My Current Project...(Teasers...)

[Wolf]

Bottomless trap (or a type of series notch) = LC across driver.
Infinite notch or tank (or parallel wired LC) = L||C in series with driver.

The tank has a way of allowing upper treble ripple through as the cap is also in series with the woofer. Usually, it's easy to suppress the ripple in the audible band. This does not mean there is no output above there ultrasonically. We just can't see the effects above 20kHz in most modelers. This can cause issues in some amplifiers as it might cause ultrasonic oscillation. Of all the amplifiers I've heard of others having over the years (solid state), I can count on one hand the amplifiers that have been reported as having problems. The normal fix is to have a 4-5 ohm resistor in series with the cap that is across the low-pass coil. It affects the tank's notch function minimally, and keeps the amplifiers in question happy. If you have a standard second order with the tank, and there is already a damping resistor in series with the shunt cap, usually the other resistor can be omitted. Just try to keep a circuit that does not trace; in - cap - cap - gnd. This will avoid the problems.

Glad this has been of assistance,
Wolf

[mattp]

I didn't post a complete x-over to try. (Yet) Here's something to look at. Here's your current xo, and my xo that uses a lot more parts. There's a voicing difference here that would be very significant. One is a bit laid back, (yours), the other is rather aggressive (mine). It's tough to know which is preferred, but once you listen, and see measurements of what you listen to, you get a pretty good idea where you want the response.

rpb,

There is a pretty significant difference there for sure. I like the added efficiency you show. I haven't used PCD in years, can you draw up a schematic and post? I may give something like this a try before I commit to anything. Weighing all options is never a bad thing.

[mattp]

Well after some fiddling today I made one minor change to the current setup, I upped the woofer L value to 3 (3.2 measured, you were right about those Jantzen coils Wolf...) which brought the break up down and along with the parallel cap (is that 0.33 a tank filter, or is that only filters that go to ground?) it flattened out and peaks at about 56 dB now. This did bring the theoretical XO point down a smidge but I am confident these tweeters will be fine.

FYI if anyone sees anything that stands out as blatantly stupid in any of my sims please let me know.

I think I will add a little more damping to the cabinets before its all said and done but after sitting here for a couple hours already tonight I really don't see an immediate need for any new changes. The room and my electronics I think are the biggest limiting factor. I need to rethink my wall treatments and add some bass traps. That will happen once the subs are done.

Stay tuned. More updates on the stands soon, then the subs will start. Is it bad when I have 3.5 sheets of MDF in the shop for two 2 cubic foot subs and I am worried it may not be enough?

Where is Joey Butts and all his ridiculous bracing when I need backup here with all my overkill lol.

[rpb]

The latest measurements I posted are the gated unmodified measurements. John has posted his corrected measurements using those files.



Sort of... I have a couple bins of XO parts up here I can pick from but I don't always have the values I need. I make due. I have gone through probably 5-6 variations of designs thus far, including my own, Johns and yours and even a bit of a combination and have been slowing tweaking here and there.



Yes I am going to do external crossovers. I got tired of removing woofers a million times to change xo's. This is so much easier.

The number of designs that have been posted is overwhelming. The biggest pain has been lugging one of these back and forth to the lab from home 4 times now. They are not small and they are not light...and they are fragile lol. I have bumped into my share of doorways etc in the past and nicked up boxes. I would rather not do that with these.

The latest I mocked up last night really brought them to life. Phase looked good, breakup was fairly well controlled and the XO moved up to about 1700. I will probably listen to this for a couple days and see if there is anything else I can do to improve more. The only changes I might make are to that breakup if it seems to be an issue...but so far I think it's ok. Time will tell, I had a couple hours on them last night.

I never expected so many designs to come forth. This is pretty cool, a million different ways to get near the same result. I wish I could mock them all up and give them each a week long session but that is just not practical. I appreciate everyone's input so far. I have learned a lot during this build that I didn't realize I didn't know...

Once I make a final design decision I will bring them back for final measurements. I will also take in-room measurements when the whole system is done and share everything.

This is what I have as of now. The breakup is the only thing that has me slightly worried as Wolf pointed out before. I really don't think I notice it but I still might try to get it lower just to be on the safe side. It could very well be audible and I just don't know what I'm listening for. Nothing stands out as obvious or "wrong" at the moment during listening tests. Image is absolutely pin-point and the sound stage is massive. I couldn't believe the difference after the first few mock ups.

Measured:



Modeled:

I didn't post a complete x-over to try. (Yet) Here's something to look at. Here's your current xo, and my xo that uses a lot more parts. There's a voicing difference here that would be very significant. One is a bit laid back, (yours), the other is rather aggressive (mine). It's tough to know which is preferred, but once you listen, and see measurements of what you listen to, you get a pretty good idea where you want the response.

[Wolf]

I think the Ceramics are wonderful sounding on their own with the right xover. I would not coat them if it were me.

I've even used notches similar to these on composite sandwich cones, to reduce the breakup as they are very stiff- even if well damped.

The only part values that enter into the ultrasonic range here would be the tank caps. The others are being low-passed and shunting breakup within the audible band.
I agree that a resistor on the tank caps might be beneficial, I just did not model them in. I do not feel the shunt legs in this simmed network will offer any negative effects on their own as used.

I knew your intentions were likely valid and true about how you felt, and that you likely had good reasons for them. However, I could not see anything wrong with what I did. Tanks are a bit troublesome ultrasonically at times, as the caps will pass signal above their value point which means ultrasonics. Some amps do not like these, but I've not come across one of my own that can't cope with their usage.

It does however give me pains that you thought of me as a know-it-all and not someone whom is willing to learn of something he does not know. I've been wrong before, and will likely be again. I do strive to be as accurate as possible. I try to learn as much as I can about this stuff- I love it, and the camaraderie.

Take care, Fred,
Wolf

[fdieck]

I looked at it more closely it probably is fine. I just don’t like circuits that present complicated loads at RF frequencies since high bandwidth amps, chip amps without out RC damping networks, and class D amps can sound bad or oscillate with complicated loads like this. I would put a few ohms in series with the caps on the parallel LC networks at least. I used to do EMI suppression design at work and mod and design power amps so this is not just paranoia. You might even even try a 5 ohm in series with 0.1 uF network across the input to the crossover to provide the amp with resistive load at high frequencies.

PS I put damping material on aluminum cones. I can’t even listen to them with the trap circuits I have tried. The SB acoustic “ceramic” driver really benefited from cone treatment. Rudy Bozak had sense enough to put latex paint on aluminum tweeters over 50 years ago.

[Wolf]

Whatever.... There no point in arguing with people that just don’t want to learn anything. If you really think that is a valid crossover because it looks good in a simulation program have fun. I started not to post it since I figured it would fall on deaf ears.

I never said I didn't want to learn anything. I just don't see your problem with it. The LC notch (trap) across a driver is swamped by the inductance of a lowpass coil. You also have the R factor of the coil involved. The phase swing is minimal, and only dips to -45 in one place and it's not a bad impedance in that region. What I did is simple, I paired a broad-band notch with a centralized notch in the same bandwidth, and the low pass coil increases the net impedance so it does not tend towards zero + DCR. It will never get that low. A lot of amps do not like the tank + trap arrangement, yet it exists, and used heavily for enough suppression.
So- what don't I know here? I do know that tank + trap will effectively suppress aluminum breakup, and I've done it several times.
Please- school me,
Wolf

[mattp]

Minimalist, updated for new files, break up pushed down, 1.8K x-o, decent phase tracking, impedance etc. The dip at 3.1K is likely the faceted baffle. I'd check off axis before investing more parts

John,

Yes that dip is the baffle and it does smooth out off axis. I might play with third order on the woofer next, I feel like I am getting really close.

[mattp]

Do you still have the measurements? Change that gating to before the first reflection, and post them. They will be much smoother. The drivers I measure are not plus or minus 3dB in the 400hz to 1k range.

The latest measurements I posted are the gated unmodified measurements. John has posted his corrected measurements using those files.

Can you borrow xo parts, and try filters, and measure them?

Sort of... I have a couple bins of XO parts up here I can pick from but I don't always have the values I need. I make due. I have gone through probably 5-6 variations of designs thus far, including my own, Johns and yours and even a bit of a combination and have been slowing tweaking here and there.

I see dual binding posts in the pictures. Do you plan to keep the x-over external?

Yes I am going to do external crossovers. I got tired of removing woofers a million times to change xo's. This is so much easier.

The number of designs that have been posted is overwhelming. The biggest pain has been lugging one of these back and forth to the lab from home 4 times now. They are not small and they are not light...and they are fragile lol. I have bumped into my share of doorways etc in the past and nicked up boxes. I would rather not do that with these.

The latest I mocked up last night really brought them to life. Phase looked good, breakup was fairly well controlled and the XO moved up to about 1700. I will probably listen to this for a couple days and see if there is anything else I can do to improve more. The only changes I might make are to that breakup if it seems to be an issue...but so far I think it's ok. Time will tell, I had a couple hours on them last night.

I never expected so many designs to come forth. This is pretty cool, a million different ways to get near the same result. I wish I could mock them all up and give them each a week long session but that is just not practical. I appreciate everyone's input so far. I have learned a lot during this build that I didn't realize I didn't know...

Once I make a final design decision I will bring them back for final measurements. I will also take in-room measurements when the whole system is done and share everything.

This is what I have as of now. The breakup is the only thing that has me slightly worried as Wolf pointed out before. I really don't think I notice it but I still might try to get it lower just to be on the safe side. It could very well be audible and I just don't know what I'm listening for. Nothing stands out as obvious or "wrong" at the moment during listening tests. Image is absolutely pin-point and the sound stage is massive. I couldn't believe the difference after the first few mock ups.

Measured:



Modeled:

[jhollander]

Note when calculating the offset remember not to mix the as measured and minimum phase. There are multiple files posted in this thread. It's interesting to note that the as measured and minimum phase files have close to the same offset.

[jhollander]

Why speculate? The T & W file is included, you can calculate the offset yourself, both in PCD and XSim.

[rpb]

Wolf you should add in the woofer offset to get the phase tracking right. The little m next to the driver S1 would indicate it's modified. My sims use a woofer offset of 1.39 inches in XSim

That offset seems excessive for a 5.25". I would speculate that it is half that.

My previous xo works well with the new files, although I changed some values to make the mid-bass measurement look smoother at what I think may be the splice point..

[rpb]

Can you borrow xo parts, and try filters, and measure them?

[rpb]

Sorry I forgot that part. Yes they are gated just above 6ms.

Do you still have the measurements? Change that gating to before the first reflection, and post them. They will be much smoother. The drivers I measure are not plus or minus 3dB in the 400hz to 1k range.

[fdieck]

The load is just fine, as the swing is minimal. You can see this in the system impedance plot. What I've done here has been done many times before.

John-
This is exactly how I designed the Zingers, and they are just fine. Single point in space, and 2 measurements on one axis.

Later,
Wolf

Whatever.... There no point in arguing with people that just don’t want to learn anything. If you really think that is a valid crossover because it looks good in a simulation program have fun. I started not to post it since I figured it would fall on deaf ears.