We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.
I'm trying to fill an array in xdata with zeros as follows:
unsigned char n; //Index to scan the array unsigned char xdata bt[100]; //The array to fill
for(n=0;n<100;n++){ bt[n]=0; }
But when the array size is bigger than 38 elements the MCU hangs. Does anybody know the reason why this happens?
look in the .map file, how much xdata?
Hi Erik! thanks for your reply. Below the Memory map: TYPE BASE LENGTH RELOCATION SEGMENT NAME ----------------------------------------------------- * * * * * * * D A T A M E M O R Y * * * * * * * REG 0000H 0008H ABSOLUTE "REG BANK 0" REG 0008H 0008H ABSOLUTE "REG BANK 1" DATA 0010H 0002H UNIT ?DT?_WAIT?SMARTNODE 0012H 000EH *** GAP *** BIT 0020H.0 0002H.4 UNIT ?BI?SMARTNODE BIT 0022H.4 0001H.1 UNIT _BIT_GROUP_ BIT 0023H.5 0000H.1 UNIT ?BI?_UART_SEND_COMMAND?SMARTNODE 0023H.6 0000H.2 *** GAP *** DATA 0024H 001DH UNIT ?DT?SMARTNODE DATA 0041H 0015H UNIT _DATA_GROUP_ IDATA 0056H 0001H UNIT ?STACK * * * * * * * X D A T A M E M O R Y * * * * * * * XDATA 0000H 03F0H UNIT ?XD?SMARTNODE XDATA 03F0H 0065H UNIT _XDATA_GROUP_ As I was wondering if the problem was caused by WDT I also disabled the WDT at the very beginning of the main with PCA0MD &= ~0x40; but it doesn't help
XDATA 0000H 03F0H UNIT ?XD?SMARTNODE
you have used the XDATA up
Yeah, you're right but this was not the problem, or was not the root cause.
When the MCU start I create the variables in xdata memory, well.
I'm not able to write in this blog messages thatb are bigger than 1 row. Incredible I I do I get Error 500 (Internal Server Error), So I will post several messages.
I'm unable to post any further message...
When I create such a big arrays into xdata the system takes to long and WDT expires causing a CPU reset. Obviously the first thing I do in main() is to disable the WDT but it's too late. So it looks like despite I've got 2k of RAM I cannot use this memory for global arrays that are bigger than 20bytes, as I I do that the WDT will reset the MCU. Does anybody know how can I avoid WDT operation to take place before I got the control into the main()? Is it possible that Silab made such a beautiful memory bank but it's impossible to use for such a thing?
It is possible to create longer messages on this forum - and notice it's a forum and not a blog.
The issue is that there have been lots of spam issues. And for some reason, the word w_a_t_c_h_d_o_g just happens to trig their spam filter, giving you the error messages you are seeing. Or you might have accidentally found some other trigger word.
I think this was one of several threads where I did try to warn about the w-dog issues but wasn't allowed to post.
I've got 2k of RAM . not off the bat, you have 1k + 1k "USB memory" I think to recall that there is some configuration that you have to do in startup.a51 that will make the "USB memory" usable as regular RAM, check the datasheet
Obviously the first thing I do in main() is to disable the WDT but it's too late That is why I tell EVERYBODY to do it in startup.a51
Hi Eric, I will try to avoid to trigger the black-listed word in this forum. Well, I have modified STARTUP.A51 including ASM instructions to disable WDT before the memory is allocated. Now everything seems to work fine, no more strange behaviour. I'm using 1K of XDATA, 128byte DATA and 128byte iDATA. I'm wondering if you can suggest some setting to gain access to USB 1K FIFO. F320 is quite powerful and despite it's age is still a winning choice against many other MCU thanks to the powerful peripherals such as ADC, DAC, PWM, SPI, and so on. However the RAM limitation still makes difference in some very specific cases. Having 2304byte of RAM would open the way to several additional applications. I've checked in the user manual and specs but I've not found any specific and explicit instruction to control the 1k FIFO. I understood that this memeory is fully controlled by USB logic hence it would be hard to control from 8051 core. There might be some instruction to push/pop data into FIFO, this could be useful but far away from a clasic RAM access operation.
I'm wondering if you can suggest some setting to gain access to USB 1K FIFO. I could read the datsheet for you, but naah
Hi Erik, I did the change, and it worked! I added few rows in ASM to disable WDT and it's working great now. The news that I've got another 1K of RAM sounds amazing for my applications. It look like this RAM is hidden by my ignorance and nothing else. I've looked trough the datasheet of Silab c8051F320 and I've found that this FIFO is under control of USB circuitry hence the USB Clock must be active when operating with such a memory as described here: community.silabs.com/.../70717 However in most of my applications where I need RAM (stand alone projects) I don't have USB clock. Any wolkaround for this?
start the USB clock, that does not start USB. be aware that using the FIFO as RAM is slower than using the 'real' RAM. even better use a chip with sufficient 'real' ram there is a plethora of possibilities