Good read. In certain situations it can be a problem as reported. There is also a mention of clearer sound output after taking measures. This is with very high SPL, unlikely in domestic rooms. Normally, a CD has a quite stiff suspension in comparison with cone speakers so I suspect that most PA systems do not experience this phenomenon as destructive, but it may affect sound quality at high SPL if not adressed. Also, it is an anecdote (one datapoint as some say) with specific and vulnerable drivers, and I wonder if there are more such stories or not. So it appears to be not only a theoretical possibility, but afaik not a much mentioned issue.
Last edited:
There are clear measured differences in impedance curves for any level, even for very low SPLs, actually. I still don't see the reasoning here. It will be all linear (expect the perception, perhaps, as @gedlee has documented).
We're indeed not talking about a point when it's destructive but when (or if) this can become an audible issue. But without actual listening tests, I don't think we can find out anything 🙂 We probably can't rule out even the possibility that it's actually beneficial... 🙂
We're indeed not talking about a point when it's destructive but when (or if) this can become an audible issue. But without actual listening tests, I don't think we can find out anything 🙂 We probably can't rule out even the possibility that it's actually beneficial... 🙂
Last edited:
Any data available for the old Altec drivers ? I would like to print one of the bigger horns for 802D (1") and 288 (1.4"),
@mabat I’m not sure if this quick and dirty test will be useful — I was just curious, and I haven’t had time to print all the parts yet (not sure when I will).
So far, I only printed the central part of the ESP plug. (Since I’m using a 1.5-inch adapter, I scaled it down slightly, and the fit at both ends is pretty decent).
For the test, I simply placed the driver on the floor, then the ESP on top, then the waveguide. It’s not a sealed setup and might shift slightly, but I repeated the positioning a few times and got similar results.
The following images compare a short adapter to the G2 and the ESP with at roughly 5°, 20-25°, and 45° angles (mic positioning wasn’t perfectly consistent). 10 cycles window:
I could later print the remaining parts of the adapter to make a more solid connection and measure the horn in a better room position. But honestly, I don’t think the results would change dramatically — the main suspect is likely already clear: the driver operating in a non-pistonic mode?
So far, I only printed the central part of the ESP plug. (Since I’m using a 1.5-inch adapter, I scaled it down slightly, and the fit at both ends is pretty decent).
For the test, I simply placed the driver on the floor, then the ESP on top, then the waveguide. It’s not a sealed setup and might shift slightly, but I repeated the positioning a few times and got similar results.
The following images compare a short adapter to the G2 and the ESP with at roughly 5°, 20-25°, and 45° angles (mic positioning wasn’t perfectly consistent). 10 cycles window:
I could later print the remaining parts of the adapter to make a more solid connection and measure the horn in a better room position. But honestly, I don’t think the results would change dramatically — the main suspect is likely already clear: the driver operating in a non-pistonic mode?
So, if I get it right, you scaled the 2" ESP to 1.5" and you used it with a regular 1.5" adapter and a G2 horn?
This doesn't have to work properly even with a perfect driver (I would need to check), but yes, in general any wavefront "non-flatness" is a real issue.
What's the driver?
This doesn't have to work properly even with a perfect driver (I would need to check), but yes, in general any wavefront "non-flatness" is a real issue.
What's the driver?
Ah, I completely forgot there's already the ESP-5038-00 (1.5").
OK, around 10 kHz can be where 1.5" drivers start to have problems... Pretty plausible then.
How does it sound? 🙂
OK, around 10 kHz can be where 1.5" drivers start to have problems... Pretty plausible then.
How does it sound? 🙂
Like a sweep pointed at the ceiling 🙂 That's all I have so far (and I also don't have the patience))))
Try the updated version, hopefully it still works as it should: https://www.at-horns.eu/
Last edited:
Extracted mabat's curves for the ND3T from post 17,467 and overlayed the A520G2 on top of the A400G2 to see the low end differences.
The 5ms gate on the measurements removes some of the low end information though.
The 5ms gate on the measurements removes some of the low end information though.
This is super duper cool. I've never seen this before, but from a theory perspective, this makes a lot of sense. The horn is an impedance matching device between the diaphragm and the air (in both directions). If we had a ideal horn (perfect impedance transformer), they would have 100% power transfer in both electrical to acoustic (loudspeaker) and acoustic to electrical (microphone). Passive networks are reciprocal, so you can't achieve high efficiency in one direction without also having efficiency in the other direction.I've ever seen this only with longish-throat ("high loading") horns. Never in a rapidly opening waveguide -
The reason direct radiators and shallow waveguides don't exhibit this is because there's a relatively large impedance mismatch between the diaphragm and the air, resulting in a high reflection coefficient (poor power transfer) in both directions (low efficiency as a loudspeaker, but also poor efficiency as a microphone).
In a long-throat horn, the diaphragm is highly loaded and closer to being impedance matched, so I think the only conclusion is that your CD is also acting as a microphone.
I'm not sure if this behavior is necessarily a bad thing. Assuming the air in the room is pretty linear, it shouldn't result in much nonlinear distortion.
This is a linear distortion in the first place, I wasn't thinking about anything nonlinear.
I'm curious what you'd expect the effect to be at low excursion/SPL and why it might be detrimental. I'll prefix this by saying that I'm far from an expert so I could easily be wrong, but I don't see how it could be a problem. As @amhenikoff points out, a horn is an impedance transformer and the better the loading is, the more power is transferred in both directions.
So, what happens to the incident acoustic energy with a short horn/waveguide that has poor loading? Most is reflected back into the room due to the abrupt change in acoustic impedance presented by the device. With better loading, less acoustic energy is reflected by the horn itself so more reaches the diaphragm. For this fraction, the acoustic impedance at the surface of the diaphragm determines how much energy is reflected and how much is absorbed. Since the electrical source impedance seen by the voice coil influences the acoustic impedance at the diaphragm surface, there should exist "optimal" electrical source impedances which minimize or maximize the fraction which is reflected.
The sound waves coming from the outside and reaching the diaphragm have a clear impact on the electrical impedance. So far we know this. To me this means they must be "shaking" the diaphragm, i.e. cause it to move differently than it would otherwise move for a given input signal. This sounds like a distortion mechanism to me - a kind of modulation, I don't know how to call it better.
^Agree. An amplifier with low output impedance will try to keep the membrane from moving. A low value parallell resistor between the driver terminals will be helpful for passive crossovers, or a inductor (as Danley has done)