Hi All,

I am building a device using the BGX13S. It is working well except that sometimes I find it awake when it should be asleep. The only way I can tell if it is awake is to send a character to it. If it responds I know it is awake.

My device is battery powered and I do not want the BGX13 to be awake when not needed. It seems to be waking up sometimes in response to connection requests from random BLE devices. The connection fails so the con_active pin is not asserted, therefore the MCU stays asleep. I have the MCU waking up regularly as a rough work-around but that is affecting battery life as well.

It would be really useful to have a pin asserted when the BGX13 is awake. Is this possible?

Thanks in advance,

Rob

Hi ReneK,

Thank you very much. Either of those will do the job.

I am using about 1,000 bytes of xdata for variables. Does Keil automatically allocate the lowest bytes for that purpose? If so there is little danger of interfering with the UID.

Rob

Providing you're using KEIL to compile, this is really easy:

unsigned long GetUid(void)
{
    return *(unsigned long xdata*)0x0FFC;
}

Maybe it's an option for you to not use a function at all, just declare a variable:

extern unsigned long xdata uid;

Then, use the ASSIGN directive to make the variable accessible, see screenshot. Now, you can use the variable where needed, for example:

unsigned long GetUid(void)
{
    return uid;
}

Hi All,

I am working with the EFM8SB2 MCUs, which have 4 kB of XRAM. The reference manual says that "A 32-bit unique identifier (UID) is pre-loaded upon device reset into the last four bytes of the XRAM area on all devices."

It would be really useful to read the UID of each device. Is there a function by which I can read these four bytes using ordinary C code?

Thanks in advance,

Rob

Maybe I was doing something incorrectly or "not done" in C. I have asked an experienced programmer if there is a convention and if I was doing it wrongly. The problem is that the code compiled OK but just didn't run. There was no error message. As soon as I reduced the number of static variables it started working again. I did it differently and added several more static variables to see what happens, ant it just keeps working.

It this rate I am becoming an EFM8 guru, with all the arcane undocumented knowledge I am accumulating.

I have filled out the survey once. It won't let me do it a second time.

I have just wasted three days on another undocumented quirk of Simplicity Studio. I posted a question in this forum and someone moved it to the hardware forum. A hardware guy tried to help but had no idea. The problem was some sort of undocumented limitation on initializing arrays; maybe (probably) a shortcoming of the "quirky" Keil compiler but maybe the crumbly old Eclipse as well. Either way it was a Simplicity Studio problem. I had to solve it entirely on my own and it took three days.

I know it's free. I would pay a reasonable amount for some support.

Robert Keith

I fixed the problem by declaring the arrays like:

uint8_t xdata crcBytes[4];

and then initializing them before using them:

    for (i = 0; i < 4; i++)
    {
        crcBytes[i] = 0;
    }

I have no idea if that is the "right" way to do it but it works.

I still put the whole problem, and three wasted days, down to "quirks" of Simplicity Studio and the Keil compiler. Much of my time for the last three months has been working around the odd behavior of the IDE. I am very impressed by the EFM8 parts, but they are let down by this development environment.

Thanks for your input.

Rob.

I fixed the problem by declaring the arrays like:

uint8_t xdata crcBytes[4];

and then initializing them before using them:

    for (i = 0; i < 4; i++)
    {
        crcBytes[i] = 0;
    }

I have no idea if that is the "right" way to do it but it works.

I still put the whole problem, and three wasted days, down to "quirks" of Simplicity Studio and the Keil compiler. Much of my time for the last three months has been working around the odd behavior of the IDE. I am very impressed by the EFM8 parts, but they are let down by this development environment.

Thanks for your input.

Rob.

OK. I've found the problem. I thought it was a problem with the compiler, and I still think so, which is why I posted my question in the Simplicity Studio section of the forum.

I rolled everything back to the night before the problem arose. The project ran with no problem. The function I was working on receives a hex configuration string and parses it into a number of variables, some of which are arrays.

When I was working on the CRC function I declared a global array like:

uint8_t xdata crcBytes[4];

However, it behaved badly and, for example, would not retrieve the last CRC byte from the register. If I declare and initialize it like:

uint8_t xdata crcBytes[] = {0,0,0,0};

It worked fine. In writing the code that broke the project, I added a few more 4 byte arrays which are useful for the time functions I need. At that point everything broke.

I replaced the old files with the newer versions until I was left with the one C file that caused the project to break. I commented out all the changes I had made between the working version and the broken one. Then I started uncommenting things. If I removed the // from a line declaring and initializing one of the new arrays it loads and runs OK. If I remove it from two lines it breaks.

The weird thing is that it doesn't matter which arrays are not commented out. It breaks if more than one is uncommented.

It works if the arrays are declared with the number of elements but not initialized.

Is there a limit to the number of arrays that can be initialized at that level? I am not an expert on C, Eclipse or the Keil compiler, but it seems odd. As I mentioned above the message regarding memory usage indicates that I have plenty of spare memory.

Can you suggest why the IDE would behave like that?

Thanks,

Rob

I already thought of that. There were 5. I removed them all and put one back.

The code does not start even to the stage of hitting a breakpoint. If I remove all breakpoints it still doesn't start.