Here's the hero -
It's a one-perimeter print with almost no infill and I accidentally left it on the sun - then it crumples immediately like this, so I resigned any proper finishing and only sprayed it a bit, mainly to close the surface.
It's a one-perimeter print with almost no infill and I accidentally left it on the sun - then it crumples immediately like this, so I resigned any proper finishing and only sprayed it a bit, mainly to close the surface.
Last edited:
Wow, FR curves on peerless betters the rosso.
Do share your comparison of the two with G2 waveguide.
Do share your comparison of the two with G2 waveguide.
Agreed. I'd like to elaborate because we can look at the two phenomena separately. I'm confident that my take on diffraction is correct, but there are holes in my knowledge when it comes to reflections back into the horn.
All diffraction and reflection-related issues manifest themselves mostly on-axis, particularly with an axisymmetric design. This is because all disturbances around the main axis of the horn reach the ear/microphone at the same time. For this reason, I'll focus on the main axis.
The dips at 5.5 kHz and its multiples (11 kHz and 16.5 kHz) are indicative of a comb filter effect. This effect is caused by destructive interference between the direct sound and the delayed, opposite polarity diffracted sound resulting from the sudden termination of the horn. Sound radiates from the throat of the horn straight to the microphone, having the shortest propagation delay. However, some sound first travels to the horn edge and then to the microphone. At the edge of the horn, the acoustic impedance suddenly drops, leading to a drop in sound pressure and a corresponding phase flip of the diffracted sound.
Pythagoras tells us that the diffracted sound travels a path that is approximately 65 mm longer than the direct sound. I simulated the effect of a delayed inverted copy of the direct sound in VCAD, with an arbitrary level for the diffracted sound of -10 dB:
Besides edge diffraction, a secondary effect of the sudden drop in acoustic impedance is that sound will reflect back into the waveguide, towards the throat. When a sound wave encounters a sudden increase in acoustic impedance (such as a wall), you get a positive phase reflection. Conversely, when sound encounters a sudden decrease in acoustic impedance, you get a negative phase reflection.
This negative phase reflection travels back towards the throat of the horn, where it reflects back towards the mouth. The reflected sound then travels to the microphone. The relative delay is much greater than that of the diffracted sound (the propagation delay is approximately twice the time it takes for sound to travel from the throat to the edge), and thus the resulting comb filter will start at a much lower frequency. In this particular horn it probably looks something like this:
To be honest, I haven't much looked into this phenomenon. I tend to focus on minimizing issues caused by diffraction, and when you've done that, reflection also isn't an issue.
The ABEC simulation does show peaks and dips at approximately the same frequencies as the VCAD sim. Therefore, you'd think these might indeed be caused by the reflection back into the horn. If that is the case, the effect appears to be less pronounced at higher frequencies. However, the ripples in the acoustic impedance graph resemble the VCAD sim irregularities quite well...
However, it doesn't really make sense to me. That's because at lower frequencies the horn provides less loading than at higher frequencies (check the throat impedance graph in the ATH report). Therefore the change in acoustic impedance is smaller at lower frequencies, and thus there will be more diffraction (waves wrapping around the waveguide) and less reflection.
Therefore, I think those peaks and dips at about 500 Hz, 800 Hz and 1500 Hz are not caused by reflection back into the horn, but by something else. I don't know what. Anybody else?
Are you questioning the purpose of your work on ATH? In my view, when we're designing a horn in ATH, the thing most of us are chasing is smoother response and smoother directivity. But those are linear effects. What kind of crazy distortion are you referring to?
I have in fact built and listened to simple conical waveguides without termination. They're very easy to make and they sound like ****. As with anything, EQ surely helps, but you can't get rid of all coloration, let alone the variance in response when you move around.
I have however only listened to a single such horn, not a pair. So I can't really say anything about imaging.
Attachments
This is the bigger Peerless (DFM-2544), A520G2/25-STD-1.
BMS 4554-16, A520G2/36-STD-1
(All the absolute SPLs in these mesurements are NOT directly comparable between the different drivers.)
(This one has a slightly smaller exit than 36mm...)
(All the absolute SPLs in these mesurements are NOT directly comparable between the different drivers.)
(This one has a slightly smaller exit than 36mm...)
To compare two different FR is
//
Would it be hard to compare to just mate the 520 horn on the CD 1,4" native interface i.e. without the exit modifier?
//
Some amazing results here. However the big question then becomes a subjective one. In my tests the difference in 'tone' between these combinations is very large - we are not talking subtleties here.
Since the number of combinations is near infinite I wonder if there is a semi-rigorous method of combining everyone's horn/cd/eq experiences in a spreadsheet? Patterns might emerge - and then maybe further tests can be devised.
Since the number of combinations is near infinite I wonder if there is a semi-rigorous method of combining everyone's horn/cd/eq experiences in a spreadsheet? Patterns might emerge - and then maybe further tests can be devised.
Or buy DSP, the waveguide works almost the same with any driver, except top octave I think, perhaps with adapters even that gets quite close between drivers, directivity wise. Equalizer would then be able to equalize the drivers responses to match each other, which means also polars would be the same for ~any combination, with possible exception on top octave. If there is still audible difference, you'd be probably looking at distortion plots or something.
This is the right application to use EQ, rarely is directivity so smooth that EQ fixes the response to all directions, but this is the case.
This is the right application to use EQ, rarely is directivity so smooth that EQ fixes the response to all directions, but this is the case.
Last edited:
Even when I've taken enormous lengths to eq combinations the same (sometimes even using audiolense .fir filters) the texture and tone of different combinations can vary wildly when for example listening to solo violin. Titanium diaphragms can be detected from the next room.
Top: ROSSO 65CDN-T in the original state + T520-36-STD-1, but I had to take the other piece from the pair I have, otherwise it would be too laborous at the moment to make the joint smooth enough.
Bottom: T520-ROSSO-STD adapter
Bottom: T520-ROSSO-STD adapter
Faital HF1440, A520G2/T520-36-STD-1
It's a strange driver. I only wonder if this is the current production line. I bought them quite early.
But of course with a little active help, it's otherwise remarkably smooth and clean.
BTW, here's the same driver (and a 65CDN-T) in EXAR550: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7645128
It's a strange driver. I only wonder if this is the current production line. I bought them quite early.
But of course with a little active help, it's otherwise remarkably smooth and clean.
BTW, here's the same driver (and a 65CDN-T) in EXAR550: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7645128
Last edited:
My favourite, 1" Selenium D2500TiNd, A520G2/T520-25-STD-1
And with T520-36-STD-1 + the extension used originally here: http://at-horns.eu/exar-story.html#d2500
(The polar angles obviously don't match, probably my mistake. The coverage patterns should be pretty much the same at least up to ~10k).
I yet need to print the extended adapters.
And with T520-36-STD-1 + the extension used originally here: http://at-horns.eu/exar-story.html#d2500
(The polar angles obviously don't match, probably my mistake. The coverage patterns should be pretty much the same at least up to ~10k).
I yet need to print the extended adapters.
Last edited:
Isn't that just the driver, not (the good) waveguide? so just matter of finding a driver that seems to sound nice, not combination of drivers and waveguides? If there is no good drivers, perhaps try with dome and good waveguide.
Short story about recent observation, kinda related:
Yesterday I was on a party, DJ playing quite loud, the PA had RCF ART712 tops, specsheet attached. Peculiar thing was that anywhere off-axis the sound was really nice and quite balanced, nice sound overall, but closer to on-axis the sound turned unbearable, very different nasty sound. From specsheet, the speaker has reasonably smooth directivity for the whole high mids. Thinking about it I could come up few possibilities what might contribute to on-axis nasty sound:
1. some hearing threshold, few more db highs moving from off-axis to on-axis made my own hearing system distort
2. the waveguide has no mouth roundover and the throat might have some issues as well as it's oval by looks so there is at least some problems with diffraction
3. since off-axis sound was nice, no hint of distortion as such, I'm not sure if it was driver distortion since relatively good specsheet directivity a distorting driver would make bad sound all around, right?
4. perhaps the woofer distorted instead, beamed the distortion
And now that I'm thinking this stuff I just write it here, I'm really tired so help me out 😀 could diffraction make existing driver non-linear distortion, HD, that could be quite inaudible itself, standout and make it sound nastier?
Simple thought experiment, please comment if it makes sense or not: if there was 2kHz tone and it's harmonics 4, 6 and 8kHz produced in the driver. These would have wavelengths of about 17cm, 8.5cm, 5.6cm and 4.2cm or so. Now if our diffraction problem was at 4cm distance (secondary sound 4cm after direct sound), it would boost the ~8cm long 2nd harmonic and cut the ~4cm fourth harmonic, right, being opposite phase. This would alter the amplitudes of original harmonic series (the driver distortion). Now take any other series and see these affect differently to the harmonics, also spatially, because the diffracting structure is static in size and shape, while wavelength varies a lot.
If I'm thinkin it right basically any driver distortion products would have constructive or desctructive interference from diffraction related secondary sounds. Since this is not music related scramble, but static physical size/shape inflicting on top of it, really magnified on-axis and making the scramble standout over the music. It could be, right? Actually, same goes to the original sound harmonics, like violin.
If it is, it means driver distortion is important, but also diffraction is, making driver distortion more audible. Good very low distortion drivers might sound much better on bad waveguide, but on a very good low diffraction waveguide differences aren't likely as stark, right, at least on low SPL while driver distortions are lower in general. Perhaps throat angle mismatch and all this kindof stuff really show up only when the driver distorts.
What you guys think?
ps. I took short videoclip with mobile of this effect, and it hears through. I try to upload it later if anyone is interested.
Last edited:
This one: https://www.wagneronline.com.au/tym.../audio-speakers-pa/dfm-2535-76697/1005441/pd/
There are still shops claiming a non-zero stock but when you contact them, they tell you it's all gone and unavailable. I would buy tens of them if there was a chance.
Anyway, those who have the drivers now also have a perfect waveguide for them 😉
There are still shops claiming a non-zero stock but when you contact them, they tell you it's all gone and unavailable. I would buy tens of them if there was a chance.
Anyway, those who have the drivers now also have a perfect waveguide for them 😉
@mabat have you also tried contact these?
https://www.ariat-tech.com/parts/peerless-by-tymphany/DFM-2535R00-08
- 2682 pcs new original in stock
I would take 2 if they were available.
BTW, approx. half year ago Peerless "announced" a new driver on their webpage:
https://www.ariat-tech.com/parts/peerless-by-tymphany/DFM-2535R00-08
- 2682 pcs new original in stock
I would take 2 if they were available.
BTW, approx. half year ago Peerless "announced" a new driver on their webpage:
Source: https://peerless-audio.com/products/
DN2035 High-frequency driver
- Outside diameter of just 55mm
- Neodymium magnet
- High output vs. size-cost ratio
- High sensitivity and low crossover point
- 1.4” VC 0.75” exit
Last edited:
I'll try an adapter specifically made for the DFM-2535 (21.7° exit angle). I wonder how much better it can be...
The 25-STD-1 adapter used for the previous measurement has 12°.
The 25-STD-1 adapter used for the previous measurement has 12°.