Looks like the new reference build for foam core enclosures.
Everything about 'foamcore' + this geometry seems proper. If you do another horn, maybe round-over the mouth 180deg for proper termination and do 5 petals so there are no parallel walls. Anyway, this looks awesome...The most tempting foam core project I've seen from a fidelity and fitness of materials standpoint.
and yes, I'm late to this thread
Everything about 'foamcore' + this geometry seems proper. If you do another horn, maybe round-over the mouth 180deg for proper termination and do 5 petals so there are no parallel walls. Anyway, this looks awesome...The most tempting foam core project I've seen from a fidelity and fitness of materials standpoint.
and yes, I'm late to this thread
Last edited:
Thanks Anthony. Something very cool about the shape and form that I also really like about it. The 5 petal horn is a good idea - does it help to reduce modes along the length?
I like the fact that one can build a 'reference' speaker that has a frequency response that is flat as a board with low THD with a $10 driver and $1 of foam core.
This one is definitely a keeper!
I like the fact that one can build a 'reference' speaker that has a frequency response that is flat as a board with low THD with a $10 driver and $1 of foam core.
This one is definitely a keeper!
IG,
I have some near cone measurements (1 in from cone). My setup uses a Panasonic WM61A capsule and the built-in mic input on the laptop which was designed for video conferencing etc so I think it has a low pass filter at around 5kHz. I am pretty sure the speaker has plenty above this sharp falloff. What I can measure though is very flat - as expected from a cochlear spiral with stuffing. There appears to be NO reflections whatsoever and the response is amazingly flat. The bump at 30 Hz is the hum from the HVAC. The drop near 330 Hz probably corresponds to the closed-ended TL length of 36 inches (1-Lambda of 312 Hz)? Look at how clean the impulse response is, and the low THD of less than -40dB, and the nice -20 to 20 deg linear variation in phase.
I wonder what that sudden dropout is? I've never seen this on a near-cone plot. Even for a far-field response, it'd be way too sharp and of too great magnitude to be baffle-setp. Hold-on, is that response with the horn on?
IG
Thanks Anthony. Something very cool about the shape and form that I also really like about it. The 5 petal horn is a good idea - does it help to reduce modes along the length?
I like the fact that one can build a 'reference' speaker that has a frequency response that is flat as a board with low THD with a $10 driver and $1 of foam core.
This one is definitely a keeper!
Yeah, 11 dollar reference kit is amazing. Since most people value items by price tag or how cool the brand/idea makes them to their friends, this design probably needs the underdog card attached to get adoptance... Name suggestions anyone? Off hand I'll submit:
The "Notaloss"
or "Notalot"
Re: mouth roundovers, check this analysis between conicals and le'cleac'h flare:
LeCleac'h horn vs Conical Horn
Re: horn petals, I can't find the graphic now and it's only partially related, but I've seen measured internal interference mapping of a conical and it seems that a 5th wall would break the modes up significantly. Some notorious horn builders in Germany experimented with large mid bass horns of petal counts between 4 and 13 and found that 11 (if I remember correctly) was the optimal number. An even numbers of walls sets up parallel reflective surfaces so I think at a minimum 5 is appropriate when trying to get a hi-fi response and 7 is probably the best balance for directivity (between 4 and inf.) given the amount of time it takes to cut petals out.
IG,
I have been scratching my head on this one too. perhaps it is the fortuitous confluence of baffle step (for a 5 in wide baffle is 337 Hz) and the TL-tuning frequency? This is measured without the horn. Regarding the TL-freq, I thought an end-sealed TL freq is set by wavelength/2 but it appears to be just 1-wavelength?
I have been scratching my head on this one too. perhaps it is the fortuitous confluence of baffle step (for a 5 in wide baffle is 337 Hz) and the TL-tuning frequency? This is measured without the horn. Regarding the TL-freq, I thought an end-sealed TL freq is set by wavelength/2 but it appears to be just 1-wavelength?
Interesting, 11 is the number of petals I chose for the horn I posted. I suspect that by this point, it's becoming close enough to circular that the differences should be small.
Baffle-step for a 5" (0.127m) baffle would be 115/0.127=~900Hz and the magnitude pretty much always is the expected 6dB. The step tends not to show up much if the mic is significantly closer than one baffle width. A closed pipe does tune to half a wavelength, but I don't know if tapering changes anything as it does on open quarter-wave TL's, where S0>SL tunes lower.
IG
I am thinking that the cochlear gradual taper and sound its absorbent Wood's horn or 'black hole' like absorbent qualities will make it perform not like a sealed TL, but more like a semi-infinite pipe. Thus it should have qualities similar to a wall mounted speaker between two large rooms. Still trying to figure out what cause the first corner at 300 Hz though. Is it gain to the right or loss to the left? If it is a loss, could it be a (dull - broad) reflection cancellation?
The horn really improves the intelligibility of human voices and I think might be a good candidate for a HT center channel. I just realized that I now have a full set of FC speakers that can make a nice HT setup. Not that I have a HT amplifier or tv to go with it
Use the two 0.53x scale Karlsonators for the front L R. The Nautaloss with tractrix for C. The 0.4x scale for the rear L and R.
Use the two 0.53x scale Karlsonators for the front L R. The Nautaloss with tractrix for C. The 0.4x scale for the rear L and R.
Room Effects
I am pretty sure the frequency shelf near 300 Hz is a room-effect. I measured the frequency response again at 12 in away and got the same drop-off in frequency response when I mounted the driver on a piece of 24x36 in cardboard in an open baffle. I also used an old driver to make sure it wasn't a breaking-in effect which it is not. So the effect is clearly not a function of the cabinet design. The good news is that the Nautaloss speaker has an even flatter response than an open baffle - confirming my suspicion that the damped spiral absorbs the back wave like an infinite space.
So what we have here is a compact speaker that has a response like the front half of an OB without the huge OB size.
1/6th octave smoothing applied. Red is Nautaloss, Green is OB, Blue is OB with broken-in driver.
I am pretty sure the frequency shelf near 300 Hz is a room-effect. I measured the frequency response again at 12 in away and got the same drop-off in frequency response when I mounted the driver on a piece of 24x36 in cardboard in an open baffle. I also used an old driver to make sure it wasn't a breaking-in effect which it is not. So the effect is clearly not a function of the cabinet design. The good news is that the Nautaloss speaker has an even flatter response than an open baffle - confirming my suspicion that the damped spiral absorbs the back wave like an infinite space.
So what we have here is a compact speaker that has a response like the front half of an OB without the huge OB size.
1/6th octave smoothing applied. Red is Nautaloss, Green is OB, Blue is OB with broken-in driver.
Attachments
Last edited:
What part of the Pluto writeup are you referring to? I am looking for confirmation of the drop at 300 Hz due to floor or wall bounce cancellation.
IG,
the dip happens to combine fortuitously with a sharp edge in the response that is predicted in the simulation. The dip is just to the left of it, and the response is already falling off at -6dB/oct due to baffle step, so what you see is a notch to the left side of the 300 Hz 15 dB cliff. In otherwords, a well placed dip in an asymmetric baseline coupled with a sharp cliff on the right. That is my theory anyhow. An easier way is to change the speaker height or room location when measuring. I don't have another driver.
Here is a simplified model simulation of the Nautaloss without room reflections:
Here is the sim with reflections:
Here is the raw un-smoothed data for the response:
the dip happens to combine fortuitously with a sharp edge in the response that is predicted in the simulation. The dip is just to the left of it, and the response is already falling off at -6dB/oct due to baffle step, so what you see is a notch to the left side of the 300 Hz 15 dB cliff. In otherwords, a well placed dip in an asymmetric baseline coupled with a sharp cliff on the right. That is my theory anyhow. An easier way is to change the speaker height or room location when measuring. I don't have another driver.
Here is a simplified model simulation of the Nautaloss without room reflections:
Here is the sim with reflections:
Here is the raw un-smoothed data for the response:
Attachments
The horn really improves the intelligibility of human voices and I think might be a good candidate for a HT center channel. I just realized that I now have a full set of FC speakers that can make a nice HT setup. Not that I have a HT amplifier or tv to go with it
Use the two 0.53x scale Karlsonators for the front L R. The Nautaloss with tractrix for C. The 0.4x scale for the rear L and R.
I was replying to this post!
The horn really improves the intelligibility of human voices and I think might be a good candidate for a HT center channel. I just realized that I now have a full set of FC speakers that can make a nice HT setup. Not that I have a HT amplifier or tv to go with it
Ha! I don't have HT amp or TV either, but I'm good with that.
Especially with 7 and 10 y.o.s in the house. Someday maybe.Now back to your regularly scheduled programming...