• State Resolved
  • Locked Locked
  • Replies 28 replies
  • Answers 6 answers
  • Subscribers 82 subscribers
  • Views 1361 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TIDA-010232: TIDA-010232

Hemant Purohit
Prodigy 10 points

Part Number: TIDA-010232
Other Parts Discussed in Thread: AMC3330, TLV6001

Hello,

 I am using TIDA-010232 as my reference design. I am not getting how the ground fault and warning for insulation are being calculated. I have gone through the design guide and got the formula to calculate the RisoN and RisoP, as well as Vp and Vn. There is no value interface in that formula which is fetched by the ADC circuit through the Microcontroller.  However, the Viso formula is there to give you the calculated value with the ADC value. I need to understand how the insulation fault is calculated.

  • +1
    TI__Expert 4845 points

    Hi,

    Thanks for reaching out. 
    I do not understand the question 100%, but let me share the flow of calculating RisP and RisoN

    1. Switch SP is closed and Vp is measured. You can use formula (18) on page 16 of the design guide to calculate the Isolation voltage which is Vp in case SP is closed

    2. Open SP , close SN, measured Vn, Again you can use formula (18) to calculate the Isolation voltage (In this case Vn)

    3. Calculate RisoP and RisoN using formulas (13) and (14)

    4. Check thresholds. If RisoP or RisoN is smaller then the fault threshold indicate a fault. If RisoP or RisoN is smaller than the warning threshold indicate a warning.

    Hope this helps.

    Best regards,

    Andreas

  • +1 in reply to Andreas Lechner
    Prodigy 10 points

    Thank you Andreas for this answer. But is there any threshold data to be used to test the circuit or verify the calculated data?

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi,

    We used the threshold values used in standards. 500Ohms/V for a warning and 100Ohms/V for Fault.


    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    I also need to understand that there is the data and calculation in the Excel provided in the name of Error calculation. What is that data for, as it is completely filled with 200k ohm for real RisoP & RisoN? 

  • +1 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi,

    In the error Excel sheet, you can enter different values for the true RisoP and RisoN in fields A535 and B535 and then see the maximum worst case error for the measured/calculated  RisoP and RisoN in fields AP535, AP536, AQ535 and AQ536. This is done on a theoretical worst-case analysis. Expected values are better.
    The conditions are specified in B3 to B20.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    How can I test this product for these case scenarios? We have a 1000V DC power module, but it is very unsafe to directly test on that. So how can I safely test this system and its reliability?

  • +1 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi,

    The test procedure we did is described in chapter 3.4 Test Setup.
    You can also start testing at lower voltages, but accuracy will be worse if the resistor dividers are not adjusted.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    Hi,

     How the resistor divider circuit value is selected as per the input voltage source. I have used equation 17 to have that value. What voltage drop we should have at each resistor in the resistor divider network?

  • +1 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi,

    The resistor divider is selected in order to have a maximum of 1V at the input of the AMC3330 in case the complete Bus voltage would be present across the resistor divider.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    I was doing the calculation of V isolation RisoP and RisoN. So the VinADC be common for both the calculations or it will be a different value to calculate the RisoP and RisoN, like VinADC_P and VinADC_N.

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi Herman,

    In general, VinADC will be different for both measurements. VinADC is just a intermediate step to calculate Vp and Vn.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    So in order to calculate Vp and Vn we need to use 2 different VinADC values, like VinADC_P to calculate Vp and VinADC to calculate Vn. Then we going to use Vp and Vn to find RisoP and RisoN.

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi Hemant,

    Yes, that is correct.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    Hi Andreas,

    I have put the formula but I don't know why the values don't make any sense to me. I am attaching the pictures of the formula and the values. The formula doesn't return a value under 300k ohm resistance at RisoP or RisoN.

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi Hermant,

    There are some mistakes in the formulas. Its not possible that a isolation voltage is greater than VDC. 

    Let's do the complete calculations, if we assume RisoP = RisoN = 300kOhm.

    Let's first calculate the expected Value for Visolation (Vp) when SP is closed like in Figure 1-5 with equation (4)

    Vp = Vdc/(1+(RisoN /(RisoP * (Rst+RinAMC)/(RisoP+Rst+RinAMC))

    Vp= 1000/ (1 + (300000/(300000*(500000+340)/(300000+500000+340)) =  384.6757 V

    Now we calculate Vn for the case the switch SN is closed like in Figure 1-6:

     Vn = -Vdc/(1+(RisoP /(RisoNP * (Rst+RinAMC)/(RisoN+Rst+RinAMC))

    Vn= -1000/ (1 + (300000/(300000*(500000+340)/(300000+500000+340)) = 384.6757V


    RisoP  = -(RinAMC +Rst) * (VDC+Vn+Vp)/Vn = -(340+500000) * (1000+(-384.6757 - 384.6757))/ (- 384.6757) = 300000

    RisoN = (RinAMC +Rst) * (VDC+Vn+Vp)/Vp = (340+500000) * (1000+(-384.6757 - 384.6757))/ 384.6757 = 300000

    Hope this helps.

    Best regards,
    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    I think there is a mistake in the formula. In equations 13 and 14 from the latest design guide, the formula used is 

    RisoP= -(RinAMC+Rst)*(Vdc+Vn-Vp)/Vn

    RisoN= (RinAMC+Rst)*(Vdc+Vn-Vp)/Vp

    But you have used 

    RisoP = -(RinAMC +Rst) * (VDC+Vn+Vp)/Vn

    RisoN = (RinAMC +Rst) * (VDC+Vn+Vp)/Vp

    Which is the correct one?

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi Hemant,

    The one in the design guide is correct. Sorry for the confusion.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    So, then my answer is correct for Viso RisoP and RisoN? Or I should go with your formula?

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Please go with the formulas in the design guide.


    See the corrected calculation below (Corrections are highlighted in blue):

    Vp = Vdc/(1+(RisoN /(RisoP * (Rst+RinAMC)/(RisoP+Rst+RinAMC))

    Vp= 1000/ (1 + (300000/(300000*(500000+340)/(300000+500000+340)) =  384.6757 V

     Vn = -Vdc/(1+(RisoP /(RisoNP * (Rst+RinAMC)/(RisoN+Rst+RinAMC))

    Vn= -1000/ (1 + (300000/(300000*(500000+340)/(300000+500000+340)) =  - 384.6757V

    RisoP= -(RinAMC+Rst)*(Vdc+Vn-Vp)/Vn

    RisoN= (RinAMC+Rst)*(Vdc+Vn-Vp)/Vp

    RisoP  = -(RinAMC +Rst) * (VDC+Vn-Vp)/Vn = -(340+500000) * (1000 + (- 384.6757) - 384.6757)/ (- 384.6757) = 300000

    RisoN = (RinAMC +Rst) * (VDC+Vn-Vp)/Vp = (340+500000) * (1000 + (-384.6757) - 384.6757)/ 384.6757 = 300000



    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    However, we have to use equation 18 for VisoN and VisoP calculation, as per our previous discussion. I have used that formula and the RisoP and RisoN values are very much different. We have to calculate the fault values as per the ADC value and not the fixed values.  

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Yes,
    as you can see in Figure 1-5, if Vp is 384.6757 V VinAMC will be 384.6757 * 340/(340+500000) = 0.26140 V 


    AMC3330 has a gain of 2 (GainAMC = 2): This mean the output of AMC3330 = 0.5228 V

    Now to convert the differential output of AMC3330 to single ended signal we use TLV6001 with a gain of 0.825 (as mentioned  below equation 19) and a reference voltage of 1.65V (Vref = 1.65, Gaindiff2single = 0.825) This means after this stage the voltage  is VinADC =  0.5228 *0.825  + 1.65 V = 2.08131


    If you use these input in equation 18:

    Visolation = (VinADC -Vref)/Gaindiff2single * (RinAMC +Rst)/RinAMC *1/GainAMC 


    Visolation =   (2.08131 - 1.65)/0.825 * (340+500000)/340 * 1/2 = 384.67 

    Which is the voltage we started at, which means the formulas work as intended.

    Hope this help.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    Hello Andreas,

     We have booted up the device and for initial testing, we have given a 30V input source. The output at ADC is 1.66V constant. I am not able to understand what is the meaning of this value. and it is not changing even if I increase or decrease the input source.

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hello Hemant,

    1.65V is the bias voltage. If you connect 0 V to the input of AMC3330 a voltage of 1.65V will be present at the ADC.
    The input voltage at AMC3330 is +/-1V. This translate to+/- 2V at the output of AMC3330. Since the ADC in the C2000 is single-ended it can only ready positive voltages.

    So we use TLV6001 to convert the +/-2V to 0V to 3.3V. The middle of 0V and 3.3 V is 1.65. so 0V at the input of AMC3330 translates to 0V at the output of AMC3330 which translates to 1.65V after TLV6001 at the ADC input.

    To see the voltage change you need to start the software. This is done in Code Composer Studio. 
    You can download the example project on ti.com (The second one called TIDCG73): 

    This software works with the F280049C Launchpad: https://www.ti.com/tool/LAUNCHXL-F280049C
    In the software you need to set a flag called "startTest".
    Then the software will turn on SP and measure VP, afterwards it will turn on SN and measure VN, during this time the ADC input voltage will change.


    If  you want to do this in smaller steps you can use the signals on J5:

    Applying 3V between GND(Pin8) and SWITCH_POS(Pin3) will turn on SP, then the voltage on the ADC will change accordingly.
    Applying 3V between GND(Pin8) and SWITCH_NEG(Pin7) will turn on SN, then the voltage on the ADC will change accordingly.

    Hope this helps.

    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    Hello Andreas,

     For current testing, we are not using the C2000. We are using our controller with ADC input. So is that sample code or tool feasible with another controller as well?

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hello Hemant,

    The code is written in C and can be ported to another MCU, but the easier way is t use the C2000 Launchpad.


    Best regards,

    Andreas

  • 0 in reply to Andreas Lechner
    Prodigy 10 points

    Where can i get the sample code for this design.

  • 0 in reply to Hemant Purohit
    TI__Expert 4845 points

    Hi,

    You can get the code on https://www.ti.com/tool/TIDA-010232

    I showed a screenshot two posts ago.

    Best regards,

    Andreas