Other Parts Discussed in Thread: LM27761
Hi,
we noticed the following issue with LM27762: the device disturbs our power supply line and its own outputs, it adds ripple in the kHz-range (e.g. at ~41kHz+mulitples).
We have seen that others reported similar issues here, like:
- https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1054414/lm27762-pfm-and-constant-switching-frequency-for-light-load
- https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/935119/lm27762-light-load-operation-of-lm27762
- https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/947755/lm27762-at-low-levels-of-output-current-it-is-generating-strong-spurious-tones-within-the-loop-bandwidth-of-the-pll
- https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/671226/lm27762-operating-mode-change-internally-on-light-loads-output-ripple
Our investigation showed that the charge pump is periodically disabled (at a frequency of around 41kHz). The fundamental switching frequency of ~2Mhz is not modified. Hard to predict that behavior from the datasheet, but we think this is meant by 7.3.3 PFM Operation.
While this ripple is suppressed at the negative output to ~0.5mVpkpk, it is propagated to the input (~14mVpkpk), and thus even to the positive output (~1.5mVpkpk; dominated by harmonics).
Our values are: R1 19k, R2 47k, R3 18.47k, R4 47k, Cin 2.2uF, Cout+ 2.2uF, Cout- 2.2uF, C1 1uF, Ccpout 10uF.
We have read that the only workaround is adding load. But doubts remain. Can pulse-skipping really be suppressed under all circumstances? What conditions must be met for reliable(!) suppression of PFM mode?
Is there a better solution to suppress the PFM ripple? Should we switch to a different device?
We require low noise output at +-1.7V at low current (<<100mA), our supply is currently +3.3V.
Regards
Markus
