- #1
Freixas
- 307
- 42
- TL;DR Summary
- I need help understanding air flow. This knowledge will be used in designing melodicas.
A melodica is a wind instrument that has a keyboard (piano-style, not computer-style). The player blows into the instrument. When a key is pressed, a passage opens. The air flows over a brass reed, making it vibrate. The reed sits in a channel a hair's-width bigger than the reed. The air flows around and past the reed as it vibrates and exits out the passage opened by the key press.
I play the melodica and I belong to groups that include performers and builders. There are some simple questions that I could answer with elaborate testing, but I thought I could save time (and learn something) by asking the experts. The answers might be helpful to the instrument makers.
The Length Question
Does the length of a melodica's mouthpiece matter? Melodica players have strong opinions on the subject,
My theory (as a non-physicist) is that, in the relatively low-pressure environment of the lungs, air is incompressible and, as long as the mouthpiece is reasonably rigid, if I blow air into one end, I will get the same amount of air exiting the other end.
I've have assumed (without any evidence or study) that the signal (air pressure change) travels through the mouthpiece at the speed of sound.
I did experiment with this by playing a melodica using 15' of vinyl tubing and comparing it to blowing directly into the instrument without a mouthpiece. I found no significant differences. If my speed-of-sound theory is correct, the latency of 15' of tubing would be 13 ms, not enough to be noticeable.
In my theory, the shape of the mouthpiece is immaterial. The mouthpiece could be narrow or broad--it would make no difference.
One added question is whether the friction in the mouthpiece would affect the responsiveness of the signal. The test case is a vibrato; if friction "smears" the signal, it will act as an averaging filter reduce the amplitude differences in the vibrato. With enough friction, a vibrato would be impossible.
15' of vinyl tubing made no noticeable difference to a vibrato, even when examining the captured waveform. I then packed a section of the tubing with aquarium floss (imagine plastic cotton fibers) to add a high degree of friction. Again, this made no difference to the vibrato.
Length, but on the Exhaust
A MIDI breath controller converts air pressure to a MIDI signal. Since we want the instantaneous pressure, the designs I'm familiar with involve a Y-fork which directs some air to a pressure sensor and some to an exhaust port. Without an exhaust port, a rapid change in pressure would not be possible (no vibrato, for example).
The distance from the pressure sensor to the exhaust (unlike the length of the mouthpiece in the prior section) needs to be short to maintain responsiveness. Add 15' of tubing to the exhaust and the vibrato rate drops to zero.
I don't have a theory that explains this.
The Y tube
Take three tubes, all with the same inner diameter, and connect them using a Y-connector. My theory is that if I blow in one tube, equal amounts of air will flow out each of the other two ends.
Now swap one of the exit tubes for one with a smaller inner diameter. My theory is that the bigger exit tube will have proportionally more air flow; in other words, the total input air flow will equal the total output air flow, but the ratio of air coming out the larger tube to that of the smaller tube will be proportional by the ratio of the cross-sections of each tube.
This is important in melodica design in that the lower the note, the bigger the reed and thus the channel it sits in and into which the air flows.
When a single note is played, the air flow is determined by the discussion in the first section above; in other words, the size of the passages in which the air flows is immaterial--the same amount of air flows out as flows in.
When two notes are played, the air flow is as described in this section: proportional to the size of the passage. The low notes would get more air than the high notes.
Air Speed as a Factor
There may be another factor I'm missing.
The lower note reeds are bigger. This means they have more length and more mass. All reeds have the potential to break, but the larger reeds are most susceptible .
The solution I've heard is to alter the size of the passage to the reeds. The chamber leading to the lower notes is shallower than the chamber leading to the high notes. The lower notes then get less air and thus vibrate less hard.
This makes no sense if we go by my thoughts in the first section: the size and shape of the mouthpiece has no effect on the air flow. What goes in one end comes out the other.
Using my thoughts about the Y tube, the stepped chambers would have an effect on balancing the relative volumes when a high note and a low note are played together, but that would be it.
There's a lot here, so let me summarize and number my questions:
I play the melodica and I belong to groups that include performers and builders. There are some simple questions that I could answer with elaborate testing, but I thought I could save time (and learn something) by asking the experts. The answers might be helpful to the instrument makers.
The Length Question
Does the length of a melodica's mouthpiece matter? Melodica players have strong opinions on the subject,
My theory (as a non-physicist) is that, in the relatively low-pressure environment of the lungs, air is incompressible and, as long as the mouthpiece is reasonably rigid, if I blow air into one end, I will get the same amount of air exiting the other end.
I've have assumed (without any evidence or study) that the signal (air pressure change) travels through the mouthpiece at the speed of sound.
I did experiment with this by playing a melodica using 15' of vinyl tubing and comparing it to blowing directly into the instrument without a mouthpiece. I found no significant differences. If my speed-of-sound theory is correct, the latency of 15' of tubing would be 13 ms, not enough to be noticeable.
In my theory, the shape of the mouthpiece is immaterial. The mouthpiece could be narrow or broad--it would make no difference.
One added question is whether the friction in the mouthpiece would affect the responsiveness of the signal. The test case is a vibrato; if friction "smears" the signal, it will act as an averaging filter reduce the amplitude differences in the vibrato. With enough friction, a vibrato would be impossible.
15' of vinyl tubing made no noticeable difference to a vibrato, even when examining the captured waveform. I then packed a section of the tubing with aquarium floss (imagine plastic cotton fibers) to add a high degree of friction. Again, this made no difference to the vibrato.
Length, but on the Exhaust
A MIDI breath controller converts air pressure to a MIDI signal. Since we want the instantaneous pressure, the designs I'm familiar with involve a Y-fork which directs some air to a pressure sensor and some to an exhaust port. Without an exhaust port, a rapid change in pressure would not be possible (no vibrato, for example).
The distance from the pressure sensor to the exhaust (unlike the length of the mouthpiece in the prior section) needs to be short to maintain responsiveness. Add 15' of tubing to the exhaust and the vibrato rate drops to zero.
I don't have a theory that explains this.
The Y tube
Take three tubes, all with the same inner diameter, and connect them using a Y-connector. My theory is that if I blow in one tube, equal amounts of air will flow out each of the other two ends.
Now swap one of the exit tubes for one with a smaller inner diameter. My theory is that the bigger exit tube will have proportionally more air flow; in other words, the total input air flow will equal the total output air flow, but the ratio of air coming out the larger tube to that of the smaller tube will be proportional by the ratio of the cross-sections of each tube.
This is important in melodica design in that the lower the note, the bigger the reed and thus the channel it sits in and into which the air flows.
When a single note is played, the air flow is determined by the discussion in the first section above; in other words, the size of the passages in which the air flows is immaterial--the same amount of air flows out as flows in.
When two notes are played, the air flow is as described in this section: proportional to the size of the passage. The low notes would get more air than the high notes.
Air Speed as a Factor
There may be another factor I'm missing.
The lower note reeds are bigger. This means they have more length and more mass. All reeds have the potential to break, but the larger reeds are most susceptible .
The solution I've heard is to alter the size of the passage to the reeds. The chamber leading to the lower notes is shallower than the chamber leading to the high notes. The lower notes then get less air and thus vibrate less hard.
This makes no sense if we go by my thoughts in the first section: the size and shape of the mouthpiece has no effect on the air flow. What goes in one end comes out the other.
Using my thoughts about the Y tube, the stepped chambers would have an effect on balancing the relative volumes when a high note and a low note are played together, but that would be it.
There's a lot here, so let me summarize and number my questions:
- If we blow air into one end of a passage, what are the physics involved in determining what comes out the other end (and when)?
- What happens in breath controller when the exhaust path is lengthened (with respect to something like a vibrato)?
- What happens in a Y-shaped path when one of the exit paths has a different cross-section than the other? Actually: is it a matter of the cross-section at the exit or the volume of the entire exit path?
