[David I Am]

Okay, here’s the sequence of events.

1. the aperture is closed. There is a pressure differential between the air distribution chest and the reed, but no air is moving. There is a static air pressure across the chest.

2. the aperture opens. No air has moved, but there is now a path for it to flow.

3. A front forms as the air immediately adjacent starts moving into the opening rushing towards the lower pressure.

4. the air next to the initial air moves to follow the previous air, advancing the front and diminishing it somewhat, expanding radially from the opening.

5. the air next to the next air moves to follow the previous air, advancing the front and diminishing it somewhat, expanding radially from the opening.

6. within milliseconds the front has moved across the entire chest, impacting first the diaphram, then the long walls, and finally the short walls and right up the mouthpiece to the source of the air pressure.

At this point consider, you have a pressure gradient – on the outside of the reed is atmospheric – there’s a differential across the reed block – then a somewhat smooth gradient of increasing pressure that could be measured all the way from the reed block to the pressure source. This pressure gradient will exist so long as there is flow, based on the viscosity of the air.

In a rigid chest, the large opening of the bass reed will allow a drop in the pressure – too much air rushing out too fast per the size of the chest, with the viscosity of the air involved the pressure will drop too low in the chest to continue to deliver enough impact air to the reed to make it start sounding. You’ll get a puff, a momentary pause, and then the fronts will have propagated up and dynamic flow conditions will set up the volume.

In an elastic diaphram chest, when that front impacts the diaphram it allows the diaphram to tighten – to maintain the pressure in front of near at near that of the elastic tension of the diaphram provides – proving a a much shorter path – from across the shortest way of the chest – to provide fill air in the critical initial milliseconds of a valve opening, while the worst of the pressure gradient is in place.

This may be somewhat unintuitive because of the high speeds involved – I just typed all these paragraphs literally describing what takes places in the time it takes me to strike a single letter. 🙂

It may help to think of it like with water which is more obviously viscous – water buffer chests/tanks are used for exactly this purpose in applications where they need a surge-blast of water all at once – such as a commercial dishwashing machine. A pressure tank is colocated with the washing machine and connected with thick pipes to its outlet. the smaller inlet and outlet are both at the bottom – it may or may not have a diaphragm, but it will have pressurized air – a bubble at the top of the tank. When the valves snap open on the dishwasher, rather than depending on the pressure gradient driving all the way back to the main water heater, instead the water rushes down the pipe from the pressure tank. It can’t provide more than 20 or so gallons of water at any one surge – but it doesn’t need to. While the machine uses the surge of water it slurped in, the smaller inlet is busily recharging the tank – just like the mouth of the player is recharging the chest diaphragm of the melodica far slower than the rush of air when a valve opens.