Smps connecting a load before the output inductor

[dr pepper]

I've thrown together a Psu with a Sg3525 its 15v 10a and operates in half bridge.
The thing works fine.
While tuning the compensation network by connecting & disconnecting a load (halogen lamp) I accidently connected the load to the rectifier diode side of the output choke, the lamp lit but over a period of a few seconds the mains lamp I have wired in series with my isolation transformer powering the supply went from a slight glow to fully on and the supply just sat there trying to start.
I wasnt expecting that to happen, is this the transformer saturating, or a result of my pants board layout.

 

[mtwieg]

The RMS voltage at the switching node is higher than at the filtered output, so connecting the load there would have increased the load power (how much depends on what duty cycle the smps was operating at). That would result in a larger-than-expected load, which must cause input current to increase, and thus the series lamp to be brighter. Hard to say without more details, but I doubt transformer saturation was playing any part.

 

[cupoftea]

I'd get a representative LTspice sim going and see whats going on...adjust the leakage inductance in the transformer.

Though of course, the half bridge does not need an output inductor at all anyway.....there are loads of half bridges out there which just use the leakage inductor in the transformer....(with a few mitigations to permit this)

https://www.edaboard.com/threads/open-loop-half-bridge-smps-really-needs-sic-fets.400843/

....there are lots of other threads on it here...but i cant yet find them

Ill send an LTspice sim of a half bridge soon and if you wish you can play with it till you see the situation youve got there.

...tell you what...the half bridge LTspice sims are somewhere in this SMPS course, which you can dig out if you want.......give us a shout if you want me to get it.....

SMPS course_Big folders
https://drive.google.com/file/d/17ij_pQLkAzzYeL_TRNci86sXHMirKHmD/view?usp=sharing

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

 

[dr pepper]

I didnt even know there was a smps course.
Thinking about it the choke removes stuff, noise, so before it there is going to be more.
What threw me was the slow invrease in bulb brightness.
I might even have the model for the sg3525 on lt spice, I certainly do have the ux3800 series.

 

[Easy peasy]

sounds like the Tx saturated ( flux staircasing ) - and I expect at least one of the mosfets has died - and possibly one or more of the output diodes

pay to check everything ....

 

[dr pepper]

All seems Ok, fets good & diodes good, transformer insulation still good.
Wearing a face shield I removed the bulb & powered direct from my isolation trans, now the halogen lamp load is the same either side of the choke visually and 1/2 a volt more on the diode side with a meter as you'd expect with a good trms meter, so this is a series bulb thing, bulbs have a non linear resistance with voltage so thats probably what I'm seeing with both load and series limiter bulbs.

 

[cupoftea]

What type of rect are you using...is it the attached kind of half bridge (LTspice attached)
LTspice is free download.

I tried putting the load the diode side and it works, but obviously pulsey current in the load...

 

[dr pepper]

Not quite the same, mine doesnt have the series choke on the primary, and the secondary is centre tapped with a double diode. one off each end of the winding.
I have Ltspice, but not the sg3525, I can sim the power stage but not the whole thing.
And that makes sense mine only behaves weird when theres a bulb in series with the mains, I wonder if theres a model for a filament bulb in ltspice.

 

[d123]

I wonder if theres a model for a filament bulb in ltspice.

Hi,

light bulbs

 

[dr pepper]

Indeed there is, thanks.
I'll mess arond with some of those models.

 

[cupoftea]

Sorry to sound like prying...and i am sure you are up on these issues.....but it may be worth checking for you...
1...Do you have pri side current sense?
2.....Are you in current mode, and if so, how are you mitigating split rail cap unbalance?....especially during overload.
3...Have you got snubbers across the output diodes.
4...If you're in voltage mode, how would you avoid split rail cap unbalance in overload if it then goes into primary peak current limiting.

 

[dr pepper]

Are you thinking I have + and - rails and one can be overloaded creating an imbalance in the core?
Maybe I misled my supply has only one rail, froma center tapped secondary, the ends of the windngs go to a To247 package dual diode.
One end of the primary goes through a series capacitor.

Mode is voltage.
No primary current sense, there is crude current limit on the o/p but no short protection, havent got that far.
There is a snubber on the primary, but not on the rectifiers, there is some parasitic overshoot but its within the rectifiers rating.