input filter and smps instability ?

[eem2am]

Hello,

I have just simulated a Boost LED Driver with switching frequency = 390KHz.

The LED current and input voltage both oscillate at around 70KHz.

Is this oscillation due to the interaction between the input filter and the
negative impedance of the SMPS.?

Here is the schematic

XXXXXXXXXXXXXXXX


Here is the oscillating Input Voltage (pink) and LED Current (green)

XXXXXXXXXXXXXX

Here is a close up of the oscillating LED Current

XXXXXXXXXXXXXX

Here is a close up of the oscillating Input voltage

Here is the gate drive waveform

XXXXXXXXXXXXXX

Here is the LTSpice simulation file (.asc)

https://www.megafileupload.com/en/file/174388/No-Overshoot-Cancellation-asc.html



,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
by the way, you may read on input filter and smps instability in the article
"the evolution of power electronics"
on switchingpowermagazine.com website ( you need to register)
see under articles>filters

 

[svhb]

It seems that the control loop has a sustained oscillation at this frequency.
What happens if you add the parasitics of the inductors and capacitors? A totally undamped filter will oscillate easily. If you do a test with a somewhat overestimated copper resistance for the filter solenoids, you can see if the filters are the reason.

What happens if you reduce the bandwith of the control loop? I don't think it is necessary for a LED-driver that the bandwith is so high that it can generate oscillations up to 70Khz.


Just out of curiosity:
Why is the injection of the output voltage needed in the control loop (R18?)

Stefaan

 

[eem2am]

thankyou stefaan,

R18 and Provide a zero......and make compensation easier...the supply is easily made stable.........it is needed becasue the opamp puts an unfortunate delay in the feedabck loop, and the overshoot on start up is too much without R18 and C4