I am building a homemade xylophone out of copper pipes. I want it to be accurate if I measure it for 5 seconds and take the average pitch. For my F4 pipe for example, when I do this, it tends to go really high (somewhere around A7 usually) and then go to around 340 Hz, which is closer to the frequency I want, but then it will alternate between 341 and 0 Hz, so my average frequency gets really skewed. Any ideas as to why? My bars are mounted at their nodes, and I've tried using a hard and softer homemade mallet. Thanks!
The main problem is your measurement, not necessarily the instrument. The Xylophone creates a lot of different frequencys. Their relative strength will change over time and will confuse most measuring devices. The only instrument I know of that does a good enough job is ears. You have to listen to the instrument and decide whether it has a good sound and is in tune. Ideally you need to listen a few meters away, and in different directions. Some of the "noise" frequencys gets less noticeable a bit away from the instrument.
I suggest you do a goodle search on "xylophone frequency spectrum", people have been looking into this already.
If you want to change the sound or timbre of the instrument you might want to experiment with different materials, different geometries, perhaps drilling holes or adding dampening or resonant materials. This might allow you to deaden all but the fundamental frequencys, but then it will not sound like a xylophone anymore. But, please, do go ahead and experiment, it is great fun. /Gunnar