Null and Void at 630 and it Hertz!

[toddalin]

Ahh, the ever continuing saga and quest for a flat, pleasing sound.

Those who have been following my "Keeper" crossover threads know that I've been fighting a (loosing) battle to get some sembalance of a flat frequency response, especially through the critical vocal range. To this end, I built crossovers that should be an ideal match for the components.

When that failed, the woofers were redone (less than 2-year old authentic JBL cones and remagnetized last Monday) and test out perfectly normal as per a 2235 specs using Woofer Tester-2. Nice strong BL factor too now (23.5)! Thanks for coming by last night Dave (Grumpy), so we could test this out. ;)

The problem appears to be a null at about 630 Hz. The null is down about 11 dB from the average signal and is fairly sharp in contour. :biting:

Thinking it may still have been the crossovers, they were rewired from their original configuration (smaller Zobal cap and bigger Zobal resistor) to pass more highs to no real avail. The 11 dB null persisted. :banghead:

The crossovers were rewired again changing both caps (from 24 mF and 33 mF on the Zobal per JBL N200B specs to 15 mF and 24 mF on the Zobal) again to no avail. (Still have the same null.)

OK, lets' be really sure it's not the crossover. When the woofer is connected directly to the speaker tap..., a 10-11 dB null at 630 Hz. :banghead:

This is smack dab in the critical portion of the vocal range. In fact, when playing pink noise, (and looking into the void), as soon as you speak, that 630 band comes on nice and strong. :blink:

We can only conclude that it is not the crossover, but some manifestation of the woofer in its environment. (Yes, I made sure that all eqs were off.)

Does anyone have any ideas on trying to determine what could cause such a drastic "suck-out" right at about 630 Hz? :help:

[toddalin]

That's weird?? Wave length at 630Hz is 21 1/2 inches. So half wave is 10 3/4

You have anything at 32" from them or some multiple where you could get a strong reflection to null it out??

What happens in the middle of the room??

How high is the center of the woofer off the floor?? If it was 16" off the floor there and back would be 1 1/2 wavelength so it would be 180 or out of phase.

Just a shot in the dark that's quite a null. That's kinda high to be a room placement issue.

Rob:)

Don't think so. :blink:

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/room2.jpg

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/room1.jpg

[toddalin]

Well maybe you do. The L300 cabinet isn't all that much different with the woofer spacing with a L200 if I remember right. Your are on pedestals that look higher than stock. What do you get when you measure them?? Figure there and back you only need 16" off the floor.

Could simply not be the issue but it couldn't hurt to check.

Rob:)


? = ~14"

I really don't think that's the problem.

[toddalin]

I connected the Behringer directly to the 2205/2235, turned the pink noise up to -13, laid the mic on the floor and the speaker pretty much exhibited the classic 2235 response pattern. :) Surely I was on to something! :hmm:

This then led me to believe that there could be a problem with the Yamaha (RX-Z9 http://www.yamaha.com/yec/products/productdetail.html?CNTID=200001&CTID=5000300) invoking it's own sonic signiture, even with all eqs turned off. OK, possible progress! :bouncy: This would make sense because I replaced the 130A shortly after replacing the A1 with the Z9.

I dragged out the old Yamaha (DSP-A1 http://www.yamaha.com/yec/products/productdetail.html?CNTID=200321&CTID=5002600) and ran the Bnger directly into the amp input by-passing all processing.

But no... As soon as I raise the mic off of the floor..., there's that drop-out in the 500-650 range. I've tried pillows on the floor and around the speaker, again to no avail. :banghead:

So, as long as the mic is on the floor, no problem (except it doesn't pick up the tweeter).

I guess I'm just going to have to lay on the floor when I listen. Got to be a better solution. :blink:

[grumpy]

yeah... thought mic on the floor might separate "floor bounce" issues from other effects...
I don't really want to haul a 4430 over :p (yet).

goofing w/ parametric EQ in the Behringer -did- put back energy in the nulled area;
wasn't in an effect or tape loop, so only noise was listened to in this mode.

Still wonder about the L200 cab... Q measured pretty high IIRC, & rear panel cancellation
distance would be about right ( case of 'glass being 1/2 full? :D ).

[Tim Rinkerman]

I wouldn't trust what a Behringer device was telling me if my life depended on it....

[toddalin]

I wouldn't trust what a Behringer device was telling me if my life depended on it....

The Bngr only confirms what the ear doesn't hear (at least in this case). The vocal suck-out is very obvious as Grumpy can atest. :banghead:

[edgewound]

Shove a big 'ol slab of R30 directly behind the woofer in the cabinet to see if that mitigates any rear cabinet wall cancellation. Is the rear cabinet wall rigidly braced?

[toddalin]

yeah... thought mic on the floor might separate "floor bounce" issues from other effects...
I don't really want to haul a 4430 over :p (yet).

goofing w/ parametric EQ in the Behringer -did- put back energy in the nulled area;
wasn't in an effect or tape loop, so only noise was listened to in this mode.

Still wonder about the L200 cab... Q measured pretty high IIRC, & rear panel cancellation
distance would be about right ( case of 'glass being 1/2 full? :D ).

If this is related to rear panel cancelation, there is something that can be done. Originally the crossovers (or something) bolted to the back brace. There are still screw holes there. A piece of thin plywood could be bolted in place and would only slightly reduce the overall cabinet volume. The reflection distance to the back of the woofer would be altered slightly.

[toddalin]

Shove a big 'ol slab of R30 directly behind the woofer in the cabinet to see if that mitigates any rear cabinet wall cancellation. Is the rear cabinet wall rigidly braced?

Typical JBL L200 cabinet. There is a brace behind the woofer that runs side-to-side along the back wall. There are also front-to-rear braces along the side walls.

[grumpy]

Shove a big 'ol slab of R30 directly behind the woofer in the cabinet:yes:


The vocal suck-out is very obvious as Grumpy can atest:yes:


I wouldn't trust what a Behringer device was telling me if my life depended on it.... Works well enough as a "for indication only" device. I've calibrated a B-mic vs a $$
Earthworks measurement mic and it's certainly flat enough to represent the wide-ish null in
the FR around 600Hz. ... and my life doesn't depend on it. :D -grumpy

(got started late & the main test was woofer BL after re-mag... other equipment was left @home.)

[toddalin]

Shove a big 'ol slab of R30 directly behind the woofer in the cabinet to see if that mitigates any rear cabinet wall cancellation. Is the rear cabinet wall rigidly braced?

If this is the case, wouldn't we also see the suck-out with the mic lying on the floor? I mean, wouldn't it be there at any mic position?

Also, it is manifested in the center channel cabinet, and that's a completely different cabinet (which is another reason I was thinking it may have been the Yamaha).

[4313B]

I don't really want to haul a 4430 over :p (yet).Bring over a regular 2235H and bolt it in his boxes if you're so inclined and test it.
I guess I'm just going to have to lay on the floor when I listen. Got to be a better solution. :blink:You're incredibly persistent, I'll say that much. I would have bagged the whole mess, sold a vette and bought a pair of Everest II's instead. You have what appears to be a really nice room and you would probably do well to invest in today's equivalent of the Hartsfield.

[edgewound]

If this is the case, wouldn't we also see the suck-out with the mic lying on the floor? I mean, wouldn't it be there at any mic position?

Also, it is manifested in the center channel cabinet, and that's a completely different cabinet (which is another reason I was thinking it may have been the Yamaha).

Todd,

Please allow me to be perfectly frank with you.

You now have a reasonably modern foundation to build a very capable system around your "2231" woofers (2205 core + 2235H recone kit).

It seems to me you still have LE175 mids and those god awful sounding(by today's standards) 075/2402 ear bleeding bullets that will never get you a pleasing high end response unless you have total HF hearing loss and need that shrill tweeter to get any semblance of "treble".

Do yourself a favor if you're going to stick with the furniture that you have.

"Upgrade" to a 2425/2426 or LE85/2420/2421 on an "OSAR" waveguide horn and a 2404H-1 or 2405 tweeter and see if that helps. To be perfectly honest with you, you are completely wasting your time with the mids and highs that you've got. While you're at it....build a crossover that that would be somewhat like a 4333. Somewhere here you have to let go of all the frustration that you are experiencing with what you've got and send good money after good components without trying to reinvent the wheel. All the tools are right here for the taking.

Get rid of those "bullets". Somembody on eBay will take them off your hands to mate with a D/K/E130 PA system.

Sorry to be so blunt...I want you to enjoy the music....not bang your head against the wall.

[4313B]

:yes:

Or just ask Zilch about the 2452H-SL and his favorite waveguide.

[johnaec]

Just for grins, what happens if you drag the speaker out in the middke of the room, (or better yet, outside), lay it on its back, and check the response with a mic hanging over it? 'Same suck-out?

John

[toddalin]

Just for grins, what happens if you drag the speaker out in the middke of the room, (or better yet, outside), ?

John

Been tempted and very possible.

[johnaec]

Been tempted and very possible.Well, (especially outside), that would immediately nail down if the problem is with the cabinet/speaker, as compared to room/placement.

John

[toddalin]

A breakthrough last night.

I decided that the 1.0 mF/0.5 mH choke/1.5 mF on the 075 creating a double hump that was not too my liking and I thought it sounded better with just the single cap and a more natural roll off. After much experimentation, I decided that retention of the 1.5 mF cap provided the smoothest transistion from the tweeter to the horn.

To wit, the tweeter circuit will first pass through the first capacitor of the horn circuit so that both the horn and tweeter signal will pass through two caps and maintain the same phase angle. The 5 and 7 ohm resistors that were left over are used in attenuation for the 075 so that the L-pad has a workable span. :D

The 4.3 mF cap was removed as the horn low pass and replaced with 2.5 mF; 1.5 mF that I removed from the horn high pass, and 1.0 mF that was used in the tweeter "cascade." This put some of the highs back into the horn without that ungodly peak at 10kHz that mixed with the tweeter creating a "super peak." :applaud:

The removed 4.3 mF was added to the 16.4 mF (20.7 mF total) that make up the horn high pass (replacing the 1.5 that had previously been placed there [for a 17.9 mF total] and now used in the horn low pass). This addition paid BIG DIVIDENDS and the horn now easily goes down to 800 Hz (over a 3 dB increase in this range and is pretty flat down to 900 Hz) whereas before it was cutting off at about 1,000 Hz. This substantially reduced the bandwidth around the suck-out and does fall within a portion of the critical vocal range. :D

Today, I'll again try to switch main/Zobal the caps on the woofer to try to get a little more top end out of it and work with the tweeter attenuation resistors to best dial them in. Then, I'll throw a big piece of insulation in them and probably call then "good." :bouncy:

[grumpy]

glad you're on to something :) ... went back and looked at data from Sept.
(couldn't imagine we wouldn't have tested this... 4313B... thanks for prodding
me to look)

Found plot of my 2235H in Todd's L200 cab but w/ 3110 as crossover.
No 600Hz suckout. At any rate, "if it makes you happy, it can't be that bad"
:) -grumpy

[pelly3s]

just for the hell of it did you try to invert the phase on the mids to see how the crossover points react differently. it may be to your advantage. Hey its a shot in the dark but it might work

[toddalin]

just for the hell of it did you try to invert the phase on the mids to see how the crossover points react differently. it may be to your advantage. Hey its a shot in the dark but it might work

Actually, yes, and I'm finding that things are overall flattest (both to the woofer and tweeter) with the mid inverted. The trough where the woofer meets the mid is about 1 dB deeper this way, but other improvments of over 3 dB are noted just beyond the trough.

Also, the overlap with the tweeter is smoothest like this. ;)

[Zilch]

When you lay the mic on the floor in a room that size, you're in essence doing a ground-plane measurement. That suggests to me that the speaker is behaving itself. Something else is up.

4313B may have some insight as to what's going on. I've never had enough room to work with that technique.

There's a section in one of the threads where he and I pursued the 600 Hz notch with nearfield measurements.

I wish Todd had kept the whole deal in one thread; it's too difficult to track the ongoing saga this way.

[This from the forum's foremost blogger.... :p ]

[toddalin]

glad you're on to something :) ... went back and looked at data from Sept.
(couldn't imagine we wouldn't have tested this... 4313B... thanks for prodding
me to look)

Found plot of my 2235H in Todd's L200 cab but w/ 3110 as crossover.
No 600Hz suckout. At any rate, "if it makes you happy, it can't be that bad"
:) -grumpy

Grumpy, if memory serves, this was the 3110 on the 2205/2235. And I though we determined that the crossover may have been bad.

In any event, we see the suck-out even with no crossover, so maybe it was a matter of mic placement?

At this time, depending on mic placement, I can minimize the appearence of the suck-out, but then get the dip at ~450 Hz as shown on your curve.

[toddalin]

Shove a big 'ol slab of R30 directly behind the woofer in the cabinet to see if that mitigates any rear cabinet wall cancellation. Is the rear cabinet wall rigidly braced?

Is R30 the best to use to this purpose? Is that the itchy 'glass used in attic insulation? I may also wedge a brace (2"x2" board) between the front and rear walls if you think this could be the problem.

[edgewound]

Is R30 the best to use to this purpose? Is that the itchy 'glass used in attic insulation? I may also wedge a brace (2"x2" board) between the front and rear walls if you think this could be the problem.


Yep...that's the stuff. A coat of hairspray or some spray glue like Super 90 can help keep the fibers from dusting out.

The virtual volume of the R30 will make the cab think it's a little bigger.

Bracing the cab won't hurt either.

I still think your problem lies with the mid and hi drivers integration with that woofer.

JBL networks weren't always designed with electrical crossover points in mind. If you do some calcs with the JBL Network components values you can end up scratching your head.

[toddalin]

JBL networks weren't always designed with electrical crossover points in mind. If you do some calcs with the JBL Network components values you can end up scratching your head.

Agreed, the ERSE crossover calculator comes up with totally different values than used in the JBL N200B.

The thought was that JBL tailors their crossovers to the actual characteristics of their components, but lot of the old stuff was produced before modern modeling techniques and may have been based on someone's own ideas on what sounded right coupled with parts availability at the time.

Is it even possible that the difference in the size/layout of the magnetic structure on a true 2235 vs an AlNiCo "quasi-2235" could change the internal way in which the sound comes off the back of the speaker and interacts with the cabinet internals?

[4313B]

Agreed, the ERSE crossover calculator comes up with totally different values than used in the JBL N200B.

The thought was that JBL tailors their crossovers to the actual characteristics of their components, but lot of the old stuff was produced before modern modeling techniques and may have been based on someone's own ideas on what sounded right coupled with parts availability at the time.
It's been that way up until very recently. G.T. has often spoken of his network cart with tons of capacitors, inductors and resistors of varying values. The general consensus is that JBL wasn't real strong in filter design until guys like G.T., M.G., D.S., and D.B.K. showed up.

[grumpy]

if memory serves, this was the 3110 on the 2205/2235. And I though we determined that the crossover may have been bad.

could have been 2205/2235 ... I was going by the annotation I put on that
plot, which I trust more than my memory of the event :).
Overall throughput on that 3110 was low for some reason... I'll open it up
and look this week, but the FR looked "normal" otherwise. Time plot (not
shown) also indicates a ~1m measurement distance.

If your setup has stabilized, let me know and i'll bring both a 2235H and
the measurement system we used before. -grumpy

[toddalin]

I've been doing more experimentation with the crossover from the center channel that I can then go back and compare to the one in the cabinet.

The biggest factor in all of this is mic placement. I've determined that 900 Hz is the peak when the horn and woofer combine and the horn is connected out of phase. Connected in phase, this becomes a trough. But other frequencies also change for the better and worse with phase change. Still, the 900 Hz frequency appears to most affected by phase.

When the mic is placed ~1 meter from the speaker, I get a flatter frequency response removing capacitance from the horn. When the horn and woofer combine, 900 Hz becomes the major peak and the removal of capacitance (from 16.4 to 8.2 mF) totally flattens the peak and brings the horn right to the level of the woofer. This produces the flattest curve..., at about 1 meter.

But, when the mic is moved to the seating area, the 900 Hz peak is fairly level with the rest of the bands, (though there are major dropouts here and there).

In this position, the horn need more capacitance, and the 16.5 value (JBL spec) is almost ideal.

While an additional 4.3 mF on the horn (20.8 mF total) extends its lower range slightly, it also exascerbates the 900 Hz peak.

This is all a compromise.



When I place the mic on the floor

[toddalin]

could have been 2205/2235 ... I was going by the annotation I put on that
plot, which I trust more than my memory of the event :).
Overall throughput on that 3110 was low for some reason... I'll open it up
and look this week, but the FR looked "normal" otherwise. Time plot (not
shown) also indicates a ~1m measurement distance.

If your setup has stabilized, let me know and i'll bring both a 2235H and
the measurement system we used before. -grumpy

I just don't recall actually putting your speaker in the cabinet at any point.

Stabilized? You mean finalized?

[grumpy]

I just don't recall actually putting your speaker in the cabinet at any point.

You're probably right.

A direct 2205/2235 vs 2235H comparison might still be of some interest to others. A final FR (1m and at listening position) would be a nice wrap up.



Stabilized? You mean finalized?

Yes, if such a thing is possible ;)

[toddalin]

You're probably right.

A direct 2205/2235 vs 2235H comparison might still be of some interest to others. A final FR (1m and at listening position) would be a nice wrap up.



Yes, if such a thing is possible ;)

No problem. I can finish the crossovers and leave one speaker out until we do it. Next week would be good.

[Ian Mackenzie]

Todd,

This has been a real saga....but your having fun!

As you mention earlier it could be the mic location and it could also be the displacement of the drivers and the topology of the network playing tricks in the vertical polar pattern.

If you can post a diagram wit measurements of the drivers in horizontal and vertical plane and the schematic I will try and model the system to account for the actual driver, the crossover and the phyiscal domain during the week.

Ian

[toddalin]

Todd,

This has been a real saga....but your having fun!

As you mention earlier it could be the mic location and it could also be the displacement of the drivers and the topology of the network playing tricks in the vertical polar pattern.

If you can post a diagram wit measurements of the drivers in horizontal and vertical plane and the schematic I will try and model the system to account for the actual driver, the crossover and the phyiscal domain during the week.

Ian

Cabinet is a standard L200 with one port blocked off, so has a slightly "laid back" array. The 2402 was added vertically above the horn. Center of woofer to center of horn is ~11". Center of tweeter to center of horn ~6".

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/L200-2235.jpg

Crossover schematic is included below:

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/Crossover_final2.jpg

[Ian Mackenzie]

Todd,

Is that grill frame or is the baffle inset?

Edit. I had a skim through the other threads and it appears the system has previously been modelled and measured. I am inclined to thinks its a reflection off a wall or even off the edge of the baffle

Ian

[toddalin]

Todd,

Is that grill frame or is the baffle inset?

Edit. I had a skim through the other threads and it appears the system has previously been modelled and measured. I am inclined to thinks its a reflection off a wall or even off the edge of the baffle

Ian

A 3/4" square wooden frame holds the grill cloth where the prior "window screen" used to be.

[Ian Mackenzie]

Hi Todd,

How deep is the frame and inset?

[Ian Mackenzie]

I'll post the preliminary simulation and model tonight.

Its quite revealing.

[toddalin]

Hi Todd,

How deep is the frame and inset?

~4.5"

[toddalin]

I'll post the preliminary simulation and model tonight.

Its quite revealing.

Is this something I should know before I button these up again? I.e., is there something that I can do to help the situation while I have the cabinets apart?

Thanks, would like to get them back together ASAP so can watch TV tonight.

Wife is in there now listening to 1/2 (left side) of a football game. ;)

[Ian Mackenzie]

Yes,

The data you pmk me shows there is a resonance mode issue in the second run that appears to be with the woofer in the box..specifically there is an impediance issue that is not as it should be. The impediance jumps from a gradual climb at 606 hz 9.49 ohms to:

10.88 ohms @606hz,
11.70 @654hz
12.20 @704hz
11.87 @761hz
10.65 @821hz
9.36 @886hz
8.44 @7.99hz

Because there was not documentation is difficult to know if the test was done without the crossover connected but it suggests you have an standing wave problem amongst other issues.

Have look at how they stuffed the 4430 and do likewise.


Todd, assuming no errors in the crossover the issue and we know you have reasonable confidence in the drivers this fairly broad band cancellation and those deep insets would add to the effect suggest strong evidence of a boz issue. This goes against what I said in your other thread but the modes and the impediance issue is too coincidental to ignore.

A few modes would not matter, but where you have a number of modes around the same general frequency it will show up.

There is also this post. With so many threads I have to say it nearly improssible to make any real sense of what is happening.

http://www.audioheritage.org/vbulletin/showpost.php?p=187083&postcount=36

[Ian Mackenzie]

Simulation:

Todd,

Here is the first draft of the model taking into account the woofer TL data from you WT files and the crossover posted above and the estimate acoustic offsets.

As usual the devil is in the details.

I amended the model for variations in the offset of the woofer and the horn by only an inch and it makes a big difference to the void around 900 hz where the horn and tweeter nodes are connected in phase. Out of phase is better depending on the offset.

I would like more accurate data of the woofer dustcap from the surface of the front baffle and the top plate of the compression driver from the front baffle.

Not included in the model is the actual compression driver impediance.

The 2235H is real data and quite accurate.

I loaded in 11 ohms impediance but below 1000 hz that will go to hell. If there is data of this driver loaded on this horn I can load the data in to the model and even edit the impediance model.

I suspect your 175 driver is possibly behaving differenctly at the low end and if it is shallower in depth it will mess up the crossover point. Try the LE85 or 2420 and see what happens. You might be surprised.

If the real data on the above points can be obtained the network can be quite possibly optimised assuming there are no other unknowns like internal standing waves.

As I point out though, this design is very sensitive to driver offset and that is the main issue. From a sound quality view point the 2344/4430 is far superior, ist has none of the above problems and you have all the design details to make it go first pop.

[toddalin]

Yes,

The data you pmk me shows there is a resonance mode issue in the second run that appears to be with the woofer in the box..specifically there is an impediance issue that is not as it should be. The impediance jumps from a gradual climb at 606 hz 9.49 ohms to:

10.88 ohms @606hz,
11.70 @654hz
12.20 @704hz
11.87 @761hz
10.65 @821hz
9.36 @886hz
8.44 @7.99hz

Because there was not documentation is difficult to know if the test was done without the crossover connected but it suggests you have an standing wave problem amongst other issues.

Have look at how they stuffed the 4430 and do likewise.


Todd, assuming no errors in the crossover the issue and we know you have reasonable confidence in the drivers this fairly broad band cancellation and those deep insets would add to the effect suggest strong evidence of a boz issue. This goes against what I said in your other thread but the modes and the impediance issue is too coincidental to ignore.

A few modes would not matter, but where you have a number of modes around the same general frequency it will show up.

There is also this post. With so many threads I have to say it nearly improssible to make any real sense of what is happening.

http://www.audioheritage.org/vbulletin/showpost.php?p=187083&postcount=36

Ian, thanks.

For the Woofer test, the woofer was not in the box or connected to the crossover. The woofer was in free air (held between my knees) for the test and driven from a laptop computer.

The two runs were just two back-to-back tests, with the second to confirm the first.

[toddalin]

Simulation:

Todd,

Here is the first draft of the model taking into account the woofer TL data from you WT files and the crossover posted above and the estimate acoustic offsets.

As usual the devil is in the details.

I amended the model for variations in the offset of the woofer and the horn by only an inch and it makes a big difference to the void around 900 hz where the horn and tweeter nodes are connected in phase. Out of phase is better depending on the offset.

I would like more accurate data of the woofer dustcap from the surface of the front baffle and the top plate of the compression driver from the front baffle.

Not included in the model is the actual compression driver impediance.

The 2235H is real data and quite accurate.

I loaded in 11 ohms impediance but below 1000 hz that will go to hell. If there is data of this driver loaded on this horn I can load the data in to the model and even edit the impediance model.

I suspect your 175 driver is possibly behaving differenctly at the low end and if it is shallower in depth it will mess up the crossover point. Try the LE85 or 2420 and see what happens. You might be surprised.

If the real data on the above points can be obtained the network can be quite possibly optimised assuming there are no other unknowns like internal standing waves.

As I point out though, this design is very sensitive to driver offset and that is the main issue. From a sound quality view point the 2344/4430 is far superior, ist has none of the above problems and you have all the design details to make it go first pop.


Agreed that driver phasing has its biggest effect right at 900 Hz. One way this becomes a peak and the other way, a null. Out of phase provides the best 900 Hz band, but also yields the biggest "hole" between the drivers at 600-700 Hz.

The 600-700 null appears with or without the horn connected, so it hardly seems that this is a case of mutual cancellation. Also the hole shows even with no crossover connected.

I tried an LE85 and 2425 with similar results. I keep coming back to the LE175s that seem to provide the "warmest" sound (and best of the three bass extension)to me.

Dust cap sits in ~1-7/8" while the slant plate sticks out about 2-5/8".

[Robh3606]

Hello Toddalin

Are you sure they used JBL cone kits when they re-did these?? The reason I am asking is the free air peak that Ian is asking about, but it could be the way they were held when you ran the sweeps. I can run one of my 2235's in the 4344 cab just to see if I get any peaks in the same range when I get home from work.

Rob:)

[Ian Mackenzie]

Hello Toddalin

Are you sure they used JBL cone kits when they re-did these?? The reason I am asking is the free air peak that Ian is asking about, but it could be the way they were held when you ran the sweeps. I can run one of my 2235's in the 4344 cab just to see if I get any peaks in the same range when I get home from work.

Rob:)

Hi Rob,

That could be but the data had aclassic bass reflex impediance peaks and the tuning was right on 29 hz.

If another oppoortunity arises the impediance run should be with the woofer mounted in the box without the crossover connected.

Ian

Ian

[toddalin]

Hello Toddalin

Are you sure they used JBL cone kits when they re-did these?? The reason I am asking is the free air peak that Ian is asking about, but it could be the way they were held when you ran the sweeps. I can run one of my 2235's in the 4344 cab just to see if I get any peaks in the same range when I get home from work.

Rob:)

Yes, OCS used real JBL cone kits.

[Ian Mackenzie]

Hi Todd,

I need to know where the compression driver top place is ie the diaphragm relative to the front panel.

[toddalin]

Hi Rob,

That could be but the data had aclassic bass reflex impediance peaks and the tuning was right on 29 hz.

If another oppoortunity arises the impediance run should be with the woofer mounted in the box without the crossover connected.

Ian

Ian

Been there done that before the recent recharge. Will send you the tests. They also show the impediance hump in the same range.

As an aside, at one time I noted that putting a resistor on the woofer took care of the problem. (Of course most people on this site poo-pooed the idea of using a resistor on the woofer.) I think this helped because the impedience hump became smaller relative to the overall impedience curve.

[Ian Mackenzie]

Todd,

I looked up the horn length and driver details in the Library.

Here is the modified simulation with the woofer and horn out of phase (tweeter also out of phase relative to the woofer).

The first curve is with the mic mid way between the woofer and horn centres.

You can see a much improved crossover function than earlier.

The second curve is with the mic on axis with the horn.

The third curve is the mic on axis with the woofer.

The fouth curve is the voltage drivers with the network loaed by the drivers. As mentioned above I have used purely resistive loads for the horn. The 2235 is from your WT T/L data.

All simulations are with the mic at 2 metres (what would be the minimum listening distance)

This effectively proves that the crossover modelled and the drivers located as are reasonable well optimised.

Unfortunately Todd the issue of the woofer response hole lay else where.

I would suggest its the location of the woofer relative to the side walls or in internal standing wave problem. The action of the dust cap could also be an issue.


Ian

[toddalin]

Hi Todd,

I need to know where the compression driver top place is ie the diaphragm relatice to the front panel.

HL-91 is 7-3/4" behind the baffle panel and driver is 3-7/8" deep. Assuming that the diaphram sits 2/3 of the way in and baffle board is 3/4", front of baffle to diaphram is probably ~11".

[Ian Mackenzie]

Todd,

I used this data:

http://www.lansingheritage.org/images/jbl/catalogs/1979-comp/page10.jpg

If your horn is the longer 800 hz horn the extension behind the baffle is 10.75 inches, add the depth of the compression driver and estimate of the top plate my guess is 13.9 inches to the diaphragm.

As discussed above I think this is reasonably accurate as the sims show.

Ian

[toddalin]

Todd,

I used this data:

http://www.lansingheritage.org/images/jbl/catalogs/1979-comp/page10.jpg

If your horn is the longer 800 hz horn the extension behind the baffle is 10.75 inches, add the depth of the compression driver and estimate of the top plate my guess is 13.9 inches to the diaphragm.

As discussed above I think this is reasonably accurate as the sims show.

Ian

You are looking at the HL-92 (used in the L300 and 4333). HL-91 is 3" shorter, or about 11" to the diaphram as I noted.

[4313B]

As an aside, at one time I noted that putting a resistor on the woofer took care of the problem. (Of course most people on this site poo-pooed the idea of using a resistor on the woofer.) I think this helped because the impedience hump became smaller relative to the overall impedience curve.That's right, one doesn't usually put a resistor in series with a woofer, oftentimes one does put one in parallel if it helps the overall situation.

In another thread I suggested removing the conjugate. You can then place a 50 to 100 ohm resistor in parallel with the 2235H. Run both if you have to.

I would suggest its the location of the woofer relative to the side walls or in internal standing wave problem.All one has to do is set their oscillator at the frequency where the impedance blip occurs and listen for any anomalies. Standing waves will sound quite nasty.

[edgewound]

HL-91 is 7-3/4" behind the baffle panel and driver is 3-7/8" deep. Assuming that the diaphram sits 2/3 of the way in and baffle board is 3/4", front of baffle to diaphram is probably ~11".

The distance between the top plates of the H-91/LE175 and 2205 is 8 inches when mounted to a 3/4" baffle...2205 mounted from the front.

[Ian Mackenzie]

So your definately using the shorter horn?

Why on earth use the shorter horn?

Here is the updated simulation with 8 inch placement of the path lengths on axis with the mic horn on axis in the first curve and mid way between the woofer and horn centres for both in and out of phase.

Its not pretty.

Ian

[toddalin]

So your definately using the shorter horn?

Why on earth use the shorter horn?

Ian


Longer horn does not fit in L200 cabinets. The driver would stick out the back ~2".

[edgewound]

So your definately using the shorter horn?

Why on earth use the shorter horn?

Ian

Because it's an L200 cabinet that came with the H-91. The H-92 is too long for the box. It's shallower than the L300.

The L200B components were 136A and LE85-16 ohm on H-91 horn.

[4313B]

Why on earth use the shorter horn?JBL didn't know everything back then, they did the best they could with what they had. Some of the guys thought they could get away with using any horn with any woofer. It was the later crowd that brought in or designed the gear required to design more advanced systems.

[toddalin]

JBL didn't know everything back then.

Blastphemy! ;)

[Ian Mackenzie]

I edited the last post with the curves.

Okay that explains a lot. I guess they had to work with what they got.

It would be useful to see how the shorter horn works in reality on the baffle and measure it down to 800 hz. Becuse its shorter I would expect the response to tappper off a bit and you may end up with a reasonable transition.

Todd,

Sorry I dont have an answer for your woofer ills.

Do the sine wave sweep and listen carefully as suggested above.

Ian

[4313B]

It would be useful to see how the shorter horn works in reality on the baffle and measure it down to 800 hz.You can do that with your 4345's. :)

I guess I could too... :hmm:

[Ian Mackenzie]

Yeah I thought of that too.

But at any rate Todd's got other issue that make this look like the icing on the cake.

Personally I would get the jigsaw out a cut a hole for the longer horn and make a neato rear housing for the driver hook up a 4333 network :).

ian

[Robh3606]

Hello Ian

If you mean the 2307 here's how one measures mounted in my 4344 cabinet with a 2425 driver. No crossovers.

Rob:)

[toddalin]

Yeah I thought of that too.

But at any rate Todd's got other issue that make this look like the icing on the cake.

Personally I would get the jigsaw out a cut a hole for the longer horn and make a neato rear housing for the driver hook up a 4333 network :).

ian

Already been considered and just yesterday had a bid in for a pair of the longer horns (that ended up going for ~$157 with one being cut up and no slant plates included). I was just going to make a hole in the back and glue a large PVC cap over the back. Wife was not in love with that idea. :baby:

Anyway, if I recall the 4333 crossover uses a propriatary choke, as do most JBL crossovers, which is a major reason why the N200B was selected.

The hole issue manifests itself even with no crossover or horn connected, so I really don't see how using the 4333 crossover would remedy this anyway. :blink:

[toddalin]

Hello Ian

If you mean the 2307 here's how one measures mounted in my 4344 cabinet with a 2425 driver. No crossovers.

Rob:)


I think the 2307 is the shorter H-91 while the 2312 is the longer H-92. I've no idea which is used in the larger cabinets that have a cone lower mid.

[Robh3606]

Still over a 6 dB change between 700-900 Hz.

So +/- 3 db Yes but look up at 1.2K. That's really as low as you should use them. I know JBL took them down to 500Hz and all but they are starting to unload below about 1k

Rob:)

[grumpy]

Think it's time for an outside test. ...can also run in-box impedance and individual
driver response curves with no crossover. -grumpy

[Robh3606]

Think it's time for an outside test. ...can also run in-box impedance and individual driver response curves with no crossover. -grumpy


Good idea you should also run the impeadence on the horn and driver mounted to the baffle.

Rob:)

[Ian Mackenzie]

Hi Rob,

Thanks for posting the 2307 and horn response, the peak at around 1khertz might even fill a few holes.

Looks usable to 800 hz if only for domstic power levels.

I agree, some more impediance runs wioht drivers in the box would prove useful.

Ian

[boputnam]

just for the hell of it did you try to invert the phase on the mids to see how the crossover points react differently. it may be to your advantage. Hey its a shot in the dark but it might workMy favorite post here.


Actually, yes, and I'm finding that things are overall flattest (both to the woofer and tweeter) with the mid inverted. The trough where the woofer meets the mid is about 1 dB deeper this way, but other improvments of over 3 dB are noted just beyond the trough.

Also, the overlap with the tweeter is smoothest like this. ;)So, why was this abandoned? :blink:


Its not pretty. Yea, and it's right at the crossover points, suggesting to me that something along the lines of pelly3s' idea needs attention.

[Robh3606]

OK here's a 2235 sweep in my box. The 500Hz and up range looks like what you would expect.

Rob:)

[toddalin]

My favorite post here.

So, why was this abandoned? :blink:

Yea, and it's right at the crossover points, suggesting to me that something along the lines of pelly3s' idea needs attention.


Nothing was abandoned! Phase has been inverted every which way, and these result in small changes (ranging from a few tenths of a dB to a few dB at various freqencies), but some of you are missing the BIG PICTURE.

I'm talking about a major suck-out of over 10 dB between about 600-800 Hz on the woofer even with no crossover connected. :biting:

The horn is pretty flat down to 900 Hz and down about 5 dB from that at 800 Hz, so the crossover is doing its job in that respect. It is an "800 Hz" crossover.

But if the crossover is supposed to start rolling the woofer off at say ~600-700 Hz..., and the woofer is already down 10 dB at that frequency..., voila a hole in the frequency response!

So the real question is what's making the woofer roll off so early? Recognize that I get the same thing in my center channel and the cabinet is of completely different architecture and I'm even using a different horn.

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/center-2235-w15gti.jpg

I just keep coming back to the woofers, but maybe the room could create such a suck out? I have moved furniture around to no real avail. Still, I don't recall seeing it using the W10GTIs/2425s sitting on top of the L200s...

:blink:

[edgewound]

Todd...

Have you tried a different amp? Maybe the Yamaha that you have has some sort of problem....or just doesn't like the 2205/2235. This is weird.

I highly doubt the room would mess around with that high a frequency...unless the room is really live, reverberant. Does the room "zing" when you clap your hands loud? Corner reflection/bounce?

Just a thought. I'm sure you're really frustrated.

[toddalin]

Todd...

Have you tried a different amp? Maybe the Yamaha that you have has some sort of problem....or just doesn't like the 2205/2235. This is weird.

I highly doubt the room would mess around with that high a frequency...unless the room is really live, reverberant. Does the room "zing" when you clap your hands loud? Corner reflection/bounce?

Just a thought. I'm sure you're really frustrated.

Yes, I too assumed that it could be the amp (RX-Z9) leaving a sonic signiture, but tried another (DSP-A1) and even Grumpy brought a JBL amp over and ran it right off the computer all with the same result.

Like I noted, I can get the classic 2235 frequency curve with the mic lying on the floor.

Room does have one wall of large windows and a wall of "French" windows, but these are covered by heavy, light-proof shades. There is also a triangular window near the ceiling. Other walls are typical wall board with some artwork. Furniture is typical overstuffed. Room is approx 28'x16.5' and ceiling slopes from 16' to 8'

BTW, there is no suck-out when you talk in the room. :blink:

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/room3.jpg

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/room2.jpg

http://www.largescaleonline.com/eimages/lsolpics/Team_Member_Pics/toddalin/room1.jpg

[pelly3s]

so i ran into this same kind of the issue working with a box a customer brought me the other day. the baffle was set back about 4 or 5 inches and when I added a piece of foam to the baffle on the sides of the woofer and it corrected the problem. i figured it was the fact that the distance the baffle was set in was about a 1/4 the wavelength at about 650hz. needless to say he pretty much just had to live with the problem if he wanted to use the boxes.

[Ian Mackenzie]

Yeah,

It could be one of the issues.

Ian

[grumpy]

I had suggested Todd might want to check that some time ago... I don't recall any specific results
(positive or negative)... easy enough to recheck next visit :)