CCS/BQ76PL455EVM: Sample code for TMS57004 and BQ76PL455EVM not working

Hi  Ruby,

The sample code is intended to "work" as-is, but is  not useful. The code is mostly examples of  configuring registers. Have you looked to see if there  is any communication back to the MCU on the RX line? 

I just hooked up an EVM to a TMS570, and used  a Saleae Logic Analyzer to verify that it should upload as-is. You can  see that  the 455 EVM receives a string  of commands on the RX line after the wake tone, and sends packets back on the TX line when a  response is requested:

Ruby,

No problem...

Starting  with the bq76PL455EVM....

1. Set the  switches  SW1 and  SW2.
   1. SW1  should be switched towards  the side  with the  text that reads "1234"
   2. SW2 should be  switched  towards "Single",  since you are not using the isolated  comms
2. The  bq76PL455A is  powered off  of the battery stack. With the bq76PL455 EVM, we can simulate the battery voltages by taking a power supply and using a simple resistor ladder to  mimic the voltage drop of a cell. The 455 needs a minimum of 12V at the top of the stack to work. In our lab we use  sixteen 100 Ohm  resistors in series, and connect the 455 to each  tap. Connecting  a  power supply to this resistor divider should work all the way down to ~12V.  If  you  don't want to get into high voltage, I  suggest using something like 20V. This will create a cell undervoltage fault condition  with the default settings, I believe. This  would  only cause the  nFault pin to trigger, though. 

Time for the  MCU...

1. Connect the MCU header pins  to  the header J3 of the 455 EVM
   1. These connections are detailed at the beginning  of the  sys_main.c file.
2. Assuming  the project has  been imported into CCS and  compiled, connect the MCU to the PC w/ the USB cable
3. Build the project in CCS
4. Run the debugger, and  hit Resume (F8). This  runs the code on the  microcontroller, and will allow you to step through the code and hit breakpoints.

I also apologize - in my earlier reply i had  forgotten to remove  resistor R8 of the MCU. After doing so, the responses should look like this:

Thanks again David.

Is there any "useful" code available to program the TMS57004 to control the bq76PL455EVM?


Regards,

Hi Ruby,

Unfortunately, we do not have any code published that is "plug and play", mostly for liability reasons. However, the example code will work well as skeleton code. The commands are already built in, it is just up to the user to decide how to use them.

Thanks again David.

I tried connecting the MCU to the EVM again and following the instructions that you outlined after "Time for MCU".
The JTAG LED (D8) blinks when I step through the code which I interprete as communication between CCS and the MCU.
However, after stepping through the entire code, neither the SCI1TX (D5) or SCI1RX (D6) on the MCU blink and there is no output on the wakeup, RX or TX pins on the EVM.
Aren't the SCI1TX (D5) or SCI1RX (D6) supposed to blink when data is being transmitted and received?
Are there specific jumper settings for J5 to J9 which may affect this?

You are correct, D5 and D6 should have activity. However, the flashes will be fast enough that it will  be tough to see. The best way to verify would be  to use a logic analyzer to see the data  transmit, or use  an oscilloscope and trigger  on the  falling  edge  of  the bq76PL455EVM RX pin.

I am using an  almost brand  new  TMS570 launchpad  - all  jumper configurations are as they were fresh from the box. The  only modification is removing R8. When removing R8, did you damage  something  on the  board or  were there  any solder bridges created? I would  verify that  the components  around that resistor were not damaged.

As far as the 455 EVM, the  only jumper that would cause major malfunction is  the comm mode  sel switch. Can  you verify  that both switches have been moved to the "single" position?

Could you share your  setup with me? Maybe I  can help find  something during the  troubleshooting process.

Hi David,

When I move both switches to the single position, the EVM does not turn on. The D22 Led remains unlit. I tried it in that position with the GUI and the GUI does not connect "Unable to open Com port".

However, when I send the first switch from SW2 to the on position and I leave the second switch in the "not on" position as seen in the picture below, the GUI connects to the EVM. 

Here is the set up, I connected the jumper wires as stated in sysmain.c

Ruby,

Both  comm mode select switches should be in the  "single" position" for proper use. You  would  still see communication with the switch in differential mode.

What  are you using  to power the  455 EVM? I do not see any connections at the BAT connector

I am using the EVM to balance a 36V battery as seen below.

Since you said that both switches need to be in the single position for proper operation and when I put both switches in that position, the GUI does not open, does that mean the board has an issue? The GUI only recognizes the board when I switch "switch 1" to the ON position.

Hi David, thanks for your patience.

I did the tests that you outlined, changed the usb to TTl cable. The EVM works as expected right now. The voltages at VIO is 5.1V in single mode and 5.34V in differential mode. The battery voltages are accurate. BAT16(s)-BAT10 are tied together.  So I have established that the board is working as it should.

Below is a screen shot of the measurement of the cells using GUI.

Cell 1, 6 and 10 are deliberately lower than the values of the other cells to simulate an unbalanced situation.

I also captured the RX and TX activity on the oscilloscope;

Yellow: TX

Blue: RX

Purple: Wakeup

However, I still have the same issue when using the TMS57004 with the EVM.

I left the EVM in single mode, connected the jumper wires as described in the code provided.

I imported the code and loaded it to the board. The wakeup pin goes high at the point in the screenshot below;

When I step through the code, the wakeup pin goes low after the gioInit(); function;

No activity on the RX and TX pin and Fault pin goes high when the code is run.

D22 illuminated

TX pin: 3.3V (Steady using oscilloscope)

RX pin: 3.3V (Steady using oscilloscope)

VIO: 3.3V

VDIG 5.34V

VREF: 2.5V

V5VA0: 5.31V

Hi Ruby,  I have  reached out to the email address you have provided with  a snippet of code that will enable  cell balancing  for  all channels - you can refer to the datasheet to see the register and what bits  to set if you wish to modify it.

Have you attached the TMS pins SCI1_TX and SCI1_RX pins to anything over 4.6V? I am wondering if the mcu pins are damaged possibly. The GIO pins are clearly working, since they are sending the wakeup command. 5VAO would be ~4.8 if the 455EVM was in sleep mode, and the other voltage references would be turned off. 

It is rubyrebz758@gmail.com

Thanks

Hi David,

I am facing the same issue as shown by Ruby. I was running through your complete conversation.

Can you also mail me the code for Cell balancing for BQ76PLEVM and TMS57004 launchpad?

Even i will check it out.

Email: ritulshah92@gmail.com

Thanks

Ritul

Hi David,

There is one command in sample code of BQ76PL455 in TMS570 launchpad

// Set cell over-voltage and cell under-voltage thresholds on a single board (section 2.2.6.1)
    nDev_ID = 0;
    nSent = WriteReg(nDev_ID, 144, 0xD1EC, 2, FRMWRT_SGL_NR); // set OV threshold = 4.1000V
    nSent = WriteReg(nDev_ID, 142, 0x6148, 2, FRMWRT_SGL_NR); // set UV threshold = 1.9000V

Now the registers 144 and 142 stores the OV and UV threshold values. Where can we see these registers in code  composer studio?

Also, tell me what code should be written to enable Cell balancing?

thanks

Ritul Shah

Ritul,

I will share the generic code here to enable cell balancing, and please refer to section 7.3.7 in the datasheet for an "english" description of how to enable balancing.

Using the API used in our example software:
nSent = WriteReg(0, 0x13, 0x08, 1, FRMWRT_SGL_NR); //configure CBCONFIG register
nSent = WriteReg(0, 0x1E, 0x0000, 2, FRMWRT_SGL_NR); //set BIST config (disable squeeze resistor)
nSent = WriteReg(0, 0x14, 0xXXXX, 2, FRMWRT_SGL_NR); //enable balance FET on X channel


The api is written as WriteReg(device ID, Register Address, Number of Data Bytes, Packet Type). In this case, the frame type is a Single (SGL) write without response (NR). Please note that you may also have to disable the balance Fets, depending on how long you have balancing enabled in the CBCONFIG register. Please note that the end user is also responsible for the balancing algorithm...


In regards to your second post, to see the registers in the 455 you must use the ReadReg function shown in our example code. If you set a breakpoint after the data is read, then you can watch the variables update in code composer studio, as the data will then be stored in the MCU as well.

Hope that helps,

Ritul,

I am glad  the last  post helped.

If you want to read from bottom board, try one of  the two ways:

// Send sample request to single board to sample and send results (section 4.2)
nDev_ID = 0;
nSent = WriteReg(nDev_ID, 2, 0x00, 1, FRMWRT_SGL_R); // send sync sample command
nSent = WaitRespFrame(bFrame, 27, 0); // 24 bytes data + packet header + CRC, 0ms timeout

Make sure  27 bytes is enough for 1 packet header byte, 2 CRC bytes, and 2 data bytes  per sampled channel. This method will have  the  pl455 immediately send the results  to the  mcu after the  ADC  conversion is complete.

Alternatively, you can do this:

// Send sample request to single board to sample and send results (section 4.2)
nDev_ID = 0;
nSent = WriteReg(nDev_ID, 2, 0x00, 1, FRMWRT_SGL_NR); // send sync sample command

,,,,,, do other things....

nSent = WriteReg(nDev_ID, 2, 0x20, 1, FRMWRT_SGL_R); // send read stored  values  command

nSent = WaitRespFrame(bFrame, 27, 0); // 24 bytes data + packet header + CRC, 0ms timeout

Either approach should work  and will  work for  reading from any single board. If  you are  going to  constantly poll the cell  voltages, you may also consider  disabling auto-monitor  do information does not "collide" and cause confusion

David,

I am using the code provided by you. I tried both by enabling and Disabling the Auto monitor. I am getting the values in bFrame as attached.

The values observed in the bytes are quite vague. I am unable to interpret them.

Please help me out.

Ritul,

I am glad it is working now.  Thank you for  the update.

bq76PL455A-Q1 Example Code.rarDavid,

I have attached my code for BQ76PL455. Please find it in attachment.

Even the EM1402EVM_code was run on this controller. The cell Voltage measurement comes the same as i got earlier. There is no change in the output voltage.