Adjusting a fixed-voltage DC/DC converter's output

I'm trying to get a 3.5V regulated DC rail from a 28V supply (...it's a weird voltage because the microcontroller's maximum frequency depends on its Vdd -- 4MHz at 2.7V or higher, and 8MHz at 3.5V or higher, and I'd like to run at 8MHz.)

I have a bunch of AP63203 chips laying around (datasheet: https://www.diodes.com/assets/Datasheets/AP63200-AP63201-AP63203-AP63205.pdf), which have a fixed 3.3V output. Before I found out about the microcontroller's clock limitation, I was planning to use one directly.

But this chip has a feedback pin. Normally this is connected directly to the Vout node (in between the inductor and the output capacitance):

simulate this circuit – Schematic created using CircuitLab

...and the VM1 voltmeter would be at 3.3V. But I wonder. If I connect it to a resistor divider, with the FB pin connected to the middle, shouldn't this drive the VM1 output to close to 3.5V?

simulate this circuit

I don't see anything in the datasheet that reads like the raised voltage on the SW or BST pin would be a problem.

One issue I do see is that the feedback voltage won't be quite 3.3V when Vout is 3.5V, because there's an internal voltage divider in the AP63203 to convert a 3.3V level at the FB pin to 800mV before comparing it to the internal 800mV reference voltage. Depending on the input current on that pin, I'd need to modify the value of the 33k resistor, or get rid of it and simply connect a different-value series resistor between Vout and FB ... but I don't see any input current specified in the datasheet. Does anyone else? Could I derive it from other specifications?

My best bet might be to just get an AP63200 (...not a 63201, because the 63200 does PFM), or any other random high-frequency adjustable buck converter, and adjust its output to 3.5V. But I'm trying to make what I have work, if possible. (Plus the 63200 and 63201 only run at 500kHz, while the 63203 runs at 1.1MHz. I like the better efficiency that the higher frequency gives.)

Is it feasible to adjust a normally fixed-output regulator like this?

Thank you!

• I'd recommend using a voltage of a little higher than 3.5 V if you want 8 MHz operation, but I tend to like having some headroom. Commented May 17, 2021 at 1:57
• Chances are that the FB pin has some internal resistor divider already. So it's input impedance could be mediocre. Using an external divider might not work as straight as intended..if you can find out about the internal values you could just put a larger cap to ground only in addition to the first circuit.. try with a poti to ground of e.g. 100k Commented May 17, 2021 at 4:41
• @Hearth - I'd like to, but several other components have an "absolute max" rating of 4V and a "recommended operating conditions" max of 3.6V. So I don't want to go above 3.6, as for some vendors this starts to shorten the life, even if it's fine for others. Commented May 18, 2021 at 2:14
• @tobalt - Yes ... although now that I'm thinking about it again today, I could just measure the resistance from the FB pin to the GND pin, to find the internal divider's sum, then use that as the bottom leg of an external divider. That will work as long as it is a real resistor divider, and not some kind of active element. Commented May 18, 2021 at 2:16