Oops didn't realize the scale was chopped off in the screenshot. 2dB step, so roughly a 15dB difference. For that particular example, I had selected an averaged frequency range of 1.6-2khz. But I found it was pretty consistently that pattern until you got up to the top octave.
Here Danibosn presented a script intended for the ESS AMT-1. I have very little experience with ATH, but I generated the mesh with that script and realized that the geometry at the throat would not really fit the AMT.
Is there a way to start the horn (at the throat) with an arbitrary angle on the sides of the rectangular throat?
To make it more clear, here is a pic of the AMT-1. The horn would start where the metal part of the "V-shaped" contour ends and would need to start with that same angle. The newer version of the AMT-1 (pic #2) looks a little different, but the horn would be the same.
Sure, just delete the first horn section from the script (Horn.Part:1), which is only a segment of a straight duct in that example (a = 0).
Then in the section that remains (you can rename it to Horn.Part:1, but it doesn't really matter), set 'a' to the desired values. This is set independently for H and for V, which are arbitrary OSSE profiles.
Then in the section that remains (you can rename it to Horn.Part:1, but it doesn't really matter), set 'a' to the desired values. This is set independently for H and for V, which are arbitrary OSSE profiles.
Last edited:
Got it, thank you.
I don't have the exact geometry of the drivers, I bought a pair but I have not received them yet.
For simulation purposes, one would need two sections (horizontally) , the first one following the waveguide of the driver itself and a second one as a continuation of that.
This would be nice, but I have to learn much more how the script works. Do you have a template perhaps of something similar?
I don't have the exact geometry of the drivers, I bought a pair but I have not received them yet.
For simulation purposes, one would need two sections (horizontally) , the first one following the waveguide of the driver itself and a second one as a continuation of that.
Last edited:
@mabat - I know you found the alignment pinholes a PITA, but is there any chance you can still have them included as an alternative assembly method in your current download? I have the earlier download version that included the alignment pin method (#15,697), but the base/mouth modifications had not yet been included (ie. post #15,731).
I had less success with the spline method for variety of reasons:
- I went with CA glue for assembly. This was an excellent choice until you get to two halves - the CA glue has too short of an assembly time and does not possess the best filler properties
- Printing the petal with the cavity resulted in slight wrapping (I'm sure partially due to my open printer) at the edges - so the petals to don't perfectly match together.
A340G2 and 13cm midwoofer close to the mouth, port firing, more parallel to the waveguide axis. As predicted by Hornresp resonance is smaller (blue)
Below is tweeter 2535 green (not in scale) and two types of port: slit expanding blue and wide open red:
Below is tweeter 2535 green (not in scale) and two types of port: slit expanding blue and wide open red:
Last edited:
I manufacture bass cabs. The goal posts are in a very different place for electric bass than for home audio. A bass cab's typically non-flat frequency response hopefully allows it to "sit in the mix well" with the other instruments: Most bass cabs don't attempt to deliver the lowest fundamentals because that "steps on" the kick drum. And most bass cabs are not loud in the same region as guitar cabs because that instigates a volume war with the guitarist(s), one which the guitarist(s) inevitably win. Guitar cabs tend to be quite loud in the octave or so above 2 kHz (ballpark), so many bass cabs are loud in the octave or so below 2 kHz (ballpark). Also some bass cabs have a horn tweeter which gives them a bit of top-end energy particularly north of the octave or so where the guitar cabs are loudest, so north of 4 or 5 kHz or so. The name of the game is "sit in the mix well", and "make the other musicians sound great". (For those of you familiar with American football, bass players are like the offensive linemen: You may not know their names, but they make everyone else look good.)
For a solo bass player, like Michael Manring, something much more like a "hifi PA rig" would be used because the entire spectrum belongs to him. Also, the goal posts for amplified double-bass are pretty close to those for home hifi, as the intention is usually "sounds like my instrument, only louder."
Personally I would be skeptical of using a cluster of four 10" woofers up to 800 Hz. The radiation pattern looks MUCH better (to me) from a single 15" woofer (unless the 10's are splayed horizontally, and I'm not aware of anybody doing that in a commercial bass cab).
Here is a 15" bass guitar speaker that imo could be used for home audio with an 800 Hz crossover, but I don't actually know who is using it in a commercial product:
https://eminence.com/collections/15-inch-bass-guitar-speakers/products/legend_cb158#specifications
Last edited:
That's interesting. What makes you think that a 2x2 matrix of 10" woofers would have a much worse radiation pattern than a single 15"? Up to a frequency where lobing starts (which would be higher than the intended use) these two should be actually pretty close, the 2x2 matrix almost having an edge, IMO.
The idea is to use a slower and gap filling glue and assemble the full petal circle before the glue dries, so all the accuracy imperfections have the chance to distribute evenly. For this the splines are good because they can hold the petals together (if the fit is good). I'm afraid this would be difficult with just the pins/pinholes.
It surprises me that you have a struggle printing the spline version. But if that's so, what can you do.
Last edited:
It's definitely a user error 🙂. I have to use slower-curing glue and your idea about assembling the full petal so imperfections have a chance to distribute evenly. The spline method is great - and I got near-perfect 4/6 joints. It's just when I tried to glue to two halves - that's where the glue on the two sides cured before I could force the parts together.
I'm going to try the spline method again.
It seems like you use a pretty fast glue. That's always a risk. Is it not possible to cut it and do it again?
In home audio I like to design with a lot of toe-in in mind (perhaps 45 degrees) such that the axes criss-cross in front of the listener, as advocated by Earl Geddes. So the designed-for listening axis is actually somewhere in the 15-25 degree range (Earl optimizes for 22.5 degrees as I recall).
This set-up geometry has advantages for off-centerline listening locations. You're probably well aware, but to recap for anyone new to "time/intensity trading", with this set-up geometry as you move off to one side of the listening area's centerline you are moving increasingly off-axis of the near speaker (whose sound arrives earlier), and increasingly on-axis of the far speaker (whose sound will be louder PROVIDED that the near speaker's response falls off smoothly and fairly rapidly as we move off-axis horizontally). The idea is that the earlier arrival time of the near speaker is offset by the greater intensity (loudness) from the far speaker in the region where the ear gets most of its localization cues from. The net result is considerably better spatial quality, as well as better sound quality, for off-centerline listners than would normally be the case. So you get a much wider "usable sweet spot" area, spatial quality still being best in the middle of course.
The issue with the 2x2 matrix of 10" woofers is the width of the central lobe. At 800 Hz, the central lobe is -6 dB at about 28 degrees off-axis, and -20 dB at about 45 degrees off-axis. A listener off to one side of the centerline is likely to be somewhere within that range, and so there will be a dip in and a bit below the crossover region in the frequency response of the first-arrival sound from the near speaker, which MIGHT be enough to throw off either the sound quality or the spatial quality. I welcome correction from anyone with more insight into the perceptual implications of this dippage in the first-arrival sound from the near speaker.
I didn't mention this previously, and this is just speculation for the 800 Hz region, but might there be some degradation of clarity from multiple very early arrival times due to the slight differences in path lengths of the four cones? I'm confident that clarity would be degraded if this was happening further up the spectrum, but not sure about the 800 Hz region. Again, I welcome insight from those who know more about this.
In contrast, a single 15" woofer has no lobing at 800 Hz and is about -3 dB at 45 degrees off-axis, so there would be no crossover-region dippage for off-centerline listeners.
(The situation is arguably even worse for the audience of a bass player playing through a traditional 410 cab. The lobing gets progressively worse as we go up in frequency, so only the few who are close to on-axis are getting direct sound free from lobing nulls at various frequencies. Of course only the occasional audiophile in the audience is likely to critique the sound quality of the electric bass at a live performance,. but a lot of bass players actually do care whether the audience gets to clearly hear what they are doing.)
Last edited:
I guess I could, but I'll try reprinting again. This time, just so that I can practice gluing - I'm going to print with limited infill and 2 layers of walls. I've gone through more than a couple of boxes of filament over the last few days 🙂.
I finally got a printer as well and currently printing first project the A460G2. 2 wall layers and 5% gyroid infil makes surprisingly durable and stiff petals, using PETG. Can't find soft spots with finger and it's stiff if tryign to bend. Dropped them to floor few times and no damage. Still about 80g of filament per petal.
It's quite much work still, more than one roll of filament per device, few hours of work with the machine, surface finishing and all, so thicker walls and perhaps more infil would be worth it to make complete device survive a drop test. For prototyping, 2wall layers and 5% infill is plenty 🙂
It's quite much work still, more than one roll of filament per device, few hours of work with the machine, surface finishing and all, so thicker walls and perhaps more infil would be worth it to make complete device survive a drop test. For prototyping, 2wall layers and 5% infill is plenty 🙂
Yea, I can imagine PETG being fairly good.
I would need to check, but I think that around 800 Hz there still must be basically only a beamwidth narrowing, i.e. no strong lobes, just a narrowed width of the main beam (which is what we want; otherwise it wouldn't be usable, of course). (?) And because it's not so much larger than a single 15", I would not expect a large difference overall. It would be close to a single 18", and those are not big differences. Certainly we could do with a single 18" in terms of midrange directivity. The bigger waveguides can be used down do 600 - 700 Hz without any issues.
Last edited:
I noticed a few complaints that there's not enough space for accessing the mounting bolts in some of the kits. I don't know where I got to this but I have a short Allen key and it works perfectly. So I would suggest just to obtain one 🙂
In some cases it's possible to use threaded rods instead of bolts and then the access to a nut with a wrench from the side is much more straightforward.
In some cases it's possible to use threaded rods instead of bolts and then the access to a nut with a wrench from the side is much more straightforward.