I'm looking to build a very simple two way speaker that incorporates a single capacitor at a ~20khz crossover point -- a 1uf capacitor. The tweeter I'm considering is ~95db and the full range driver complementing it is ~86db. I'm thinking that with a 6db per octave rolloff both drivers should achieve roughly equal output at around 5khz, which is fine by me and provide some added brilliance in the upper frequencies -- though maybe too much, we'll see. I'm wondering whether certain types capacitors have better distortion characteristics for an application like this and if so which parameters would indicate this? Furthermore I presume the smaller the capacitor value/volume, in an application like this, may yield better results -- less resistance caused by inductive reactance?
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Single Capacitor Two Way Crossover -- distortion??
- Thread starter milezone
- Start date
Do you mean 20 kilo Hertz crossover?
- Thread Starter
- #3
Yes ~20 kilo hertz (edit) filter? A clever(?) employee at an audio shop in Japan, Koizumi Musen suggested it and I think it's an interesting idea given the right tweeter // driver combination.
Last edited:
I am guessing you want to make a simple capacitor filter divider crossover at ~2kHz. Not 20kHz. I've seen a bipolar capacitor anywhere from ~4.7 micro Farad for a 8 Ohm tweeter to ~10 micro Farad on a 4 Ohms tweeter being used for a 2kHz crossover point. The impedance of the tweeter and the desired crossover frequency will determine the capacitor value.
- Thread Starter
- #5
I appreciate the suggestion however 20khz is what I have in mind. Maybe that number is slightly high, but given a more sensitive tweeter relative to the full range driver, and a 6db per octave rolloff, I think it may work. I'll have to experiment and see how it sounds. I think 1mf-1.8mf is what I'm going to start with. That said the unflat nature of the response of the design I'm describing is no doubt bad -- however if my ears say it works, the simplicity of this, and the potential benefit of using a smaller capacitor?, may make it worth my while. I prefer upper end brilliance -- the Genelec flavor -- perhaps this is a simple way to approximate that sound signature.
Comparing Genelecs 8030s to KH120s for example, both measure flat however the Genelecs present much more detail, shimmer, and brilliance compared with the nasal, underwater, and uninspiring sound of the KH120s in my opinion. I presume this has a lot to do with the material, mass, sensitivity and damping characteristics of the tweeter -- less damping = more brilliance?? -- so I figure it's fairly safe to experiment with what's intuitive, but from an equalization approach, to try to approximate what I like.
Comparing Genelecs 8030s to KH120s for example, both measure flat however the Genelecs present much more detail, shimmer, and brilliance compared with the nasal, underwater, and uninspiring sound of the KH120s in my opinion. I presume this has a lot to do with the material, mass, sensitivity and damping characteristics of the tweeter -- less damping = more brilliance?? -- so I figure it's fairly safe to experiment with what's intuitive, but from an equalization approach, to try to approximate what I like.
Last edited:
I doubt you can even hear 20kHz. Maybe what you want is a super tweeter. Designed for the very high of high frequency of the hearing range. Or at best a parametric EQ so you can dial in the parameters for the tweeter. It works very well. Active crossovers for home and car audio(used in a home system) are not expensive these days and with a additional amp for the tweeters you can have full control of your brilliance.
A single capacitor will not create a crossover, just (presumably) a low-frequency rolloff to the tweeter, while the woofer runs full-range. The woofer will roll off naturally but I'd expect its distortion to be very high because of the high frequency range it would be required to cover.
The accepted upper limit of average human hearing, for young people, is about 20 kHz so putting the frequency corner there means the tweeter will be doing very little in the audible range.
Aligning amplitude and phase between drivers to prevent frequency response aberrations would seem to be problematic but hopefully you've figured that out already.
The accepted upper limit of average human hearing, for young people, is about 20 kHz so putting the frequency corner there means the tweeter will be doing very little in the audible range.
Aligning amplitude and phase between drivers to prevent frequency response aberrations would seem to be problematic but hopefully you've figured that out already.
- Joined
- Oct 1, 2018
- Messages
- 826
- Likes
- 1,226
I'm more than 20 years old now and can still hear 24kHZ 
- Thread Starter
- #9
I concur..
Last edited:
You really should use the active crossover route and dial in the parametric EQ for best effect. I did it on a tri-amp'd home system and I had very brilliant cymbals and horns and such. It was great. Way better than dialing in a capacitor divider at 3db/decade and with phase issues galore like @DonH56 advised.
- Thread Starter
- #11
I'd be using the tweeter in conjunction with a full range driver. The tweeter is ~12db more sensitive than the full range driver and thus at a 6db per octave rolloff will match the output sensitivity of the full range driver around 5khz ideally.
As for aligning amplitude and phase, I have not figured that out. Any suggestions?
- Thread Starter
- #12
I agree I prefer that design from a performance standpoint. However I'm looking to design as simple a speaker as possible with no active crossing or bi amplification -- two ncore amps and a d90 dac -- and a full range driver in an undetermined enclosure, with a supplementary tweeter, that's relevant from about 2khz up and presumably providing more detail than the full range driver is capable of reproducing.
The phase relationships are a big can of worms. I would worry more about getting the desired effect with the proper gear rather than solving phase relationship issues. You will have phase relationship issues regardless of which route you choose. But active crossover is the best.
Im completely lost.
You want a tweeter to take over at 20kHz, which as you say, is inaudible.
You will never match the woofer to the tweeter just by plonking some components together.
Might I respectfully suggest you read some speaker design books before taking this further. Its just not that simple Im afraid.
D
Deleted member 2944
Guest
I think this clever employee at the audio shop in Japan is not very clever. I suggest to abandon his/your idea and implement a PROPER crossover between your FR driver and super-tweeter. It shouldn't be that difficult.
Dave.
I suggest to abandon his/your idea and implement a PROPER crossover between your FR driver and super-tweeter. It shouldn't be that difficult.Dave.
- Thread Starter
- #17
I appreciate the suggestions and I know it's counterintuitive to all you folks. I still plan to explore the idea. And I remain curious about distortion related to capacitors in a crossover? Can adding more components in this context -- a passive crossover -- reduce distortion? Not to my knowledge though I'm not sure?
- Thread Starter
- #18
It's not taking over at 20khz. If it were a 4pole crossover it would be a different story. It's a highpass filter that's 12db more sensitive than the full range driver at 20khz. Hopefully that clarifies things a bit. The 12db sensitivity difference is extreme. A 6db difference in sensitivity with (edit) filter cutoff frequency at 12khz say is perhaps more reasonable. Though just proposing an idea for a very simple speaker design that I hadn't considered.
Last edited:
By adding in a inductor and a capacitor you can get a steeper slope but you will have more a complicated phase relationship with the driver. That is all beyond the scope of your ability from what I gather. You're not going to solve the phase relationship issues anyway so I suggest forget about those and just go for the crossover frequency you desire. Perhaps you just want 2 tweeters in each speaker. It will give you lots of brilliance. Parts express has a wide selection of pre-made crossovers for reasonable prices.
Im afraid it doesnt clarify anything.
Im sorry but you seem to have a few ideas mixed up here.
The drivers are vastly different sensitivities. You will need to match the two. That cant be done with just a cap. So what is the crossover frequency? with a first order slope they will overlap considerably. You will almost certainly end up with a massive null.
