Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

[SOLVED] Control loop BW limited by the flux balancing of CPM controller ?

Status
Not open for further replies.

CataM

Advanced Member level 4
Joined
Dec 23, 2015
Messages
1,275
Helped
314
Reputation
628
Reaction score
312
Trophy points
83
Location
Madrid, Spain
Activity points
8,409
Hello everyone,

Isolated converters (except half bridge) controlled via current mode control, there is no problem of flux imbalance in the transformer.
However, in order for the control to achieve this, i.e. have symmetrical magnetizing current shape, the "control reference" must be the same during both the positive and negative intervals of the current waveform, or, in other words, during the whole switching period of the waveform of the transformer.

By contrast, the current mode control, has a sampling frequency of 2*switching freq of the waveform at the transformer (fCMC=2*fXFMR).
In other words, the control loop must keep its signal constant during 2 periods of its sampling frequency (fCMC).

That being said, looks like it is useless to design the BW of the control loop regarding the sampling frequency of the CMC, but instead, it might be focused to half the sampling frequency.

I am not able to find any documentation regarding this in order to check the mentioned logical assumption, which is desirable since lower crossover frequency brings disadvantages.

Any comment on this is appreciated !
 

Given that current mode overcomes the inherent LC output filter pole - it will generally give fast(er) response,

your comment of " the control loop must keep its signal constant during 2 periods of its sampling frequency (fCMC) " is not strictly correct.

The control variable (really the peak current + slope comp) can vary a little over a full period - just not in a wild fashion, it should be at a rate low enough not to cause current loop issues, a continuous rising signal (up to max) is OK, the current loop has to "catch up" a little but this is easily done and the currents in the Tx are bounded. Similarly this signal can fall very rapidly without upset, ( e.g. to zero in less than one half switching cycle ).

Many newbies set the volt loop too fast and this can interfere with the current loop, as you suggest, however the "type 3" compensator ( I hate that label ) with feed-forward or "phase-lead" RC in the volt sensing line can generally make a system fast enough for most purposes.

Some split cap damping in the feedback network on the controlling op-amp is a good idea too.

Try and make the current loop as fast as possible with the lowest amount of filter cap on the current signal, and a high quality op-amp if you are amplifying a shunt.

Of course reducing the output L ( and to a lesser extent the C ) will help in allowing the fastest possible volt loop.

Once you have built a system you can measure the response of the loops and see where you can speed things up.

It is possible to analyse with pen & paper ( & calculator ) or using LTspice on a fast computer - so you know where you will be before building, but this is usually best case and the real world introduces delays which affect voltage speed of response.

Good luck...

- - - Updated - - -

p.s. generally the theoretical fastest loop response for any circuit is half the switching frequency (<= unity gain )

practical circuits can achieve 8-10% of the switching freq relatively easily ( but with some effort), and ~ 20% or greater with very experienced design effort.

Anything higher than this requires a minimum of noise entering the control and very fast op-amps with little phase lag at the control frequencies concerned - this is where an opto-coupler is generally too slow to allow such fast control (phase lag).
 
  • Like
Reactions: CataM

    CataM

    Points: 2
    Helpful Answer Positive Rating
However, in order for the control to achieve this, i.e. have symmetrical magnetizing current shape, the "control reference" must be the same during both the positive and negative intervals of the current waveform, or, in other words, during the whole switching period of the waveform of the transformer.
Basically not. The average current must be balanced. During a set point ramp, you rather want continuous variation of each half wave.
 
  • Like
Reactions: CataM

    CataM

    Points: 2
    Helpful Answer Positive Rating
Some split cap damping in the feedback network on the controlling op-amp is a good idea too.
I can not picture how "split cap damping" is implemented, can you explain it a little bit ?
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top