TMS320F28377D: Hi-Z Trip-Zone state

Hi Gennadi,

My first thought is this sounds like there may be an issue with your PWM configuration.  Could you post your initialization code for that PWM module which is experiencing the issue?

Are there any other circuits on this pin which may be driving a voltage on this pin? Is it possible to isolate it from the rest of your system?

Regards,

Kris

here is our TZ initialization:

EALLOW;

// Enable trip zone
epwmAddr->TZSEL.bit.CBC2 = TZ_ENABLE; 
epwmAddr->TZSEL.bit.OSHT3 = TZ_ENABLE; 

// Pulse for clear CBC TZ
epwmAddr->TZCLR.bit.CBCPULSE = 1; 

// When trip zone event happen
epwmAddr->TZCTL.bit.TZA = TZ_HIZ;
epwmAddr->TZCTL.bit.TZB = TZ_HIZ;

// Clear trip zone flag
epwmAddr->TZCLR.bit.OST = 1;
epwmAddr->TZCLR.bit.CBC = 1;
epwmAddr->TZCLR.bit.INT = 1;
EDIS;

*****************************************************************************

and here GPIO initialization for PWM's:

EALLOW;

GpioCtrlRegs.GPEPUD.all |= 0xFFFE0000; 
GpioCtrlRegs.GPFPUD.all |= 0x000001FF; 
GpioDataRegs.GPECLEAR.all |= 0xFFFE0000;
GpioDataRegs.GPFCLEAR.all |= 0x000001FF; 
GpioCtrlRegs.GPEDIR.all |= 0xFFFE0000; 
GpioCtrlRegs.GPFDIR.all |= 0x000001FF; 
GpioCtrlRegs.GPEMUX2.all &= 0x00000003;
GpioCtrlRegs.GPFMUX1.all &= 0xFFFC0000; 

GpioCtrlRegs.GPEMUX2.bit.GPIO145 = 1; // Configure GPIO as EPWM1A
GpioCtrlRegs.GPEMUX2.bit.GPIO146 = 1; // Configure GPIO as EPWM1B
GpioCtrlRegs.GPEMUX2.bit.GPIO147 = 1; // Configure GPIO as EPWM2A
GpioCtrlRegs.GPEMUX2.bit.GPIO148 = 1; // Configure GPIO as EPWM2B
GpioCtrlRegs.GPEMUX2.bit.GPIO149 = 1; // Configure GPIO as EPWM3A
GpioCtrlRegs.GPEMUX2.bit.GPIO150 = 1; // Configure GPIO as EPWM3B
GpioCtrlRegs.GPEMUX2.bit.GPIO151 = 1; // Configure GPIO as EPWM4A
GpioCtrlRegs.GPEMUX2.bit.GPIO152 = 1; // Configure GPIO as EPWM4B
GpioCtrlRegs.GPEMUX2.bit.GPIO153 = 1; // Configure GPIO as EPWM5A
GpioCtrlRegs.GPEMUX2.bit.GPIO154 = 1; // Configure GPIO as EPWM5B
GpioCtrlRegs.GPEMUX2.bit.GPIO155 = 1; // Configure GPIO as EPWM6A
GpioCtrlRegs.GPEMUX2.bit.GPIO156 = 1; // Configure GPIO as EPWM6B
GpioCtrlRegs.GPEMUX2.bit.GPIO157 = 1; // Configure GPIO as EPWM7A
GpioCtrlRegs.GPEMUX2.bit.GPIO158 = 1; // Configure GPIO as EPWM7B
GpioCtrlRegs.GPEMUX2.bit.GPIO159 = 1; // Configure GPIO as EPWM8A
GpioCtrlRegs.GPFMUX1.bit.GPIO160 = 1; // Configure GPIO as EPWM8B
GpioCtrlRegs.GPFMUX1.bit.GPIO161 = 1; // Configure GPIO as EPWM9A
GpioCtrlRegs.GPFMUX1.bit.GPIO162 = 1; // Configure GPIO as EPWM9B
GpioCtrlRegs.GPFMUX1.bit.GPIO163 = 1; // Configure GPIO as EPWM10A
GpioCtrlRegs.GPFMUX1.bit.GPIO164 = 1; // Configure GPIO as EPWM10B
GpioCtrlRegs.GPFMUX1.bit.GPIO165 = 1; // Configure GPIO as EPWM11A
GpioCtrlRegs.GPFMUX1.bit.GPIO166 = 1; // Configure GPIO as EPWM11B
GpioCtrlRegs.GPFMUX1.bit.GPIO167 = 1; // Configure GPIO as EPWM12A
GpioCtrlRegs.GPFMUX1.bit.GPIO168 = 1; // Configure GPIO as EPWM12B
EDIS;

*****************************************************************************

And now the hardware circuit on the output:

see the attached screenshot 1. I did some more tests:

Test without 240 Ohm  => the signal goes low if TZ event (see screenshot 2)

Test with 240 Ohm  but without 220pF => the signal goes to 1.7V if TZ event (see screenshot 3)

Screenshot 1

Screenshot 2

Screenshot 3

 

I still don't know what a problem is. Is it the 74ACT16244DGG driver input?

I would highly suspect the 74ACT16244DGG driver input. It sounds the hi-z is working exactly as we'd expect it to with the external circuit disconnected.  If you disconnect the 240 ohm and check the voltage on the driver input side, I'm guessing you will see a voltage greater than or equal to 1.7 V.  Can you check that?

Regards,

Kris

Hi Kris,

yes, you are right. If I disconnect the 240 ohm I see 1,76V on the input of 74ACT16244DGG, but just if the input was high before disconnecting. Do you have some idea why? What is the input circuit of 74ACT16244DGG (I can't find it in datasheet).

If we have 1 kohm pull-down on the input it seems to work properly, but 1 kohm is to small for our application and is large load for DSP output.

Hi Gennadi,

Can you check to make sure the pull up resistor on the C2000 device is disabled? This is done via the GPxPUD disable register where 'x' is the GPIO port of the pin.

Also, can you describe the process you are going through to remove the 240 ohm resistor? Is the driver input physically disconnected from the pull down resistor on the C2000 side, and therefore the input is floating?

Thanks,
Kris

Hi Kris,

1) yes, the pull up on the C2000 is off

2) yes, I disconnected one input of 74ACT16244DGG, so there is no physical connection to the 10k pull down 

3) if I connect it physicaly to 3.3V and then disconnect it again -> I can meas 1.76V on the input (screenshot 1)

4) If I don't connect it to some voltage potential -> I can meas 0V on the input (screenshot 2)

screenshot 1

 

 

screenshot 2

 

Regards,

Gennadi

Hi Gennadi,

In an effort to get you a resolution faster I have requested this post be moved over to the analog forum. They should be able to help you with this debug quite a bit faster, since it seems we have ruled the C2000 device out as the issue. Please let us know if there is any other assistance we can provide.

Regards,
Kris

Thanks Kris and hi Gennadi,

To get us started on troubleshooting the buffer, can you post the schematic section surrounding the 74ACT16244?

Hi Emrys,

her are the schematics screenshots surrounding the buffer

Regards,

Gennadi

Thank you Gennadi,

I don't see the power pins for the 74ACT16244, however I assume it is being powered from a 5V rail and has appropriate bypass capacitors.

Is it possible to get a measurement of the supply current (Icc) and a scope shot of the output of this device when the 1.76V input occurs vs when it does not? It is possible that the input voltage slowly passing the device's switching threshold could cause internal oscillation or a linear response.

I noticed that the RC circuit on the input is filtering very high frequency content - is it possible to remove this filter from the circuit? If not, I believe it would be possible to achieve a very similar roll-off frequency by removing the 220pF capacitor and increase the series resistance from 240 ohm to 10kohm - utilizing the small input capacitance of the CMOS inputs as part of the low-pass filter (Ci = 4.5 pF typical).

Hi Emrys,

unfortunately we can not meas the Icc current. We have the same issue if we remove the filter circuit from the input and just have 10kohm pull down resistance directly at the input. The buffer works properly just with smaller pull down resistance (1kOhm). I have passed the problem to my hardware colleagues. Do you have some Idea for recommended pull down resistance at the input?

Thank you

Gennadi

Hi Gennadi,

Since those are CMOS inputs, a 10k pull-down should be more than sufficient. I want to get some of those parts into my lab to do some tests. It will take some time though since they aren't standard devices we have in local stock.

Hi Emrys,

Thank you. I hope you can reproduce it in your lab.