SMPS Control Loop Design ..... need help

[abcsantosh]

    Hi all, I want to design the control loops for SMPS topologies. (Flyback, Forward, Buck, Boost, Half Bridge, Full Bridge, Sepic, Pfc Boost) I know moderate technics of compensating the control loops. I know that once you learn the basics to compensate for one topology, it will applicable to other topologies too, with some changes.
    Especially, suggest me the feedback technics using opto isolation with TL431 and optocoupler. Also I want to design variable control on output voltage and current. Show me how can I achieve Variable current control or Voltage control or both simultaneously.
    Suggest me the starting point where can I start.

Have a Good Day       

[tAhm1D]

Hi abcsantosh,
For feedback, the TL431 acts as a variable "zener" or "shunt regulator" which turns on the optoisolator at a specific voltage. Say, you have 16v unregulated output, but you want to regulate it for 12v, you have to set the TL431 as a 13.4v zener (12v for output + 1.4v for optocoupler LED). So, if your output rises above 12v, the opto turns on, and controls the compensation/feedback pin of the PWM controller to change duty cycle as required.
For constant current / variable current, you can use a shunt resistor. The voltage across this will be fed to an operational amplifier. Calculate voltage across it from V=IR. Say you placed a 0.1R resistor and you want 10A constant current. Then the voltage across this will be 1v. You can amplify this with the opamp and compare it against a reference. Say, you have a 5v reference and the 1v is amplified to 5v (non-inverting amplifier). If your voltage exceeds 5v (ie, current exceeds 10A), you reduce the duty cycle of the PWM controller using the opamp and viceversa.
Hope this will help you.

[abcsantosh]

    Hi I got what you say. Now tell me where to start this whole feedback theory. Any book in particular will be very useful. Do I have to study op amp in details, or basics will be enough ? What to you suggest ?
    If somebody show some step by step feedback loop stabilization procedure, it will be very useful.

Have a Good Day

[tAhm1D]

Hi,
op amp is a must for  understanding and designing whole of Electronics. Without understanding it, you cannot proceed much further in Electronics.
For theoretical knowledge regarding Feedback,you can study the book by Marty Brown. But for applicational knowledge, you have to study the internal structure of the pwm ic and its datasheets. Particularly, I can not remember any book, reading of which you can earn mastery in that but with constant study and practical application, you will understand it properly.

[titi]

Hello,

Here is some documents about the using of TL431 on feedback SMPS.

AN-32 TOPSwitch-GX Flyback Design Methodology: http://www.powerint.com/sites/default/files/product-docs/an32.pdf
Solution Optimizer for SMPS Applications: http://www.sts.tu-harburg.de/pw-and-m-theses/2006/sint06.pdf

Best regards.

[qolpa]

hi,

AN1327/D  -  Very Wide Input Voltage Range, Off-Line Flyback Switching Power Supply

isolated flyback application with 3845. feedback with tl431 and opto.  have a look at it.

regards

[masa]

good info using the TL431

[arash_tah]

Hi abcsantosh
For designing control loop in SMPS you need to be familiar with control teorem. Then what u need is to obtain a perteration model of SMPS and then design controller parameter for example u can use BODE and phase margin criteria to adjust the parameter of controller.

[abcsantosh]

Hi masa6614

pls, share the original book from which you gave TL431 document.

Switch-Mode Power Supplies Spice Simulations and Practical Designs (With All Chapters).

It is very useful.

Have a nice day

[itp]

Hi,
Is this is the book

http://rapidshare.com/files/126720242/Switch-Mode_Power_Supplies_Spice_Simulations_and_Practical_Designs.rar

Size: 15Mb

Chapters:
1. INTRODUCTION TO POWER CONVERSION
2. SMALL-SIGNAL MODELING
3. FEEDBACK AND CONTROL LOOPS
4. BASIC BLOCKS AND GENERIC SWITCHED MODELS
5. SIMULATIONS AND PRACTICAL DESIGNS OF NONISOLATED CONVERTERS
6. SIMULATIONS AND PRACTICAL DESIGNS OF OFF-LINE CONVERTERS—THE FRONT END
7. SIMULATIONS AND PRACTICAL DESIGNS OF FLYBACK CONVERTERS
8. SIMULATIONS AND PRACTICAL DESIGNS OF FORWARD CONVERTERS
Regards
Itp

[carbontracks]

Beyond basic voltage mode control, the math gets pretty tedious...  On the bright side, more sophisticated techniques usually tend to be more stable inherently, so a lot of the math becomes irrelevant.  I'd recommend going with current mode control, which is very easy to compensate (but may be tricky to implement with isolated outputs).

[medik]

Great info on smps. Please, can i implement this using PIC? I'm thinking of replacing the PWM I.C and implementing the function in software using PIC But the major problem i foresee is the initial power supply to the PIC. Anybody done this before?

[carbontracks]

Great info on smps. Please, can i implement this using PIC? I'm thinking of replacing the PWM I.C and implementing the function in software using PIC But the major problem i foresee is the initial power supply to the PIC. Anybody done this before?

Powerup may be an issue depending on the type of converter.  For buck, the output can't be powered until the switching starts.  With boost, some power will flow even if the switch is off (this may not be a problem).

When using a microcontroller, the bigger issue is usually latency in your control loop.  If you're trying to run at high frequency, then just a few microseconds of delay can be completely intolerable.

[solutions]

Great info on smps. Please, can i implement this using PIC? I'm thinking of replacing the PWM I.C and implementing the function in software using PIC But the major problem I foresee is the initial power supply to the PIC. Anybody done this before?

If you do the math (yes there's math unless you're a SPICE-Monkey), you'll need some HUGE passives to slow things down enough to control them with a micro.  Been there, done that analysis, a few weeks ago actually.

I think the power supply to the PIC is the LEAST of your worries if you are serious about dumping the PWM chip and writing assembler, let alone the fantasy of writing the PWM in C (or BASIC), taking compiled output from Labview, or worse, using a visual-code compiler to generate code for you.

Sometimes you have to swallow the pill and get dirty with appropriate hardware.

[Gallymimu]

if anyone ever used SMPS TEA1716 from NXP?

you should be starting a new thread for this question rather than hijacking a very old unrelated thread.