Dayton Tweeter Tests and Modifications

First Measurements

David L. Ralph

4 July 2001

This first set of measurements were taken to form a basis for any modifications. I wanted to eliminate box influences so that any non-linearities would be attributable to the drivers to make it easier to see what was happening. I took the largest piece of uncut mdf I had on hand (39" by 49", about 1m x 1.2m) and cut a mounting hole in the center. This one is 11" in diameter (for later use with large drivers). I then cut a ring which fit this opening and has a 7" diameter hole in it. With this I can easily make discs with the cutout for any tweeter (and mid up to 6-1/2") I might want to test using small pieces of mdf.

It is possible that this setup could have intrinsic resonances, but with tweeters I think that they will be minimal because of the large size. I have an accelerometer which I will use later to test for resonances.

All of the measurements were made at a distance of 0.5m. This helps to lower the frequency at which the measurement is still valid.

The raw unmodified frequency responses of the tweeters show the relative uniformity of 3 out of 4. Unit D, which had an odd impedance measurement, shows a comparable SPL variance from the others.

I chose Unit C to be the representative example for multiple tests. I expect to leave it untouched throughout this process as a fixed reference, with the exception of opening it for inspection of ferrofluid.

The graphs here are for this unit on the flat panel. There is a characteristic dip from 3 to 5 khz which seems to be typical of this model. This will be considered when selecting baffle dimensions and positioning of the driver on the baffle.

These are Cumulative Spectral Decay (CSD or waterfall) plots of two of the drivers as received. Unit C is a typical response while Unit D is the atypical one.

A new full set of measurements will need to be made later and should look different because I will open them all for inspection. I am rotating the domes while set in the magnet assembly in an attempt to evenly distribute the ferrofluid. It seems to work reasonably well, but care must be taken to prevent damage. The photos I will post later show the initial and adjusted fluid distribution.

Unit D is the one with the unusual impedance and SPL responses. Note the one rather severe resonance at 5khz. Even though this is at the same point in the others, it is fairly quickly attenuated in them, though not enough to prevent some impact on the initial SPL. The full size graphs also have a 50db scale which shows that this resonance endures for a rather long time. I hope to isolate the cause and correct it.

These are the on-axis impulse responses of all four units overlayed and expanded for easier viewing. The full size graphs show the full impulse response of each unit overlayed. Two more graphs show the impulses from Unit C only, using the on-axis and +/- 15 dgree responses overlayed.

This is the on-axis step response of Unit C. The full size graphs show the step responses of both Unit C and D.


Getting Started First Measurements First Modifications Accelerometer CSDs