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.

Loop compensation - how to compensate without loop analyzer

Status
Not open for further replies.

Magnethicc

Member level 3
Member level 3
Joined
Feb 7, 2022
Messages
59
Helped
1
Reputation
2
Reaction score
6
Trophy points
8
Activity points
793
Hi Everyone,
I am asking the following:
Is there any article or a practical approach to compensate a converter (AC-DC / DC-DC / BiDirectional, etc.. etc..) without knowing the transfer function?
Every piece of information I found they assume I have the transfer function of the converter but in practice if we don't have a loop analyzer then the compensation (at my workplace atleast) is made by using past compensators, seeing if this oscillates/overshoots too much --> change component values based on the experience of the veteran engineers.
And even after the converter is stable we put it in 60C degrees ambient and again if things oscillate change component based on the experience of the veteran engineers.

I was wondering if anyone have a good explanation on how to compensate by feel and not by math:
1. which compensator to choose?
2. what every resistor and capacitor in the compensation network is going to help with?
3. what are the trade-offs for a pole/zero higher/lower in frequency?
4. what is most likely the culprit in different types of the converter responses?
etc.. etc...

Thank you.
 

In this couse, in the folder "switchingpowermagazine" there are some pdfs which answer your qus

SMPS course_little folders
https://drive.google.com/file/d/1WvElhDNfHI0EBei4rouSNEdWWTO1oCtJ/view?usp=sharing

Hit it with a full to no load transient and vice versa.
See if it rings and settles.
The ringing frequency is the loop bandwidth.

Also, if RHPZ, then you must cross over at 3 times less frequency
Also, crossover should be less than fsw/10
Also, be wary of voltage mode...the C output filter can ring.......

Practice adjusting comp components on sims to get some idea.
 

it does depend some-what on the exact converter and exact control loop you are starting from - it is pretty easy to do a back of the envelope calculation on the power stage to see where the gain rolls off and phase changes - this can then influence the design of the controller to compensate for the power stage effects - peak current mode control can make things a lot easier here. Some capacitors ( electrolytic ) can have lower ESR at higher temps which may be why you see the oscillation at higher temps ( less damping ).
 

Also, with voltage mode PSU's i believe you should keep your feedback loop bandwidth away from the output LC resonant frequency....either well above or below it......sometimes you see people slowing the feedback loop down thinking they are getting more stable...but they then approach the LC output filter frequency and go unstable.
 

One rough rule of thumb, looking at response in time domain (on a scope) -

1649758616872.png


This is rough in that your compensator may be more complex than simple 2'ond order
network. But it is reflective of overall stability.

Of course you could always spice sim the design and look at response/phase margin.
Even over T and V.....

Regards, Dana.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top