That's not good. There is going to be overlap on those. Just look at the transfer functions alone. Not to mention the rising response of the Seas, the break up peaks, the lack of BSC.
Granted, that's with polarity reversed. Otherwise it's even uglier. Scrap the whole thing and start over.
An externally hosted image should be here but it was not working when we last tested it.
Granted, that's with polarity reversed. Otherwise it's even uglier. Scrap the whole thing and start over.
Hi mate
I'll open up the cross-over, and send you a picture asap.
I will say that one of the design briefs that I gave to wilmslow audio was for them to design a minimalist crossover in order to maximise efficiency, but this is my first project - so I'll take it from here.
I am looking forward to re-beuilding the crossover - first timne for anything!!!!!!!!!!
thanks to everyone for the input so far
Lohan
Even a minimalist XO can do better.
Just looking at transfer functions,
woofer 1.5mH, 12uF, 0.5mH
tweeter 6uF, 0.5mH
That gives at least a flat line with ~20db down at 3khz on the woofer. And that's just 2 seconds of inputed jibberish. Again, only transfers. Impedence would wipe that out.
Good luck.
Just looking at transfer functions,
woofer 1.5mH, 12uF, 0.5mH
tweeter 6uF, 0.5mH
That gives at least a flat line with ~20db down at 3khz on the woofer. And that's just 2 seconds of inputed jibberish. Again, only transfers. Impedence would wipe that out.
Good luck.
Wilmslow do not claim to produce finalised crossovers, just starting points. They say they simulate BSC.
loheswaran, your drivers likely do not match in directivity at the crossover point. There will be more off axis energy in the tweeter at and just above the crossover point if the crossover was designed using on-axis plots. As a starting point, I would begin to reduce the size of the tweeter capacitor and listen to the results.
I think it should be possible to improve a bit on that, without going crazy
but a few additional components would needed
a few resistor here and there
slightly different cap calues
but it might still fail
much depends on whether you are willing to follow instructions
another issue may be that I/we might be bashed for even trying
proper communication is essential
for a first try I suggest to mount a small series resistor on tweeter
about 1ohm, more or less
and report what happens
and you could try to mix the caps
use 8uf from woofer on tweeter
and 12uf from tweeter goes to woofer
and you may have a problem with tweeter offset
I reckon woofer and tweeter are mounted flat on vertical baffle
you could try and tilt the whole speaker
but a few additional components would needed
a few resistor here and there
slightly different cap calues
but it might still fail
much depends on whether you are willing to follow instructions
another issue may be that I/we might be bashed for even trying
proper communication is essential
for a first try I suggest to mount a small series resistor on tweeter
about 1ohm, more or less
and report what happens
and you could try to mix the caps
use 8uf from woofer on tweeter
and 12uf from tweeter goes to woofer
and you may have a problem with tweeter offset
I reckon woofer and tweeter are mounted flat on vertical baffle
you could try and tilt the whole speaker
Replace the 12 uf capacitor that's in series with the tweeter with an 8, then try a 6. This may reduce the peak centered around 3 khz. You can try some inexpensive caps until you find one that gives the smoothest overall response and then replace it with a more expensive one if you think that matters.
Trust me Lohan, you cannot design crossover for this combo in a brutal way as you can with ScanSpeak drivers.
First, check on the woofer whether an LC-trap is already installed. This trap is a must. They should be on the woofer not on crossover PCB.
The woofer is crossed too high. It is BETTER to allow the tweeter to have a bit of transfer problem than to have the woofer resonance untreatable. It is not difficult to make it better just by lowering the crossover point (same order).
I will help you design a crossover that I believe will better this. Personally, for myself I will prefer an expensive complex crossover, but it will take a lot of effort and may not be your preferred approach. I will just use the same filter slope with what you already have but with lower crossover point. You will have L-pad (or R before first C if it is not too difficult).
ofcourse it wont work, just like that
doesn't cost anything to try it
and it might tell what to do next
but do not quit yet
if nothing else, you might still learn something
You got that right. It's bad in just about every way possible. But Linkwitz uses those same dirvers to make one of the best sounding speakers in the world. Cross LR4 active at 1400 Hz., put the appropriate notch on the W22 breakup, good to go, even in a box. Takes maybe 5 minutes to program it into a miniDSP (half an hour first time through) . . . including correcting acoustic offset and adding BSC.
Throw the old crossovers away and do it right . . . those drivers deserve it.
Here is what I have done so far. I give some design notes:
1) I always prefer a resistor in front of first capacitor to pad down tweeters. here I use 2.2 Ohm. But I show an L-Pad option (2.2 Ohm + 21 Ohm). The benefit is in the better sonic/dynamic. The drawback is (if audible) rather rolled off top end.
2) I choose first order for the woofer because I cannot find a way to make both drivers "in phase" with 2nd order (3rd order might have been better if 1st order doesn't work, but needs real drivers to listen to). The acoustical slope is 3rd order (Butterworth) as you can see from the chart.
First order like this and in phase response between 2 drivers will give you a very good sonic, and a speaker that will easily disappear (but pay attention to how you install your tweeter). But without actually hearing it, it may give you peak sound (but may be acceptable for most beginners). In this case, redesigning with increased inductor (from 1.8mH to 2 or 2.2mH) can be done.
3) I have tried to make use your available components with little success. The 0.5mH on the tweeter is a must (I couldn't use your 0.4mH). Bigger is better to reach the woofer that has to be crossed low enough (Here I aimed for 1K7)
4) I didn't go into extreme precision for this design, because without actually hearing the driver and trusting the FRD (I took it from RJBaudio site), I don't think it is necessary.
5) You can try to use both of your inductors (2mH) to check how this will sound in general.
Attachments
Box Tuning
If you look at the chart above, it seems the low end response of the woofer is too weak. I may be wrong but I don't think this is real. This is just what the FRD file is (I didn't check further).
You might think that a dedicated BSC circuit is needed but that will be expensive. Indeed this is more suitable for 3-way, where the W22 is used as midrange!!!
If an elevated response is preferred from 200Hz down to 40Hz (to increase bass perception), the port length can be reduced. In my simulation, with 7cm port diameter, changing from 19cm to 12.5 port length will only increase F3 by 1 Hz. (Blue is if you are a bass/rock lover)
If you look at the chart above, it seems the low end response of the woofer is too weak. I may be wrong but I don't think this is real. This is just what the FRD file is (I didn't check further).
You might think that a dedicated BSC circuit is needed but that will be expensive. Indeed this is more suitable for 3-way, where the W22 is used as midrange!!!
If an elevated response is preferred from 200Hz down to 40Hz (to increase bass perception), the port length can be reduced. In my simulation, with 7cm port diameter, changing from 19cm to 12.5 port length will only increase F3 by 1 Hz. (Blue is if you are a bass/rock lover)
Attachments
Uhmm, because you think that the manufacturer knows something about its strengths and weaknesses? See the datasheet:
http://www.seas.no/images/stories/excel/pdfdataheet/e0011_t25cf002_millennium_datasheet.pdf
Power handling for the Millennium is stated with a 12 dB/oct filter at 2500 Hz. If you go lower, you need steeper crossovers. I would not go below 2000 Hz, period. The risk is that heat builds up in the voice coil and destroys it over some time. I lost a couple of expensive Scanspeak Revelators by trying a 1800 Hz crossover frequency instead of 2400 Hz. Sounded fine initially, but even with 96 dB/oct slopes they died a slow heat death over a few weeks. Their power handling is stated at 2800 Hz, which tells you something about the likely size of the safety margins here.
Last edited:
all these advices seem more than proper, but actualy will not heal the real problem here..
if it was me, i would maybe just -if i had parts available- add 0.75 ish mH coil || 30 uF ish capacitor || to 2.5 ohm ish resistor, and this all paralell with the woofer.. At least it would reduce that nasty peak. 1.5 ohm ish resistor in series with the tweeter to chill it a bit.
But the actual problem is simply, that suitable freqvency range for these 2 drivers simply do not overlap enough for a proper design. It might just worth to get a sealed type mid unit, and convert to 3 way system.. Crossing the W22 at 1 Khz, and then the tweeter at over 3K should do the trick.
A nother approach, would be to just get other type of tweeter.. one that can go lower, so the W22 can be crossed under 2 Khz.
I em not sure witch one would be proper solution, as it would end up in the purchase of more drivers.
I will try to thinker a more or less proper corssover for the curent set of drivers,
but in order to do that I need to know the dimensions of Your exsisting boxes.
Suposedly correction circuitry of impedance for both the tweeter and the woofer has to be designed, baffle step compensation, L-pad, and probably a 3rd order LR crossover.
Not a minimalistic design, and I can not guarantee it will work 100%, only thing is sure it can just bebetter than what You have now..
In the event if You go by the route described above, I would suggest to build the crossover and other stuff from low cost cheapoo parts first, so You do not spend a lot at the first run.
Actualy the thing is, that experimental crossovers rarely work for the first time.
We are on absolutely at the edge of safe operation area with the tweeter, and on the edge of well behaved response for the woofer.
The choosen drivers are good ones, but not a good pair sadly.
But, since this is what we got, this is what we must make the best out of.
So, waiting for the cabinet dimensions, and hoping for the best.
I em sure we can improve this thingy. Just do not get lost between all the advices.
None of them is bad, they are all helpful additions, but it will take a bit more effort to deal with the problem.
if it was me, i would maybe just -if i had parts available- add 0.75 ish mH coil || 30 uF ish capacitor || to 2.5 ohm ish resistor, and this all paralell with the woofer.. At least it would reduce that nasty peak. 1.5 ohm ish resistor in series with the tweeter to chill it a bit.
But the actual problem is simply, that suitable freqvency range for these 2 drivers simply do not overlap enough for a proper design. It might just worth to get a sealed type mid unit, and convert to 3 way system.. Crossing the W22 at 1 Khz, and then the tweeter at over 3K should do the trick.
A nother approach, would be to just get other type of tweeter.. one that can go lower, so the W22 can be crossed under 2 Khz.
I em not sure witch one would be proper solution, as it would end up in the purchase of more drivers.
I will try to thinker a more or less proper corssover for the curent set of drivers,
but in order to do that I need to know the dimensions of Your exsisting boxes.
Suposedly correction circuitry of impedance for both the tweeter and the woofer has to be designed, baffle step compensation, L-pad, and probably a 3rd order LR crossover.
Not a minimalistic design, and I can not guarantee it will work 100%, only thing is sure it can just bebetter than what You have now..
In the event if You go by the route described above, I would suggest to build the crossover and other stuff from low cost cheapoo parts first, so You do not spend a lot at the first run.
Actualy the thing is, that experimental crossovers rarely work for the first time.
We are on absolutely at the edge of safe operation area with the tweeter, and on the edge of well behaved response for the woofer.
The choosen drivers are good ones, but not a good pair sadly.
But, since this is what we got, this is what we must make the best out of.
So, waiting for the cabinet dimensions, and hoping for the best.
I em sure we can improve this thingy. Just do not get lost between all the advices.
None of them is bad, they are all helpful additions, but it will take a bit more effort to deal with the problem.
Look at the Pass Labs SR2. It uses Nextel drivers rather than magnesium ones, but you could easily do something similar with the Millenium, a W15 and the W22.
An externally hosted image should be here but it was not working when we last tested it.
Instead of trying to reinvent the wheel without the skills nor the tools, the safer way is imho to use a proven design with those drivers.
This one is by Seas, and i hope they know better their job than their dealers...
THE ART OF SOUND PERFECTION BY SEAS - Trym
This one is by Seas, and i hope they know better their job than their dealers...
THE ART OF SOUND PERFECTION BY SEAS - Trym
Instead of trying to reinvent the wheel without the skills nor the tools, the safer way is imho to use a proven design with those drivers.
This one is by Seas, and i hope they know better their job than their dealers...
THE ART OF SOUND PERFECTION BY SEAS - Trym
Hmmm you are right. Murray Zelligman should know better about these drivers. I have built one of his design in the past.
It is a surprise how he came to the decision to marry these 2 drivers. But I know that the lower midrange from a 8" woofer is so fascinating, something that almost impossible to achieve with smaller diameter midrange. I believe that in the future there will be more design using this combo?
Not a surprised the Trym is crossed at 1K6. I started with 1K8 (by intuition) then had to go down to 1K7. I almost withdraw my proposed filter because someone mention about the danger of possibility to burn this expensive tweeter. I'm still not sure, but from my experience expensive tweeters have good power handling. Well, I don't know.
My proposed crossover uses almost similar component values. It means that while building the proven design from Murray, the first order filter can also be tried. If it work without intolerable peak, it must have strength. The over exposed midrange is not a problem ime, it is suitable to be used in TL enclosure.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.