I assume you are talking about the NeoCD2.0 right? Take a look at the distortion measurements for this tweeter, I think Zaph made some. Than decide what level of harmonics is still acceptable and pick your XO point and slope. In general I think you should be ok with a 3.5KHz LR4 XO, in the mean time I lowered my XO point to 3.1KHz LR8 and am still very very happy with the system...
Did a quick search and found it : Zaph|Audio
scroll down about 80% of the page and you will find the Harmonic distortion of the NeoCD2.0
At 3KHz F2 & F3 are about 55dB down, F2 rising slowly after that (to -45dB at 20KHz) F3, F4 & F5 go down to -95dB at 20Khz, all relative to the test signal of course
scroll down about 80% of the page and you will find the Harmonic distortion of the NeoCD2.0
At 3KHz F2 & F3 are about 55dB down, F2 rising slowly after that (to -45dB at 20KHz) F3, F4 & F5 go down to -95dB at 20Khz, all relative to the test signal of course
Found this CSD plot on the PE forum, it suggests something nasty is happening around 1.5KHz. Probably a resonance, so I guess you want to avoid crossing over any lower than ~3KHz LR4. There are a lot more measurements in that thread...
Fountek NeoCd3.5H Horn Tweeter
Hope I am not breaking any forum rules by posting a link to another forum
, going to read the rules now...
Fountek NeoCd3.5H Horn Tweeter
Hope I am not breaking any forum rules by posting a link to another forum
, going to read the rules now...Attachments
Last edited:
thanks again Mark, i think i read at least some of that thread, i agreed with the waterfall and the spike at 3k, which id hoped to avoid in a 4k or so crossover, as it turns out im at 3.5k so its not ideal. I did have a weird 4th order using the last stage as a notch, C-L-C-shunt LCR that takes the slope further, but im not sure its a great idea from my normally minimal approach.
I think that if you tackle that 1,5 - 2KHz resonance your on the right track my gut tells me the stuff you see at 3KHz is a harmonic, a notch + a "conventional" filter is not a bad idea actually look at this: Active Filters scroll to 7.3 - Cauer (Elliptic)
It's not "minimal" but it could be a good idea...
Only way to be sure is to experiment and measure and listen.
Edit: because you are using a passive filter look very closely at the impedance of your filter, make sure it doesn't drop below a level your amp can't handle!
It's not "minimal" but it could be a good idea...
Only way to be sure is to experiment and measure and listen.
Edit: because you are using a passive filter look very closely at the impedance of your filter, make sure it doesn't drop below a level your amp can't handle!
Last edited:
Again thanks! I hadnt spotted the THD plots before...
Its not too bad at around -50dB down from 2.2k and up.....2nd and 3rd being fairly similar... Interesting... With the elliptical filter im at -35dB due to the filter, so another 50dB is probably more headroom than the noise floor of my amplifier. In the 4th order its somewhat less, but its got to be ~20-25dB.
Im pretty sure THD wont be a problem then.
Thanks again
Torch is passing by soon! Maybe ill get up early to watch it...:
Its not too bad at around -50dB down from 2.2k and up.....2nd and 3rd being fairly similar... Interesting... With the elliptical filter im at -35dB due to the filter, so another 50dB is probably more headroom than the noise floor of my amplifier. In the 4th order its somewhat less, but its got to be ~20-25dB.
Im pretty sure THD wont be a problem then.
Thanks again
Torch is passing by soon! Maybe ill get up early to watch it...:
That's a bit of a different slant. Worth thinking about. If 5K second order was reasonable, one might be able to use a full range mid. It sounds like that is not a good choice. So much for the published charts and recommendations for 1500Hz.
If I were to play with at 10K and up just to get the idea, what would be a good choice/value to play with? The ribbons I have heard, I was not impressed with. They seemed metallic and edgy. I know it may not be a fault of the technology, but the execution or specific driver as we sure have a list of domes that fall under that description. It is just that I can play with a lot of $50 to $100 domes for the price of one ribbon. I am a cheap SOB after all. That's the only reason I have the budget to have a hobby like this in the first place.
to the best of my knowledge the Neo1.0 is the cheapest, then the 3.5H which i bought. Neo1.0 has a higher low end cut off, but response otherwise similar to the 3.5H, around 6db rise in last octave. the waveguide was the clincher, else i wouldve tried the 1.0.
I couldnt find anything cheaper lol
neocd1.0 around 35-40
neocd3.5h around 40-45
I couldnt find anything cheaper lol
neocd1.0 around 35-40
neocd3.5h around 40-45
Last edited:
What recommendations to use them down to 1500Hz ? Dome tweeters don't even work properly down to 1500Hz (despite the claims of some manufacturers) so why would a ribbon ? Many ribbons are flat down to below 2Khz but that doesn't mean they should be used there.
For a small ribbon the size of an AC G2 I would recommend crossover frequencies ranging from 3-5Khz, 3rd or 4th order electrical. I currently use 4Khz 3rd order to a full range driver.
The crossover frequency is partly dictated by the size of the overall system - if you're only making a small 2 way bookshelf 3Khz is probably fine as maximum SPL is going to be limited by the small midbass driver, but if you're using it as part of a large 3 way system with high SPL output a higher crossover frequency is beneficial to minimise the power dissipation and IM distortion of the tweeter.
Personally I wouldnt bother with the "super tweeter" concept, crossing over at 10Khz, I've tried it and not been impressed with the result. Putting a crossover frequency that high just doesn't work due to time alignment and driver spacing issues, I would stick with the 3-5Khz region for crossing over a ribbon.
What ribbons have you heard ?
No, believe it or not, I actually can play CD's on my 20K Krell/Bryston system!!
I noted the bad sounding MP3s didnt make most of those tweeters sound bad while the great measuring HDS sent me running. How you get that I only use MP3s to judge tweeters escapes me. Maybe this works better: I couldnt get some tweeters to sound bad no matter what I played on them, even MP3s.
I shared info on the factual use of some drivers because if the HDS tweeter can sound harsh and sterile to a few people in various applications while the SB29 sounds good to so many without much work, then measurements alone dont tell the story. I'm sure someone doing a search will come across it and possibly like the input instead of reading other guys typing away about how zaph measured low distortion on the low end of the HDS then it must sound great....yet they never heard it.
Thanks for your input.
I tend to agree with Simon.
Im hoping to minimise the impact of the CSD spike at 1.5k, and (as another poster said) the possible harmonic spike at accompanies it at around 3k.
I have optimised slightly different sims of the crossover for between 3.5k and 3.75k. I considered crossing higher, but Im not so sure the polars would match up so well above, say 5k or 6k, even with a different midwooofer.
Even crossing 2nd order as I am for initial testing the sound is impressive. Not as flat as the G20SC dome tweeter, on axis or off. But more precise sounding. The peaks of mid HF ripple in the plot I posted before are audible, and the filter indeed should at least be 3rd order at around 4k or 4th at maybe 3.5k. The output in the midrange is just too audible, which does detract from the woofer. AT low levels its presence is felt as it sounds surprisingly clean, despite the worrying spike on the CSD, but at higher levels I darent push it.
Despite the botched test crossover, It does very well.
I am totally new to ribbons, and I had no pre-conception of what their sound is really, so excuse the audio jargon.
The Visaton G20SC I am testing against are £35 tweeters, maybe a little more. These are pretty good looking on the plots, better than many soft domes Ive seen, and I refuse to pay hundreds for a ring radiator or other expensive dome. These things were £40 each. I think much of the issue I have with them, could be reduced a lot by crossing higher than 4k, AS WELL. Its a 20mm dome, so its polar 'bloom' or whatever would be mitigated a little.
That being said,I don't think there is a cheaper way to try a ribbon tweeter, except the NeoCD1.0.
Im very confident that a real crossover would help though....
Im hoping to minimise the impact of the CSD spike at 1.5k, and (as another poster said) the possible harmonic spike at accompanies it at around 3k.
I have optimised slightly different sims of the crossover for between 3.5k and 3.75k. I considered crossing higher, but Im not so sure the polars would match up so well above, say 5k or 6k, even with a different midwooofer.
Even crossing 2nd order as I am for initial testing the sound is impressive. Not as flat as the G20SC dome tweeter, on axis or off. But more precise sounding. The peaks of mid HF ripple in the plot I posted before are audible, and the filter indeed should at least be 3rd order at around 4k or 4th at maybe 3.5k. The output in the midrange is just too audible, which does detract from the woofer. AT low levels its presence is felt as it sounds surprisingly clean, despite the worrying spike on the CSD, but at higher levels I darent push it.
Despite the botched test crossover, It does very well.
I am totally new to ribbons, and I had no pre-conception of what their sound is really, so excuse the audio jargon.
The Visaton G20SC I am testing against are £35 tweeters, maybe a little more. These are pretty good looking on the plots, better than many soft domes Ive seen, and I refuse to pay hundreds for a ring radiator or other expensive dome. These things were £40 each. I think much of the issue I have with them, could be reduced a lot by crossing higher than 4k, AS WELL. Its a 20mm dome, so its polar 'bloom' or whatever would be mitigated a little.
That being said,I don't think there is a cheaper way to try a ribbon tweeter, except the NeoCD1.0.
Im very confident that a real crossover would help though....
Last edited:
Hey mondogenerator, do you mind sharing a design x-over?
I mention before that I order also the same ribbon as you have, just to try them with my "dust collecting" JX92s drivers...
Sure i can post the crossover. It is only a simulation using the frd. I have a 4th order and cauer type filter for the Neo3.5, and im not sure it will be optimal with the jordans, but it would be a guide to start with, and tweak by experimentation. If this is ok, then i shall post the circuit.
It's a hard truth to swallow, but more often than not "bad" recordings are not really bad recordings, but simply recordings that have content that expose particular flaws in the speakers.
In particular harshness / edginess in the upper mid/treble region seems quite common on many pieces of music, especially modern pop where its been excessively processed (hyped up treble, compression, all sorts of special effects etc) and can sound pretty awful on many speakers where other songs may sound fine.
For years I just assumed there were some recordings that simply sounded harsh and poor quality, blaming the songs, but I've come to realise that harshness on a recording is almost always the speakers and very rarely the source material, even amongst modern over-processed stuff.
If you have a resonance in the treble at say 10Khz then a song that doesn't excite that particular resonance much may sound perfectly fine, but another song may hit that resonance hard, for example if the song has exaggerated treble it may be unlistenable due to that resonance, but the song gets blamed. Same for sibilance, which is generally caused by resonances in the 6-8Khz region.
I think it was when i first got my ribbon tweeters that I suddenly realised that many of the songs that I'd heard on many other speakers and written off as poor quality recordings were actually NOT harsh sounding, even the ones that had excessive treble (pop mainly) sounded crisp but clean. Quite a surprise to me. There was still the odd song here and there which sounded a bit harsh but it was a big improvement.
Fast forward many years later, still using the same tweeters but putting a lot more effort into crossover, measurement, careful EQ etc, and I had noticed that there was a small 1dB bump at around 8Khz (mouth diffraction of the waveguide perhaps) and another small 1dB bump at around 15Khz. (cavity resonance behind the foil is my suspicion)
In the overall frequency response measurement they both look very benign, hardly noticeable, but since I have a digital PEQ I thought "why not", and carefully EQ'ed out those two small peaks with a 1dB 1/3rd octave notch each, and sat back to listen. I was very surprised to find that the few remaining songs in my collection that still sounded a little bit edgy particularly at higher volumes now sounded absolutely fine, no hint of harshness at all.
In fact I haven't been able to find anything in my collection now which sounds "edgy" even highly processed stuff which I would have previously written off as badly mastered.
The moral of the story I think is that even benign looking blips in the frequency response due to resonances can at higher frequencies cause harshness or edginess which is very song dependent, and that any potential resonances or glitches in the response of the tweeter need to be examined individually for audible effect, something that is very easy to do with a PEQ.
So I would say that yes, listening to "bad" recordings is a valid test of a speaker. Whenever I see reviews of speakers that say "this speaker is very revealing and makes good recordings sound great and bad recordings sound terrible" that screams out to me resonances in the upper mid or treble region...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.