TOPswitch --- AC-line ripple on the DC output
I used PI-Expert to design a 29V, 2.5A universal input (85-265V, 50-60Hz) power supply (design attached). The prototypes exhibited poor voltage regulation and 120Hz (2x line freq) ripple in the output under load at 115V. (Good regulation/ripple @230V.)
After reviewing some of the App Notes, I revised the feedback path to be comparable to AN-32, figure 6. This resulted in significant improvement in regulation (acceptable, but not perfect.), but the AC ripple is still unacceptable under heavy load. (About 3Vp-p at full load.) The ripple is worst at high loads and low AC-in, which is no surprize, since this is when the ripple on the bulk rail is greatest.
What components in the feedback path affect the filtering of this low-frequency noise?
Is there another feedback path that will improve compensation at these low frequencies?
Comments
The changes based on the app note included reducing R11 to 1K. Also... Removing R9/C10 and adding a resistor (Rx), 3.3K in series with C15.
The ripple is synced to 2x the line frequency, but could be an oscillation excited by the ripple on the bulk rail.
Removing Rx results in more instability with oscillations near 200Hz.
Is there any way to add Rx to the simulation in PI-Expert?
You may be getting to much ripple into cathode of TL431 -- try this circuit and see if it works any better:
Remove R12 and C19
add 2K in series with C14
add a series RC lead network across R14 which would be 20K in series with 0.01uF
Place sense pick-up of R14 on other side of L2
Is ripple line frequency related or is it oscillation? If it is line line related you can increase loop gain by reducing R11 perhaps to 1K.