Designing an 8" 2-way using the SB23NRXS45-4 and SB29RDC-C004

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
We may have to agree to disagree on this one, but I don't see any practical point in using an arbitrary power value of 100W to decide if the current flow is excessive for the amp.
I agree. But I do this for two reasons that may not be obvious...

1) The current graph on XSim does not refine itself below 1 amp per graticule. So we spike up the power to get a useable scaling on the graph. If it had a 50ma and maybe a 10ma graticule I'd be perfectly happy at the standard 1 watt level.

2) XSim doesn't recognize driver overloads or potential damage... so we have to do this efficiency monitoring step manually.

I think both should be fixed because preventing these excessive and sometimes unexpected currents gives us much better design outcomes.

In this context, I accept that speakers are inherently inefficient. I am all for diminishing this to as little as possible but I am also willing to live with the inefficiency if it gets my drivers to the target responses I want for them and therefore a better sounding speaker.

Speakers yes... crossovers no. A crossover, properly designed, can pass all of the amplifier's current to the drivers with only trivial losses. I will agree that some inefficiency can be acceptable but to be honest, I would choose a different driver before letting it get out of hand.

I don't know if you've had much to do with Infinity speakers but some of them are known as amp killers with crossovers so complex their impedance dips into fractions of an ohm. So yeah, it's a real concern.

Thank you for asking... and thank you for an interesting rebuttal.
 
Last edited:
@Douglas Blake and jReave. I just wanted to pop in and thank both of you for sharing your thoughts on crossover design using Xsim.:D This is good stuff! Please keep it going even if it means starting a new thread. I’m hungry for this kind of knowledge.

Best Regards, and be safe
Rich
 
@Douglas Blake and jReave. I just wanted to pop in and thank both of you for sharing your thoughts on crossover design using Xsim.:D This is good stuff!

Now that you mention it we are sort of hijacking someone else's thread here... If you want to ask questions maybe start a new thread and I will see what I can do to answer them... hopefully our friend @jreave will join us.
 
Last edited:
Hey Rich, good to see you around again. I hope that you and your family are well and getting through this crisis comfortably and safety and without too much difficulty or hardship. I hope you are getting some good time in keeping your ears happy too. :)

Happy as always that someone is following along and getting some usefulness out of the strange musings that go on inside my head. By all means start a new thread if you have some further questions on the topic. I added another thread on this same topic to the stickied thread "Favorite threads and posts..." located up top.

All the best and stay healthy.
 
Back on topic, I took some more time to have another go at the xo and attached is about the best I could come up with. Still a few tiny little problems but hey, what can you do?

This time I added in some target curves (in green) for an LR4 xo at 1600Hz which is nicely acieved this time with 3rd order electric on the tweeter and 4th order electric on the woofer. Tweeter polarity is now reversed. Phase alignment is perfect right at the xo point but diverges a bit as you go plus and minus in frequency.

Current draw and watts dissipated are with 40W this time corresponding to where xmax is achieved on the woofer and producing about 103dB at 1m. The current draw is high at about 100Hz and through the tweeter inductor L3 going to ground but remain under the max 5amps that you can get from a 100W amplifier. Not perfect but better than last time.

The 1 resistor in the tweeter circuit actually sees more power than one would think so if the speaker is actually going to see these kind of SPL levels, doubling up resistors in parallel to achieve more power capability might be a good idea.
 

Attachments

  • XSim xo SB23 + 29.jpg
    XSim xo SB23 + 29.jpg
    24.6 KB · Views: 232
  • XSim FR SB23 + 29.jpg
    XSim FR SB23 + 29.jpg
    192.7 KB · Views: 233
  • XSim Z SB23 + 29.jpg
    XSim Z SB23 + 29.jpg
    72.2 KB · Views: 228
  • XSim A SB23 + 29.jpg
    XSim A SB23 + 29.jpg
    64.8 KB · Views: 210
  • XSim W SB23 + 29.jpg
    XSim W SB23 + 29.jpg
    62.6 KB · Views: 219
Hi jReave,
It is so great to hear from you! Hope all is well with you and your family. This crisis has not impacted my life to much. Well, it has delayed my Parts Express order. I’m trying to get a few crossover components to go along with existing parts I have on hand in order to get these 3 way stand mount speakers playing soon. My patience is running out. I have lost count on how many sessions of shellac I have applied to these cabinets. It is time to call the finish good enough and begin soldering some components.:)
Give me a day or so and I will post photos of these speakers along with Xsim docs so you and @Dougas Blake can critique my design.:D It will be fun. I really don’t have specific question(s) in mind other than continue the movement you guys have taken in pushing into more advanced Xsim uses. I really appreciate the crossover design philosophies you both utilize. I’m sure I will have some questions come to mind once this discussion commences.

Best Regards,
Rich
 
Thank you. As I suspected it's resonating quite strongly at about 2.7khz

Load up the DXO, open the current panel and get curves for R5 and L3... The thumbnail shows what happens if you open C3 or R1. Note the scale on the current plot... Yep, that's 45 amps.

I gotta say, that's a first for me.


What does this have to do with crossover design, other than as a warning not to run a crossover without the driver hooked up?
 
What does this have to do with crossover design, other than as a warning not to run a crossover without the driver hooked up?

Don't know for sure, just yet...

Any unloaded low or high pass LC filter becomes a series resonant circuit, and eddy currents between the two components can be quite strong... In this case the resonance is strong enough to actually affect current in the filter that is not coming from the amplifier and I would imagine that would have a deleterious effect on sound quality.

I took my message about that down, while I figure it out...
 
Sorry for the confusion ... it took a bit to suss this one out.

The first thumbnail shows what I was obsessing about. It shows that c2, l3 and c3 were very close to resonance and L3 was actually handling more current than was being taken from the amplifier. In my radio days we would call that "ringing" or "near resonant" behaviour and it represents some degree of instability that I was thinking might actually be audible in a speaker.

The second thumbnail shows a re-worked tweeter circuit that doesn't ring... I also added a little RC filter to even out the very high frequency response a bit.

There's nothing particularly wrong with @JReave's design ... it's just that I'm not fond of parts getting a mind of their own :cool:
 

Attachments

  • Resonance.JPG
    Resonance.JPG
    185 KB · Views: 226
  • 8-2way v3.JPG
    8-2way v3.JPG
    336.7 KB · Views: 217
  • 8-2way v3.dxo
    54.2 KB · Views: 30
A little confusing. Isn't resonance a normal part of a crossover? Do you indicate a difference between resonance and ringing?

High and Low pass LC filters are series resonant circuits, when operated with out a load and they will behave like one given the chance. (See the thumbnail)

Resonance in a series LC circuit is characterised as a very low resistance (fractions of an ohm) and the circuit itself appears as a resistor to the other components around it.

Below resonance it appears as a capacitor
Above resonance it appears as a coil

Near resonance this can fluctuate ... or ring.

This instability could be audible in some cases and you don't want the crossover getting a voice of it's own and colouring the sound.
 

Attachments

  • series resonance.JPG
    series resonance.JPG
    158.9 KB · Views: 213
  • series resonance.dxo
    2.1 KB · Views: 26
Last edited:
Douglas, I generally don't take any xo seriously when it is only showing system phase and not the individual driver phase. I think this should be especially paid attention to in this thread because it was one of the OP's original concerns.

The fact that your woofer response is actually greater than the summed response around 1000Hz tells me that phase alignment between the 2 drivers isn't actually very good. If you have a thing about current, I have a thing about phase alignment. Also the tweeter response doesn't look like it is maintaining the 4th order roll-off any more although it's a little difficult to tell without the target curve added in. Again, asking as little as possible from the tweeter below about 1600Hz I think is vital to the success of this design.

But I like the way that you've used the parallel capacitor in the tweeter circuit to tame the slight rising response. Nice.

Allen, yes indeed, my files include baffle effects on both the the woofer and tweeter as well as a modified woofer impedance response for the woofer in a vented enclosure, so a full 6dB of baffle step loss although I'm not 100% sure that is what the OP will need.

For christiank24, I keep forgetting to add this in for you.

To trace FR and impedance files try this: FPGraphTracer : fprawn labs

And to add in baffle and box effects to your files try Response Modeler since you already have Excel.

This thread may help in those respects as well: So you want to design your own speaker from scratch!
 
I had a quick look at this too. I totally agree with jReave on phase, the little experience I have tells me that proper phase tracking is very important for sound and depth perception. Without measurements of drivers in the actual baffle, a suitable XO is not possible to get right. IMO the baffle should be stepped or possibly tilted to align the acoustic centers.

I would think crossing the tweeter low (for directivity reasons?) will give distortion problems at anything more than moderate levels, and possibly blown tweeters if played too loud. I think I would rather sacrifice the directivity and cross the tweeter higher, or add a midrange. I heard some 8"+1" speakers that sounded really nice even if the XO was 'theoretically' too high, but I guess it depends on the drivers dispersion, the room etc. EDIT:Forgot to mention waveguide, as mentioned before, a pretty neat solution to aligning the drivers, directivity, and also power handling for crossing low.

When it comes to resonances, I think an important aspect (maybe most important), especially when crossing the tweeter low, is the resonance frequency of the tweeter. The higher the 'lookback impedance' of the tweeter is in the resonance region it will let the tweeter resonate more freely at it's resonance frequency, and this can sound really bad in some cases. Lookback impedance can be corrected with LCR, or by the topology in the filter. Also a high Q in a filter is something I try to avoid.

There are also problems with woofer resonance when crossing the woofer low. It starts interacting with the series coil in the XO causing a bump in FR in the region above the woofer Fs (in box). I recently had some good results with bass quality when I tried LCR to flatten the woofer impedance peak. Takes large coils and caps though..
This can easily be observed in Xsim if you look at the bass response with large coil and with shorted coil. Funny thing is that the simulated diffrence looked small, but the perceived improvement was significant in my case.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.