Electronic – Powering microcontroller with a capacitor

If your wires/traces are thin little buggers it might happen that one voltage leads on the other in your implementation with 10uF to the order of tenths of volts. But I do mean thin and long.

On a 3.3V source voltage a charge differential of 0.4V (smallest maximum difference between the two in the constraints) on two 10uF capacitors each with 1 to 10 100nF caps parallel to it means a comparatively huge difference in resistance (and/or inductance) of the tracks/wires going to it.

Since VDDIN is allowed to be slightly below VDDIO and VDDIO is allowed to be slightly below VDDIN for all the given constraints, that makes the problem sort of balanced, so you need not look one way or the other.

Just make sure your traces can support the current requirement, keep your capacitances sane (which 10uF still is) and try to keep the large capacitors the same type and physical size, if the specifications allow for that, to make sure ESRs and such are in the same neighbourhood. If your traces are "fat" enough to support the current requirements of the MCU without a noticeable voltage drop, they're going to be "fat" enough to charge 10uF caps quick enough.

However, the best solution is always to do what they say in the datasheet examples, if they drive it with a single 3.3V -or similar- source (which I think would be likely), use their connection example and guidelines. If they want Cap X and Inductor Y for filtering, use those unless you know for sure your solution will work too.

You have a large number of options.

a) Your link to motorised cap is indeed way over the top. However, if you can buy, or rescue from a junked radio, a vane-type variable tuning capacitor, you can buy a geared stepper motor for a few $£euro from fleaBay, and couple them yourself.

b) Although you say you only need 10% variation in C, it is usually unwise to use a variable capacitor as the entire tuning C. Far better to used a fixed capacitor, and connect a variable capacitor with a relatively large capacitance swing to it with a small capacitor. This has a number of advantages.

The stability and resolution of the resulting combination is far better than that of the variable capacitor alone.

The RF voltage across the variable element is lower than that across the combination. This is not so important if the variable is a vane-type, but could be very significant if the variable is a varactor.

If the variable is a varactor with a low Q, the Q of the combination will tend to the Q of the fixed capacitor, which you could choose to be higher.

c) You can adjust the inductor, perhaps by moving a ferrite core in and out of a solenoid, perhaps by making the inductor from two solenoids in series, and rotating one to change their mutual inductance, so changing the total inductance.