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/UC3845: TL-431 Voltage divider resistor value

Part Number: UC3845
Other Parts Discussed in Thread: LMV431, TL431

Tool/software: WEBENCH® Design Tools

Hi;

I am working on Flyback offline SMPS using UC3845.

I have used TI WEB-BENCH online design tool, this tool automatic generated schematic as per my parameter requirement.

I confused when I saw TL-431 side, I failed to understand how TL-431 ref pin resistor value came?

Vo = 5V

As per TL-431 equation =

Vo = Vre (1 + (Rupper/Rlower) , TL-431 act as 10V Zener , but my output is  5V

How this tool calculated Rupper and Rlower resistor divider network ?

I have attached screen shot also, please check.  

I have also uploaded pdf file of design tool.WBDesign4.pdf

  • Hi Jaydeep,

    Thanks for contacting us.We have forwarded the query to our experts.

    We shall keep you updated on the same.

    Thanks and Regards,

    Umayal

  • Hi Jaydeep,

    For this schematic, LMV431 is being used instead of TL431. You can check it by clicking on the shunt symbol in WEBENCH schematic

    LMV431 as a shunt regulator is behaving like an error amplifier of the feedback voltage, The divided output voltage is compared to shunt regulator reference voltage( ~1.24 V for LMV431 in this example) by its internal error amplifier and the output is given to the opto. 


    Hence the calculation is :

    Rtop = (Vo*Rbot/Vref) - Rbot

    Vref= 1.24 V, Vo=5 V,

    Assuming 1000 uA flowing into sense resistors, Rbot ~ (1.24/1000e-6) ~ 1.24 k.

    Rtop can hence be calculated from: Rtop = (Vo*Rbot/Vref) - Rbot

    Thanks
    sahil

  • Hi; Sahil

    Thank you so much for your quick reply.

    Yes Its my mistake, I assumed that voltage regulator was TL-431.

    I have seen TI webench generated report, I have attached one screen shot regarding my question.

    As per report, Primary peak current is 414 mA, and as per current sens resistor formula , Value of 

    Rsns  = 1 volt / Ipk.

    In my case requirement is 2.41 Ohm . But in the schematic, it is 1.78 Ohm. How it possible ? 

    Rcs and Ccs may be for suppress leading edge spikes, but how I find this components value ?

    And what is the purpose of transistor used near the IC ? , Is it for slope compensation ? 

    Highlighted portion is like black box for me, please help.

  • Hi Jaydeep,

    For primary current sense resistor, we need to take some margin for over-current for calculating Rsns(20-30 percent).

    Rsns = Visense/(1.25*Ipk) = 0.9/(1.25*0.414) ~ 1.74, take Visense to be the minimum value of 0.9 volt( worst case) from datasheet.

    Yes Rcs and Ccs is for leading edge spike suppression, this can be selected by keeping time constant of Rcs and Ccs to a low value, in this case it is kept around 400 n,
    Fixing Rcs to 1k, hence Ccs ~ 400 n/1000 ~ 400 p

    Qsc and Rsc is for slope compensation, this can be referred from datasheet as well,

    let us know if anything else, 

    Thanks
    sahil

  • Dear Sahil;

    Thanks for your quick reply.

    Yes, if it is for slope compensation then where is compensation capacitor ?, Compensation cap is require for coupling AC only to Isens pin, Otherwise it will add dc offset into current sens signal . I have attached TI datasheet application schematic. 

    Also in my case maximum duty cycle is 0.12 . Generally slope compensation require when duty is higher than 0.5, then why web bench schematic consist slope compensation network even in low duty cycle ? 

    Please help. Can you please provide mathematical steps of Rramp and Cramp for this attached webdesign ?

    I also cant understand what is the purpose of D21, C21 and R21 at  output side ? Looking for your support.


  • Hi; 

    Are you there ???

  • Hi Jaydeep, 

    R21, C21, D21 is external soft start circuit, I will get back to you on the compensation question by tomorrow, 

    Thanks

    sahil

  • Hi Jaydeep, 

    You can remove the slope compensation circuit(Qsc/Rsc) as D<0.5 in this case, 

    However , if you still want to find Rramp, it can be estimated from these equations:

    Sn(inductor rising slope) = Vinpkmin*rsns/Lpri;

    Sosc(oscillator slope) = 2.3*fsw/Dmax. Dmax is 0.47 from datasheet

    M = 0.75; required slope compensation

    Rramp ~ (rcs*Sosc)/(M*Sn)

    sahil

  • Hi Sahil;

    Thank you so much for your help..

    Can you please provide more  information about your previous answer?

    How R21; C21 ; and D21 act as soft start circuit element?

    why this require?  and how I can select value of each?

    Because I have never seen this type of circuit in secondary side. Generally external R ; C and Transistor network use at primary side near IC OSCILLATOR pin for soft start event...

    Please provide guidance regarding this secondary side soft start network. Please also provide steps for selection of these components; So it will become easy for us to make circuit  in any other Output voltage requirement or any other SMPS.

  • Hi, Jaydeep,

      the soft start circuit is generally needed for the power supply to start up with limited current stress.

      For example, if the flyback starts without the soft start, the primary side current would reach its maximum value at first switching cycle, due to the peak current mode control. After the primary side switch turns off, the secondary side has no voltage to discharge the inductor. Therefore, the primary side current stays almost the same when the next switching cycle starts. In the second switching cycle, due to the leading edge blanking, the inductor current would rise above its desired maximum value. This behavior will continue until you generate some output voltage to help discharge the inductor current. To avoid this, the primary side peak current is desired to keep lower during the start up.

      For the soft start circuit, the principle is easy. The TL431 circuit's output controls the primary side peak current, which is corresponding to the power delivered to the load. When the TL431 output voltage is high, there is little current in the opto, which demands the large peak current and large power delivery. When the TL431 output voltage is low, there will be large current in the opto and the peak current and power delivery will be low. The added circuit uses the capacitor to keep the TL431 output low during the start up to reduce the primary side peak current. Once the voltage on the capacitor is created, the diode becomes reverse bias. The added circuit would have no effect to the circuit operation.

    Let me know if this explains the principle of the operation.

    Thanks.
    Bing

  • Dear Bing;

    Thank you so much for your reply ,

    Sorry but till I cant understand the working procedure of circuit.How soft start circuit make TL-431 output to low during startup? What is the purpose to use schottky diode at there ?

    Please also explain how can I choose each components value in  ? Which mathematical steps require ?

  • Hi, Jaydeep,

      The R-C-D circuit for the soft start is trying to use the C to keep the TL431 output voltage low during the startup process.

      The C is discharged to zero before startup, therefore, once the whole power supply starts, the C voltage rises from 0.

      The schottky diode is trying to keep the voltage on TL431 output as low as possible because it has less voltage drop comparing with a regular diode.

      The diode selection is easy, you can pick any low voltage diode that meet your requirement. There is no high voltage or high current in this case, a 10V 100mA diode should do the job.

      For the R and C, you want to choose a right time constant to achieve the good soft start performance. Generally, this is a try and error process. But you can start to have the initial value and based on the performance to adjust the R and C values. generally, you want the C to be at least 10 times larger than your compensator capacitors. and the R-C time constant can start with 10ms and based on the result, you can adjust the parameters. Larger C and R makes the soft start time longer. You might have trouble to start up within the time you required. The smaller R and C makes soft start shorter, but you might see too much stress on the system. You can play with the values and get settled down easily.

      Let me know if this is clear.

      Thanks.
    Bing

  • Hi Bing

    Thank you so much for detail reply.

    • 1)  Question regarding reflected voltage

    I have already uploaded TI-WEB BENCH Report , From  the report I have found Dmax and NPS Turn ratio.

    I have already known 5 VDC output  . So from Flyback equation I can get reflected voltage information.

    In this case Turn ratio = 4.2 & OP = 5 VDC

    I want to know Is it possible ?  This is too low  level  reflected voltage , Is it ? In general people considered 50 - 70 V reflected voltage  for 5 VDC output &120 or 150 V reflected voltage for 12 VDC output

    As per report , Operating mode is CCM .

    I know low reflected voltage lead toward low turn ratio and we can eliminate problem of fitting inductor winding within effective core area of coil former. But here It is too low 23 V

    2)  Rsens

    Another thing is TI Enginner Sahil replied me regarding Current sens resistor question . (You can see in my previous conversation content with Mr. Sahil

     Is it right way to find Resistor  value ?

    Because in Denominator You considered 1.25 times higher than actual peak current . Is it right way ?

    Thanks

  • Hi, Jaydeep,

     I am glad the answer helps.

    regarding your new questions.

    1. the reflected voltage. There is no limitation on the reflected voltage, you can make it large or small. However, there are some tradeoffs to consider, so that the reflected voltage needs to be chosen carefully.

    a. When the reflected voltage is large, you are putting more voltage stress to the primary side switch. Therefore, higher voltage rating device has to be used. This could increase the cost and hurt the performance since normally the high voltage device doesn't behave as good as low voltage device.

    b. When the reflect voltage is low, there is less stress on the primary side, but more stress on the secondary side diode. Therefore, you have to choose higher voltage rating diode, which in turn hurts the efficiency and cost.

    c. The reflected voltage changes the transformer turns ratio, which determines the duty cycle of the converter. At different duty cycle, the primary side and secondary side RMS current and peak current would be different, which would cause the efficiency different.

    The selection of the reflected voltage is tradeoff between all three considerations. There is no right or wrong answer.

    2. The current sense resistor. I agree with Sahil. You need to consider the component tolerance and the efficiency losses. You always want to have some margin in the design so that your converter is able to deliver more power than its maximum power. The 25% extra current is a good practice.

    Let me know if you have any further questions.

    Thanks.
    Bing

  • Dear Bing;

    Thank you so much , your suggestion is very help full for me.

    Can you please provide suggestion about selection procedure of wire Gauge (AWG) ?

    1) Is Wire Gauge size depends on Peak value or RMS value ?  

    2) How can I select AWG size for primary / secondary or Auxilary ?

    3) How I can check number of turns with this wire gauge fit or not within winding area ?  Using core Area Product ? or Using fill factor ? Can you please provide detail explanation with examples

    3) Dear I have been trying to make SMPS , and I have following specification and calculation . Can you please check ? Please also provide suggestion regarding my mistake.

    Dmax = 0.1785

    Vref = 50

    Vo = 5V ,1 Amp

    Efficency = 0.8 

    Vin (AC) = 180 to 350

    Cin =6.2 uF

    Vbulk(min) = 230 VDC  (= Vbulk max  - Vripple)

    Fsw = 65.151 KHZ

    Lpri = 0.00318 H

    Ripple Factor Krf = 0.65

    CORE = RM 8

    CORE MATERIAL = N87

    MFG : TDK/EPCOS

    TYPE =GAPPED CORE

    Inductance Factor = 400 nH

    Effective Core Area =64 mm*mm

    Bs = 0.3 T

    air gap s = 0.18 mm

    AN (Winding cross section Area ) = 28.4 mm*mm

    Npri = 90

    NPS =9 

    NS =10

    NAUX = 23

    Ipk (pri) = 251.28 mA

    Ipri(RMS)=68.69 mA

    Isec(pk) =2.2 Amp

    Isec(RMS) =1.292 Amp

    How can I select wire size for above specifications ? Also need to fit it in transformer .

    Thanks you in Advance 

  • Hi, Jaydeep,

      I am glad my answer helps.

      Here are some information for your new questions.

    1.Is Wire Gauge size depends on Peak value or RMS value ?  

    The wire gauge is determined by mainly the RMS value. The idea behind it is to control the heat created by the wire or the copper loss. It is more complicated than just RMS value because the frequency also plays a role here. The skin effect would push the current flowing on the surface of the wire, increasing the wire size wouldn't help too much when the frequency is high. Normally people would move to the smaller wire  using multiple strands or using Litz wire.

    2. For aux winding, normally there is not much current flowing through it, we normally chose the wire by continence. If a wire size can help to fill the window, or the same size as primary or secondary side wire. If none of these works out, just use the smallest wire available because it can help to take less spacing.

    3. For the design, I believe Webench also have the transformer design tool that helps you to design the transformer. You can click on the transformer and select the customized one. It should give you the transformer structure with the wire sizes.

      Let me know if you get any further questions.

      Thanks.

    Bing