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.

LM3478: AN-1990 Compensation for Current Mode Control SEPIC

Part Number: LM3478

Hi team,

Customer has some questions

  • Good
  •  I have some problems with AN-1990 Compensation for Current Mode Control SEPIC
  •  Converters as the calculations I get are nothing like the ones shown in AN-1990.
  •  1 the results of the coefficients. Appendix B (7)
  • DC0 6.728e-13 --DC1 1.874e-16
  • DC2 2.732e-21--DC3 2.357e-26
  • DC4 3.557e-31--DC5 1.7657e-31
  • DC5 1.1654e-41
  • Appendix B (8)
  • N0=8.9843e-13;
  • N1=1.5273e-18;
  • N2=1.0377e-22
  • N3 =2.016e-28
  • N4 =1.958e-33
  • N5 =6.650e-36
  • N6 =6.457e-44
  • The parameters of the AN-1990 sheet are used
  • 2---the transfer function (6) in Matlab (6)
  • G= ((NCC*)/(DCC*RSN))*ADC;
  • 3 I'm not sure, but Cv1 is 0 I don't know if it's an error or but that equation the result is 0.

Any bug or help would be great.

 Finally if you have a correct sample of the calculations in matlab I would appreciate it.

Here is AN-1990

https://www.ti.com/lit/an/snva405a/snva405a.pdf

So it only occurs to me that I post by my "exemple.txt" and the attached image and have a professional compare, with
 AN-1990 Compensation for Current Mode Control SEPIC
 Converters
 With the calculations(6), APPENDIX A-
 APPENDIX B
 ,since the results do not match.
 And mainly CV1 =0.
VIN =5;   %--PARAMETERS OF THE LM3478
VOUT= 5;
ROUT =10;
COUT =0.0001;
RCOUT =0.05;
L1 =0.000033;
L2 =0.000033;
CS =0.000001;
FSW =400000;
RSN =0.02;
RSL =2000;
VREF =1.26;
GM =0.0008;
RO =47500;
VSL =0.092; %---END PARAMETERS
D =0.5;
RF2 =10000;
RF1 =29700;
ADC=0.180;
LM=(D^2)*L1+((1-D)^2)*L2;
MC=(VSL+0.000040*RSL)*FSW/RSN;
T2 =1/FSW;
TM=(T2/2)*(2*MC+(VIN/L1)+(VIN/L2));   %--------
s = tf('s'); 
CC0 =L1*L2*LM;
CC2 =L1^2*L2^2*CS;
D0=ROUT*((1-D)^2); %---Δ(s)coefficients of(4)
D1=LM+((1-D)^2)*RCOUT*ROUT*COUT;
D2=LM*(RCOUT+ROUT)*COUT+((1-D)^2)*(L1+L2)*ROUT*CS;
D3=L1*L2*CS+((1-D)^2)*(L1+L2)*RCOUT*ROUT*CS*COUT;
D4=L1*L2*(RCOUT+ROUT)*CS*COUT;

N0 =5*10; %---(3)Nn(s);  %--From(2)ND(s)
N1 =VIN*RCOUT*ROUT*COUT-(D^2/(1-D)^2)*VIN*L1;
N2 =VIN*(L1+L2)*ROUT*CS-(D^2/(1-D)^2)*VIN*L1*RCOUT*COUT;
N3 =VIN*(L1+L2)*RCOUT*ROUT*CS*COUT-(D/(1-D)^2)*VIN*L1*L2*CS;
N4 =-(D/(1-D)^2)*VIN*L1*L2*RCOUT*CS*COUT;
NN0 =D*(1-D)*ROUT; %--From (3)NN(s)
NN1 =D*(1-D)*RCOUT*ROUT*COUT;
NN2 =(1-D)*L2*RCOUT*CS;
NN3 =(1-D)*L2*RCOUT*ROUT*CS*COUT;
CD0=(VIN*L1*L2)/(1-D); %--The coefficients of (7) 
CD1=L1*L2*LM*TM+(D/(1-D))*((1-D)*L2-D*L1)*VIN*L1*(T2+(L2/ROUT*(1-D)));
CD2=(VIN*L1*L2/(1-D))*((L1+L2)*CS-L1*T2*(D^2/ROUT*(1-D)));
CD3 =L1^2*L2^2*CS*TM;
CV0=(1-D)*L1*L2;
CV1=D*L1*(LM-D*L1)*T2;
CV2=(1-D)*L1*L2*(L1+L2)*T2;
NC0 =CC0*N0;  %---coefficients of (8) 
NC1 =CC0*N1;
NC2 =CC0*N2+CC2*N0;
NC3 =CC0*N3+CC2*N1;
NC4 =CC0*N4+CC2*N2;
NC5 =CC2*N3;
NC6 =CC2*N4;
DP0 =CD0*D1+CD1*D0-CV0*N1;   %---The coefficients of (7)
DP1 =CD0*D2+CD1*D1+CD2*D0-CV0*N2-CV1*N1-CV2*N1;
DP2 =CD0*D3+CD1*D2+CD2*D1-CD3*D1-CV0*N3-CV1*N2-CV2*N1;
DP3 =CD0*D4+CD1*D3+CD2*D2-CD3*D1-CV0*N4-CV1*N3-CV2*N2;
DP4 =CD1*D4+CD2*D3+CD3*D2-CV1*N4-CV2*N3;
DP5 =CD1*D4+CD3*D3-CV2*N4;
DP6 =CD3*D4;
DCC =DP0+DP1*s+DP2*s^2+DP3*s^3+DP4*s^4+DP5*s^5+DP6*s^6;
NCC =NC0+NC1*s+NC2*s^2+NC3*s^3+NC4*s^4+NC5*s^5+NC6*s^6;
G= ((NCC)/(DCC*RSN))*ADC;
bode (G);
Could you help?
Thanks & Regards,
Vivian
  • HI Vivian,

    Thank you for reaching out but I am immensely confused first of all what do they mean by Matlab file? I do not think there is any Matlab file available on the TI website.

    second of all, the coefficients from AN-1990 what does he mean he does not get like the ones in the AN-1990, those are formulas so, I am very confused could you please elaborate and clearly write the question instead of copy pasting the customers last message as I am not aware or in the loop of what were you talking about.

    Thank you 

    BR,

    Haroon

  • Hi Haroon,

    Thanks for your quick reply.

    These contents are not the incomplete description or final information of the customer, but the complete and detailed description. 

    I have forwarded your reply to the customer. I hope he can reply to your question or re-describe her question as soon as possible.

    Thanks for your help.

    Regards,

    Vivian

  • Hi Vivian,

    Thank you for your reply, I see. The question seems to be not very elaborated, 

    Looking forward to her response then.

    Thank you

    BR,

    Haroon

  • Hi Haroon,

    Customer replied:

    "first of all what do they mean by Matlab file? I do not think there is any Matlab file available on the TI website"

    --About the matlab file I shared it in (.TXT) format instead (.M).
     Why did I copy the code into a file (.TXT)
     THE TXT FILE CONTAINS THE CODE IN MATLAB.
     But I share the (.m), from matlab.
    when I mean matlab file, I mean ".m".
    First.
     Ask about transfer function, not sure anymore
     If in equation (6)?.
     VOUTN =Ncc(s)/
                      Dcc(s)RSN * Vc
     The question is (S) represents "S1"
     ,"which is the sum of iL1 and iL2" o represents "s" of the transfer function.
    Or it just (s) doesn't represent anything and is part of DCC AND DCC AND I'm just confused S1 WITH S.
    Second.
    Cv1 = DL1(LM - DL1)T2 is the only parameter that is always 0, is it normal?
    NOTE. THE SAME PARAMETERS SHOWN IN THE DOCUMENT ARE USED.
    forget about coefficients
    Could you help?
    Thanks & Best regards,
    Vivian
  • Hi Vivian,

    I am sorry but I am not an expert of MATLAB, could she please ask me what her question about something specific is?

    BR,

    Haroon

  • Hi Vivian,

    I am sorry we could not help so far with this issue.
    Has the question already been resolved, or is there still something we can help with from our side?

    Please let me know if there are any specific questions on SEPIC design compensation.

    Best regards,
    Niklas

  • Hi Niklas,

    Since there is no follow-up response from the customer, you can close the post first.

    Thanks for your help.

    Best regards,

    Vivian