Forums

Jerry,

I can think of two possibilities:  1) isolation or  2) noise in the system  

1) There should be isolation between your computer (especially if it is a desktop computer) and the F28335. Buying an isolated emulator and plugging it into J2 of the docking station is one option.  (http://e2e.ti.com/support/microcontrollers/tms320c2000_32-bit_real-time_mcus/f/171/t/87408.aspx)

2) It sounds like you've wired the Docking station to your custom board.  I would make sure that these wires are short and each signal wire is twisted with a ground if possible.


Thank you,
Brett

Hi Brett,

Thanks for your quick reply. I also thought noise might be the possibility, could you give more details of how to eliminate noise?

Now before i go for the option one, i want to check if the wire connection of the docking station is correct or not.

The only connection between the Docking station and the AC/DC converter board are:

6 PWM signal wires

6 ADC signal wires

3.3V wire and its ground wire (We use the 3.3V of the Docking station to supply the chip SN74LVCC3245PW on the converter board, SN74LVCC3245PW is a butter which transfer the PWMs from 3.3V to 5V because the voltage level of PWM output from DSP is 3.3V)

With this connection, we can build, download and run the program without any error, and the PWM output and ADC output values are all correct. It is just when supply high voltage for the converter that CCS no respond and collapse eventually. Could you see anything wrong with this connection?

Kind regards,

Jerry

Jerry,

Based on what your description, your problem is likely isolation (option 1).

For option 2, I was mentioning that you can add a grounded wire to each signal wire that goes from the experimenter kit to the converter board.  This will help shield your signal connections, decrease the possibility of noise getting into your system, and improve the quality of these signals significantly.  Also make sure the wires are as short as possible.

(for example, attach a wire to experimenter kit ground, wrap the wire tightly around one of the PWM connector wire, and then attach the other end to the SN74LVCCC3245PW ground.  Repeat for the other signal wires.  Wrap the 5V supply wire as well.)


Thank you,
Brett

Hi Brett,

We use a isolated USB but the problem is still exist.

After further test, we found that the DSP and CCS work well if the line current of the converter is almost sinusoidal. But CCS collapse when the current is quite distorted (get many harmonics). The question here is the ADC of the DSP should definitely measure any kinds of voltage or current, so why the current would influence the DSP and CCS system?

Kind regards,

Jerry

Jerry,

What emulator are you now using?


Thank you,
Brett

Hi Brett,

We are still using the On-board USB emulator of the Docking station. What we do now  is use another isolated USB instead of using the original USB from TMS320C2000™ Experimenter Kit. But the CCS was still disconnected soon after the system runs.

Kind regards,
Jerry

Jerry,

Does the power supply to C2000 device remain steady when the system is running?  

Do you have the two Grounds (C2000 and system) connected?

Regards,
Daniel 

Hi Daniel and Brett,

We use the oscilloscope to observe the 5V and 3.3V of the Docking station, they were quite stable during the system is running.

Speaking of the grounds, I am a little confused. As I discussed with Brett before, we use the PWM and ADC of the DSP, so there were two kinds of grounds in the custom board, one is the ground of PWM (let us denote it as digital ground), another one is the ground of AD signals processing (transfer the output of voltage sensor or current sensor to 0~3V, let us denote this ground as analog ground), and these two grounds are connected with inductor. we connect the digital ground to the GND pin of Docking station.

Now our questions are:
1. Though there are several pins marked "GND" in Docking station, but after check the schematic, there are actually three kinds of ground, i.e. DGND, AGDN, and GND-ISO. What exactly are they, respectively? I could not find any TI docs mentioning them. More importantly, which ground of Docking station should be connected to the digital ground of my custom board?
2. For ADC, should the ADCLO of the DSP connected to the analog ground of our custom board? (We never connect them before during the experiment)

Kind regards,
Jerry

Jerry,

1. If you look at the Docking Station's schematics and the controlCARD's schematics, you will see that the analog (ADCGND)and digital grounds (DGND) are connected together on the controlCARD.  GND-ISO (which links to the "Gnd" pin on J3 of the Docking Station) is only used by the SCI transceiver on the F28335 controlCARD.  I would recommend not using this GND-ISO pin.  
2. In a system where the 'F28335 resides on your application baseboard, yes.  However, in the controlCARD system, the controlCARD ties the grounds together.  You should not need to do anything other than tie your PWMs + ADCs + 3V3 + GND to your converter board. 

Having only one combined ground plane simplifies layout at the expense of more analog noise.  The greater analog noise was deemed to be minimal, so in the controlCARD we've tied them together in the controlCARD.


Thank you,
Brett

Jerry,

I believe that the CCS crashing is related to the emulator connection failing during high power operation.  This is because you said that-

1) C2000 MCU continutues to run properly (only CCS disconnects) with AC power applied
2) CCS does not disconnect when no AC power is applied
3) CCS disconnects during high noise situations (non-sinusoidal currents)

Have you always used an isolated USB or did you try non-isolated USB initially?

Regards,
Daniel

Hi Brett,

Thanks for your kindly reply these days. The CCS disconnection seldom happened when we use a laptop as a host.

Kind regards,
Jerry