What crossover type ? - Page 9

fatmarley · 6th March 2013, 06:36 AM

You keep saying how bad this driver is and yet the Snell k gets good reviews every year at the hifiwigwam show (Snell A got the best reviews this year).

I'm busy now, but i'll have a look at your circuit diagram later and try it in LspCAD. Would sending you the files make things easier? If you can model a better crossover, i'd be more than happy to try it.

system7 · 6th March 2013, 06:47 AM

If it looks like a pig, and acts like a pig, it probably IS a pig!

For all that, you have high-efficiency units here, with all their high mechanical Qms which means low energy storage. Something Joachim Gerhard likes.

I have no doubt that there are some good qualities in those drivers, but flat frequency response (and phase) is not one of them. Might be worth reading Troels' high efficiency design section. You seem to have a "Curate's Egg" here.

I only throw ideas out. It's your baby!

AllenB · 11:58 AM

I don't think these drivers are being used correctly and maybe that is the problem.

Take the M21 for example. Cone breakup appears to be taking hold at 1400Hz but behaves itself reasonably until 2k, where it becomes unreasonable. The maximum theoretical crossover frequency for this driver is probably 1800Hz, but everything above 2k ought to be way down. Crossing lower, maybe 1400Hz would make that simpler.

However you're now faced with a narrowed radiation pattern and require a waveguided tweeter (unless you want to cross the woofer on the south side of 1kHz and cross to another cone or dome). At the angle that your waveguide appears to have I'd guess the optimum crossover frequency would be down closer to 1400Hz, but I have doubts that this waveguide is capable of doing the right thing below even 3kHz (anyone have better info on this w/g?).

It's important that you don't read too much into the 0 degree plot by itself. To a point it's reasonable when the woofer is crossed low enough, and crossed to another cone or dome on a flat baffle but once this changes the 0 degree response alone won't show any of this, and it also stops showing the true response of the speaker.

fatmarley · 8th March 2013, 06:14 AM

Quote:

Originally Posted by system7

I think you need some inspiration here to make a silk purse out of a sow's ear. This woofer is really a horror with its puny magnet and soft surround...

The underdamped tweeter is causing that resonance at 4kHz and the subsequent 7 kHz dip. I did much better on phase with a regular SC10N tweeter.

For all that, the model that works best with a high Qts woofer in a box that is effectively too small is to use LESS bafflestep and a lower order filter. Troels' CA18RLY design was a good start. I ended up with something quite well-behaved and tweeter phase was much better. The higher crossover is just something you have to live with. These drivers are very difficult and I would have given up on them long ago.

Again, the red components are select-on-test. I modelled it as a freestander, away from the wall. What else can you do?

Sorry for the delay Steve, i've only just had the chance to give you a proper reply.

I've tried varying degrees of baffle step and I know that I can't get away with much less than I have already, so your crossover's not going to work i'm afraid. If you design one that has the same amont of baffle step as mine then i'll give it a go.

Also, I've got a bone to pick with you

- I was looking through Trolls stuff last night and wondering what to do with my Seas 29TTF/W, when I came across THIS. If you scroll down too "SEAS CURV FINAL CROSSOVER" you will see that Trolls used the exact same "random" notch as me to form the baffle step. I guess i'm not a stupid as I thought

fatmarley · 8th March 2013, 08:23 AM

Quote:

Originally Posted by AllenB

I don't think these drivers are being used correctly and maybe that is the problem.

Take the M21 for example. Cone breakup appears to be taking hold at 1400Hz but behaves itself reasonably until 2k, where it becomes unreasonable. The maximum theoretical crossover frequency for this driver is probably 1800Hz, but everything above 2k ought to be way down. Crossing lower, maybe 1400Hz would make that simpler.

However you're now faced with a narrowed radiation pattern and require a waveguided tweeter (unless you want to cross the woofer on the south side of 1kHz and cross to another cone or dome). At the angle that your waveguide appears to have I'd guess the optimum crossover frequency would be down closer to 1400Hz, but I have doubts that this waveguide is capable of doing the right thing below even 3kHz (anyone have better info on this w/g?).

It's important that you don't read too much into the 0 degree plot by itself. To a point it's reasonable when the woofer is crossed low enough, and crossed to another cone or dome on a flat baffle but once this changes the 0 degree response alone won't show any of this, and it also stops showing the true response of the speaker.

Hi Allen,
Yes a lower crossover point would be better for the woofer but the slopes would have to be steep for the tweeter to handle it and I don't want to try any new tweeters.

I've got a three way project that i'm excited about starting, so I don't want to go too much further with these speakers.

fatmarley · 8th March 2013, 08:35 AM

I need to tidy up the crossover and hot-glue the components down. When i've done that i'm going to take them to a local guy to compare to his World-audio-design KLS3. We did a comparison back last year (before I was using the AR-XO) and mine were terrible compared to his, with naff all bass and vocals that didn't project at all (high frequencies were ok though). It was obviously a lack of baffle step that was causing the problem (my listening room is a tiny box compared to his), so hopefully we will notice a big improvement this time.

I'll report back the verdict, good or bad...

AllenB · 08:46 AM

Quote:

Originally Posted by fatmarley

I guess i'm not a stupid as I thought

Well said. Umm, I was going to take Steve to task over this the other day and decided against it but since you brought it up again

I'm also a believer in learning process. Doing things a certain way because that is right. Train a newcomer this way and they stand to branch out to the harder stuff with both enthusiasm and experience.

...I got into electronics at a young age and stayed steadily with it, but I had only encyclopaedia and a local library to gather from and it was slow. When I made the move to college I'd sometimes notice my peers standing behind me watching, and the teacher would be in front of me with a slightly raised eyebrow and his arms poised at the fire extinguisher. (j/k)

If you understand enough to feel truly confident then you should understand how to remain 1 step behind any danger, and 2 when you're not sure.

I was going to mention a tweeter crossover to Steve. It uses only an inductor and a resistor. I think it helps to make a point.

AllenB · 8th March 2013, 09:29 AM

Quote:

Originally Posted by fatmarley

Hi Allen,
Yes a lower crossover point would be better for the woofer but the slopes would have to be steep for the tweeter to handle it and I don't want to try any new tweeters.

I've got a three way project that i'm excited about starting, so I don't want to go too much further with these speakers.

Fs is 750 IIRC (I can't find your links at the moment) and it was recommended for use above 2k. I don't know why that is but lookng at the harmonic distortion plots, 2k was the showpoint. 2nd and 3rd harmonic are not the worst distortions and there was only a few percent further down. I'd be willing to try 1k5.

Maybe third order electrical to be on the safe side, or even second with care. I've not owned a ceramic tweeter though. (make sure the filter is effective through the impedance peak and below.) I also think the waveguide can be helped with a further roundover at the mouth. Mainly, I think because a crossover at 3k is a little easier to fault.

fatmarley · 8th March 2013, 09:11 PM

Yes it's been a bit of a steep learning curve but i'm enjoying it and it's quite satisfying when the speakers start to make nice sounds.

Ok you've twisted my arm - I've been playing with LspCAD and modelled a crossover at 2k. Frequency response is very similar to the current one and phase tracking looks good, so i'll build it and report back...

Oh, yes the tweeter Fs is 750Hz and it's a soft dome (not ceramic).

AllenB · 9th March 2013, 02:41 AM

I used to run Vifa and Scan-Speak domes within the first octave of their resonance using just a capacitor (and peak compensation/padding resistors) for a number of reasons, some of which I don't believe in the same as I used to but I know it can be done with care.

If you do go down to 2k, you'd be able to (or need to) go to a larger woofer. The roundover on the waveguide was suggested to hold its pattern lower and may help even if you stayed at 3k, but would become more necessary below that.

It may be worth designing the crossover using your 30 degree response plots.

6th March 2013, 06:47 AM	#82
system7 diyAudio Member Join Date: Aug 2008	If it looks like a pig, and acts like a pig, it probably IS a pig! For all that, you have high-efficiency units here, with all their high mechanical Qms which means low energy storage. Something Joachim Gerhard likes. I have no doubt that there are some good qualities in those drivers, but flat frequency response (and phase) is not one of them. Might be worth reading Troels' high efficiency design section. You seem to have a "Curate's Egg" here. I only throw ideas out. It's your baby! __________________ Good Night, and Good Luck. Best regards from Steve in Portsmouth, UK.

6th March 2013, 11:44 AM	#83
AllenB diyAudio Member Join Date: Oct 2008	I don't think these drivers are being used correctly and maybe that is the problem. Take the M21 for example. Cone breakup appears to be taking hold at 1400Hz but behaves itself reasonably until 2k, where it becomes unreasonable. The maximum theoretical crossover frequency for this driver is probably 1800Hz, but everything above 2k ought to be way down. Crossing lower, maybe 1400Hz would make that simpler. However you're now faced with a narrowed radiation pattern and require a waveguided tweeter (unless you want to cross the woofer on the south side of 1kHz and cross to another cone or dome). At the angle that your waveguide appears to have I'd guess the optimum crossover frequency would be down closer to 1400Hz, but I have doubts that this waveguide is capable of doing the right thing below even 3kHz (anyone have better info on this w/g?). It's important that you don't read too much into the 0 degree plot by itself. To a point it's reasonable when the woofer is crossed low enough, and crossed to another cone or dome on a flat baffle but once this changes the 0 degree response alone won't show any of this, and it also stops showing the true response of the speaker. Last edited by AllenB; 6th March 2013 at 11:58 AM. Reason: Typo

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Crossover Optimization.........what of type component swap boards do you use?	Gavinator68	Multi-Way	13	23rd October 2011 10:48 PM
Klipsch type E balancing network (crossover) pair	jmillerdoc	Swap Meet	0	3rd June 2011 03:20 AM
Snell Type 1 Crossover Rebuild	lt_texan	Multi-Way	31	22nd January 2011 11:44 PM
Least Audible Midrange Crossover Type?	cuibono	Multi-Way	67	12th September 2010 08:37 PM
Capacitor type for crossover	ThSpeakerDude88	Multi-Way	23	4th August 2007 09:00 AM

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

6th March 2013, 06:36 AM	#81
fatmarley diyAudio Member Join Date: Aug 2003 Location: Cheltenham	You keep saying how bad this driver is and yet the Snell k gets good reviews every year at the hifiwigwam show (Snell A got the best reviews this year). I'm busy now, but i'll have a look at your circuit diagram later and try it in LspCAD. Would sending you the files make things easier? If you can model a better crossover, i'd be more than happy to try it.

8th March 2013, 08:35 AM	#86
fatmarley diyAudio Member Join Date: Aug 2003 Location: Cheltenham	I need to tidy up the crossover and hot-glue the components down. When i've done that i'm going to take them to a local guy to compare to his World-audio-design KLS3. We did a comparison back last year (before I was using the AR-XO) and mine were terrible compared to his, with naff all bass and vocals that didn't project at all (high frequencies were ok though). It was obviously a lack of baffle step that was causing the problem (my listening room is a tiny box compared to his), so hopefully we will notice a big improvement this time. I'll report back the verdict, good or bad...

8th March 2013, 09:11 PM	#89
fatmarley diyAudio Member Join Date: Aug 2003 Location: Cheltenham	Yes it's been a bit of a steep learning curve but i'm enjoying it and it's quite satisfying when the speakers start to make nice sounds. Ok you've twisted my arm - I've been playing with LspCAD and modelled a crossover at 2k. Frequency response is very similar to the current one and phase tracking looks good, so i'll build it and report back... Oh, yes the tweeter Fs is 750Hz and it's a soft dome (not ceramic).

9th March 2013, 02:41 AM	#90
AllenB diyAudio Member Join Date: Oct 2008	I used to run Vifa and Scan-Speak domes within the first octave of their resonance using just a capacitor (and peak compensation/padding resistors) for a number of reasons, some of which I don't believe in the same as I used to but I know it can be done with care. If you do go down to 2k, you'd be able to (or need to) go to a larger woofer. The roundover on the waveguide was suggested to hold its pattern lower and may help even if you stayed at 3k, but would become more necessary below that. It may be worth designing the crossover using your 30 degree response plots.