Yes, thanks. It indeed looks like TH4001 clone but here in a full wood version (top and bottom).
What you can expect from the horizontal radiation polar was found by our collaboration with @DonVK, @fluid using AKABAK:
It took some time for us to find an appropriate way how to model the different channels inside the horn and we tried many different models with or without interfaces/subdomains. The result is with a 2" round driving source and an appropriate adapter (round to square).
The main difference of TH4001 to Yuichi is that Yuichi has the horizontal flare added which helps to avoid midrange issues. It is not entirely perfect but reasonable good for such a horn.
It took some time for us to find an appropriate way how to model the different channels inside the horn and we tried many different models with or without interfaces/subdomains. The result is with a 2" round driving source and an appropriate adapter (round to square).
The main difference of TH4001 to Yuichi is that Yuichi has the horizontal flare added which helps to avoid midrange issues. It is not entirely perfect but reasonable good for such a horn.
Last edited:
You may notice with some horns that the greater they load the more critical it is to stay away from the cutoff, but there are other factors. It depends how you cross. You're not the first to want to cross it higher than this.
I currently doubt that the 725 Hz dip is horn related.
Just look at the curves of the 18sound ND3N attached to this horn with a special adapter:
The curves have been separated within the sw. The more the mic is away from the mouth you get comb filter effects from floor/celing/etc.
I suggest that you take two additional measurements with the mic very slightly before the mouth plane (1-2 cm) centred with mic pointing to the centre channel and slightly asymmetric with mic pointing in line to one of the inner fins. This will show us the native driver/horn response except the baffle loss of the horn.
I also suggest that you start with an x-over point around 750-800 Hz at least 2nd order.
Just look at the curves of the 18sound ND3N attached to this horn with a special adapter:
The curves have been separated within the sw. The more the mic is away from the mouth you get comb filter effects from floor/celing/etc.
I suggest that you take two additional measurements with the mic very slightly before the mouth plane (1-2 cm) centred with mic pointing to the centre channel and slightly asymmetric with mic pointing in line to one of the inner fins. This will show us the native driver/horn response except the baffle loss of the horn.
I also suggest that you start with an x-over point around 750-800 Hz at least 2nd order.
This is impressive. The directivity seems very well controlled (enough but no too much) across such a wide frequency range.
Thank you for the suggestions. I will resume measurements once I receive the throat adaptor - which should be soon. There is little point in starting a new filter until throat matters are settled
. And I plan on a x-over point somewhere between 725 and 775 as I found this to be a sweet spot for the TAD woofer.In the mean time, I have done cursory listening using the old filter. The sweet spot is huge compared to the JMLC, as you are all aware. In particular, and I am very happy about this, I can listen closer to the speakers without the soundstage disintegrating.
There was not enough top end to my liking, so I dusted off some old tweeters to try a 3 way configuration for fun. Here are some pics.
For tweeter crossover I just mashed together components I had lying around. Second order, 0.1 mH and 0.82 uF. As I develop the new midrange filter, I can increase the tweeter level by choosing a different screw on the board for higher cap values. Note that this does not mean that I am yet sold on a 3 way and I still do intend to EQ the driver and listen to a proper 2 way configuration.
Fun experiment. More later when the throat extenders arrive.
Last edited:
Measurement distance
As pointed out by docali the dip measured at 725 Hz was a room effect. Thanks for the tip!
Here are measurements from flush to the TAD horn mouth, up to 1 m distance, in 20 cm increments. The dip of (non-)concern is clearly seen appearing at 60 cm (24 in) distance (brown curve) but does not show up at the horn mouth (top curve). So it is caused by room reflections, not by the horn.
As pointed out by docali the dip measured at 725 Hz was a room effect. Thanks for the tip!
Here are measurements from flush to the TAD horn mouth, up to 1 m distance, in 20 cm increments. The dip of (non-)concern is clearly seen appearing at 60 cm (24 in) distance (brown curve) but does not show up at the horn mouth (top curve). So it is caused by room reflections, not by the horn.
The Eminence HA14-2 conical throat adaptor has arrived. Its dimensions are:
Input: round diameter 1.4 inch (35.56 mm)
Exit: round diameter 2 inch (50.8 mm)
Length: 47 mm
Angle: 9 degrees (per side)
Adding to it the 20 mm length of the round-to-rectangular TH4001 adaptor results in 67 mm for the total length of the composite adaptor. The impedance obtained using both the "stubby" 24 mm adaptor (dashed curve) and the new longer adaptor (solid curve) are shown in this plot.
As docali expected, the difference is substantial and the loading provided by the longer adaptor should now be closer to that provided by the JMLC design.
Also, the exit angle of the Radian 745NEO Be is not 10 degrees as specified on the spec sheet. Is is 4 degrees per side according to a Radian engineer that Lynn Olson spoke to a while back [cf. 2012-10-27 3:07 pm post in the "Beyond the Ariel" Thread]. Peering into the driver's exit through the HA14-2 adaptor I confirm that the angle is way less than 10 degrees indeed, in fact less than 9 degrees, and on the order of 4 degrees appears very plausible.
Input: round diameter 1.4 inch (35.56 mm)
Exit: round diameter 2 inch (50.8 mm)
Length: 47 mm
Angle: 9 degrees (per side)
Adding to it the 20 mm length of the round-to-rectangular TH4001 adaptor results in 67 mm for the total length of the composite adaptor. The impedance obtained using both the "stubby" 24 mm adaptor (dashed curve) and the new longer adaptor (solid curve) are shown in this plot.
As docali expected, the difference is substantial and the loading provided by the longer adaptor should now be closer to that provided by the JMLC design.
Also, the exit angle of the Radian 745NEO Be is not 10 degrees as specified on the spec sheet. Is is 4 degrees per side according to a Radian engineer that Lynn Olson spoke to a while back [cf. 2012-10-27 3:07 pm post in the "Beyond the Ariel" Thread]. Peering into the driver's exit through the HA14-2 adaptor I confirm that the angle is way less than 10 degrees indeed, in fact less than 9 degrees, and on the order of 4 degrees appears very plausible.
I can show a good example what "good" horn loading is doing for the drivers damping. One of my favourite drivers is the 18s ND4015Be. The following picture demonstrates the difference of free-air impedance to horn loading with a special adapter round 1.5" to 50x40mm TH4001 throat entry:
red curve is free-air.
The corresponding frequency response near mouth:
This is one of the top drivers in the market also with a shorting ring (very low inductance increase). And the phase plug design is optimal for such horn as it nearly provides an exit of only a few degrees (almost planar).
red curve is free-air.
The corresponding frequency response near mouth:
This is one of the top drivers in the market also with a shorting ring (very low inductance increase). And the phase plug design is optimal for such horn as it nearly provides an exit of only a few degrees (almost planar).
Loading isn't so much about constraining the diaphragm, it's more about radiation efficiency.
Agreed that the pp is very similar to ND4015Be. But to be honest most modern drivers follow a high flare rate pp and I would call a coherent pp more or less "older design" although I think that it fits perfectly for the kind of horns in this thread.
And the 1.5" exit is not really "modern" ;-)
And the 1.5" exit is not really "modern" ;-)
A good example indeed.
My own measurement below (JBL 2450J with Truextent Be dia + Arai A320FL horn):