Need help in using TX_SDK_V1

Hi Charles,

 

       Thank you very much for your reply!

       If I use +-48V for VPP and VNN, what is the best value of VSUB?

       The datasheet only say VSUB should <=VNN. And I see your reply that 'Yes, the output can overshoot more than 10% if you are connecting it to the transducer. When that happens, we recommend VSUB to be 10% (depending on the overshoot) lower than VNN."

       So VSUB should <=-52.8V. And is there a best value? Is VSUB=-65V the best value?

       Thank you so much!

Frank

Hi Charles,

       Got it.

       Thank you so much!

Frank

Hi Charles,

       We are using the TX_SDK_V1 envaluation board with the powering up sequence as: GND,VSUB(-50V),3.3V,1.8V,+5V,-5V,+10V,-10V,+50V,-50V. However, when we power up the +50V, the board is damaged. We found the -5V and 3.3V are both short current to GND.

       Is there anything that we did wrong？ How could this happens？

      Thank you.

Hi Frank,

Sorry to hear your board is damaged. You follow the power sequence correctly.

What is the position of J_SUB? Since you are using separate power supply for VSUB & -50V, there should be a short between pin 2 & pin3 of J_SUB (page 5 of user guide)

Hi Charles,

           Thank you very much for your reply.

           Yes, the short is between pin 2 & pin3 of J_SUB. We think we didn't do anything wrong according to the user guider.

           We've bought all of the main chips on the board. But before we change new chips onto the board, we should find the cause of the damage.

           Our power sequence was : GND,VSUB(-50V),3.3V,1.8V,+5V,-5V,+10V,-10V,+50V,-50V.

           1. Does +50V and -50V need to be powered on simultaneously?

           2. Does -50V need to be powered on before +50V?

           3. Is any possiblity that the Power Jitter damage the chips when +50V is powerd on? But 3.3V and -5V are short to GND, that's weird.

           We are stuck by this problem and can not move on. Really need for TI's advice.

          Yours sincerely,

Frank

Hi Frank,

1) The +50V & -50V don't have to power on simultaneously.

2) -50V doesn't need to be power on before +50V as long as Vsub is powered on first.

3) A little noise on the power supply shouldn't damaged the devices.

Sorry that you are stuck. I'll try to help as best as I can.

Did you set the current limit on all your supplies, especially VSUB? If you set current limit on VSUB to be less than 10mA, you might damaged it part during power up. To be on the safe side, I would set the current limit to 100mA for VSUB.

Are you setting the LM96570 in CW mode? You can't be in CW mode if your VPP & VNN is +50V & -50V, respectively.

How are your supplies set up? Are they similar to the attached setup?

TX_SDK_V1 LM96551 Lab Bench Power Supply Setup.ppt

Hi Charles,

Thank you very much for your reply.

1. We set the current limit on the supplies, but none of them are less than 100mA. The current limit for VSUB is 1A.

2. We didn't program the LM96570 in CW mode. And the board got damage when powering up. We didn't make any settings on the GUI Control Panel on PC.

3. We use the following supplies set up. Their GND are connected. We connet all of the supplies to a Switch Board and connet the switch outputs to the TX_SDK_V1 board. We use the Switches to control the power up sequence. And we use a LED for each Switch to indicate whether  the power supply is on or off.

  We have tested the power up sequence of the Switch Board before connet to the TX_SDK_V1 board. And after damage, we also checked the Switch Board, the power up sequence is right.

        Thank you.

Frank

Hi Frank,

Can you draw me the schematic of your switch board? Specifically, how are you connecting the -50V power supply to VSUB and VNN (-50V)?

Thanks,
Charles

Hi Charles,

      Thank you very much for your reply.

      And I'm sorry for reply so late, we went to a conference for several days.

      The schematic is as follows:

      And we takes all the main chips off the board, and find that:

     1.  The VSS(-5V) of LM96530 is short to GND

     2. The VLL(3.3V) of LM96551 is short to GND

     We don't know why these two chips are damaged.

     And we have a question: when the power supply is off, should the output of the supply be 0V or be open circuit?

     For our switch board, the output is open circuit when the switch is off. And we think the direct ouput for a power supply may be 0V when it's turned off.

     Thank you.

     Yours sincerely,

Frank

     

Hi Frank,

Are you sure your schematic is correct? All the diodes are forward bias. Can you measure the voltage at all the supplies when not connecting to the TX_SDK_V1 board?

Regards,

Charles

Hi Charles,

           Thank you for reply.

           Yes, we are sure the schematic is correct. The diodes are lumious diodes, just to indicate us whether the supply is on or off.

           We measured the voltage at all the supplies before connecting to TX_SDK_V1 board, and checked all the voltage are connect and the sequence are all right.

           Thank you.

Frank

Hi Frank,

I don't know what could happen to damage the board if you follow the user guide & power up sequence correctly.

If I were you, I would take out and replace all the active devices on the TX_SDK board to make sure there are no shorts. I would then power up the TX_SDK board without your switch board first. If it's fine, I then power up the board with your switch board.

Regards,

Charles

Hi Charles,

           Thank you for reply.

           Here is the new update:

           1. We have taken all of the chips(6 chips:LM96551, LM96530, LM96570, 74AC08MTC, LMV321, FXL4TD245BQX ) off the board, and we checked that there are no short between all the powers and GND on the board.

           2. We take 6 new chips onto the board, and follow the powering squence, but when we turn on -50V, the 3.3V are overloaded.

           3. We take all the 6 chps off, and take a new LM96570 onto the board, with the other 5 chips not mounted, as shown in the following picture. According to the schematic of TX_SDK board, the LM96570 only take 1.8V and 3.3V. We take the following powering sequence:

           step0. We checked that, all the powers on the board are not short to each other or to the GND.

           step1. Turn 3.3V on.               3.3V takes neally 0mA.

           step2. Turn 1.8V on.               1.8V takes about 22mA.

           step3. Turn VSUB on.             Nothing happens, that is normal.

           step4. Turn -50V on.               The 1.8V and 3.3V are both overloaded. And when we take LM96570 off, we find the 1.8V and 3.3V of LM96570are both short to GND.

            So this is really strange, the  LM96570 only takes 3.3V and 1.8V. And there is only LM96570 on the board. And we measure the impedence between 3.3V and -50V is several MOhm,  impedence between 1.8V and -50V is also several MOhm.

             And when we did the test, we use the power supply's button to turn on/off the power, without our switch board.

            4. We changed another LM96570 and follow the steps in 3. The same thing happens.

            So the problem is narrowed to the LM96570.

            Till now, 4 pieces of LM96570 are damaged.

            That is strange, the powering of -50V damaged the LM96570, we can't believe that.

            Any possibilities for that?

            Thank you.

Frank

And does the TX_SDK_V1 board must be connected to the FPGA board before powering up?

And does the TX_SDK_V1 board must be connected to the FPGA board before powering up? I think it's not a must.

Hi Frank,
The -50V supply only goes to the LM96551 and the VSUB supply goes to LM96551 & LM96530. I have no idea why it would effect the LM96570. It doesn't make any sense.
You are correct that the FPGA board doesn't need to be connect to the TX_SDK board during power up.
Regards,
Charles

Hi Charles,

      We'v found the reason why the chips are damaged and we'v solved the problem.

      The switch we use is a mechanical switch, and when we power on, the +50V power goes up very quickly. Since there is a capacitor about 100uF between -50V and GND on the TX board. There is a large current on the GND when power up. Since the 3.3V GND and -50V GND are connected and the 3.3V GND is near the board and the +50V GND are on the far-end. The pulse current makes a big Voltage jump on the 3.3V GND, about 4.5V. So that makes the 3.3V up to 7.8V which damaged the chip.

     We tried to make the 3.3V GND and -50V GND conneted in a Star Topology, which makes the jump only a little smaller, still has a voltage about 7V.

    So, we take away the mechanical switch and use the power on-off button directly. Using the Power on-off button, the output of the power goes up much slower that using the mechanical switch. So the pulse current is very small, so the Voltage is very small. That solves our problem.

     Now, we have used the TX board to drive our ultrasound probe successuly. ^ ^

     Thank you very much for all your help!

     Yours sincerely,

Frank

Hi Frank,

I'm glad that you found and fix the problem.

Regards,

Charles

I'm already designed power module to supply for this board.