Standing waves in a resonance tube I am doing an experiment about standing waves in a resonance tube. I use a bucket of water, a waterproof tube (open at both ends), and a frequency generator app. I have two set-ups, A and B:
A. Setting the frequency to be constant and dipping the tube into the bucket of water until I hear resonance that signifies standing waves. I record the effective length of tube at which this occurs, from the part where the tube touches the water to the other end. Then, I compute the corresponding wavelengths.
B. Setting the length to be constant but with increasing frequency. I record the frequency at which the standing waves occur and compute the wavelengths.
Using the formula for the speed of sound $v =$ wavelength x frequency, I have solved the experimental speed of sound. Now, using the room temperature to solve for the theoretical speed of sound, I found out that the Set B resulted to less deviation from this theoretical value compared to the case of Set A. Is this expected in general? What do you think are the main sources of error?
 A: You are using the same apparatus to either, 1 detect a maximum loudness whilst altering the length or, 2 detecting a maximum loudness whilst altering the frequency.
From your post it would appear that method 2 allows you to estimate the maximum loudness with greater precision which is perhaps no surprise as I think that I would find it easier to twiddle a knob to change the frequency in "comfort" and estimate the frequency for maximum loudness as oppose to moving a tube up and down whilst keeping it vertical and setting the tube at estimated maximum loudness.
I do not know how you analysed your data but you as you have not mentioned an end correction, $e$, for the tube which is a correction for the fact the the displacement node (pressure node) does no occur exactly at the open end of the tube.
This means that for the fundamental resonance position $L+e= \lambda/4$ where $L$ is the physical length of the tube and $\lambda$ the wavelength of the sound.
$\dfrac {\lambda}{4} = \dfrac cf \Rightarrow L = \dfrac {4c}{f}-e $ where $f$ is the frequency of the sound  and $c$ the speed of sound.
So a graph of $L$ against $\dfrac 1f$ has a gradient of $4c$ and an intercept on the $L$ axis of $-e$.
Note also that if you do not account for the end correction it produces a smaller error in the estimate of the wavelength the larger the wavelength (the smaller the frequecncy) is.
A: It's hard to comment on a specific experimental setup, but I am not surprised by your result.  In B, the machine can control the value of the frequency much more precisely than you can manually submerge the tube into the water to a certain length in A.
One thing to try calculating is, if your length measurement is off by say $\mathrm{0.1 \, cm, 0.5 \, cm, \,or\, 1 \, cm}$, what % change does that produce in the speed of sound result?
Another thing to assess is whether your Method A results seem to be consistently too high, too low, or randomly scattered around the true value.
