Thank you, Vineeth. I would say it probably calls for DSP but otherwise it's just exemplary. I only wonder if we will ever be able to shape the low-frequency response in advance, based on the parameters of the driver. Here's a BMS 4550 in two different horns: https://www.diyaudio.com/community/...he-easy-way-ath4.338806/page-686#post-7446831 - based on that I would say it must be possible but honestly at the moment I don't have a clue how to do it. This interaction remains a mystery to me.
The low-frequency gain is still remarkable - around 520 Hz the absolute sensitivity is still about the same as for 10 kHz. I assume it can be around 100 dB SPL/2.83V@m, maybe even more. You should have no issue crossing even this small driver around 600 Hz, IMO.
BTW, looking at the construction of BMS 4550, I think it could benefit from a throat plug as well. The initial outlet from the phase plug seems conveniently small enough for this.
BTW, looking at the construction of BMS 4550, I think it could benefit from a throat plug as well. The initial outlet from the phase plug seems conveniently small enough for this.
Last edited:
Since the EXAR 400-25 is still quite close to what I used originally (at least comparing the throat impedance), here's a summary of what was measured so far (all at random SPL levels - not directly comparable, unfortunately):
Faital HF108:
Peerless DFM-2544:
Peerless DFM-2535:
Selenium D2500Ti-ND:
There's obviously some similarity between the BMS4550 and DFM-2544 regarding the shape of the LF response.
Faital HF108:
Peerless DFM-2544:
Peerless DFM-2535:
Selenium D2500Ti-ND:
There's obviously some similarity between the BMS4550 and DFM-2544 regarding the shape of the LF response.
Thanks @mabat, for the discussion and all the data you posted above.. 
It is interesting to see the low-end frequency response lift trends differing between different drivers..
I plan to cross this horn to a 15-inch Faital pro 15PR400 woofer in the region where there is a good enough directivity match and the distortion is still low.
I have a pair of DFM 2544s in hand. But I plan to try the SB audience Rosso 65CDN-T next with the 1-inch adaptor and check how good it is compared to the BMS.
Another interesting aspect is the beam width of this horn (somewhere around 70 degrees narrowing to 60ish degrees by 10kHz) is narrower above 1kHz than any other speakers I have. So it would be interesting to hear how it sounds in my room at the MLP which is only 2.5 to 3m away from the speakers.
In the frequency response data, there are small ripples spread throughout the frequency range probably indicating some sort of reflections from the horn interfering with the direct sound.. So another interesting aspect will be to see if it causes any audible issues at louder volumes..
It is interesting to see the low-end frequency response lift trends differing between different drivers..
I plan to cross this horn to a 15-inch Faital pro 15PR400 woofer in the region where there is a good enough directivity match and the distortion is still low.
I have a pair of DFM 2544s in hand. But I plan to try the SB audience Rosso 65CDN-T next with the 1-inch adaptor and check how good it is compared to the BMS.
Another interesting aspect is the beam width of this horn (somewhere around 70 degrees narrowing to 60ish degrees by 10kHz) is narrower above 1kHz than any other speakers I have. So it would be interesting to hear how it sounds in my room at the MLP which is only 2.5 to 3m away from the speakers.
In the frequency response data, there are small ripples spread throughout the frequency range probably indicating some sort of reflections from the horn interfering with the direct sound.. So another interesting aspect will be to see if it causes any audible issues at louder volumes..
From the data it looks like a reflection delayed 2 - 2.5 ms (~0.8 m) and attenuated at least by 25 dB compared to the direct sound. I don't think this could be a reflection within the reach of the horn, but who knows.
I can't wait for this. Actually, I have the 65CDN as well and I plan to test it with all the 1" and 1.4" waveguides but I'm just terribly slow...
But at least I have the petals for the first EXAR550 printed (probably my most accurate prints so far) -
Last edited:
Here is the data for SB audience Rosso 65CDN-T 1.4inch exit CD on the EXAR 400 with a 3D printed throat insert to make it a 1inch exit
In below plots, please ignore the absolute SPL levels. The seem low around 50 to 60dB SPL in plots since I had not done SPL calibration for the mic (ISEMcon EMX-7150). In reality they should be hanging around at least 80+ dB
Angular measurements upto 90 degrees
On axis response + impedance curve
VituixCAD plots
I just wish that break up was a bit milder.. (Obviously, not because I can hear it, but just so that it might look nicer to the eyes )
PS: The version of EXAR 400 I printed does not have holes for the screw insertion. So I had attached the driver to the horn for this measurement session using insulation tape..
In below plots, please ignore the absolute SPL levels. The seem low around 50 to 60dB SPL in plots since I had not done SPL calibration for the mic (ISEMcon EMX-7150). In reality they should be hanging around at least 80+ dB
Angular measurements upto 90 degrees
On axis response + impedance curve
VituixCAD plots
I just wish that break up was a bit milder..
(Obviously, not because I can hear it, but just so that it might look nicer to the eyes )PS: The version of EXAR 400 I printed does not have holes for the screw insertion. So I had attached the driver to the horn for this measurement session using insulation tape..
Last edited:
What do you mean it doesn't have mounting holes? Did I let out something like that?
Not searching for that.finally found the holy grail?
- But the goblet printed:
Last edited:
Looks like there is a little ditch between each section and the each sections seems a little convex - or is it just an illusion?What do you mean it doesn't have mounting holes? Did I let out something like that?
Not searching for that.
- But the goblet printed:
...
//
The walls are curved in this cross section, the rest is due light.
(All those little blobs are due to the 'nearest seam position' option of the slicer - I find it a lot easier to sand this off than to have an aligned seam somewhere.)
Here's the whole waveguide. There's a visible line where the petals are attached to the base.
(All those little blobs are due to the 'nearest seam position' option of the slicer - I find it a lot easier to sand this off than to have an aligned seam somewhere.)
Here's the whole waveguide. There's a visible line where the petals are attached to the base.
Last edited:
It has 10 segments, as it's easier to assemble it as two halves (as the final step). It only now occured to me that I can split it into two halves even with an odd number of segments... Maybe next time, I could try nine, or even alternate segments of different angular spans. But I doubt it can be noticeably better than the existing EXAR 400 presented above, which is already very smooth. It's almost like that in the listening window there's no diffraction at all.
BTW, is there a sound reason why an odd number of segments should be advantageous?
BTW, is there a sound reason why an odd number of segments should be advantageous?
Last edited:
Seeing also the individual responses would help here, as the actual acoustic crossover seems even higher. It must be the woofer that's narrowing here, since the waveguide itself has nothing but a smoothly rising DI with frequency. I only don't quite understand why would you want to cross it so high.
With a passive crossover, this can still be tricky. Anyway, it should be possible to do much better than this.
With a passive crossover, this can still be tricky. Anyway, it should be possible to do much better than this.
Last edited:
Your comments will be helpful, specially if you can comment on the low end quality/limits and if you can compare with other horns. I am also interested in the Rosso 65CDN-T, but even more in larger drivers like the old JBLs to go down as low as possible with the horn (400 or 550).
@swak: I am not an expert in this and @mabat knows better about the performance of these horns and what drivers can be used. But here is a performance comparison between my measurements of the Rosso 65CDN-T and the BMS 4550 driver on the EXAR 400 horn.
Measurement set up: The aim was to measure different parameters in such a way that both the drivers when attached to the horn are producing about the same SPL level around 1-2kHz region. So I adjusted the input to the drivers such that it resulted in 86dBSPL at 1m resulting in the following frequency response curves produced at 40cm away from the horn mouth.
Raw frequency responses
BMS 4550: Distortion plot
Rosso 65CDN-T: Distortion plot
Responses after drivers are EQd flat from 700Hz up (no additional high pass filter added)
BMS4550: Spectrogram once EQd flat
Rosso 65CDN-T: Spectrogram once EQd flat
Based on the above data, I think both drivers are good on this horn. They can be crossed over steeply around 600Hz I think. But the low end performance especially below 1kHz looks better on the Rosso 65 CDN-T. There is no distortion hump for it like the BMS. The top end performance above 13kHz looks marginally better on the BMS in the spectrogram since the breakup resonance seems a tiny bit more subdued comapred to the Rosso. But in actual listening, I doubt whether that will matter to some one. In fact people may even like it..
I dont know about any of the other drivers you mentioned.. Maybe others can help here..
Measurement set up: The aim was to measure different parameters in such a way that both the drivers when attached to the horn are producing about the same SPL level around 1-2kHz region. So I adjusted the input to the drivers such that it resulted in 86dBSPL at 1m resulting in the following frequency response curves produced at 40cm away from the horn mouth.
Raw frequency responses
BMS 4550: Distortion plot
Rosso 65CDN-T: Distortion plot
Responses after drivers are EQd flat from 700Hz up (no additional high pass filter added)
BMS4550: Spectrogram once EQd flat
Rosso 65CDN-T: Spectrogram once EQd flat
Based on the above data, I think both drivers are good on this horn. They can be crossed over steeply around 600Hz I think. But the low end performance especially below 1kHz looks better on the Rosso 65 CDN-T. There is no distortion hump for it like the BMS. The top end performance above 13kHz looks marginally better on the BMS in the spectrogram since the breakup resonance seems a tiny bit more subdued comapred to the Rosso. But in actual listening, I doubt whether that will matter to some one. In fact people may even like it..
I dont know about any of the other drivers you mentioned.. Maybe others can help here..
