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.

dc/dc stability problem

Status
Not open for further replies.

franticEB

Full Member level 3
Joined
May 10, 2010
Messages
152
Helped
1
Reputation
2
Reaction score
2
Trophy points
1,298
Activity points
2,540
Hi, i've realized a prototype board with a DCM flyback converter with these features:
Vin=50V
Vout=250V
Pout=50W
Fsw=100Khz

I've connected thi prototype with laboratory power supplies and the converter seemed stable to vary the input voltage and the load.

Now i had to import that converter on the ultimate board and i connected to others board's power supply in this way:

f1.png

The converter is not stable yet, and after numerous attempts the only way to make the smps stable is to do this connection

f2.png

What may have gone wrong in this final board? Is there any common mode noise that make my pwm driver unstable?
Could you help me?
 

Well, you're ruining the voltage domain isolation that way.

I think the problem is that the "GND250" and the feedback
connected to it, does not return to the evident "250V"
low side and to your feedback and optoisolator have (if
wired as shown) an incomplete current loop and no local
ground to complete it.

 

No it is my mistake in drawing the schematic... GND250 is actually connected to feedback circuitry.

I noticed also that when the connection beetwen GND50 and GND250 is not present the optocoupler output to pwm driver COMP pin oscillates. Instead when there is that grounds connection there is no oscillations.
 

OK, but is GND250 as-shown, also connected to the secondary
and filter cap as I sketched? I see it at the feedback block.

If it's as I sketched it, next questions are (1) whether the output
is simply noisy enough that the FB>opto -should- have some
switching behavior; (2) at the opto, is "oscillation" same freq
as switching fundamental, or 2X, or small signal and lower
frequency than PWM freq (3) is opto "oscillation" full scale and
driven by Feedback block; (4) What is the nature of the COMP
pin, does it provide enough pullup current that the opto can be
a negligible phase-lag contributor or is the opto in fact driven
into saturation and causing a relaxation oscillation (not small
signal)?
 

The oscillation seems a triangular wave but i have to measure its frequency...

The comp pin provides 5k pull-up and the feedback is the classical one with TL431 network.

I'll be more precise later...
 

Have you tried a .01 MF capacitor instead of your short? Switching current at high speed will give all sorts of ringing, for instance every time the 250V transistor conducts, the 50V will drop, is this affecting the 12 V circuit. Why is this floating wrt to 50V?
Frank
 

you have far too much CM signal across your opto, put a scope across it (kathode to emitter) and see for your-self... such a system with floating gnds generated from flybacks will give a novice problems, unless you use a quality opto with a shiels and/or a bigcap across the C-E on the opto...
 

Agree with Easy P.
There is far too much common mode noise, more than the opto can isolate.

Several solutions suggest themselves, an opto with an internal electrostatic screen should offer much better high frequency common mode isolation.
I cannot at this stage suggest a specific replacement.

Placing a pair of electrostatic screens between primary and secondary in the flyback transformer should help a lot. Ground one screen to the input side, and the other screen to the output side. That would be my first choice.

A quick and very dirty band aid solution might be to replace the "red link" with a suitable high voltage capacitor.
It would tie both sides together at high frequency, but still maintain dc isolation.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top