Great idea. I think it would best if comp driver was donated to you and print the horn to keep measurements consistent. The donator of the compression driver pays return postage and a donation to cover printing costs and they now have an optimized version of a waveguide they can print for themselves. Or they can pay for postage and printing for a pair of waveguides shipped back in the original oversized box that they posted the compression driver.
Otherwise you could find a bunch of people on the forums that are experienced and meticulous in measuring loudspeakers and will follow the measurement protocols.
May or may not yield information that is useful for producing an optimization model but would be very useful for adding to the knowledge pool and potential improvements to the software.
Actually, measuring a compression driver for this purpose wouldn't be very hard. Among the objects to print there would be a small mounting and rotating jig (with angle stops, etc.), so virtually the only thing to be careful about would be setting the prescribed mic distance and on-axis direction. The mic even doesn't need to be calibrated in any way, as this has no effect on the results.
I can't do what you suggest.
I can't do what you suggest.
In that case I agree with you. Best done as a group effort.
I would be happy to contribute.
Yes of course take it at your own pace. It is very interesting work you are doing.
BTW, do we see any non-axisymmetric modes in practice? I mean, with a driver rotated on the mouting flange, do (e.g.) the horizontal polars change in a noticeable way? Has anyone ever observed this?
- I suspect that the cut-off frequencies will be lower for these modes but their actual levels too low to bother in real drivers.
- I suspect that the cut-off frequencies will be lower for these modes but their actual levels too low to bother in real drivers.
Last edited:
Don Keele claims to have done that - many years ago - and found significant differences. I think that drivers today are much better than way back when, but it's still a good question. No experience myself.
A rocking mode of the diaphragm would be the first non-axi mode. That would have a very poor acoustic coupling, basically a dipole in the compression chamber - no net change of pressure.
A rocking mode of the diaphragm would be the first non-axi mode. That would have a very poor acoustic coupling, basically a dipole in the compression chamber - no net change of pressure.
Last edited:
The principle is straightforward. As a "linear" system the modes are uncorrelated and hence you will have a frequency response for the zeroth and the first modes as separate frequency responses. The first mode will be a negligibly small amount at LFs to simply very small. So yes, in theory the frequency response is complex.
I have done calculations both ways -i.e. reconstructing a sound field from measurements, complex and real. There is very little difference at LFs and a noticeable amount at HFs. It's all up to the complexity of going to complex for the SVD. If you have code that can do that then why not go complex.
I have done calculations both ways -i.e. reconstructing a sound field from measurements, complex and real. There is very little difference at LFs and a noticeable amount at HFs. It's all up to the complexity of going to complex for the SVD. If you have code that can do that then why not go complex.
Calculation with complex numbers is no problem.
- I'm finishing the "reference waveguide" design. As it is, it should be easy to print in one go, without any supports (those who print know very well what I mean). It will compose of two pieces - mounting plate and a waveguide body. (For 1" drivers this is ⌀170 x 50 mm, not necessarily the final one.)
- I'm finishing the "reference waveguide" design. As it is, it should be easy to print in one go, without any supports (those who print know very well what I mean). It will compose of two pieces - mounting plate and a waveguide body. (For 1" drivers this is ⌀170 x 50 mm, not necessarily the final one.)
Attachments
