Playing with a 2426

[Ian Mackenzie]

I would like to see a curve of the 2420 with the D8R2425 on 2344.

Here's my result taken some time again.

Ian

[Mr. Widget]

...I discussed this with Giskard and it should not work 100% correct but it should work.

I am no absolutist, but not working 100% is usually not working in my book.

Hey Ian,

What is your vertical scale?

Here is my quasi-anechoic measurement of one of Zilch's stock 2425Hs w/ 3135 and 2344.

Widget

[Ian Mackenzie]

Sorry I don't recall, I think it was 12 per division, even so its very smooth on a properly mounted baffle at 1 metre.

regards

Ian

[Earl K]

Guido,

Your plots do show that the aluminum diaphragm has reduced that (nasty) peak at 4KHz....

- Yes , regarding that peak around 4K on the 2344.

- Since you and Zilch are just "all-over" re-engineering the 2344 / & all other 2 way designs with new parts ,,, I think it's time you whipped up a "tuned" series type LCR filter ( shunted across the driver just before R5 on the 4430 network ) to hammer down that 4200hz peak . The "how-to" particulars are buried in the "Project May" area in the "bandpass" thread . Giskard & Widget can give you testimonials to just how effective these filters can be when they are zeroed in . Recently, I whipped up a "low-Q" type out of spare parts to "hollow-out" a bulge on a 2441/horn combo I've been working on ( I didn't really know what I was doing - but it worked very well despite that handicap ).

Give it a try .

[Mr. Widget]

I corrected differences in scaling and changed the color and superimposed my 2425H on 2344 Quasi-anechoic plot from the post above and placed it on top of Guido's. It is fairly similar to his 2425 plot. Some of the differences could be due to different pot settings as well as unit to unit variance.

Widget

[Zilch]

Response:

Blue = 2426 with fresh D8R2425 on 2344, Clear extended HF-Response
Red = 2426 with fresh D16R2421 on 2344

If you normalize them, 2421 has better response, both high and low....

[Zilch]

Since you and Zilch are just "all-over" re-engineering the 2344 / & all other 2 way designs with new parts ,,, I think it's time you whipped up a "tuned" series type LCR filter ( shunted across the driver just before R5 on the 4430 network ) to hammer down that 4200hz peak .

I think we're abandoning N3134/35, at least the HF section, in favor of newer crossover topologies for use with 2431 and PT waveguides.

We're still at the "gross" stage in this, but we may be able to get away without any additional notch filters. Wishful thinking, perhaps....

[Guido]

I am no absolutist, but not working 100% is usually not working in my book.

OK, it will work!
I see no reason why not.

[B&KMan]

Hello Guido,

What news on this faboulus comparison ???

What diaghrams you prefers objectively (mesured) and for subjective listening ?? with your 4343

For the 4343 (2420) what is the final best choice ???

As you know I have a problem of my old 2420 because the pressure is really drop in arround of 8K.

If I look the response curve on your test, the 2421 is look the better linear response (arround 5-6 dB variation).

Thanks.

Jean.

[Earl K]

Jean,

If I look the response curve on your test, the 2421 is look the better linear response (arround 5-6 dB variation).

- I believe you are misreading and misinterpreting Guidos posted curves. That "lighter green" curve was added by Mr. Widget. It's a 2425H on a 2344 horn . It does have 5 to 6 db better UHF response when compared to the older 2420 ( or the 2421b for that matter ). The curve of the 2420 or the 2421b diaphragm with this network (N3134/5) does show UHF droop. This is to be expected when these diaphragms are measured with the wrong network .

- These response curves are misleading you . This N3134/5 network was specifically designed to work with JBLs' low impedance 8 ohm diaphragms ( like the 2421a or the 2425h ) not the higher impedance 16 ohm diaphragms such as the original 2420 or the 2421b or a 2425j . It's dangerous to draw conclusions from these posted graphs. For an effective comparison; the high impedance diaphragms should have been combined with a 3134 type network that had had, all its' values reworked for use with the higher impedance ( 16 ohm ) diaphragms.

As you know I have a problem of my old 2420 because the pressure is really drop in arround of 8K.

- IMO, you are better off to look elsewhere for a fix to that response hole that is occuring between your horn driver and your 2405 UHF driver .

- I see Giskard came back this summer out of exile and provided you with a bunch of clues as to how to proceed to fix your problem . That was nice of him . I guess you didn't pick up on those clues . FWIW; for a Canadian, changing diaphragms will be your most expensive solution. OTOH, learning to manipulate crossover design by understanding the theory will ultimately be much cheaper and more rewarding.

- What diaghrams you prefers objectively (mesured) and for subjective listening ?? with your 4343

- For the 4343 (2420) what is the final best choice ???

- Well, I too will wait for Guidos answer. After all, he did eventually get rid of his 4343 speakers .

[B&KMan]

Jean,
- Well, I too will wait for Guidos answer. After all, he did eventually get rid of his 4343 speakers .

Thanks for reply Earl,

It is sense for me but the variation of impedance it is really big change in signature of spectrum ???

do you have mesured any motor with 16 vs 8 ohms diaphrams for comparison ??

I understand this is not exactly same set-up but maybe the reactance is look a little bit same ??

I know the final solution is corrected the network but because the 2420 is more than 20 years, maybe it is interesting to adaptive network solution in fresh and more accurate diaphrams ???

I'm not feel to extend for couple of years this research...



Jean.

[Earl K]

Hi Jean

but because the 2420 is more than 20 years, maybe it is interesting to adaptive network solution in fresh and more accurate diaphrams ???

- I realize you are searching for some solutions to the mismatch between your new network and your older ( non recommended ) components .

- If I were you, I wouldn't jump to conclusions that your 2420s are worn out and the source of all your problems . Afterall, Zilch has showed recently that his old , venerable le85 ( 2420s ) actually have quite extended response in the top and bottom end / when compared to some others ( like the 2426 ).

- So, I would recommend measuring the frequency response of your 2420 drivers without any of your crossover in the circuit. Measure with those 2307 horns on the drivers using pink noise . Set your RTA to some resolution that is recognizable to Zilch and myself. I use pink noise high passed at around 300 hz for this sort of testing. Test with the 2308 lens off . I do this test at about 1 foot away / on axis to the horn mouth / and only loud enough to get a respectable reading .

- When you have done this / publish some pics of the response curves here, so that we can try to gauge the health of the diaphragms in those drivers . I have 2307/8 combos here - but unfortunately, no le85/2420 drivers / just 2410/2425 types etc.

[B&KMan]

first thanks for reply and offert to check response...

Hi Jean


- I realize you are searching for some solutions to the mismatch between your new network and your older ( non recommended ) components .

- If I were you, I wouldn't jump to conclusions that your 2420s are worn out and the source of all your problems . Afterall, Zilch has showed recently that his old , venerable le85 ( 2420s ) actually have quite extended response in the top and bottom end / when compared to some others ( like the 2426 ).

well I perform completely electrical test with supervision of Mr mackensie and no problem in electrical response. (really close to theorical) so next step is the membrane ...

- So,

1---- I would recommend measuring the frequency response of your 2420 drivers without any of your crossover in the circuit.
2---- Measure with those 2307 horns on the drivers using pink noise .

3---- Set your RTA to some resolution that is recognizable to Zilch and myself.

4---- I use pink noise high passed at around 300 hz for this sort of testing.
5---- Test with the 2308 lens off . I do this test at about 1 foot away / on axis to the horn mouth / and only loud enough to get a respectable reading .

- When you have done this / publish some pics of the response curves here, so that we can try to gauge the health of the diaphragms in those drivers . I have 2307/8 combos here - but unfortunately, no le85/2420 drivers / just 2410/2425 types etc.

===== what is ( 4 ) pink noise in around 300 Hz ????

pink noise is white noise but attenuation by the number of frequency include in each band 1/3 octave.

===== it is necessary to put resistance in serial ?? or in parrallele ???

===== what resistance to put 8 ohms or 16 ohms ????

thanks for more complete details before I unbuild again my speakers....

[B&KMan]

Hello again,

I perform test

pink noise 60 secondes at 200 to 200K Hz for 1/3 octave analyse RTA

and white noise 1000 averages records at 200 to 200K Hz for FFT analyse.

this is result in 4 mesures.

I mesure 2420 and 2405 in this I save another open speakers

let me know your feeling Earl , zilch and any other members.

thanks for all.

(I persist to sign this problem is better information on thread 4343 but as you wish earl.)

[B&KMan]

look the response of UHf is increase and the 2420 is stable (little drop...)