doesnt solve the original request, but might be preferable to using the SDK or Code Bundle examples to set up FAIM. Note that FAIM has a limited number of programming cycles, so be careful!

[kbk1 · Accepted solution]

Thanks

The Flash Magic tool could be an easy way to do it.

In our case, we use a simple testprogram to test the board during production, and this program then contains the FAIM operation to enable I2C ISP.

u32FAIM_data = 0x68000000;
FAIMWrite( 0, (uint32_t)&u32FAIM_data);

I am quite sure that the FAIM register is "valid" from factory. This means that it uses UART as default ISP interface, and you need to change this to I2C.

br

Kaare

[brendonslade]

Hi Kare,

I'm 99% sure the manual ought to say "invalid or unprogrammed" for FAIM in table 3, section 3.4 - you shouldnt have to go through that extra step for new parts. I'm checking this with the design team now.

regards

Brendon

[brendonslade]

Ok - looks like the UM is correct ... you have to set FAIM to enable ISP mode over I2C (as Victor stated).

[kbk1]

OK, thanks.

It’s not a big deal in my situation, because we will need to program a test software for productions test anyway, and this is done by SWD/JTAG.

And this program simply calls the needed line of code to set the FAIM register.

Thanks

Kaare

[victorjimenez]

Hello Kaare,

Just to confirm, you are right. The FAIM register is valid from factory.

Regards,

Victor.

[victorjimenez]

Hello Kaare,

Regarding your questions please see my comments below.

So, how do I access the FAIM registers, to change it to use ISP over I2C?

I recommend you use the example FAIM_and_Clocks  of the CodeBundle for the LPC845, to change the value of the bits 31:30 of the FAIM word0. If you are using MCUXpresso IDE you can find the code bundle in the following path: C:\nxp\MCUXpressoIDE_10.2.1_795\ide\Examples\CodeBundles

Isn't it possible to program the LPC84x by ISP/I2C without having to access the MCU registers first?

Unfortunately no, it's not possible to program the LPC84x by I2C without changing the FAIM values first.

Hope it helps!

Victor.

-----------------------------------------------------------------------------------------------------------------------

Note: If this post answers your question, please click the Correct Answer button. Thank you!

-----------------------------------------------------------------------------------------------------------------------

[kbk1]

Hi

I have now tried out the FAIM example for the LPC845.

It is fairly simple, but i can't figure out how it works (and it does not work):

I use Rowley Crossworks, but I simple copied the necessary source to my main.c, to try it out:

When i call FAIMRead like this:

   FAIMRead( 0, (uint32_t)&read_data);

...it returns OK, but "read_data" is never updated by the call to FAIMRead.

Do I do anything wrong, or have I misunderstood something?

Here is the source I copied from the FAIM and clocks example:

struct sIAP
{
uint32_t cmd; // Command
uint32_t par[4]; // Parameters
uint32_t stat; // Status
uint32_t res[4]; // Result
};

static struct sIAP IAP;

// Pointer to ROM IAP entry functions
#define IAP_ENTRY_LOCATION 0x0F001FF1

enum eIAP_COMMANDS
{
IAP_PREPARE = 50, // Prepare sector(s) for write operation
IAP_COPY_RAM2FLASH, // Copy RAM to Flash
IAP_ERASE, // Erase sector(s)
IAP_BLANK_CHECK, // Blank check sector(s)
IAP_READ_PART_ID, // Read chip part ID
IAP_READ_BOOT_VER, // Read chip boot code version
IAP_COMPARE, // Compare memory areas
IAP_REINVOKE_ISP, // Reinvoke ISP
IAP_READ_UID, // Read unique ID
IAP_ERASE_PAGE, // Erase page(s)
IAP_READ_MISR=70,
IAP_READ_MISR_EX=73,
IAP_READ_PAGE_FAIM=80,
IAP_WRITE_PAGE_FAIM=81
};

// IAP Call
typedef void (*IAP_Entry) (uint32_t *cmd, uint32_t *stat);
#define IAP_Call ((IAP_Entry) IAP_ENTRY_LOCATION)

int FAIMRead (uint32_t adr0, uint32_t adr1)
{
IAP.cmd = IAP_READ_PAGE_FAIM; // READ FAIM PAGE
IAP.par[0] = adr0; // FAIM page number
IAP.par[1] = adr1; // Destination RAM address to store the read value
IAP_Call (&IAP.cmd, &IAP.stat); // Call IAP Command
if (IAP.stat) return (IAP.stat); // Command Failed

return (0); // Finished without Errors
}

[kbk1]

Ah !

It was the compiler optimization that made weird things happen (or not happen).

It works now, thanks !

We have decided to make a testprogram to download at production / test, which changes the FAIM to use I2C port for ISP interface.

It was not what we hoped, but it is doable.

br

Kåre

[kbk1]

Hi Victor 

Thanks for the fast response

But I don't like your answer :smileyhappy:

The example you point me to, shows how to change the FAIM from the LPC845 MCU itself.

This means that I have to have firmware in the MCU.

But, I don't want hat.

The reason to boot the LPC845 from a remote host via ISP, is to be able to boot the LPC845 directly from the factory.

I don't want to first program a dummy firmware into the chip, just to change the FAIM register setting.

As I read it, I have to use the UART to change the FAIM. This must be done in the factory, then.

And then I have to have the UART RX/TX exposed somehow.

And I must implement this in our production test facilities (which I hoped was not necessary...)

This was definitely not what I had in mind, and will complicate things a lot.

br

Kaare