Hi,
I've been testing a Webench design for a 2.5A 12V output supply using the LM25117 evaluation board. We want it to operate down to 13.5V input (and up to 40V) so I fitted the values generated by Webench accordingly.
I am however seeing classic sub-harmonic oscillation at the low input voltage end (i.e. alternating wide and narrow SW pulses) up to around 14V, and since I'm using the evaluation board I doubt it's a layout or noise problem.
I'm wondering if some form of adaptive slope compensation may be necessary to avoid sub-harmonic oscillation at very high duty cycles. The LM25117 seems amenable to modifying the slope externally with, say, a switched current source, and we have an FPGA in the system which I can use to generate the necessary timing pulse to generate a piece-wise slope. Do you think this might be worth pursuing, or alternatively do you have any application notes etc that might help me tackle the problem?
