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.

WEBENCH® Tools/TPS65400: Compensatrion network calculactions in Webench do not corespond to data sheet, no bode plot in webench

Part Number: TPS65400

Tool/software: WEBENCH® Design Tools

I am designing using TPS65400, to make question easier lets focus on SW3 .

My parameters are Vin=5V, Vout =1.8V, Iout 1.8A, Cout =2x 10uF 1mohm ESR

From webench I am getting Rfb1=20k, RfB2=16k, Rcomp=9.09kohm, Ccomp=1.8nF, Ccomp2=39pF

Using eqation for Rc from data sheet page 24 Table 3 Rc=(2*pi*fc*Co*V0)/(Gm*Gmps*Vref), and assuming as data sheet says Gm=120uA/V, Gmps=10A/V and Vref =0.8V I am getting Rc=27kOhm while webench calculates 9.090kOhm.

Then of cource since Rc is wrong all othe calculations ( Cc and Croll) are wrong.

But even assuming Rc=9090ohm, acc to same data sheet Cc=(Rl*Co)/Rc= (1*0.00002)/9090=2.2nF not 1.8nF given by webench?

Also  Cc2=Croll=(Resr*Co)/Rc=(0.0005*0.00002)/9090=1.1pF not 39pF given by webench?

Could you please explain the discrapencies and advise what equations should be used for calculations.

Also is there a chance to run simulations (especially bode plot) in webench for TPS65400 , I can not find options to run them in the menu?

Thanks Sebastian

Since Rc i sused in all othe requations all ot

  • Hi Sebastian,

    The equations used for compensation used are same as in datasheet for this device. The differences are in the parameter values being shown in the datasheet (Fig 17 on page 24) with Gm=120uA/V and gmps=10A/V and what is actually there in the electrical characteristics table 7.5 on page 8 where Gm=133uA/V.  The gmps value is actually different for output 3/4 as compared to outputs 1/2 as they have different power stages and the numbers we have are gmps(outputs 1,2) = 14.49 and gmps(outputs 3,4) = 7.5. Along with this, the output capacitance value might be different than 20uF as we also do DC bias voltage derating based on manufacturer data when we have it in our database otherwise we use linear derating. With the linear derated model, for a 5V output, a 10V rated output cap with nominal value of 44uF will have 22uF value.

    In order to simulate bode plot, you need to click on a single channel and then click "Sim" tab. Please see attached screenshot below:

    Regards,

    Amod

  • Hi Sebastian,

    Another parameter I did not mention but is key is the choice of crossover frequency Fc. We are using a little more conservative estimate based on an approximate dominant pole and esr zero location and we limit initial Fc max value to 50kHz. Of course, you can update the R-C values using the Re-Comp tool to suit your needs.

    Regards,
    Amod

  • Hello Amod, thak you for your answer.

    I tried using parameters you have suiggested:

    Vin=5V, Vout =1.8V, Iout 1.8A, Cout =2x 10uF dereated to 14uF ( webench value ) 1mohm ESR, Gm=133uA/V, Gmps=7.5A/V, fc =50KHz

    Rc=(2*pi*fc*Co*V0)/(Gm*Gmps*Vref) gives 9.9kohm still not 9.090 kohm but I am not sure if this not close enough so happy to accept this.

    To check the correctness I jump to another channel SW1 , where I wantto have 1V at 2.6 A so parameters are:

    Vin=5V, Vout =1V, Iout 1.8A, Cout =2x 10uF dereated to 14uF ( webench value ) 1mohm ESR, Gm=133uA/V, Gmps=7.5A/V, fc =50KHz

    Rc=2.9k ohm not 3.48kohm given by Webench

    What do you think where the difference can come from?

    I guess something is still not right also looking at Cc2=Croll=(Resr*Co)/Rc=(0.0005*0.000014)/9090=7.7pF not 39pF

    On the Bode plot the sim button is not available:

    Sebastian

  • Hi again,
    I made a mistake above,
    When calculating Rc for SW1 with Vout =1 V I have chosen wrong capacitance, since we operate at 1V dereation gives capacitance of 18uF,
    so Rc=3.667kohm, closer to 3.48 but still not there.
    Sebastian
  • Hi Sebastian,

    I can take a look at the detailed calculations. If you can please share your project and send the public link here, i can look at your design details for these calculations.

    For the simulations, once you click on the channels on the left and then see the Sim button on the top?

    See step below is to click on any channel and then you should observe the sim tab. Bode plot simulation is the first one in the dropdown. Please share your screenshot if you are unable to see this same type of screenshot. Please use the "Attach file" feature to get your screenshot through.

    Regards,

    Amod

  • Hi Amod,

    Just noticed my screenshot (with missing sim button) was not attached:

    Also sharing the project is not a problem at all :

    https://webench.ti.com/wb5/OpenPublicSharedProject.jsp?id=E2E34E576E1BDE4A&s=c

    Thanks in advance

    Sebastian

  •  I can see the screenshot somehow does not get through, so trying to attach it as file.

    Hope this works

    Sebastian

  • Hi Sebastian,

    Thanks for sharing the design. I ran the calculations again and found slight differences in gmps and derated Cout value which may not show up in the UI due to rounding. I am sharing the details of the values below in the screenshot:

    The snapped values are the ones being used in the actual design. 

    I am not sure why the simulation tab is not visible for you once inside a "channel" view. If you can please confirm that you see for a new design, then it might help debug this issue. Another fallback option is to use the PSPICE average model under: www.ti.com/.../toolssoftware

    Regards,

    Amod 

  • Hi Amod,

    Thanks that it is great we have Rc covered.

    Now I would like to look at 

    Cc2=Croll=(Resr*Co)/Rc=(0.0005*0.0000132)/9128=0.723pF not 47pF given by Webench.

    On the subject of Sim button: yes I can see it for any new design of single regulator where I do not have to select a section of it.

    What do you think

    Sebastian

  • Hi Sebastian,

    Cc2 has an issue in the calculations in that it uses an estimated ESR value in the equation. It should be using the actual ESR and in that case, 0.723pF is correct. Basically, you just need a little capacitance to ensure good gain margin and high frequency noise reduction.

    On the sim button: Not sure why the new design has a sim tab. But, if you see it on an individual channel level (TPS65400/1, TPS65400/2, etc.), then you can simulate bode plots on those channels. The results should match very closely with "Opvals" and "Re-Comp" bode plot values.

    Regards,
    Amod
  • Hi Amod,
    Thanks so Cc2 is explained.
    Sim button I was probably not precise, I am saying I can get it for any other regulator apart from TPS65400.
    For TPS65400 it is stable situation regardles on the design that I am not getting sim button.
    Sebastian
  • Hi Sebastian,

    We will need to look at the simulation issue in depth to figure out what is going on. It could take some time to debug and see if there is a fix for it. I would recommend trying out the PSPICE average model on the www.ti.com/.../toolssoftware page if you have access to the simulator.

    Regards,
    Amod
  • Ok that is finr.
    Thanks you very much for your help
    Sebatian
  • Hi Sebastian,

    I am unable to reproduce the simulation button issue for TPS65400.

    Kindly note that TPS65400 is PMU IC (Power Management Unit) consisting of 4 Buck channels. Please find below snapshot from Data-sheet FYR.

    WEBENCH also has implemented this device in a similar way where "PMU1 Total" represents total PMU IC whereas "PMU1_CH1", "PMU1_CH2", "PMU1_CH3" and "PMU1_CH4" are the individual channels [BUCK1, BUCK2, BUCK3 and BUCK4] of TPS65400.

    Currently there is no way to run electrical simulation for TPS65400 at system level [i.e PMU1_Total]. Instead you can run Bode simulations for individual channels. Just select any of the channels (Ch1 is selected) in below image and select on "Sim" option [see below image]

    Somehow for your shared project, we are unable to get "Sim" Option even for individual channels.

    Could you please try recreating new project from http://www.ti.com/product/TPS65400  and let us know if you see "Sim" option for individual channels ? Alternatively you can use below shared project for your input/output requirements.

    https://webench.ti.com/wb5/OpenPublicSharedProject.jsp?id=C7C0A630F6BEA174&s=e

    Thanks & Regards,

    Harish

  • Hello Harish

    I am glad you can see the issue in the simulation I have shared.

    I can confirm that in the one you have shared, the sim button is visible so I can simulate , maybe it depends on type of the account used?

    I have recreated the new project from link as you pointed and still no sim button for individual channels.

    Sebastian

  • Hi Sebastian,

    Thanks for confirming that you are able to view simulation for the shared project. You can use shared project to modify the design as per your requirement and perform Bode simulations. Kindly let us know in case you face any issue or have any query, we will be happy to help you.

    We have taken note of the issue (sim button not visible when you are creating design). However the fix for this would take some time.

    Best Regards,
    Harish
  • Thanks for all your help
    Sebastian