TMS320F280039C: Digital Control (RampCtl) resetting constantly

Nicolas Murguizur

Prodigy 210 points

Part Number: TMS320F280039C
Other Parts Discussed in Thread: C2000WARE

Hi,

I want to generate a ramp using RampCtl.h. I have included the file from C2000ware into my project.

I initialized the RMPCNTL object to:

Then fclRampControl(&RampControlHandler); is run every 1ms.

Debugging, I can see that RampDelayCount increases at every fclRampControl call. SetpointValue increases whenever the time comes, as shown in the following picture:

But right after this increase, if I click "Step Over" to control the execution of the code, it stops at the line shown in the following picture:

If I understood it properly, the compiler decided that that zone was to be left out, as indicated by the shaded lines, but it was not left out and not even that but also it is being run. If I step into, I get into the piece of code showed in the following picture:

Where this function compares 2000 and what should be 0.00001525f, but for some reason it indicates 0 while debugging, maybe due to casting.

My question is, why is the RampCtrl behaving like this? Id there any initial configuration I missed? Why is the MCU running code which the compiler shades?

Thanks and regards.

over 3 years ago

0 Peter Luong over 3 years ago

TI__Mastermind 21901 points

Hi Nicolas,

I'll have our CCS folks look at this issue and speak on why this may be happening in CCS. Appreciate your patience on this

Regards,

Peter

0 JohnS over 3 years ago in reply to Peter Luong

TI__Guru**** 163035 points

Nicolas,

The shading is done by a source scanner built into Eclipse and not the compiler itself. It works on the information it has and can be inaccurate. In the third capture my bet is that the editor is confused and does not know if FLT_MAN_DIG and DBL_MANT_DIG are defined.

I will see if someone else can answer the other questions.

Regards,

John

0 George Mock over 3 years ago

TI__Guru**** 253360 points

Nicolas Murguizur said:
Where this function compares 2000 and what should be 0.00001525f, but for some reason it indicates 0 while debugging

Just for now, presume this is a display error in CCS. Does this particular call to fminf return y?

Thanks and regards,

-George

0 Nicolas Murguizur over 3 years ago in reply to George Mock

Prodigy 210 points

Hi George, it returns y, which should be a small number, but it is 0.

Additionaly I get this warning during compilation:

Description Resource Path Location Type
<a href="file:/C:/ti/ccs1200/ccs/tools/compiler/dmed/HTML/225.html">#225-D</a> function "fminf" declared implicitly .ccsproject /MotorController line 112, external location: C:\ti\c2000\C2000Ware_MotorControl_SDK_4_00_00_00\libraries\utilities\math_blocks\include\v4.3\rmp_cntl.h C/C++ Problem

I wonder whether it could be related as it affects fminf function.

+1 George Mock over 3 years ago in reply to Nicolas Murguizur

TI__Guru**** 253360 points

This diagnostic ...

Nicolas Murguizur said:
function "fminf" declared implicitly

... means you do not #include <math.h>. That means the fminf function is presumed to return an int, when it actually returns a float.

Please let me know if this change resolves the problem.

Thanks and regards,

-George

0 Nicolas Murguizur over 3 years ago in reply to George Mock

Prodigy 210 points

It resolved the issue, thanks George. It might be a good idea to modify rmp_cntl.h from the digital control library to include math.h, given that it depends on it.

0 Whitney Dewey over 3 years ago in reply to Nicolas Murguizur

TI__Guru 54765 points

I've filed a bug for the MC SDK to add the #include.

Thanks,

Whitney

C2000™︎ microcontrollers

C2000 microcontrollers forum

TMS320F280039C: Digital Control (RampCtl) resetting constantly