Cross-over schematics for the 4660

[Jan Daugaard]

Does somebody have the schematics for the 4660 cross-over?

I'm not sure what the difference between the 4660 and the 4660A is (it could be the woofer 2225 vs 2226), and I need the schematics for a 2226 woofer and a 2425/2426 compression driver on a 2346 horn.

Thanks in advance!

[4313B]

Ok

[Jan Daugaard]

Hi Giskard,

thanks very much for the schematic -- 4660A_Page_1.jpg contains just what I was looking for.

It appears that L3 has a switch with 5 positions. Do they represent 5 choices of induction between 0 mH and 3.3 mH?

And what does NC stand for?

[4313B]

It's a tapped autotransformer. You'd have to measure the taps if they aren't specified in the schematic.

No Connection.

[Zilch]

It's an odd note about the 2426 impedance. Looks like they switched to the 16-Ohm "J" in the "A" version, and advised swapping out diaphragms in the originals when required....

Ooops!

[spkrman57]

Seems like they would both work unless I am missing the boat on this one.

Not like I haven't been wrong before!!!!

Ron

[spkrman57]

Will using a 8 ohm or 16 ohm 2426 make a noticable difference with the 4660 crossover schematic!

If nothing nothing else, the HF horn output would be off a few db!

Hoping someone will chime in here!

Giskard would be able to answer this question, someone talk him out of retirement.

Ron

[johnaec]

Will using a 8 ohm or 16 ohm 2426 make a noticable difference with the 4660 crossover schematic!'Seems to me the 16 ohm version would be about 3db down from the 8 ohm version, but I'm not sure that exactly applies in this case. 'Curious how they use the same crossover in both versions. I wonder if they went to the 16 ohm to better match the level between the two different LF speakers?

John

[Zilch]

Someone needs to model the HF section of the crossover with actual impedances of the two versions of the driver. From previous experience, there's more than just attenuation involved, as the different impedances shift the filter operating points proportionately. There's a HF boost loop there, too.

I'm not comfortable doing that with the tapped inductor in there; there's considerable discussion on modeling those in other threads here, which I don't understand.

Another option, which I might try, is to measure the actual performance of both drivers with an RTA and/or CLIO. That's a beast of a task, though, disassembling one of those things to get at the driver and yank the crossover. I DO want to swap in some 2226's, tho, to free up the 2225's for conversion to 2235's.

They're the final stage of the "Quick & Dirty" project. August, maybe.... :p

[spkrman57]

Would it change much, or does the other attenuation/HF compensation circuit components keep the load to the crossover constant regardless which one would be used(8 or 16 ohm).

Ron

[Zilch]

Proportionately, according to the ratio of impedances. In two-pole filters, it also messes with the filter Q substantially. Look at your crossover component selection charts in the Parts Express catatlog....

This is why I termed the note "Odd," above. It's as if they made a mistake building units with 2425H drivers. It may also be that the original crossovers are actually different. The change is linked to a specific crossover model number.

I can get impedances with WT2, but definitive response curves will have to be measured quasi-anechoically using CLIO. RTA will reveal gross differences only, due to inherent averaging.

Both practically and academically, it'd be a worthwhile exercise. I'm just not there yet. Gotta BIG 4688 subwoofer I need to load to play with them first.

If you've got the stuff, try it yourself. We'll compare notes.... :D

[Lancer]

I would have to measure the #4 tap on the 3.33 mH inductor and I don't feel like dragging it out and measuring it right now. Maybe later.

[Zilch]

I would have to measure the #4 tap on the 3.33 mH inductor and I don't feel like dragging it out and measuring it right now. Maybe later.You want me to generate impedance curves on WT2? Or do you know the impedances well enough already?

I have two each 2425H's (if that's what's in 'em) and J's here, plus a pair each of 2426J and 2427H I could mount on one of those horns.

[Gonna be a bear bustin' those painted-on horns loose. Gotta do it eventually, tho....]

[spkrman57]

I bought mine and it came without cd to load into my pc.

3 days later the cd rom arrived, I could not wait to see how she runs.

Bummer!!! The damn unit never clicked in and did not perform any functions, just ran and ran without getting anywhere. The figures just kept changing but it kept telling me connection failure or internal fault.

I have to send back to PE and wait for new unit to come back. I liked the old one even if it was antiquated, just thought the new one would be so cool.

Sorry to hijack this thread, just wanted to know how your(Zilch) WT2 is working for you.

Ron

[Lancer]

Bummer!!! The damn unit never clicked in and did not perform any functions, just ran and ran without getting anywhere. The figures just kept changing but it kept telling me connection failure or internal fault.

I had the same problem. It turned out to be a USB issue. I can't remember if I hit my computer with a rubber mallet or simply pushed it off the table onto the concrete floor to get it to work. Maybe I just unplugged all the USB crap, turned the machine off, then back on and plugged the USB garbage back in. It works now.

You want me to generate impedance curves on WT2? Or do you know the impedances well enough already?

I was just going to dig up the inductor, measure the tap points and then model it with the standard JBL text fixture, 8 ohm and 16 ohm loads. I thought I had one sitting behind me but it turned out to be the 1.55 mH from the LX80, N333, 3133 and LX300. The other tapped autotransformers are downstairs in a box somewhere. If you want to post the impedance files for me to look at later that's fine.

[Zilch]

My WT2 works fine. I was somewhat disappointed to discover that I can't work with the program, like look at files and stuff, unless the device is plugged in.

Upon inquiry, I was informed that was their approach to copyright protection.

No big deal for me 'cause I only have one computer, but I'd prefer not to have a bunch of gear hangin' off it all the time.... :(

[Lancer]

Here's what I got using the typical -6, -8, and -10 dB taps with 8 and 16 ohm "standard test fixture" loads.
Actual impedance curves of the CD/horn systems would be more revealing. I don't think WT2 can take voltage drive measurements so you'll have to use something else. I think SoundEasy can. I think SoundEasy can also pound out an electroacoustic equivalent circuit. I could be wrong, you'd have to ask TimG.

[Zilch]

Interestingly, the rolloff frequencies don't shift (not by the octave I'd expect with conventional LC), but the Q's sure do.

I'd try both driver impedences and see which gave the best response, probably. Lemme see how tough it's gonna be to yank that stuff and haul it over to Mr. Widget's.... :)

[Zilch]

4660 crossovers evidence charring where the parallel 39-Ohm resistors overheated. This pair, though not "A" version, has 2425J's in it. Phase of HF driver is reversed from schematic - see yellow/blk going to the 3.9-Ohm resistor there.

RTA response curves of 8-Ohm (top) and 16-Ohm 242xx drivers, 2346 horns, and the least burnt of these two crossovers. :p

8-Ohms is better, I'd say, but I'd want to try a bunch of drivers to be sure. Looks like the difference in response follows Lancer's voltage drive curves nicely....

I tried measuring the tapped inductors there (right) but get bogus results. They're 3.3 mH across all right, but the measurements at the tap don't add up to that. :(

The leads from the unused taps are clipped short and twisted together. A bad idea, I'd think; they should not be connected. That relies on the enamel alone to prevent them from shorting. Looks like it held up through the fires there, though....

Insight: The N4660 crossovers are the 3110A circuit with mid set to minimum and HF boost set to max. That crossover was spec'd for 16-Ohm HF drivers.

Bonus tracks: Third RTA curve is with 8-Ohm 2426H running on N3134 (the 4430 crossover) set to "0". I can dial them nearly flat with more HF boost. There's a familiar signature there, tho, too much response at 8 - 10 kHz, sooooo ...

Fourth curve is using "AM" crossover from Quick and Dirties (raised 5 dB). Mighty pleasant sounding, I don't think I'll be givin' high priority to rehabbing them N4660's. :p

For the 16-Ohm drivers, maybe a tweaked version of N200t3 with HF boost? Sometimes I get lucky.... :hmm:

[Lancer]

I tried measuring the tapped inductors there (right) but get bogus results. They're 3.3 mH across all right, but the measurements at the tap don't add up to that. :(They won't add up. Please run three WT2 arbitrary plots. Set the sweep from 500 Hz to 2 kHz and the sweep points to 6. First data file should be the two end leads of the tapped autotransformer. Second and third data files should be from the tap to each of the end leads. All the leads should be taken out of the circuit for measurement.
The leads from the unused taps are clipped shortYeah, I wish JBL wouldn't do that! :banghead: If they are cut wrong you end up with an open circuit and have to solder them together.
Insight: The N4660 crossovers are the 3110A circuit with mid set to minimum and HF boost set to max. That crossover was spec'd for 16-Ohm HF drivers.You are obviously taking the time to study numerous schematics and examine their similarities/differences and impact. :)

[Lancer]

JBL 1.55 mH tapped autotransformer.
Five leads.
For our purposes:
Inside lead labeled IL
First tap labeled T1
Second tap labeled T2
Third tap labeled T3
Outside lead labeled OL

IL to OL = 1.53 mH @ 1 kHz

IL to T1 = 0.37 mH @ 1 kHz ~ -6.0 dB
IL to T2 = 0.53 mH @ 1 kHz ~ -4.5 dB
IL to T3 = 0.78 mH @ 1 kHz ~ -3.0 dB

OL to T1 = 0.41 mH @ 1 kHz ~ -6.0 dB
OL to T2 = 0.27 mH @ 1 kHz ~ -7.5 dB
OL to T3 = 0.15 mH @ 1 kHz ~ -10.0 dB

Target specs were -3, -4.5, -6 and -6, -8, -10

[Lancer]

Are you going to run the measurements as asked above and post them Zilch? :D

[Zilch]

Yup. This evening, yet.

Re: 1.53 mH inductor: So, for different attenuation ranges, you just invert the inductor?

[There are people MUCH more clever than me doin' this.... :p ]

Thanks to Techbot for gettin' the complete Everest white paper posted. :thmbsup:

[Zilch]

The really smoked one. They're lucky the church didn't burn down.

1) Top to bottom
2) Top to tap
3) Tap to bottom

[Zilch]

Same results as I got with the meter, basically, that I thought were bogus:

[Zilch]

Looks like about 30W into the driver would concurrently exceed the 20W rating of the two parallel resistors. How hot they gotta get to set the glue ablaze?

There's no evidence of overheating on the inductors themselves, thankfully.

[It's dumb to put hot melt glue or silicone RTV on power resistors, of course. HEAT SINK would be a better idea.... :banghead: ]

That'd be pretty loud, tho. They were heapin' a might of praise at 1 kHz to accomplish ignition.... :p

[Zilch]

Crazy horn:

[Jan Daugaard]

Hi fellow forumites,

Matthias and I intend to buy the single 4660 / Everest horn for sale on Ebay. The sole purpose is to make glasfiber casts of it.

It is too early to give the specifics of colour, weight, price, and shipping from Germany, but the casts will be made available to our fellow forumites in due course.

Please don't outbid me ...

Jan D.

PS. Three months ago, we won another auction for an Everest horn, and Matthias paid for it, but the seller never shipped it.

[Zilch]

Well, CRAP! :banghead:

I thought that had been worked out for Matthias.... :(

[Lancer]

The really smoked one. They're lucky the church didn't burn down.

1) Top to bottom
2) Top to tap
3) Tap to bottom

Same results as I got with the meter, basically, that I thought were bogus:Ok, so wired in one way you end up with a 10 dB drop (0.3 mH secondary), wired in opposite and you end up with a 3 dB drop (1.7 mH secondary).

[Lancer]

Looks like about 30W into the driver would concurrently exceed the 20W rating of the two parallel resistors. How hot they gotta get to set the glue ablaze?All those resistors in front of the tapped autotransformers take a beating. I've seen 4333's with their 7.5 ohm 40 W resistors burned to a crisp.

Here's the impedance curve of a 2344A/2425H with the 1.55 mH inductor (-6 dB tap) from an L300/4333 without the 7.5 ohm resistor and with the 7.5 ohm resistor. One can see it keeps the preceeding series capacitor from going haywire.

[Zilch]

I have some new resistors on the way, and I'll try that when I install them. Good trick to tame the impedance there and, I assume, get requisite additional attenuation (about 3 dB?), as well.

These parallel pairs read 18.2 and 18.4 Ohms, respectively. Can't read the actual values on them, if it was ever there. Should read 19.5 Ohms if they were actually 39 Ohms per the schematic; being fried like this would likely change their values, tho.

I'm using single 18-Ohm aluminum N/I ones as replacements, which I'll mount on heat sinks. Paired 40-Ohms might be a better choice, now that I think more about it.

Not gonna beat myself up over it, tho, as I have no clue at this point what I'm gonna ultimately DO with these things. The horns themselves are sounding good with better crossovers, and if they go back in these boxes, they'll likely get 2235H's mated in them instead of 2225/26.

Maybe I use them for backyard theater.... :p

[WTPRO]

Hi everyone,

I noticed that some of you are using the WT2 and thought I would post an FYI that an FAQ and an alpha version of the next WT2 software release has been posted to www.woofertester.com (http://www.woofertester.com). Just keep in mind that the alpha code is a testbed for new features. Help for now is in the FAQ. Neve the less, I think a few of you will like it.

- FAQ: A number of examples have been posted showing things like how to test multiple driver, dual voice coil drivers, effect of series/parallel resistors and more.

-WT2: Box simulation with data overlay. This is a new child window where th idea is to overlay the measured phase and impedance to that of the simulator. If they match, you are going in the right direction. Works for free air, sealed and vented boxes

- WT2: Frequency dependent Rem/Xem. Ever noticed how the older T/S models did not produce acurate phase and impedance at the higher F's? The addition of freqeuncy dependent Rem/Xem in series with the electrical tank fixes this. The WT2 will measure -and- now model this.

-WT2: Improved USB support where Windows was making a mess of things. In 1.01 getting the WT2 to be recognized sometimes involved mucking around with the multimedia device manager. This should be solved now.

- WT2: The left mouse button will now set an arbitrary frequency and the right mouse button will log that data point into the present data buffer. These data points are not part of the T/S test, but they can be used to smooth (beutify) the curves or zero in on something of interest.

- WT2: The addition of an external parallel resistor (to keep the current driver happy) and some math now allows the WT to measure capacitors.

Note: The alpha code will be getting updated from time to time. The about dialog (also pops up on start) includes build time information.

Best regards,
Keith Larson

[email protected]
DSP and Analog Consultant
Lincoln, Ma 01773

(a C&S Audio Partner)

[Zilch]

Here are CLIO response results for 4660 crossover using pairs of 2426H and 2426J drivers on 2346 "Everest" horns. Measurements taken perpendicular to horn face:

[Zilch]

Thanks again to Mr. Widget for sharing his equipment and expertise.

[Not to mention his patience with the Zilchster.... :p ]