GRS 12SFPC-B VISATON R10S and DTW72

The Usual Suspects

Photo 1

An attempt to obtain damn good sound without the expense of boutique prices.

An MTM arrangement.

The mids are Visaton R10S units and the tweeter is a Visaton DTW72. The woofer is a GRS 12SFPC-B. It would have been nice to have used the Visaton W300-8 but at $91, it would have deterred the system's being called budget.

The loudspeakers and crossover components used here are all available at Parts Express

Woofer GRS 12SFPC-B

Midrange VISATON R10S

Tweeter VISATON DTW72

The photos below show the narrow and wide baffled high frequency sections, the latter to simulate the effect of the front baffle if the MTM section were installed in the bass enclosure. The tall high frequency section is due to their being originally used for as an MMTMM arrangement in a previous design.
PHOTO 1	PHOTO 2

Figure 1

The Crossover

A simple first order three way design. Note the polarity reversal of the midrange section. This will be explained later in figs 3 & 4

It should be noted that the 0.3mH inductor will start the high frequency roll-off around 3500hz. This was necessary to obtain a flat response since the midrange high frequency output is quite linear to a little above 10khz. High pass filtering the mids at 8khz adde3d to the low frequency roll-off of the tweeter causing a rise in the response of several dB, regardless of polarity of the mids or tweeter.

Second or third order filters would have reduced that rise but the added expense and complexity would have defeated the original intent, simple and inexpensive.

Figure 2

These response curves show the difference between the 4.5" baffle, green and the 12" baffle, red. The grey curve is THD and is remarkably low at 0.17% above 300hz with the exception of the peak just below 4khz of 1.26%. This oddity initially was thought to be caused by the polarity reversal of the midrange units but the next figure refutes that.

This pair of curves shows that there would be little difference between having the treble section on top of the enclosure or isolated within the larger one.

It seems that it would simplify construction to have the treble section on top. The narrow band 3dB peaks at 500hz and 750hz seen in the red curve should be unnoticed unless one does a somewhat instant A-B comparison with a switch.

Figure 5

Thirty degrees off axis shows little difference between the narrow 4 inch baffle, light blue and the 12 inch baffle, dark blue. Compare this with fig.2 which shows the same comparison on axis.

Note that the dip around 12khz is gone and the rest of the response seems a little smoother. This will allow the speakers to be placed perpendicular to a rear wall without having to angle them inward toward the listener.

Figure 5a

The following three curves are gated responses to remove most of the effects of the room. The green and blue curves are ON axis with the 4" and 12" baffles, resp. The red curve is with the 4" baffle but 30^o off axis. This is very similar to those responses in fig.5. The disappearance of the 12khz dip here shows that the dip is indicative of the on axis response, gated or not. Fig.5 (above) is also a non gated response at 30^o off axis. The slight downward slope above of the higher register is expected.

NOTE

The references "#xx check" in curves 37 and 38 refer to responses measured several days ago and checked again to verify that mic position or other parameters in the system had changed. They didn't.

#37 green is #35 check 1w1m Xover4 4"baffle midrevpol R1R2 1:30-3oclk gtd on axis

#38 blue is #34 check 1w1m Xover4 12"baffle midrevpol R1R2 1:30-3oclk gtd on axis

#39 red is #37 1w1m Xover4 4"baffle midrevpol R1R2 1:30-3oclk gtd 30d off axis

Figure 6

Near field, orange and the response at 1w1m. The near field¹ level was adjusted to match that of the 1w1m level at 200hz. It was felt that at this point, the response of the woofer with its low pass filter would be unaffected by the midrange output with its high pass filter, both at 800hz

Note that at 60hz, the woofer near field output is about 4dB above the level at 200hz. This is very close to the modeled response of BassBoxPro, fig.7

Figure 7

1. Near Field. Response made with the microphone about 2mm from the center of the diaphragm. Upper practical frequency limit is about 400hz.

Presented by Don Keele to the Audio Engineering Society, 1973.

That paper (pdf) can be found via THIS LINK

There may be some who have concerns about the low impedance in the midrange section. This can be alleviated by adding a 4W resistor in series with the midrange. This resistor should be placed after pin 2 of the LP-1 pad and the negative terminal of the midrange parallel pair. The negative terminal on the midrange units is the input terminal of the parallel pair, since these fellas are wired reversed polarity. (ref. fig.1)

OK. The first effect that comes to one's mind is the shift in crossover points. Initially, with 4W in the mid section, the crossover frequencies are 800hz and 3500hz. Adding 4W will cause the lower crossover frequency to shift to 423hz and the upper crossover frequency to shift to 4244hz. This would seem to cause a serious effect on the linearity of the system's frequency response. To find out, a response was run with the system with the added 4W. It did change a little in the mid section but not enough that it couldn't be corrected with the L-pads. Initially, the L-pads were set at 1:30 o'clock and 3:00 o'clock for the mid and tweeter, respectively. These were reset at 3:00 o'clock and 12:00 o'clock for the mid and tweeter, respectively.

The following figures are those of the crossover (4N) with the notch filter and a suggested component layout. (Disregard the red Ln at the bottom of fig.11; it is an editing error.

Again, note the reversed polarity of the mid-range units. Despite the conventional normal polarity in a first order filter, a flatter response was obtained with the polarity reversal. See fig.4. This is possibly due to interference from the woofer which has a good response to 3khz. A larger low pass filter was tried with little effect. While a larger low pass filter of second order may have produced a satisfactory result, the extra cost was unjustifiable.

Figure 12

This figure shows the effect of the notch filter. RED is with the filter and GREY is with the filter bypassed. It was noticed that with the filter bypassed, the sound was quite satisfactory, at least to my aged ears. However, at levels approaching 95dB, female vocalists and violins can sound harsh.