Well done for even estimating uncertainties, and being critical of what they are telling you. You're in good company, many scientists who calculate uncertainties fail to include all the sources of possible error the first time round. Many publications of fundamental physical constants show differences way beyond their error bars.
My question is, does the cut off frequency of a circuit can be altered by the frequency that we are subjecting the circuit itself?
No. Though see later for my assumptions.
You'll improve your question greatly by including
- (a) most importantly your list of sources of uncertainty in the measurement,
- (b) also very importantly your estimate of their magnitudes
- (c) the calculation for how you went from that list to the final uncertainty.
Some sources of uncertainty are systematic and frequency independent, they will affect the scaling of the answer, like the scope timebase and signal source reference frequency accuracy (making the assumption that their front panel measurements faithfully represent the correct ratio from their reference frequencies). What are the component tolerances for the theoretical answer?
Some sources are frequency dependent, like the scope and squarewave rise time, or the jitter between squarewave edge and the scope sampling clock (assuming a digital scope), or flatness of the squarewave top (which can go wrong if either scope or signal source are AC coupled).
The scope will have amplitude noise from both its input broadband noise, and its ADC quantisation noise. The source will also have amplitude noise. Both source and scope will have jitter noise on their timebases.
A 'time constant' model is just an approximation. What are the expected magnitudes of inductance for the components? I am assuming they are very small, at audio frequencies, and so insignificant, hence I make that assumption above in my answer 'No'. Are they insignificant? Or a better question, 'at what level do they contribute to the answer?' How have you accounted for the contribution of cables between source, device under test, and scope?
There are some uncertainties I've omitted deliberately. There are probably others I've forgotten, or not been imaginative to think of. It's not unreasonable to keep digging into what possible errors there could be, as if the error will be truly insignificant, it's usually easy to show that with a very rough calculation.
