why is SMPS control such a big secret?

[grizedale]

Hi,

I am trying to find out at what frequencies the poles and zeros of a control-to-output transfer function of a current mode, DCM flyback occur.

Also, i am seeking the control-to-output transfer function of this converter.


However, googling the entire internet with phrases such as............

"pole and zero location for discontinuous mode flyback in current mode"

..reveals NOTHING whatsoever.

The DCM current mode flyback is probably the most common of all offline SMPS's.

There are literally billions of them in the world.

....and yet the entire www contains nothing about their most important aspect......their control

why is this?
why is it such a secret?

 

[FvM]

You'll find some calculations, e.g. a derivation of control-to-output function for DCM flyback in current mode in the 1984 Unitrode Design Seminar, particularly in Appendix C.

1980 - 2003 Unitrode Design Seminars
**broken link removed**
http://www.ti.com/lit/ml/slup071/slup071.pdf

 

[grizedale]

Thanks,

Do you know of a book that gives the error amplifier small signal transfer functions when TL431/Opto is used as the error amplifier.?

Basso book has it, but i wish to double check the given formulae.

Does the Erickson book do it for TL431/Opto?

 

[grizedale]

Also, regarding the above "slup071" reference from Unitrode.

...supposing i were using a unitrode UC3845..............

https://www.ti.com/lit/ds/symlink/uc3845.pdf

...in a current mode flyback in DCM, then what would the parameter "Vc" be? (page 5 of slup071)

Also, the parameter "Lf" is not explained.

Also, the slup071 is ok but does not give the transfer function for the TL431/Opto based error amplifier

 

[lm2577]

...in a current mode flyback in DCM, then what would the parameter "Vc" be? (page 5 of slup071)

Also, the parameter "Lf" is not explained.

Lf - this is the inductance of the secondary winding. ie Lf = Lp * (Ns / Np) ^ 2
For UC3845: Vc - this is the voltage at the input of the comparator, ie, output voltage of error amplifier, divided by 3.

Also, the slup071 is ok but does not give the transfer function for the TL431/Opto based error amplifier

**broken link removed**

 

[grizedale]

Thanks,

But i think if it was the secondary winding , they would have called it "Ls"?

....Also, the output of the error amplifier is variable......it can be as high as the reference voltage (5V), or it can be 0V.
-so i dont think that is what "Vc" means.

Also, that article on Closing the Loop does not give the small signal transfer function of the Error Amplifier comprising TL431/Opto.

 

[FvM]

In slup071, Lf simply means inductance multiply switching frequency. You possibly need to refer to other seminar documents to understand the relations.

Vc is simply the input voltage to the pulse width modulator. Because it's scaled with a parameter like maxVc or Vs (which represents the sawtooth generator magnitude), it's always related to a 0-100% duty cycle range and can be easily recalculated for a different place of the control loop.

TL431 and optocoupler transfer function is a different topic, not particularly related to DCM flyback. As far as I remember, it has been discussed in different Unitrode seminar documents. There are also previous threads at edaboard.

I would primarly see the said documents as a support when deriving the transfer characteristics yourself.

 

[grizedale]

For a UC3845, i am not sure where is the "input voltage to the pulse width modulator"

...is it the output of the error amplifier...
..or is it the inverting input of the PWM comparator.

...the peak to peak amplitudes of these two points is different.

https://www.ti.com/lit/ds/symlink/uc3845.pdf

 

[FvM]

I just realized, that the mentioned sawtooth generator amplitude doesn't apply to current mode. For current mode, slup071 uses a scaling factor K = maxIp/maxVc. The term doesn't clarify at which point Vc is measured. If you refer to the comparator input, K = 1/Rshunt, for the amplifier output, the divider ratio must be multiplied into it.

 

[grizedale]

So , when using the UC3845 for DCM Current mode flyback,...

if my peak primary current is 2Amps,
-and the current sense resistor is 0R5........

Then K = 2

and Vc = Max Ip/K = 2/2 = 1

Is this correct?

 

[FvM]

Then K = 2

and Vc = Max Ip/K = 2/2 = 1

Is this correct?

maxVc = 1V

 

[grizedale]

I appreciate that Max Vc is 1V .....but i cannot believe that the DCM , Current mode flyback transfer function is not changed by changing the current sense resistor.

...and the equation on pgae 5 of slup071 does not show the sense resistor anywhere.

-this just does not make sense.

-The size of the current sense resistor must affect the transfer function.

 

[FvM]

The resistor affects the gain, as it did in your example, where K is 1/Rs.

 

[fendertunes]

I didn't read any of the posts above.

To answer your question, SMPS control is not a secret. The best way to go about it is pick a controller (other than Power Integrations junk) and design a type 2 feedback circuit. There are a lot of resources for it. But the easiest way to go about it is use the spreadsheet calculator tool most vendors provide to design your feedback.

 

[FvM]

I didn't read any of the posts above.

In other words, you are claiming to post off-topic. Really bad.

 

[fendertunes]

In other words, you are claiming to post off-topic. Really bad.

It's not off-topic. I was responding to the poster's original question.

 

[FvM]

Your post takes up the title of the original post but doesn't refer to the specific question. This happens in many threads and isn't a big thing. I also believe, that many contributors do basically the same, not reading the other posts. But they don't tell.

 

[dick_freebird]

I don't think it is so much a secret, as that there is not a
single answer - details of implementation drive the values.

And of course Googling a quoted phrase can yield you no
result if nobody published anything exactly the same. On
the other hand searching for

dcm flyback pole zero

yields a whole lot of material (164000 hits). First link is a
Fairchild app note, second is from OnSemi, etc.

 

[grizedale]

im sorry but details of the exact implementation are not required because many current mode chips are set up similarly and the modulator gain would be the same....and at the end of the day, a dcm flyback is a dcm flyback and they have the same poles and zeros as each other.

Noe of the app notes are really any use for DCM flyback.....ive searched thousands and all are o flittle use.