...Lee Neidow to John Whittaker
On the other hand, if a consideration of the BG driver being 'underdamped'
affects one's decision to purchase I'd think having some sort of objective
measure would be nice as opposed to "the ever present bullshit to which
audiophiles are exposed."
...John Whittaker to Lee Neidow
INTRODUCTION
In the beginning there was me and my BGs (and the Mighty DynaPleats). Then Rudi Blondia wrote to the Bass/DIY List requesting assistance in measuring his new Carver 60" and BG RD75 drivers. I had an IMP/MLS instrument and a large backyard, and was interested in starting a baffle development program for my own RD75s. It seemed a perfect way to get the ball rolling.
On Saturday, May 17, 1997, we met and proceeded to make 41 measurements (setups) on the assorted planar drivers. It was a good day in terms of work accomplished. Unfortuately, the day became windy, and that became troublesome later on in the afternoon. The Carver drivers were measured when the day was least windy. Comparing the data to that obtained indoors at later times, it seems remarkably clean, without the reflections that plague indoors measurements. The raw impulse responses are presented for inspection and analysis; and, indeed form the basis for much of this comparison.
Of the 41 measurements made, 6 were choosen for their equivalency in situation. Two Carver 60" planar drivers and two BG RD75 planar drivers are represented. There are two conditions explored, with and without baffles. The raw data will be made available upon request.
The question to be answered is (from above): "My biggest concern with the BGs is their susceptibility to low-end artifacts resulting from their underdamped configuration." - that is, which driver is underdamped compared to the other. This "concern" comes from a well-known individual on the Bass/DIY List (who is invited to reveal himself); and, from Rudi Blondia, who has repeated the opinion to me that the RD75 driver is underdamped in comparison to the Carver 60" on several occasions.
On October 28th, Mr. Purl in commentary to a post of mine to the Bass/DIY List said the following in regards the question:
"One nearly ineluctable inference that arises from inspection of the data on the BG and Carver "ribbons" is that Carvers show what appears to be the traditional mass and compliance poles that all finite-mass drivers are heir to. The mass of the moving system is very low indeed, so I am surprised to see what looks like a pole in the region of the upper hundred Hertz for so light a moving mass. The simple and elegant tensioning system of the Carver affords welcome damping throughout its range of operation. The BG seems crude by comparison, and indeed its data graphs show higher distortion floors than the Carver equivalents."
What might "underdamped" mean in this matter? Let us keep this simple, since surely there will be experts coming forth to give us 'better' answers. When a loudspeaker delivers excess energy over time, from an impulse excitation, that is an example of being underdamped. When it 'rings' when 'struck' (by an impulse), that is because it is underdamped. Maybe there are other definitions that apply? I have not been made aware of them, even though I have asked - the experts.
Comparision is the essence of this evaluation. Like situations, like measurements - head-to-head, mano-a-mano. This is probably the fairest comparison of the Carver 60" planar driver and the BG RD75 planar driver that will be made available. So let us look together at what data we have and makeup our own minds on the matter of which driver is underdamped. There may never be another opportunity for such a comparison!
TECHNICAL CONSIDERATIONS
A wonderful irony is that the IMP/MLS amateur measurement system is better equiped to explore impulse related issues than are more expensive systems like the LAUD which have slower sampling rates. The IMP has a sampling rate of 61200Hz compared to the LAUD at 48000Hz. This means that for every 61.2 samples the IMP collects, the LAUD will collect 48 samples, hence the IMP has much greater time resolution, and is better suited to analysis of the impulse response. Risetime details revealed by the IMP could be expected to be 'washed out' by the LAUD system.
Both drivers were measured in the same position in my backyard, using a Yamaha M-630 amplifier (170W/ch @ 6 ohms) for amplification. The gain settings on the amplifier remained the same for all measurements. Both drivers were measured on-axis with the microphone at approximately the mid-point of the driver. The Carver drivers were mounted high on the test assembly and the microphone height was at 1.21m (47.5"), the RD75 drivers stood at ground level and the microphone was at 1m (39.3") height. Photographs of the test rack and drivers are displayed on a preceeding page in the 'Outdoor Measurements' section.
The same boards were used for the baffles for both driver types. The additional 1.5" width of the Carver baffle is due to that driver's greater width than the RD75. In the case of the 'unbaffled' measurements there w as actually a small baffle of sorts created by the mounting method. For both driver types, that mounting method produced frequency response notches - which are detailed elsewhere.
All measurements occurred in the timespan from 11:14 A.M. - 1:39 P.M. As noted previously, the wind was much less a factor in the measurement of the Carvers. The RD75 drivers were measured with and without passive notch filter installed. The Carver drivers had no notch filter available at that time.
In order to maintain compatability with other measurements in this study 432 point (7 mSec) data sets were extracted from the 4096 point primary data sets. The data was then FFT'd in a 4096 point domain in which the remainder of the points (after 432) were set to the value of the last data point. The resulting data set was windowed with a half-Bingham window to reduce low frequency FFT artifacts. Both 1/3 and 1/12 octave smoothing was applied to some of the resulting frequency responses. Primary frequency responses in which NO smoothing has been applied are also presented.
Waterfall frequency response plots, in addition to the raw impulse responses, are the primary means of evaluation of the question of the RD75 driver being underdamped - compared to the Carver 60.
CONCLUSIONS
The Carver 60" driver is more sensitive than is the RD75 driver, but apparently less linear over it's range. The RD75 driver has significantly better high frequency performance. Inspection of the unbaffled frequency responses of both drivers shows that mid-band the Carver 60" is possibly +3dB more sensitive than is the RD75. A conjecture is that the Carver's additional sensitivity benefits will be reduced when equalization is used to linearize it's frequency response.
By a small margin the RD75 driver is seen to have superior risetime than the Carver driver. One Carver driver seems defective in the area of risetime - compared to it's mate in this comparsion. One RD75 driver shows a very narrow range, high frequency ringing, which the other RD75 does not.
By inspection, of both the impulse and waterfall frequency responses, the Carver 60" driver is significantly more underdamped than is the RD75 - both drivers being judged without the required passive notch filters for attenuation of their cavity resonances. It must be understood that neither driver can be used without some form of attenuation of their cavity resonances. Given that use of a passive notch filter with the RD75 has been found to deteriorate the low frequency (>1000Hz) damping, it might be extrapolated that the Carver 60 would become even worse in regards its low frequency damping when a passive notch filter is used with it.
It can be seen that addition of the passive notch filter to the RD75 driver greatly affects its entire response, not just the range of the cavity resonance. The impulse 'tail' is greatly reduced with use of the notch filter. Presumably the same effect would apply to the Carver driver when using a similar configuration passive notch filter.
As a conjecture, it seems reasonable that the addition of the notch filter's resistance/impedanceto the amplifier/speaker circuit reduces the control of the amplifier upon the driver. This seems a compelling reason to utilize electronic equalization for control of the cavity resonance rather than the passive notch filter.
That one Carver driver and one BG driver seem to have performance anomalies, that might be considered defects, is a disheartening and sobering revelation for the prospective buyer.
| NO BAFFLE | Impulse Response | Frequency Response | Waterfall FR |
|---|---|---|---|
| BAFFLED | Impulse Response | Frequency Response | Waterfall FR |
| OTHER | Calibration |
Offaxis Response.
The Active Filter Response.
The RD75 Dipole Baffle Study - Table of Contents.
Acoustic Line Source Research - Table of Contents.
