Isn't a special thread a good place to express interest? Here it will slip through the cracks quicky, IMO.
I'm not sure if it will work or not.
It's just not clear to me what's limiting the driver's ability to cross low, even with a nice horn/WG like yours @mabat.
Is it the limited x-max that increasing the distortion below 800Hz, or is because of Fs (~600Hz on my sample)... my suspicion is the limited x-max because distortion jumps not at Fs but at 500Hz...
Being able to use a high pass filter for acoustic LR4@ 800Hz would be nice...
It's just not clear to me what's limiting the driver's ability to cross low, even with a nice horn/WG like yours @mabat.
Is it the limited x-max that increasing the distortion below 800Hz, or is because of Fs (~600Hz on my sample)... my suspicion is the limited x-max because distortion jumps not at Fs but at 500Hz...
Being able to use a high pass filter for acoustic LR4@ 800Hz would be nice...
Last edited:
I think the limit will always be the excursion, but nobody will tell you what excursion is needed for a given SPL in a particular horn. The only way I know how to get an idea about how low a driver is capable to go is where the distortion - all the harmonics - start to rise rapidly for a given setup and SPL requirements. You don't have to think much about Fs per se, IMO.
Last edited:
tktran303's measurements of dfm2335 measured on a small horn SB Audience H225, 1 inch entry, 225 x 225 mm:
- https://www.diyaudio.com/community/...constant-directivity-horn.405050/post-7714907
Have you tried a larger horn?
- https://www.diyaudio.com/community/...constant-directivity-horn.405050/post-7714907
Have you tried a larger horn?
Last edited:
Still very new to Ath, but I've always wanted a half decent portable Bluetooth speaker and figured it would be a good project to learn with. There's still some issues to work out with the prototype but overall I'm quite happy with it. Because the horn cutoff was about 1100hz, I sized the woofer taps accordingly to match that cutoff with an acoustical crossover allowing for an 80hz LPF to flatten the 6.5" drivers response. Using a DH350 compression driver paired with 2 6NDL38s and 2 SB Acoustics passive radiators. Cheers mabat for the development.
At DigiKey, the minimum order is 120 pieces at $42 (€38.60) . That is a $200 smaller investment for an additional 10 pairs.
I suppose it would depend on which side of the pond the majority of interested buyers live. to minimize shipping costs.
I'll look for a GB thread.
Bill
Speaking of the smaller Peerless CD, this is measured with a longer shaping plug:
There's still some issue (loss of output) above 10 kHz, not predicted by the simulation, I suspect some mechanical problems.
... Ah, now I see the problem -
This is not good, something is obviously wrong / not as predicted.
So much for the shaping plugs, I think I'm going to leave it for another while...
There's still some issue (loss of output) above 10 kHz, not predicted by the simulation, I suspect some mechanical problems.
... Ah, now I see the problem -
This is not good, something is obviously wrong / not as predicted.
So much for the shaping plugs, I think I'm going to leave it for another while...
Last edited:
It may be that the two channels have some internal loss, each different. Next time I can try to insert some damping into one or the other. I'm trying to simulate it now, by giving different weights to the individual channel sources, and I can see a similar pattern. That could still save the day. Not holding my breath though.
Also, the vane may not be perfectly air-tight. It would need to be immersed into a thin epoxy or something, I guess.
Funny thing is that it sounds still good.
Also, the vane may not be perfectly air-tight. It would need to be immersed into a thin epoxy or something, I guess.
Funny thing is that it sounds still good.
Last edited:
Stuff in the last octave probably doesn't audibly impact musicality all that much, I imagine. But, a straight line looks much better than a droop or wiggle, of course. 😉
This is the longer shaping plug -
It has an exponentially increasing cross section area. The low-frequency gain is not negligible.
The BEM sim says the directivity can be the same as of the shorter one, provided that everything works as expected 🙂
It has an exponentially increasing cross section area. The low-frequency gain is not negligible.
The BEM sim says the directivity can be the same as of the shorter one, provided that everything works as expected 🙂
Last edited:
Thinking about it, the central-thorn version may be fundamentally better, since it is basically made of only one channel - no summing problems.
No. @mabat 2nd gen horns next.
I think he’s right. Distortion jumps up at 500Hz, not 600Hz. So i think it’s excursion, and unrelated to Fs.
How would one go about plugging in simulated radiation impedance and WG parameters into a formula for calucaling SPL?
https://courses.physics.illinois.edu/phys406/sp2017/406pom_sw.html
https://courses.physics.illinois.ed...r_Piston/Baffled_Circular_Piston_F_Theory.txt
https://courses.physics.illinois.ed...ustic_Monopole/Acoustic_Monopole_F_Theory.txt
Something along the lines of getting a curve that scales rho0 with frequency?
(It's informative that the datasheets for those Peerless drivers disclose Sd and Xmax: 2535: 10.9cm2 , 0.1mm ; 2544: 16.2cm2 , 0.3mm I guess at least the Sd is a good general number for the driver size category. So generally speaking one would expect +3.5dB at a given X from the 50% larger Sd of a 44mm driver compared to a 35mm one, 50 to 100% increase in Xmax would be another +3.5 to 6dB.)
Quite possibly too soft and runny. Looking at descriptions of other products: https://www.polymers.co.uk/blog/prevulcanised-natural-rubber-latex I think that one is also a prevulcanized dipping latex type of product. Prevulcanized sounds to me like the final consistency is predetermined. (Not that one would want to do a high temperature cross-linking process, and not sure how would that then be applied as a coating.) There's another one that's called gel, maybe a better option.
But I would want to try to make a kind of blend with it that resembles bitumen rubber. But not with bitumen but a bitumen analog, based on guessing. Maybe paraffin wax softened with small amount of petroleum jelly, as analogs for the waxy and soft fractions of bitumen. (Not sure how that would resemble the stickiness, also that jelly is not gummy.) Or say beeswax and castor oil for the same kind of thing. (And stickiness could come from rosin, probably.) Maybe bentonite would work as the emulsifying agent to blend it into the latex., it's used in bitumen emulsions.
You can do a relatively similar horn in Hornresp and get the excursion required quite easily.
There are dampening paint on compounds made for ships and for duct work. I have used them when working on acoustical treatments for rooms where we had too much noise via duct work. They are quite effective. No need to reinvent the wheel.
The problem is not that it couldn't be simulated, but that seldom you have all the input data at hand.
But even if the required excursion was reliably known, what does it tell about the capabilities of a particular driver? What portion of its available excursion we can use without audible issues?
As we move lower in frequency, there is typically a sharp onset of the required excursion to keep a constant SPL, like in the simulation example below of an extended throat, including a simple CD model.
I think that this will also be the frequency below which the distortion will just ramp up quickly. That's probably to fastest way how to get this information - to measure the harmonic distortion.
Last edited:
I think that until we know the audible relation to measurement data in detail, the only thing we can do is to attack the issue as engineers and just remove as much non wanted artefacts as possible in order to achieve an as clean signal as possible from the aspects we know: level SPL, level vs. frequency, any order harmonic distortion level, level of non harmonic energy addition, phase linearity, dispersion and all this as a function of time.
Set up the parameter matrix (the battle field so to say) and go for a 20/80 approach - 20% effort for 80% results...
Just skip audibility and dig deep in signal quality...? We don't know the fine details anyways it seems so why bother..
Personally, I'm sure it will pay off - no doubt. And get much further actually than what "listeners" can achieve..
//