TMS320F280025: S19 question

George,

TMS570 could generate S19 file with 0x25 byte count .

Any way to generate S19 file with  0x25 byte count  for C2000?

Customer use ETAS INCA tools for calibration, in this tools the S19 data bytes  just support  10 or 16  or 20 or 32 or  64.

If byte count is 0x25  that means data bytes is 32 per line. 

if it is impossible to control the number of of bytes per line in the hex utility output for C2000. That means ETAS INCA tools cannot support C2000 device.

When I try a simple example, I get S3 records with a byte count of 0x25.

C:\examples>type stub.c
int main() {}

C:\examples>cl2000 stub.c -z -o stub.out
<Linking>

C:\examples>hex2000 -m3 --romwidth=32 -o stub_hex.m3 stub.out
Translating to Motorola-S3 format...
   "stub.out" .text ==> .text
   "stub.out" .cinit ==> .cinit
   "stub.out" .reset ==> .reset

C:\examples>more stub_hex.m3
S00600004844521B
S3250000004028AD0000FF69561F5616561A2940761F00002902761B76400118520061427622F6
S3250000005028A9011D28A800000901611B76C0011D29046F0F9B0024A9DF016C04290424A895
S32500000060DF011EA6F7A1248606A781A109011EA724A96303FF5C3B0459A9DF010009FFEC53
...

I don't know how you get a byte count of 0x22.

Thanks and regards,

-George

Sorry, but I made an error in my last post.  I used --romwidth=32.  You probably use --romwidth=16.  When I use that, I get a byte count of 0x23.  But I think we have found a possible solution.  Try changing to --romwidth=32.

Thanks and regards,

-George

George,

I just let romwidth as default , like following , so get 0x22.

"C:/ti/ti-cgt-c2000_20.8.0.STS/bin/hex2000" --diag_wrap=off --motorola=3 -o "epwm_ex1_trip_zone.S19"  "epwm_ex1_trip_zone.out" .

Are you sure,  we can set --romwidth=32 for C2000 device?

Huihuang,

George is out today and will be back in the office tomorrow.

Regards,

John

There is a problem with the command ...

Huihuang Chen said:

"C:/ti/ti-cgt-c2000_20.8.0.STS/bin/hex2000" --diag_wrap=off --motorola=3 -o "epwm_ex1_trip_zone.S19"  "epwm_ex1_trip_zone.out" .

When --motorola=3 is used, the default --romwidth=8 and default --memwidth=16.  This means you get two output files, one more in addition to epwm_ex1_trip_zone.S19.  One file has the high bytes, and the other has the low bytes.  For more details, please search the C28x assembly tools manual for the sub-chapter titled Partitioning Data Into Output Files.  You should see a diagnostic similar to ...

warning: Data is being written to auto-generated file file.m01

The typical fix is to add --romwidth that equals the default --memwidth.  In the case of C28x systems, that means --romwidth=16.  That produces a file that meets the specifications of Motorola S3 format.  However, the length of each line is 15 16-bit words, which you see on each line as 0x23.  

You can get a length on each line of 0x25 by using --romwidth=32.  The generated file also meets the specifications of Motorola S3 format.  

Unfortunately, I do not know why changing --romwidth changes the length of each line.  I'll have to get back to you on that.

Thanks and regards,

-George

Please understand that the specification of Motorola S3 format does not require that one line of output be a specific length.  Different lengths are allowed.  For details, please search the C28x assembly tools manual for the sub-chapter titled Motorola Exorciser Object Format.  

It turns out there is a loose, unintentional, relationship between --romwidth and the length of one line of output.  As we have seen, --romwidth=16 emits lines of length 0x23 and --romwidth=32 emits line of length 0x25.  Either length meets the requirements of the specification.  While this is unlikely to change in future releases of hex2000, because it is not a strictly defined part of the specification, a change is possible.  Generally speaking, the best course is for the tool that reads the hex file to be flexible regarding the length of each line.

Thanks and regards,

-George