Had a feelingI was even thinking you might add series resistance to exacerbate the problem and prove the point, but your idea sounds best.
I do have a Yamaha receiver with a noisy selector switch. I guess I could press that into service, but I have a feeling I won't like the sound of it. That receiver is thoroughly mediocre.
Hmm... How does one figure the values of R and C for the total system? Measure DCR at the speaker terminals? But how to figure the capacitor? Would I need an inductance meter?
--
So does not have the pronounced high freq roll-off found in a CD horn? Or is it likely this tractrix horn+driver has a pronounced high freq roll-off anyway? (I suspect it does...)
If you talk through a big horn you know what it sounds like, louder than life. The base will be pronounced and the general sound resolution will go down. But that is not the problem with your speakers. It's the crossover, and I have given the fix. But hey, we live in a free country, so any fix goes... Maybe you have the wrong brand of stuffing in your speakers? Maybe they are of wrong color? Have you checked your room temperature lately? Are the screws loose and rattling? Is your woofer made of a suspicious grey matter?
That's true. So the reason for the long roll-off into the low end of the tweeter response is to counteract that(?). Common practice in crossovers meant for horn tweeters?
Hi,
I appreciate you supplying me with a solution, I really do. I want to try it, but will need to build a new crossover in order to do it. That means buying parts, which I wasn't intending to do, or rummaging around to see what value chokes I have.
The thing I don't understand is, how can you tell those values and changes will work if we don't even know the impedance curves or the frequency response curves of the drivers involved? I know they're not even close to flat.
--
Zobel correction for a typical systen (lower mids upward) often requires a series LCR circuit across the speaker near the crossover, but may require more. You'd need to measure, or at least you could make a good estimate if you could get an impedance plot from somewhere else.
Why not give your EL34 amp a go?
Go back and read my post. I called it an estimate and I brought its accuracy into question. I'm not sure I appreciate this kind of talk, strawberry.
Well, I went by intuition and put 200 ohms across the tweeter. No series resistor on the tweeter, just the parallel resistor. Nothing scientific about it, just figured that if 100 ohms deadened the tweeter too much, maybe 200 ohms would be OK. I listened for a little bit and liked what I heard. Then a friend came over, we got to talking, had dinner, and the time went.
Now it's too late to play loud music, but I did wire in the EL34 amp. I'm listening to a mid-'50s mono Miles Davis CD, and it sounds quite nice, but I won't really know until I can play various test pieces at decent volume.
So that's where things are at the moment. Have to work in the office tomorrow, so won't be home until 6pm or so.
--
Now it's too late to play loud music, but I did wire in the EL34 amp. I'm listening to a mid-'50s mono Miles Davis CD, and it sounds quite nice, but I won't really know until I can play various test pieces at decent volume.
So that's where things are at the moment. Have to work in the office tomorrow, so won't be home until 6pm or so.
--
not sure what it does
horns are strange creatures
but 100-200ohm is quite large
I suspect it might change something Q related, in the xo
or maybe it relates to some kind of high frequency amp 'issues'
but I dont much about either
not related at all, seriously
tho I have a funny one about resistors in crossovers
all in series with other paralel components, be it caps or inductors
when I go from big to smaller resistor, and approaching a value slightly above the nominal speaker impedance, I know im getting closer with other related parts of the xo, or the whole xo in general
and when the opposite happens, and resistors are getting smaller than nominal speaker impedance, I know Im moving away from the ideal situation
but this rarely happens, which I actually find a bit strange
I know it sounds weird, and many might say 'rubbish'
but all my 'experiments' have shown this 'tendency'
at least with my homebrew 'method'
I think this one may be marginal. You're quite right, the useful range of reactance is likely to be from ten times, to one tenth of the load. Beyond this the effects will probably be too great or too small to be useful.tinitus said:
At the edges of the tweeter's passband the impedance is higher. A 100ohm resistor could cause a drop of two-thirds of a dB when placed in parallel with an 8 ohm resistor. Take that in a worst case scenario of driving it with a current source. In that case the resistor could possibly cause a few dB of change at the edges of the passband.
It is obvious that this is a step in the right direction but is not hitting the mark (the joys of tweaking
). It may be making two changes at the same time (upper, lower, not to mention phase at the crossover). I may have overlooked something on the thread, but I'm curious whether it is dropping the top end of the tweeters response or the lower end which is giving the greater improvement? And on that note, does rongon have access to an equaliser?
hmm, yes ofcourse, with this 100 or 200ohm resistor it might actually be the small inductance that have some effect
I used to have an EQ.
I think I could borrow one of those dbx active digital crossovers for pro audio use ("DriveRack" something or other). I would need to figure it out, but I'll bet it can be programmed to act as an EQ, no?
--
I don't think the problem is down in the bottom of the tweeter's range (around 2kHz). I think it's higher up, in the "sizzle" frequencies. More "ssss" than "shhhh" -- if that makes any sense.
I agree that the effect of that large a resistor is pretty subtle, and in theory I shouldn't be able to hear it, right? But I can hear it. I first added the resistor to one speaker, noted a slight change in comparison, so added it to both. Definitely reduced the apparent sibilance. You can tell right away on announcers' voices over FM.
I also switched to the EL34-triode push pull amp, which has about 6dB of global NFB. That should lower output Z of the amp a little. I guess I noticed a tightening up of the sound. I mean, it all sounds "good." It's just that I was trying to get that last little bit of goodness out of these Klipsch spkrs, without a complete re-design. Of course, that's when the trouble starts...
In the end, I guess these spkrs are not a good match for my push-pull 2A3 amp.
Just got a pair of Mark Audio Alpair 7 extended range speakers yesterday. Maybe I need to bi-amp, with the Alpair 7's from like 200 Hz on up, and woofers underneath. 2A3s for the full-range, EL34's or solid state for the woofers. I could put RC filters in the inputs of each power amp. Maybe with cathode followers for buffers, nice hi-Z load for the filters... with 85dB/1w mid-tweets and 93dB/1w woofers, putting a volume control on the woofer amp should allow me to balance the two.
--
I think I could borrow one of those dbx active digital crossovers for pro audio use ("DriveRack" something or other). I would need to figure it out, but I'll bet it can be programmed to act as an EQ, no?
--
I don't think the problem is down in the bottom of the tweeter's range (around 2kHz). I think it's higher up, in the "sizzle" frequencies. More "ssss" than "shhhh" -- if that makes any sense.
I agree that the effect of that large a resistor is pretty subtle, and in theory I shouldn't be able to hear it, right? But I can hear it. I first added the resistor to one speaker, noted a slight change in comparison, so added it to both. Definitely reduced the apparent sibilance. You can tell right away on announcers' voices over FM.
I also switched to the EL34-triode push pull amp, which has about 6dB of global NFB. That should lower output Z of the amp a little. I guess I noticed a tightening up of the sound. I mean, it all sounds "good." It's just that I was trying to get that last little bit of goodness out of these Klipsch spkrs, without a complete re-design. Of course, that's when the trouble starts...
In the end, I guess these spkrs are not a good match for my push-pull 2A3 amp.
Just got a pair of Mark Audio Alpair 7 extended range speakers yesterday. Maybe I need to bi-amp, with the Alpair 7's from like 200 Hz on up, and woofers underneath. 2A3s for the full-range, EL34's or solid state for the woofers. I could put RC filters in the inputs of each power amp. Maybe with cathode followers for buffers, nice hi-Z load for the filters... with 85dB/1w mid-tweets and 93dB/1w woofers, putting a volume control on the woofer amp should allow me to balance the two.
--
Last edited:
anything you do with xo's will be audible
and most of it will also have audible 'sidefffect'
hard to say if its one or the other you hear
I think thats how its always is
meaning, it is possible to have a component not really doing what its supposed to, but its sideeffects may still be perceived as postive
Yes, I hear what you're saying. There comes a point when you have to decide how much effort you want to put into experimenting for experimenting's sake.
I find speakers -- especially crossovers -- just too complicated to fathom from a technical angle. I certainly ain't no engineer, and I quickly get to a point where the math gets too intense for me. Add the need to understand large scale, physical resonances and oscillations created by electro-magnetic mechanical vibrators, and I'm lost.
So much about speaker design seems to be a matter of applying things learned from previous experience to the problem at hand.
The speakers sound better than when I got them. That's good enough for me. I certainly understand the crossover better than when we started. You guys have been great. I hope this wasn't too much of a wild goose chase.
Thanks!
--
They both have their effects. One is more annoying, but the other is more, um...maybe ear-burning verses scratching or spitting.
Your crossover seems to have some influence up high, so if you wanted to reduce some of the top end you could add some parallel resistance (less than the 100ohms) until it feels like you've gone a little too far and killed the sizzle, then put a capacitor in series with it (all still in parallel with the tweeter) whos reactance equals the resistors value in the middle of the passband as this will release the effect on the low end.
Raw theory suggests that 3dB is the minimum distinghishable quantity, but we can notice less than a dB and when our system is well balanced, a small amount can make a considerable difference.
Neither of these ideas is likely to be of more importance than fixing your sibilance, in my opinion. I wouldn't suspect them as the culprit, but it may be possible.
100 ohms actually made a big difference, easily noticeable, so I guess on the order of -2 or -3dB (?). Wacky, but that's what happened. Tinkly bells and the upper partials of cymbals got soft, like a fabric dome tweeter, but were still there.
With a nominal crossover point of 1600 Hz on the low end, and 20kHz on the high end, I guess the middle of the passband would be 9kHz or thereabouts, right?
--
Oh, and what I meant about bi-amping was taking the tweeters and internal crossovers out of the Klipsch speakers and only using them for the woofers and cabinets. Using RC filters at line level (in between preamp and amps), cross them over low (maybe 250 to 500 Hz) and use a different driver (Alpair 7) in a little cabinet as a mid-tweeter. Use a separate volume control and gNFB on the woofer amp to adjust for speaker sensitivity, and so on.
--
Last edited:
Half-way between 1600Hz and 15,000Hz is 5,000Hz (it's a logarithmic scale). You might try 330nF, give or take. If this cuts out the effect of the resistor too much, then increase the capacitance a little, then reduce the resistance a little, back and forth in steps.
With regards to adding a smaller midrange, you should remember that your tweeter has a more narrow dispersion that most likely matches the 10" driver at 1k6Hz. If you add a smaller mid in between, you'll drive your crossover frequency up a few kHz, increasing room interaction and considerably changing the crossover.
With regards to adding a smaller midrange, you should remember that your tweeter has a more narrow dispersion that most likely matches the 10" driver at 1k6Hz. If you add a smaller mid in between, you'll drive your crossover frequency up a few kHz, increasing room interaction and considerably changing the crossover.
Hey Allen, thanks for persevering with this!
Ca. 5kHz might just be where the problems are.
I have a good selection of resistors and capacitors in that range. Will try it.
The "midrange" I had in mind would replace the tweeter entirely. No more horn tweeter, no more passive crossover. Those Alpair 7's are "full range" aluminum cone drivers (4" cones) with Fs about 70-something Hz, high freq's respond out to 35kHz. They're all the rage over in the Fullrange forum. Only 85dB/1w sensitivity, so bi-amping mandatory. If I decide to do that, I'd be completely giving up on this speaker as a 2-way w/ horn tweeter system once and for all. For now, I'm trying to learn some tricks to tame this tweeter...
Thanks again for the ideas. I'll let you know what the parallel cap does.
--
Ca. 5kHz might just be where the problems are.
I have a good selection of resistors and capacitors in that range. Will try it.
The "midrange" I had in mind would replace the tweeter entirely. No more horn tweeter, no more passive crossover. Those Alpair 7's are "full range" aluminum cone drivers (4" cones) with Fs about 70-something Hz, high freq's respond out to 35kHz. They're all the rage over in the Fullrange forum. Only 85dB/1w sensitivity, so bi-amping mandatory. If I decide to do that, I'd be completely giving up on this speaker as a 2-way w/ horn tweeter system once and for all. For now, I'm trying to learn some tricks to tame this tweeter...
Thanks again for the ideas. I'll let you know what the parallel cap does.
--
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.