TMDSRESDCKIT

Thom,

The pdf files in C:\TI_F28xxx_SysSW\ResonantConv\~Docs shows the jumper positions. For the on board emulator stage it is ok if you do not put any jumpers, it should work.

Unfortunately, a few early version were shipped without programmed ftdi chip, and I assume that you solved the programming issue.

The best way to verify the board functionality is resistive divider feedback. Make sure that J12 and J13 are shorted to each other, any J7-8 is at VfdR position.

For some reason you might be triggering the TZ tripzones which immediately shut downs the PWMs, would you please delete the TZ part from ResonantConv_Pwm_Drv.c file, and disable the protections:

   EALLOW;
   (*ePWM[n]).TZSEL.bit.OSHT1 = 1;    // one-shot source
   (*ePWM[n]).TZCTL.bit.TZA = TZ_FORCE_LO;    // set EPWM1A to low at fault instant
   (*ePWM[n]).TZCTL.bit.TZB = TZ_FORCE_LO;    // set EPWM1B to low at fault instant
   (*ePWM[n]).TZEINT.bit.OST = 1;     // Enable TZ interrupt
   EDIS;

I suggest real time debugging instead of disabling it.

Regards

Bill

If you follow the instructions in resontantConv.pdf, the option for real-time debug is grey.  I chose to stop the target, enable real-time debug, and reset CPU.   This resulted in CCS locking up (not responding as a windows application).

I have killed the CCS app with task mgr and checked for stray cc_app but not rebooted my machine.

I'd rather not use real-time debug if it is not stable in this version of CCS (3.3.83.19) on this xds100 embedded emulator.   If this means that I cannot set startup=1 in the watch window, then I will move forward with changing startup=1 in the source code - unless you know a reason this doesn't work.

I am willing to buy the non-limited CCS v3.3 to get real-time debug if I can convince myself that the framework and 2808 app is a good starter-ware base for doing a rapid proof of concept prototype.  

I have eliminated the TZ code (compiled it out) and have set Startup=1 in the src.

Still no PWM or Vout.   In the watch window, after a refresh I have Gui_Vout[1] = 20.8 but I have not chased down how Gui_Vout[1]  is related to output voltage (though I can guess).

Thom

PS:

J12 and J13 are shorted across the pins next to TERM1, and I've changed J7-8 to VfdR position with no PWM resulting.   Here's watch window contents after refresh:

StartUp  1 
Gui_Vin  19.1113 
Gui_Vout[1] 0.0
Duty[1]  0.5 
DeadBand[1] 30
Period[1] 850
max_per  1000
min_per  250

I have placed two channels of the oscope on PWM1 and PWM2 and have single stepped the code in main() and have not seen the PWM signals "oscillate".  I have seen them change state from hi to lo during one of the code steps through some PWM config.   I've also put a breakpoint in the ISR and it is actively running.    Given all the problems I have had (wrong CCS shipped in the box, unprogrammed ftdi usb chip, I'm suspicious of something being wrong wit the board.

It is powered with 19.13v, its drawing about 129mA when running the resontantConv app.

On the motherboard, LED LD1 is on, LED LD2 is on.   on the controlcard LD1 is green and LD2 is flashing red.   The power switch to the board is on, the power switch for the power section is on.   The controlCard is seated well in the connector.   There is a 24ohm load on TERM1.   there are no jumpers on the emulation section.   the other jumpers are described above.   

Any suggestions?

Waiting on a Luminary Micro eval kit, I tried pruning the code down to evaluate just the PWM peripheral.   Along the way I found that 'real time debug' mode could be enabled if you follow this sequence:

Debug->Halt

Debug->Disconnect

Debug->Enable Real Time

Debug->Connect

(do not do a Debug->Reset CPU or CCS will hang).

This procedure resulted in some success running the first example with INCR_BUILD=1, as long as Startup=0 was in the code and Startup=1 was done in the watch window.   Single stepping in the debugger with startup=1 in the code would also result in PWM starting up (most of the time).   Startup=1 in the code, and performing Debug->Run would result in PWM not starting, even with TZ code (current limit) commented out.

With Vin=20.28v, Vout=10.92 with 24ohm load resistor (Current draw from 20v power source is 429mA).

Note: this procedure does not always work, there is something 'yet to be determined' that prevents  the app from correctly starting the PWM.   After changing to INCR_BUILD=2 and not having success, reverted back to INCR_BUILD=1 and could not get PWM output after several tries including board power cycling, etc.   The C2808 looks great from a hardware standpoint, but from a software standpoint, it's unclear about the stability of the dev envioronment.   Next step will be to look at CCS v4 on 2808, Luminary LM3S8962, and dspPIC.