Quest for an electrovalve acting at audible frequencies

[njiruk]

For a musical acoustic application we would need a very fast linear proportional 2/2 way air electrovalve that acts at audible frequencies with a minimal phase lag.
The valve capacity should be at least: 5 l/min, the response time should be smaller than 0.5 ms and hysteresis should be smaller than 1%. Leakage wouldn't be very disturbing, it may be up to 10%. The valve may be either normally closed or open.
The output pressure is about 100 kPa, and the input pressure depends on 2 possible valve designs:
1) subcritical functioning:
p_in = 110 kPa.
2) supercritical functioning:
1000 kPa > p_in > 400 kPa ( preferably 400 kPa, to minimize flow turbulence)
(both designs are possible as both in- and output pressure can be measured in real-time)

A thorough research on the electrovalve market turned out in no results. The combination of short response time and high flow rates is obviously not evident.

However, considering typical electrovalve designs, I was wondering why there aren't any valves that regulate a large orifice by a electronically direct acted diaphragm so that a wide opening will be created. In order to minimize the pressure load on the diaphragm, the orifice in contact with the diaphragm should be long and small.

I was thinking to use a loudspeaker (tweeter) membrane for the diaphragm. I joined a figure that roughly demonstrates the set-up.
Given the resonant frequency of the tweeter: 1,5 kHz, the response time will be sufficiently short.
Further, as the membrane displacement is about 0.1 mm peak-to-peak, and the orifice circumference is 80 mm, the created valve orifice surface is 8 mm^2. Compared to a clarinet reed opening, this would allow the desired flow rate.
Or am I overlooking something!?

 

[MikeHalloran]

The control needs to apply a constant current bias to close the valve against the supply pressure, if the supply is fed to the interior of the tube. Opening the valve then requires removing the current, making it work like a relaxation oscillator, or reversing the polarity of the current, probably necessary to get the speed you want with inductance that you can't avoid.

If the supply is fed to the exterior of the tube, like in an air horn, I think the control dynamics get more difficult.

Most people who search (in vain) for super-fast valves like that, also want them to shut off bubble-tight. Yours at least has a prayer of meeting your specifications.



Mike Halloran
Pembroke Pines, FL, USA

 

[gerhardl]

If you want to check further, I believe that you will find some of the fastest acting valves at manufacturors of air valves for cars and buses. Even if this is fast operating valves, and in milliseconds area, I seem to remember that this is some decades off your target. Capacity not checked.

You could for instance contact the Norgren factories. They are sub-supplying both special solenoid valves and proportional valves for the car industri. You will probably have to contact directly, some of the data for special valves and their properties might not be avaiable directly on the internet.

.. apart from this I agree with Mike: such an animal does (regrettably) not exist. (I would be delighted to be corrected!)

 

[77JQX]

Does it have to be one valve? Could you get the effect you're after with an array of valves?

 

[njiruk]

Thanks for your responses!

@MikeHalloran
The air supply is fed to the exterior of the tube, however, I think the problem is similar to the inverse case, so the remarks you gave about the inverse case also apply here.
It should be mentioned that the proposed construction is meant for a subcritical functioning, so for a pressure difference of 10 kPa.
The orifice slot, could have a width of about 0.2 mm, so that the pressure difference acts on a surface of 16 mm^2, to become a load of 0.16 N.
The spring stiffness is about 4600 N/m, so that the resulting displacement would be 0.035 mm, which is small enough compared to the total displacement of 0.1 mm, also, in the context of my application...
I also evaluated a (static) finite element simulation of the membrane with this load and it turned out that its deformation is negligible (0.01 mm of displacement between the coil and the membrane at the orifice level).

@gerhardl
In a primary set-up I used a Bürkert 2824, which turned out to bear too much hysteresis, and also the response time was too big. I also looked at valves for the car industry, they had no superior characteristics... I'll give it a try at Norgren though.

@77JQX
I also thought about this. It may be an array indeed! However, the valve outputs should be close enough, the maximum shared surface should be about 3x3 cm^2.
I looked at MEMS valve arrays, while their response times are very very gratifying, generally the capacity is too small. The most recent articles I read about it seemed to go in the good direction, but we're still too far from a concrete use... especially considering the prices...!
I think it would be interesting to combine intermediate size valves with sufficient response time, and a combined sufficient capacity! However, I did not found any of such small valves...