Blog Post Actions
Group Actions
Engagement
  • Author Author: martinvalencia
  • Date Created: Date Created
  • Views 1297 views
  • Likes 6 likes
  • Comments 10 comments
Related
Recommended

TI Hercules RM46 Three-Phase Generator With Transformers Part 1

martinvalencia
INTRODUCTION

My goal is to design a device that generates a three-phase signal starting from a voltage in DC.

I have a 12VDC power supply and through H bridges, power supply of the transformers in the SCOTT configuration, both transformers with their special connection send me the signals R, S, T plus a Neutral.

Between my airplanes is to vary the voltage of the power supply with a reducing reducer converter.

The system will be controlled by a microcontroller Hercules RM46, this internal has an ETPWM module, which allows me to generate PWM signals with special characteristics, such as bandwidth, phase angle, synchronization in PWM, among others.

Attached

Part 2:

TI Hercules RM46 Three-Phase Generator With IGCM15F60GA Part 2

Microcontroller:

LAUNCHXL2-RM46 Hercules RM46x LaunchPad Development Kit | TI.com

User Guide  

Technical Reference Manual

Facebook:

Electronaplica

BLOCKS DIAGRAM

image

*The image of the transformer is referential

 

CIRCUIT ELECTRICAL BACKGROUND - SCOTT CONNECTION

This is a contribution of Frank Rodrigo Coaquira Molleapaza (Electrical Engineer).

It is a connection of two single-phase transformer to convert a system of three-phase voltages in biphasic and vice versa.

 

 

This arrangement is due to the American engineer Charles F. Scott, who invented it in 1984 while working at the company Westinghouse and is based on the well-known fact that the three-phase star system composed voltage between two phases is in square with the voltage Simple third phase.

 

image

Equation for output V2a:

image

Equation for output V2b:

image

Fragment of text, extracted from the book:

 

Pag 243. “Maquinas Electricas – Jesus Fraile Mora”

 

Conclusion: from the previous fragment, it tells us that you must have two alternating signals of the same frequency with a phase shift of 90 degrees between them.

 

About the Microcontroller

The HerculesTm RM46x LaunchPadTm Development Kit is an inexpensive evaluation platform designed to help you get started quickly in evaluating and developing with the Hercules microcontroller platform. The LaunchPad Development Kit is based on the IEC 61508 SIL 3 certified RM46L852, which is a lockstep ARMRegistered CortexRegistered-R4F based MCU with integrated safety features and peripherals such as two 12-bit ADCs, programmable High-End timers, motor control peripherals (eQEP, eCAP, ePWM), USB, Ethernet, MibSPI and serial communication interfaces. Hercules RM MCUs can help reduce the development effort of IEC 61508 functional safety applications.

 

image

 

ETPWM Enhanced Pulse Width

The enhanced pulse width modulator (ePWM) peripheral is a key element in Modulator controlling many of the power electronic systems found in both commercial and industrial equipments. These systems include digital motor control, switch mode power supply control, uninterruptible power supplies (UPS), and other forms of power conversion. The ePWM peripheral performs a digital to analog (DAC) function, where the duty cycle is equivalent to a DAC analog value; it is sometimes referred to as a Power DAC.

 

As we can see in the image, you can see that ETPWM module uses two output pins of the microcontroller, side A and B, both outputs can be with the same polarity or one in reverse of the other.

 

 

The resolution of the module is of nanoseconds, generating a high resolution in the dutty cycle and in the dead band.

 

image

 

 

Code Example CCSV7

 

int main(void)

{

/* USER CODE BEGIN (3) */

  etpwmDeadBandConfig_t pwm_deadband_config; // declaracion de variable

  etpwmInit();

 

  /*Configuracion del angulo de desfase entre los PWM generados*/

  etpwmSetSyncOut(etpwmREG1, SyncOut_CtrEqZero);

  etpwmSetSyncOut(etpwmREG2, SyncOut_EPWMxSYNCI); //SyncOut_CtrEqZero);

 

  etpwmDisableCounterLoadOnSync(etpwmREG1);

  etpwmEnableCounterLoadOnSync(etpwmREG2, 8300, COUNT_DOWN); // seting del angulo de

 

  /*Configuracion de la banda muerta*/

  pwm_deadband_config.halfCycleEnable = false;

  pwm_deadband_config.inputmode = PWMA_RED_PWMB_FED;

  pwm_deadband_config.outputmode = PWMB_FED_PWMA_RED;

  pwm_deadband_config.polarity = Invert_PWMB;

  /*Iniciando la carga de los valor de la banda muerta 1000---> 510 microsegundos*/

  etpwmSetDeadBandDelay(etpwmREG1, 1000, 1000);

  etpwmEnableDeadBand(etpwmREG1, pwm_deadband_config);

  etpwmSetDeadBandDelay(etpwmREG2, 1000, 1000);

  etpwmEnableDeadBand(etpwmREG2, pwm_deadband_config);

 

 

  while (1)

  ;

  /* USER CODE END */

 

 

    return 0;

}

 

 

Pin Conection

File:LAUNCHXL2 TMS57012 RM46 REVA.pdf - Texas Instruments Wiki

image

 

 

 

Schematic

image

In the schematic there is a simple inverter bridge excited by transistors and using MOSFET to perform the energy switching

Two of these bridges are used for the project.

 

IMPLEMENTATION PHOTOS

In the following images you can see the implementation using only one transformer.

image

image

image

image

  Output voltage waveform

image

  Voltage spectrum

image

 

 

Operation Videos - Spanish

First part of the operation

Second part of the operation

Parents

  • Good project. I'm looking forward to the further parts.

     

    There is a small problem with the circuit you show, though, and it could be improved.

     

    Here's the 'shoot through' current in one half of the bridge when the PWM changes state. (This is from a simulator, so reality will be a bit different, though I'm finding that MOSFET models do seem to be fairly good.)

     

    image

     

    image

     

    That occurs because the MOSFETs have considerable gate capacitance and the time constant (in conjunction with the pull-up resistor) for moving the gate up is much different to pulling it down with the transistor, so there's a small period of time when both are conducting. (Try it with a simulator, looking at the gate voltages, to get an idea of what's happening.) The effect is also quite different with the two edges of the PWM because of the lack of symmetry with the drive.

     

    The circuit works on the bench because these MOSFETs can take that sort of current when pulsed (it's only for a few uS), the 60Hz repetition rate is very low (so the dissipation is low), and the PSU is designed for 30A (and, presumably, doesn't dip too much).

  • in reply to jc2048

    That could likely be solved by configuring the deadband in the PWM module. I've done that for half-bridge, not for 3-phase though.

     

    There's a HET example that generates 3 phase signals from the timer module. That's complex though ...

Comment Children
  • in reply to Jan Cumps

    Does this Hercules board have enough PWMs for deadband control?

     

    Another solution would be to rework the 'driver' (at the moment it's just a level converter from 3.3V to 12V), so that both MOSFETs turn off quicker than they turn on and the switching times of both are similar.

    Personally, I'd look to do both.

     

    An advantage of improving the MOSFET drive would be if Martin wanted to move on from square wave drive to consider sinewave operation [where the PWM would have to have a much shorter period and where you wouldn't want a 20uS deadband].

  • in reply to jc2048

    Jon Clift wrote:

     

    Does this Hercules board have enough PWMs for deadband control?

     

     

    It can be independently (or dependently) set for each of the 6 or 7 complementary PWM modules,

     

    image

    Here's a signal that I created with one of these PWM pairs

    image

  • in reply to Jan Cumps

    It is true that etpwm, is very powerful Hercules has enough modules to control a three-phase bridge, the example code I put above could help to control the triphasic.

    Once I get some GaN samples for high power, I will try to produce the three-phase signal in this way.

    Saludos amigo!