This depends on a lot of factors and is complicated. You shouldn't guess at this. Search the web for crossovers. You'll see you need the specifications for the drivers to put into a formula for high and low pass.
I would roll the woofer out at some point. I don't think it's a good idea to run a 12" woofer without taking it out. A woofer this big doesn't do mids very well and could operate out of its passband.
Go to the Scanspeak sight for more information.
BTW, you are using a simple 6db crossover. Of course this doesn't mean the speakers would be phase correct or time aligned. Good luck.
You might need an inductor on the woofer. What are the other cap values and why do you feel they are not the right value?
I've had the speakers for 30 years and their sound is extraordinary (due in no small measure to the lack of complex crossovers and the full-range woofer).
The existing capacitor values are impossible to see as they are glue-gunned to the cabinet beside the drivers and I'm not a soldering person.
I need to order the new Dueland VSF caps which will take 2 weeks to get to me and then I will have a technician take the speaker apart and solder in the new caps.
I wanted to know what others thought about the best values for the new caps but if its difficult, I'll have to just the repeat the existing capacitor values.
If the sound has been excellent with the existing values, why would you change them? Keep in mind- When the new caps are installed, it will take a number of hours(perhaps 200+) for them to sound right.
The reason I may change them is that the Scanspeak Mid/Woofers and Tweeters were substituted a year ago for the original Hokutone drivers.
Whilst they sound OK, their characteristics (eg impedance and X-over frequencies are almost certainly different.
Does 200uF for the Mid/Woofers and 1.766uF for the Tweeters sound reasonable?
If those speakers are 8ohm: the 200uF cap will cross at 100Hz and the 1.76uF over 10kHz. If they are 4ohm drivers: 200uF= 200Hz, and 1.766uF=somewhere over the rainbow, I mean- over 20kHz. I seriously doubt the mid-range actually has the ability to reproduce that range of frequencies accurately. If it does, I'll buy a truckload of 'em from you! If the tweeter is a nominal 16ohm piece, it would come in at 5kHz, but the cap on the mid might as well not be there at 16ohms. Or are those the values you were thinking about substituting? Most people use 8ohm drivers when building their own systems. It would be good to know with what you're dealing. Do you have a multimeter? Here's the formula for capacitance per your impedance and cut-off desires(for a 6db/oct slope): C1(in uF)= 1,000,000/6.283 X speaker impedance X desired x-over freq.
The Tweeter is 6 ohm and the Mid/Woofer is 4 ohm.
The Tweeter at the moment has a 2.2 uF cap and the Mid I believe has a 10 uF cap.
Should I stick to these values?
Well- That would bring your tweeter in a little above 10kHz, and your mid- right about 4kHz. The mid isn't doing much midding that way. I'd say 80uF on the mid(500Hz) and about 5.5uF on the tweeter(about 5kHz). I suppose if you were to go with an iron core inductor on the woofer, you wouldn't have as much insertion loss(though it may become non-linear at high input levels). If I were doing it, I would use either Solo or Alpha-Core air core inductors(very low insertion loss and phase shift, very transparent). The value of the coil would depend on the impedance of course(8ohm=2.5mH, 4ohm=1.27mH for 500Hz). I suppose a .13mH to roll off the mid at 5kHz wouldn't hurt either. You may find yourself having to pad the tweet and/or mid to match their outputs with the woofer.
Thanks again Rodman.
Whilst the formulae indicate that with the 10uF on the Mid it should only come in at 2000Hz, it patently doesn't.
The speakers have always sounded effortlessly brilliant as all my audiophile friends agree.
I don't want to mess with that distinctive balance and by inserting new Dueland VSF caps for the existing cheap 30 year old ones, I hope to hear improvements.
The Scanspeak tweeter comes on song at 1500Hz so I don't understand your thoughts on pushing it up to 5KHz?
I thought if I left the tweeter at 2.2uF and increased the Mid/Woofer to 40uF (compared to the 10uF it's had for 30 years?......it might improve the overall sound without affecting the character that I love?
What are your thoughts on that?
If you like the sound: Leave it as is! You've no one to please but yourself. Based on the impedances you gave, the values/points/slopes should be as noted. I was suggesting what generally works well for x-over points. Especially with a 12"/5 1/4"/and tweeter. Happy experimenting/listening!
One final question Rodman,
If I change the cap value on the Mid/Woofer from 10uF to 40uF, what is the possible risk/downside as your best guess?
Mr H- No risk at all. It would bring the mid in at about 1kHz which is probably closer to the the neighborhood of the 12's natural roll-off. See if you can find the data on your 12". Often that will supply you with a freq response plot. Check it's 3db down point(upper) and design accordingly, if you don't want to use an inductor.
And a 3.5uF on top should bring the tweeter in about 7500Hz.
Is this OK do you think or should I be aiming for lower than this?
I'm not certain of the response of your midrange, never having seen the data. BUT- Most don't operate smoothly up to 7500Hz. If you look at this response: (http://www.madisound.com/catalog/PDF/scanspeak/12m_4631g00e.pdf) You'll note how erratic the response becomes after 5k. In the case of this mid: I would use it between about 700 and 5kHz where it's response is the smoothest. You may be able to find data for your particular driver, and choose roll-off points accordingly. On the other hand: Most tweeters are designed to function smoothly from somewhat below 5k to above(what's generally considered) audibilty.
The Mid starts to drop at 5K Hz as you say so I think a 5.6uF on the Tweeter would bring it in at 4740Hz.
Any risks in that?
Not at all. Just add the coil I mentioned in the previous post to roll the mid's top-end off, which will eliminate that peakiness from the reproduction and avoid a major overlap in reponses. Don't forget an inductor on the 12" to avoid the same in the woofer/mid region. It may take you a while to adjust to the new-found accuracy of the speakers and don't forget: The caps are going to take a while to sound right too(maybe 200+ hours of play-time).
You're the only one in 3 forums to actually help me and not disparage my speaker's lack of X-over and suggest I keep the drivers and trash the cabinets.
It's amazing how some people can so clearly ' hear' a speaker from its specifications....and they must think I've never heard speakers with properly designed X-overs?
There's not a single one I've heard, that sounds like live music.
My cabinets sound like the original Martin Logan CLS but with bass.
Those speakers (obviously without X-overs).......continue to be my reference.
I haven't had any electrostatics since my Acoustat Model III's. But- I almost bought a pair of ML Sequels a few years back. A local store was blowing their last pair out for 1/2 price(whatta temptation that was). I've been using modded Maggies for mains for a number of years now. My bi-amped transmission-line woofers just work too well to replace, and most ML's are hybrids. NOW- If I could find a pair with blown woofers for cheap.... OH- BTW: If your mid-range rolls off smoothly around 5k(and is already 3db down at that point), you could get away w/o an inductor on it. I try never to underestimate another's ears, unless they give me an obvious indication that they don't know what they're hearing(or can't hear at all, as the case may be). I hope some of this helps you out, and: Happy re-designing!
Just wanted to thank you again for your help.
I soldered in some cheap caps of 40uF for the mids and tried both 4.7uf and 5.6uf for the tweeters while I waited for my Duelund caps to be made and sent.
Established that the 40uF and 5.6uf were good values and when the Duelunds arrived I wired them in.
The sound is now extraordinary.
The Duelunds are just crystal clear with no audible grain at any frequency.
The overall balance of the speakers is (to me)....seamless.
There are still no inductors (coils) or resistors so just 1 cap between the Mid and 1 cap between the tweeter.
Until the so-called 'speaker designers' listen to what minimal cross-over complexity can actually achieve, they will keep dishing up boxes which may measure well but which sound little like 'real' music.
The fact that so many audiophiles seek out speakers OTHER than boxes stuffed with cross-overs, should indicate that the formulas and methodology have failed.
Can anyone name a 3 way moving-coil 'box' speaker designed 20 or 30 years ago that people still listen to?
Yet Quad Electrostatics, Klipschorn horns, Accoustat panels still cut it.
Hi Mr H: I'm glad those values worked out so well for you. I'll bet they sound even better after a couple hundred hours of music. As far as older box speakers that sound like "real" music: The Dahlquist DQM-9 is an excellent system still. Actually built by Magnat(a German co.) with a very complex crossover, they still deliver a wonderfully open, accurate sound with great imaging. I bought a pair for my son some years back(we restored them, without any actual "mods"), and he's found no reason to replace them yet. There are still a number of the old LS3-5A systems out there cranking out accurate sound/excellent imaging, so- they DO exist! Glad I could help. Enjoy the music!!
I tried to help you;) Didnt tell ya to shit can anything. Pretty hard to rework a network of a custom speaker that has no info available but what little you listed. You should be happy anyone took there time to post. And I design loudspeakers and yes we mostly use simple networks with costly parts or use no crossovers at all with some designs. Just that most 12in woofers have rising responce mostly arround 2khz up. So most 12in and larger woofers need at least a basic crossover to reduce this peak. I know of only a few that dont and these are costly alnico mag woofers aaa few vintage. Thus the inductor sugestion which would be sudo 1st order no phase shift -6db slope. And still might improve your sound quality.