JBL 4346 DYI with 2123H

[lordoflight]

Dear All,

I am new in this forum and first off all I would like to thank you all for your passion in JBL vintage speakers. I believe they are outstanding. I am currently building a 4346 JBL with the following speakers:
LF 2235H (reconed on a E140 basket with a mass ring of 51 grams)
MF 2123H
HF 2425J with 2307 horn dyi
UHF 2405

Considering the above, I have the following pleas to you guys:
1. How can I integrate the 2123H speaker in this build? I want to use the CC crossover scheme provided by Mr. Giskard here http://www.audioheritage.org/vbulletin/showthread.php?31638-New-Charge-Coupled-Crossovers-in-the-house&p=319098&viewfull=1#post319098
I will not bi-amp!
2. Does anyone have all the specs of the 2307 horn. I want to machine a pair. I mean every dimension... including the arches. Sadly I don't have an original so I can make a mold after the horn.

Thank you so much for your input.

[lordoflight]

Hi guys,

I'm back again although no reply to my post so far... :(

Nonetheless, can you tell me what would be the best choice for a crossover for my DYI project? I want the best of the best.
A charge Coupled version or a plain high end cap version?
An also could you indicate the best schematics for such recommended crossover?

Thank you very much in advance.

[bldozier]

Hi guys,

I'm back again although no reply to my post so far... :(

Nonetheless, can you tell me what would be the best choice for a crossover for my DYI project? I want the best of the best.
A charge Coupled version or a plain high end cap version?
An also could you indicate the best schematics for such recommended crossover?

Thank you very much in advance.

Any plans.

[.:hb:.]

1. How can I integrate the 2123H speaker in this build? I want to use the CC crossover scheme provided by Mr. Giskard here http://www.audioheritage.org/vbulletin/showthread.php?31638-New-Charge-Coupled-Crossovers-in-the-house&p=319098&viewfull=1#post319098That's not easy. 2122H and 2123H aren't that much alike. The former is much more a bass than the latter. 2123H is a lot louder and a true midrange.
To my knowledge, something like this has been done on a 4345. perhaps someone around here knows about that.

[martin2395]

I have 4343 with 2235H, then 2123H coupled to 2450J (2-inch) - charge coupled 3155 network works perfectly.
Consider bi-amping to lose the "wooly-ness" of the foam edged "15, highly recommended.

[Chas]

I have had good success using the 4344MkII crossover network (charge coupled) with a 2123 as a substitute/replacement for the 2122 and its associated network.

I highly recommend you bi-amp @290-340 HZ fourth order, the difference is significant.

Note the above is based own listening only, I have not performed any measurements.

[lordoflight]

Hi guys,

After much tinkering I decided to use the LF and HF sections of the 4344MKII network and the HF and UHF sections taken from Mr. Giskard 3155 but I decided to take out the battery. This means no CC crossover and no bi-amp. I used instead Mundorf Evo Oil capacitors bypassed with Vishay 0.01 microF and high quality mundorf coils and resistors. I will upload a picture of it soon.

[lordoflight]

here is the picture of the crossover

[lordoflight]

Behold my babies! The l-pads are big Visatons. The crossovers are encased in big amplifier boxes. You can check my previous post for the components in the crossover. I used Jentzen JAD cables 2.5 mm OFC. The binding posts were ordered from china. They say its copper but in fact is brass. The cables were soldered with Jentzen silver solder. the drivers are:

LF - JBL 2235 - Actually a 2235 cone on a E 140 chassis

MF - JBL 2123

HF - JBL 2425 J with 2307 horn

UHF - Beyma 21CP/F

Any thoughts?

[1audiohack]

SWEET!!

Are you happy with them?

Barry.

[lordoflight]

I'm now tinkering a bit with some real time analyzers. It seems that I have a dropoff in the 10kHz range of about 2 db which is a lot. The rest seems pretty linear apart from the bass which is 1 db higher than the rest. I will make more tests with different polarities to see if there is a phase problem or not. Overall they sound OK yet a little damped. I don't know why. I will try to change the amp cables to place some silvers instead my current coppers. But the music is soothing and calm. It does not hurt my ears. Still I would like a bit more details. I will try in the weekend with some DACS to work on my source. The amp is a very capable Accuphase e360 which I like alot. So if anyone has any suggestions for improving sound please go ahead an shoot.

[lordoflight]

I have one question. To my ears the speakers sound best with all the LPADS FULL OPEN. Is there a mistake anywhere? It should not be normal. Otherwise the speakers sound a little dull and recessed.
Can anyone give me a qualified explanation?
Thank you

[mech986]

The 2425J is 16 ohm so may have reduced output compared with the 3155 network, IIRC. I don't know the Beyma tweeter but it too may have lower efficiency with the 3155 network.. BTW, you should get the 2308 slant acoustic lenses over the horns for better dispersion and less beaming.

if there are any resistors is series with the HF and UHF drivers, you may need to reduce them for more output.

[Ian Mackenzie]

Hi guys,

After much tinkering I decided to use the LF and HF sections of the 4344MKII network and the HF and UHF sections taken from Mr. Giskard 3155 but I decided to take out the battery. This means no CC crossover and no bi-amp. I used instead Mundorf Evo Oil capacitors bypassed with Vishay 0.01 microF and high quality mundorf coils and resistors. I will upload a picture of it soon.

I assume you meant LF and mid sections of The 4344k11 network.

Because you have used 2 different network this likely hood of the voltage drives being balanced is remote.

Why did you not use the complete 4344mk11 network?

I recommend buying the 2308 off eBay and installing then reassess before ou o anything.

The lense alters he response of the HF horn. It would honk off axis without the Lense.

[lordoflight]

I was instructed to use this hybrid solution by a member of this forum. Thank you for the suggestion of getting the lens 2308. I will. However it seems to be a phasing problem. I never got the time and space :)) to play with them since I have a 2 month old miracle sleeping . I'll keep you posted on the developments...

[Ian Mackenzie]

Okay,

I am amazed you have time for this with your new arrival!

Firstly, l can simulate for you the voltage drives of your hybrid in Leap 5 to evaluate the overall network phase and voltage drives and compare with the vanilla 3155 network. I might be able to look at this over the weekend.

I have the 4344mk11 network so l am familiar with its operation.

Secondly, Looking at the drivers sensitivity of the Beyma slot looks identical but the impedance curve may not be the same as the 2405? In your post you mention a 2 db drop off @ 10 Khertz? Are you referring to the horn only or the slot only or the overall response with the Lpads full?

If you can please provide clarity?

If this comment relates to the level of the Beyma only we can then focus on the adjustment range of the L pad.

If the drop off relates to a notch in the response please advise?

Some commentary of why you think this is a phase issue would assist in evaluating the situation.

The thing to understand with a 4 way is setting up the levels for each driver can be a challenge as can the phase of each driver. It's easy to wire up one of the driver. As l recall the mid, hf and uhf should all be in phase. I assume you have previously checked this? The polarity convention of the Beyma slot maybe different to the 2405.

[lordoflight]

I really appreciate you taking the time for little ol' me. I am trying to finalize this project and enjoy the music.

So, in order not to be any misunderstanding I am attaching the schematics og mr Giskard network frim which I have used the Hf and Uhf sections. The mids and lf are from the mkII.
Let me get back to you with the drop off issue. I want to replicate the test with all drivers full on.


In the meantime if you can Identify a voltage issue in the network i will be most grateful.


getting back to you also on the phasing issue. I know the Beyma is different than the 2405 And I wired it properly.

[Ian Mackenzie]

If you wish take a pic of the analyser for each test and up load for the overall response L pad full.

Then the phase reversal of the HF (only), then phase reversal of the Beyma only.

Then if you can the individual response of the mid, hf and uhf.

[Ian Mackenzie]

Okay


Can you check that the L pads are correctly connected?

On the L pad 3 is the input from the crossover, 2 goes to the driver , 1 is ground.

I have previewed the schematics and it needs be looked at in context that the Giskard schematic was prepared for the 2122 mid driver and the 2235 or 2245 woofer.

The mid driver is less sensitive than the 2123 mid driver.

I think the woofer used in the 4344mk11 is slightly more sensitive than the 2235 and is close to the 2245 woofer.

I will need to check the voltage drives but L pad on your mid driver may not have the same range of adjustment as the HF and UHF. This relates to your comment that up full is best.

[lordoflight]

Please see the graphs

pink noise full on lpads

1 all drivers together
2 mid only
3 hf only
4 uhf only
5 lf only

the dropoff is At arround 1200 hz.

Checked on l-pads connections they are OK.

[Ian Mackenzie]

Okay

High level over view.

The 1200 Hz dip is typical of these designs.

The magnitude of the dip will vary with the actual location of the Mic in the vertical plane.

The aim of the 4344mk11 mid filter is to equalise a rising response in the driver much like a bump and ensure a smooth transition to the horn.

The voltage drivers are tailored to compensate for driver intereaction in the crossover region.

The mid only response you have provided does not look right overall .

I recommend you carefully check at this stage that the Mid filter is correctly constructed. This is a complex filter with a number of elements. Check the resister values and locations , earth connections, Inductor values and locations.

That said the response above 1200 Hz is elevated and perhaps as a result of the mid filter not working correctly,

[Ian Mackenzie]

Please find attached the voltage drives into the standard 8 ohm test load.

The legend shows the colour :

Red is the Mid in the 4344mk11 filter
Blue is the HF with the 3155 Giskard HF filter
Purple is the HF with the 4344mk11 HF filter
Yellow is the UHF with the Giskard filter
Magenta is the UHF with the 4344mk11 filter.


The important point is these are voltage drives NOT the acoustic response or a representation of how the individual acoustic responses sum at the mic position.

The HF and UHF voltage driver are different for both the HF and UHF filters.

However the overall levels are similar.

The UHF filer for the 4344mk11 is lower and this is part of the design as the HF filter allows the horn to extent further into the UHF band (than the 3155 Giskard filter).

Some of these variations are made from empirical desicisions based on actual listening and voicing of the specific drivers in use. Note the HF driver in the 4344mk11 is titanium diaphragm dusted with aquaplas .

Before reading too much into these curves please review and check the mid filter.

[Ian Mackenzie]

If you have an opportunity to do more measurements disconnect the woofer and connect an 8 ohm resister or a similar to the woofer network

I have a feeling the woofer is effecting your measurements

[lordoflight]

I checked schematics and i have one question.

The first two 2.4 ohm resistors are in parallel or in series? because i thought they are in parallel and i replaced them with one 1.2 ohm resistor. is that ok?


secondly is it ok to replace the three parallel resistors 36 36 and 60 ohm with only one equivalent or the parallelism of resistors has a reasoning in the whole schematics? also i will not biamp.

Never got the time to repalce the wooder with a resistor and repeat the test but i will do it as soon as i can.

[Ian Mackenzie]

I checked schematics and i have one question.

The first two 2.4 ohm resistors are in parallel or in series? because i thought they are in parallel and i replaced them with one 1.2 ohm resistor. is that ok? Yes


secondly is it ok to replace the three parallel resistors 36 36 and 60 ohm with only one equivalent or the parallelism of resistors has a reasoning in the whole schematics? Yes also i will not biamp.

Never got the time to repalce the wooder with a resistor and repeat the test but i will do it as soon as i can.

Okay

[Ian Mackenzie]

Okay

I am curious what the mic position?

[Ian Mackenzie]

Evidence

http://www.audioheritage.org/vbulletin/showthread.php?26415-Beyma-CP21-F-tweeter

If we can confirm the sensitivity is down 5 db that is an easy fix

While this is a factor I would like to see the overall response somewhat flatter overall.

If you can provide detail of the iPhone app and how you did the measurement l can try and emulate your results.

Unfortunately l don't think l can source the Beyma driver easily to evaluate it.

[boputnam]

...I would like to see the overall response somewhat flatter overall.

If you can provide detail of the iPhone app and how you did the measurement l can try and emulate your results.It'd be nice to know/see the response curve of that iPhone mic. I've never seen one, but know for certain it cannot match an Earthworks M30... :skeptical:

[boputnam]

Ian: I just found this work by Faber Acoustical (http://blog.faberacoustical.com/2009/ios/iphone/iphone-microphone-frequency-response-comparison/). Unfamiliar to me so I cannot vouch for them or their work. But this response curve of built-in mics is interesting...

71193

Not knowing his model number, some of that observed hi-end boost, >3kHz, could be attributable to the mic itself...

[Ian Mackenzie]

Hi Bo

Good point. I had not considered that!

I was thinking if a try and to the same type of measurement here at home with my iPad mini or my iPhone 6s using the same app.

I could then compare to my Lms precision measurement (a gated sine wave ) and work out if it's a measurement "thing" or something else.

I must say some of those iphone analysers are quite nifty for what they are!

[grumpy]

With with the studio-six app, this little guy is still no earthworks mic, but a lot less hassle then a mic/stand/cables/audio I-O interface/computer :)

http://www.parts-express.com/dayton-audio-imm-6-calibrated-measurement-microphone-for-tablets-iphone-ipad-and-android--390-810

[Ian Mackenzie]

Thanks Grumpy

Agree we need to validate the measurements if possible.

[Ian Mackenzie]

While we are waiting on details of the mic location it lools like the response from 1200 Hz up is uniformly shelved some 6 db approximately.

I am theorising that is a level mis match on the HF and UHF filters or there is a lot of bleed from the woofer /mid cone as this is pink noise and not a solo tracking single frequency measurement.

Lets ee what our friend comes back with.

[lordoflight]

Never thought my issues will attract so much help. Thank you guys you are really something...


i used RTA app from onyx. i have an Iphone 5s. i took the measurements as follows

the full drivers at about 90 cm in front of the speaker

each individual driver was measured on axis at about 30 cm.

if you give me indications i can repeat the measurements with your parameters.

[Ian Mackenzie]

Hi

Was the full drivers measurement on axis with a particular driver?

[lordoflight]

no, somewhwre between the mid and the woofer

[Ian Mackenzie]

Okay,

Ideally (if practical) the measurements should be centred with the baffle at a height level with the bottom edge of the horn.

You might need to put the iPhone on a chair or a camera tripod.

Alternatively turn the enclosure upside down and raise the iPhone off the floor until it's at the right position.

The reason is the horn and slot have quite narrow vertical coverage.

[Ian Mackenzie]

I would try a do a full driver measurement at 2 metres distance.

This will help the wave fronts of the individual drivers merge in manner as heard by the listener.

You might get some interference from the room at low frequencies but we are interested in the response above 700 hertz.

[lordoflight]

2m in front of the baffle full on lpads between hf and uhf below the horn

[Ian Mackenzie]

Okay looks a lot more like it

What is the distinction of the two curves ?

Those curves are quite typical of this system.

The system response is not a magazine response curve.

My suggestion is to get a pair of the 2308 on EBay , they are about 120.00 new from Kenji.

The lense will modify the subjective and objective performance.

I recommend you considered obtaining the Dayton mic in Grumpys post.

This will provide more meaningful measurement as you can calibrate the mic.

Thanks

[Ian Mackenzie]

Evidence the data sheet on the Beyma slot

https://www.hifisound.de/out/media/pdf.php?r=DX0ABCtQAtByZFYfkCq2xvQvXxOj0yeiEzrn6WPe PJ3L3UgwnJGmD3Nr8UgYCLzS8RQfg7YnVhTJiycNsLkdVQ%3D% 3D.pdf

[ivica]

Evidence the data sheet on the Beyma slot

https://www.hifisound.de/out/media/pdf.php?r=DX0ABCtQAtByZFYfkCq2xvQvXxOj0yeiEzrn6WPe PJ3L3UgwnJGmD3Nr8UgYCLzS8RQfg7YnVhTJiycNsLkdVQ%3D% 3D.pdf

Hi,

1.
Using 2308 lenses would reduce higher FR-response due to the spreading off-axis response especially over 4kHz, that would reduce the response , as I have remembered about 3~4 dB.

2. Due to the geometrical differences between VHF and UHF drivers, a kind of comb-filter effect would be evident, especially round 8kHz~12kHz when passive network without time delay compensation is applied. Even with the time compensated electronic network, due to the drivers baffle mounting a kind of the same effect can be get, but I believe that all of that would be hard to be recognized by the listener.

3.
I think a kind of adequate measuring equipment has to be used not as mob-phone (garbage-in->garbage-out).

4.
I think that firstly drivers sensitivity has to be adjusted, and after that I would not expect any great problem at all as any of the used driver is of high quality and very linear, may be the best of any.
5.
I would prefer 2441/45/46/50 & 2311 instead of 242x & 2307, as such combination seems to me more natural sound in the lower VHF FR-spectrum (may be due to the larger diaphragm surface)

regards
ivica

[Ian Mackenzie]

Hi Ivica

Thank you for your insights.

Those Beyma slots are cute.

Did you ever get around to using them?

[lordoflight]

Hi guys. Thanks again for all your input. i will get a Dayton mic myself and make some measurements. However it seems that something eludes me. It is the sound that is not that clear. As it is a fundamental issue like for example the cables or the drivers are veiling the sound. I was under the impression that the sound of these
monitors should be clearer. Is it something that I miss? should i change the
diaphragms of the 2425's should i change theninternal cables? or this is the end sound of the apeakers and i should get along with that?

the thig is that i heard the spekers the guys at krs were rebuilding abd they sounded very crisp and clear. Not mine.... Please help

Thank you for your input.

[lordoflight]

Dear ivica,

Please see my comments inline.

thank you for your answers

Bogdan



Hi,

1.
Using 2308 lenses would reduce higher FR-response due to the spreading off-axis response especially over 4kHz, that would reduce the response , as I have remembered about 3~4 dB.

so should i purchase the 2308 lenses or not. Would that improve o. the sound clarity?

2. Due to the geometrical differences between VHF and UHF drivers, a kind of comb-filter effect would be evident, especially round 8kHz~12kHz when passive network without time delay compensation is applied. Even with the time compensated electronic network, due to the drivers baffle mounting a kind of the same effect can be get, but I believe that all of that would be hard to be recognized by the listener.

is there anythig here that may be improved?

3.
I think a kind of adequate measuring equipment has to be used not as mob-phone (garbage-in->garbage-out).
i will purchase a Dayton calibrated mic.


4.
I think that firstly drivers sensitivity has to be adjusted, and after that I would not expect any great problem at all as any of the used driver is of high quality and very linear, may be the best of any.


How can I adjust
the
drivers sensitivity?
5.
I would prefer 2441/45/46/50 (tel:2441/45/46/50) & 2311 instead of 242x & 2307, as such combination seems to me more natural sound in the lower VHF FR-spectrum (may be due to the larger diaphragm surface)


Is there a possibility to improve these 2425's ?
regards
ivica

[boputnam]

However it seems that something eludes me. It is the sound that is not that clear. As it is a fundamental issue like for example the cables or the drivers are veiling the sound. I was under the impression that the sound of these monitors should be clearer. Is it something that I miss? What are you accustomed to listening to, and at what gain, and it what setting? I'm not sure why these would sound "veiled" even if the iPhone mic is presenting an impression of a false boost of the HF - it's not that dramatic.

[ivica]

Dear ivica,

Please see my comments inline.

thank you for your answers

Bogdan

Hi Bogdan,

1.
Using 2308 lenses would reduce higher FR-response due to the spreading off-axis response especially over 4kHz, that would reduce the response , as I have remembered about 3~4 dB.

so should i purchase the 2308 lenses or not. Would that improve o. the sound clarity?

I would suggest You to use 2308 lenses when a horn like 2307/2312/2311 is used. The VHF driver would become much more linear FR response, producing less beaming of higher frequency range


2. Due to the geometrical differences between VHF and UHF drivers, a kind of comb-filter effect would be evident, especially round 8kHz~12kHz when passive network without time delay compensation is applied. Even with the time compensated electronic network, due to the drivers baffle mounting a kind of the same effect can be get, but I believe that all of that would be hard to be recognized by the listener.

is there anything here that may be improved?

I think that You do not need to do anything about,
Remember that for a decades JBL 43xx series were used in the most of audio studios, even though such effect have been present.
Using more expensive active network with time alignment possibilities, and filters with higher order slopes would make such comb-filter effect less pronounced

4.
I think that firstly drivers sensitivity has to be adjusted, and after that I would not expect any great problem at all as any of the used driver is of high quality and very linear, may be the best of any.


How can I adjust the drivers sensitivity?

Using L-pads and proper values resistors in the applied network. So a kind of measurement microphone (I think there is USB type from the "MINI-DSP" or other company) and sound card, or proper CD made disk as a source....
After I would suggest You to make off-axis measurements too. Something about 15~20 deg off-axis is a good approximation, of much more elaborated measurements .

5.
I would prefer 2441/45/46/50 (tel:2441/45/46/50) & 2311 instead of 242x & 2307, as such combination seems to me more natural sound in the lower VHF FR-spectrum (may be due to the larger diaphragm surface)


Is there a possibility to improve these 2425's ?

I do think that any improvements with the mentioned driver & horn combo can be get.

regards
ivica

[lordoflight]

I was listening to proac supertowers with the same setup based on my Accuphase E 360.

The sound was very crisp and clear on my Proacs unlike on the jbl's. The voice for example is very recessed on the jbls and i cannot find that vibrato that is usually present in the voice or in the classic guitars, the breath of the singer, the decay of the chords. The sound is damped somehow. I feel like for example if i change a cable it would clean up the sound. If i put the amp directly on the 2425 without the crossover, the sound seems more crisp and clear, but through the filter the sound gets muddy. I hope I am making myself understood.

[Ian Mackenzie]

Is your Proac the Mk1 or Mk11?

I agree the subjective aspect of the JBL performance is not up to your expectations.

Can you please elaborate on the JBL you heard at KRS?

[Ian Mackenzie]

I have been looking your crossover picture.

I assume all the filters are on this board?

But l cannot visualise the capacitors for the 2123 per the 4344mk11 crossover?

I can see the 3 30uf capacitors for the Low pass filter.

But I cannot see the 2 36uf. Capacitors, one series and one shunt in the mif filter.

If l am not following your layout let me know but it look like you have somehow included the capacitor values from the 3155 in the mid filter.

Again if l am not following your layout let me know.

But as l commented previously in your measurement of the mid filter, this would explain comments to some extent as the 2123 is overlapping the 2435 horn hence your comment on phase issues.

The crispness or instrumental clarity comes from the mid cone. If l am correct your mid cone is going too low and too high.

To validate this Are you about to take clearer pictures and best sketch on a sheet of while paper the layout with the position and values of the crossover parts.

If we can validate the implementation of he crossover then it much more time efficient to tune the system correctly.

Below l have attached the 4344mk11 crossover, your crossover layout and the 3155 crossover for clarity.

The last pic (my dog) l hope will help relieve any frustration.

Your mid filter should have the same equivalent values of all Inductors, capacitors and resisters as in the top section denoted by MF

[Ian Mackenzie]

For comparison I will do a measurement on my 2123 with the 4344mk11 filter this week.

[lordoflight]

the two 36 microF are in fact pairs of two 18 with a 0.01 bypass. They are near the litz aircoil.

and thank you for the dog. Its really cute.

the mid filter and low filter are full 4344mkii. everything should be in there. If I missed
anything I would be more
than grateful to point that out for me.

[lordoflight]

here is the clearer picture.

[lordoflight]

Is your Proac the Mk1 or Mk11?

I agree the subjective aspect of the JBL performance is not up to your expectations.

Can you please elaborate on the JBL you heard at KRS?


It was an mki with titanium hf drivers from MBQuart. However it was very sibilant at times. I had to tMe that sibilance using some copper cables.

as for the krs speakers, is what i heard on youtube. Never been to Japan:) But they sounded clear and crisp even in the youtube recording

[boputnam]

If i put the amp directly on the 2425 without the crossover, the sound seems more crisp and clear, but through the filter the sound gets muddy. I hope I am making myself understood.You are understood - thanks for the patience.

I hate to bring this up, but so far Ian is curious about intra-cabinet phase issues. Your post makes me curious about inter-cabinet phasing. So, does this symptom occur with mono (L or R, alone), or only when you are in stereo (L/R) mode?

And, apologies - a full 50% of the time I inevitably run into systems that are either completely, or by element, unintentionally out-of-phase. :(

[lordoflight]

You are understood - thanks for the patience.

I hate to bring this up, but so far Ian is curious about intra-cabinet phase issues. Your post makes me curious about inter-cabinet phasing. So, does this symptom occur with mono (L or R, alone), or only when you are in stereo (L/R) mode?

And, apologies - a full 50% of the time I inevitably run into systems that are either completely, or by element, unintentionally out-of-phase. :(


Thanks for the input. is there any way to put the speakers in phase both intra and inter so to say :). the polarity of the drivers and of the filters are a little misleading :)

[lordoflight]

lf red wire - from the - to the red post on the speaker
mf red wire - from the 2 on the lpad to the red post on the speaker


hf red wire - from the - on the network to the red post on the speaker
uhf red wire - from the - on the network to the black on the speaker. beyma has normal polarity.


lf and mf they are supposed to be in antiphase?
i'm confused

[Ian Mackenzie]

It was an mki with titanium hf drivers from MBQuart. However it was very sibilant at times. I had to tMe that sibilance using some copper cables.

as for the krs speakers, is what i heard on youtube. Never been to Japan:) But they sounded clear and crisp even in the youtube recording

Okay

Apparently the mk11 with a soft dome is far better
I am sure the Youtube was impressive but you were listening to your pc speakers.

Subjectivity you are comparing a finished hifi speaker to an in finished pro monitor.

Without the 2308 lense it is not going to work properly.

I have a pair of nice metal 2308 if you are interested.

[Ian Mackenzie]

the two 36 microF are in fact pairs of two 18 with a 0.01 bypass. They are near the litz aircoil.

and thank you for the dog. Its really cute.

the mid filter and low filter are full 4344mkii. everything should be in there. If I missed
anything I would be more
than grateful to point that out for me.

Okay but how can we help you validate your implementation if we cannot follow your layout?

There is a problem some where so if you can sketch the layout that would be great.

I plan to run a series of measurements tomorrow with the iPhone and the Dayton mic.

I suggest that you only listen to one speaker and provide feedback

[Ian Mackenzie]

lf red wire - from the - to the red post on the speaker
mf red wire - from the 2 on the lpad to the red post on the speaker


hf red wire - from the - on the network to the red post on the speaker
uhf red wire - from the - on the network to the black on the speaker. beyma has normal polarity.


lf and mf they are supposed to be in antiphase?
i'm confused

All the mid hi and uhf are in phase with each other.

The woofer is wired red to red terminal

The mid hi and uhf are wired red to black terminal

[lordoflight]

I killed it. i found the issue. it was a phasing issue. everything was messed up. i decided to use Ian's indications and look at the schematics not the wiring. Long story short the sound is now clearer and crisper is like i purchased another pair of speaker. It sounds as it should.


I am sooooo happy! Thank you Ian, thank you Ivica, thank you Bo for your phasing suggestions.

i will make new measurements to see if there is a difference now.

Happy Easter everybody.

Bogdan

[Ian Mackenzie]

That's good news.

It will only keep getting better and better now.

[boputnam]

All the mid hi and uhf are in phase with each other.

The woofer is wired red to red terminal

The mid hi and uhf are wired red to black terminalThat is true with this homemade crossover. Typically, vintage JBL networks are solid colors to Red speaker posts; anything black stripes connects to Black posts.


I killed it. i found the issue. it was a phasing issue. everything was messed up. i decided to use Ian's indications :yes: and look at the schematics not the wiring. Long story short the sound is now clearer and crisper is like i purchased another pair of speaker. It sounds as it should.


I am sooooo happy! Thank you Ian, thank you Ivica, thank you Bo for your phasing suggestions.

i will make new measurements to see if there is a difference now.

BogdanThat's terrific, Bogdan! :hurray:


That's good news.

It will only keep getting better and better now.Dammit - this was just getting interesting!
Oh well, off to the next JBL triage...:wave:

[Ian Mackenzie]

That is true with this homemade crossover. Typically, vintage JBL networks are solid colors to Red speaker posts; anything black stripes connects to Black posts.

That's terrific, Bogdan! :hurray:

Dammit - this was just getting interesting!
Oh well, off to the next JBL triage...:wave:


We're not done yet!

Hang around Pal we need your wisdom with these hand held rta thingies.

[Ian Mackenzie]

I suggest a program for peak performance:

1. Test and learn rta measurements and Calibration (Bo)
2. Test & Confirm filter functionality (lan)
3. Adjust and trim L .Pads (lan & Bo)
3. Evaluate overall in room response with tra (Bo)
4. Install 2308 lense and re test

[boputnam]

I suggest a program for peak performance:

1. Test and learn rta measurements and Calibration (Bo)
Well, those of you Dayton Plug-in Mic fans, here's the calibration file that came with Ian's recent purchase - I stuck it into Excel, just for fun.

71277

There's a total of 256 filters, varying from 0.55 Hz to 534.1 Hz in width. Safe to say, I've never (ever...) seen a correction approached like this. I shudder to think of the phase response...

Now, it's up to Ian to figure out #2 - to measure that his on-board mic is actually flat!

[lordoflight]

Now I need indications on:

How to tune the l-pads
and
How to tune the bass ports (i made mine adjustable. One cadboard tube sliding in another cardboard tube)
How can i determine the best frequency for the best bass reaponse on these speakers? is there a simple way using a multimeter or based on box resonance calculators?

As I know the volume is hard to determine given the padding inside the box. Please help.

I will purchase the 2308 lenses andinstall them and the dayton mic to repeat tests.

Also i realized that the left mid speaker was badly repaired and has a ton of glue at the neck of the membrane where it meets the spider on the outside. I am afraid to look below the cap. Thing is i will have to purchase another 2123 dont know if it worth reconing with aftermarket recone kit.

[boputnam]

How can i determine the best frequency for the best bass reaponse on these speakers? is there a simple way using a multimeter or based on box resonance calculators?

As I know the volume is hard to determine given the padding inside the box. Please help.


WinISD (http://www.linearteam.dk/?pageid=winisd)

I used this a lot, back-in-the-day. There may be something more current (Widget ?), but this used to do the trick...

[Ian Mackenzie]

Now I need indications on:

How to tune the l-pads
and
How to tune the bass ports (i made mine adjustable. One cadboard tube sliding in another cardboard tube)
How can i determine the best frequency for the best bass reaponse on these speakers? is there a simple way using a multimeter or based on box resonance calculators?

As I know the volume is hard to determine given the padding inside the box. Please help.

I will purchase the 2308 lenses andinstall them and the dayton mic to repeat tests.

Also i realized that the left mid speaker was badly repaired and has a ton of glue at the neck of the membrane where it meets the spider on the outside. I am afraid to look below the cap. Thing is i will have to purchase another 2123 dont know if it worth reconing with aftermarket recone kit.

Okay one thing at a time.

There is no point adjusting he L pads until we get the bass tuning right.

The tuning of these woofers (2235) is generally in a net internal volume in the order of 5 cu ft 3 and tuned to 34 hertz.

We can simulate the tuning quite accurately if we know some basic information.

So l can calculate the port length can you please advise :

The internal enclosure dimensions?

The external dimension of the midrange enclosure?

The internal diameter of your ports?

(If you have calculated the net internal enclosure volume for the woofer please advise)

The enclosure stuffing should be around 2.54 cm on sell internal surfaces except the front baffle.

[Ian Mackenzie]

Well, those of you Dayton Plug-in Mic fans, here's the calibration file that came with Ian's recent purchase - I stuck it into Excel, just for fun.

71277

There's a total of 256 filters, varying from 0.55 Hz to 534.1 Hz in width. Safe to say, I've never (ever...) seen a correction approached like this. I shudder to think of the phase response...

Now, it's up to Ian to figure out #2 - to measure that his on-board mic is actually flat!

I plan to use a precision analyser, the LMS to measure the driver curves and simulate the system in Leap. I will do the iPhone FFT and if we can match that with LMS curves uand with our friend's system l then the implementation is good.

In greater depth the mid filter in the 4344mk11 is entirely different to the 3155 as is the 2123 from the 2122.

So we need to ensure the phase response at the mid and horn crossover point is optimal.

The crossover can look like it sums flat but phase can be sub optimal.
I can do that remotely.

The RTA is not loudspeaker design tool in that it cannot driver phase/filter phase transform plots as far as l am aware.

As stated the intent is to evaluate the 4344m11 and blended Giskard 3155 networks.

I will report on the findings and take the next steps.

As far as flatness is concerned I think it's about the adjustment of the Lpads absolute levels of the Lf, Mf, Hf and UHF given its a mixed bag.

Graphic or Parametric eq is out of scope so we are dealing with the accuracy of SPL measurements boost or cut.

Assuming the Lpad plate of the 4344mk11 is realistic and given the 2235 is about 93 db ref 1 watt versus the 4344mk11 ref sensitivity of 95 db (the mid Lpad needs to be adjusted down to -minus or cut 3.5 db from full.)

So as a starting point the iPhone needs to be able to resolve a 3.5 db difference with the SPL meter function. All we are doing is turnings up the mid LPad full then cut 3.5 db.

The HF and uhf LPads can be referenced to the 4344mk11 voltage drives from the input voltage to the network.

This will get the system dialled in.

Given the nature of the horn response plots it's difficult to determine even from smoothed curves the actual level ie do you take the average or the peaks of the horn to compare with the mid? I cannot imagine doing this with pink noise on an RTA. We are talking 1/2 db
Window pass / fail.

So l prefer the SPL difference calculation and rely on the safe harbour of the prior engineers to get the mid , horn and slot to match.





This is a close starting point for the mid/LF level match.

[lordoflight]

I just ordered the Dayton IMM6 mic. When I will receive it we're in business.
Meanwhile I will take again the measurements of the box to give them to Ian for the ports calculations.

When we start to set out the L-pads I am expecting step by step instructions not to miss anything.

I will be your puppet Ian. :)

[lordoflight]

big box inside dimensions are
125 cm
46 cm
33 cm

please assume 1.5 to 2 cm padding back and laterals


small box exterior dimensions

23.5 cm
28 cm
30 cm

vents interior diameter 10 cm the inside cardboard tube 10.5 the outside cardboard tube.

[Ian Mackenzie]

big box inside dimensions are
125 cm
46 cm
33 cm

please assume 1.5 to 2 cm padding back and laterals


small box exterior dimensions

23.5 cm
28 cm
30 cm

vents interior diameter 10 cm the inside cardboard tube 10.5 the outside cardboard tube.

Okay

You have 170 litres.

The effect of the stuffing is to make the box appear 10% larger and then we allow for the space occupied by the woofer, bracing and the ports say 10% and the box is 170 litres net or 6.00 cu ft3.

For large boxes this is a generalised assumption but is +-5%.

The ratio of woofer Vas to the enclosure volume is such we can live with 5%.

If you put more than simple shelf braces in your enclosure let me know.

You advise you have a passive crossover so the woofer has a small dcr resistance in the series inductor.

The simulator will do the rest

Okay adjust the length of both ports to 27 cm in length. (28 Hz)

Try that and see how it goes.

If you feel you prefer a more aggressive bass adjust the ports to 22 cm (30 Hz)

Either Tuning will yield a response down to 32 hertz.

Much of this will depend on your room.

[Ian Mackenzie]

I just ordered the Dayton IMM6 mic. When I will receive it we're in business.
Meanwhile I will take again the measurements of the box to give them to Ian for the ports calculations.

When we start to set out the L-pads I am expecting step by step instructions not to miss anything.

I will be your puppet Ian. :)

In the short term set up your iPhone for SPL with the signal generator set to 800 hertz

Place the mic immediately in front of the mid cone (10-20 cm) held steady with a stand.

Start the test generator.


With the Lpad of the mid up full adjust the Lpad until the reading falls by 3.5 db.

Keep the total spl under 90 db to avoid clipping the meter.

You will find any movement in the vicinity of the mid cone will effect the reading so it will take a few tries to get this right.

If your SPL does a digital read out then it will be more accurate.

Take note of the actual SPL value and distance of the mic from the baffle.

Do exactly the same for the other speaker.

Raise the mic until it's centred in front of the horn at the same distance

Adjust the sine wave frequency to 4000 Hz.

With the HF Lpad up full turn down the Lpad until the reading is the same SPL as the mid cone.

Do the same for the other speaker.

Repeat for the slot at 9000 Hz

This will give you ball park adjustment and both loudspeakers will be exactly the same.

You will really notice the precision of this approach with the sound stage as you have 4 drivers you are trying to adjust. It will jump right out at you!

It's futile to try and do this by ear without a meter to measure and record the changes.

Given the mic is not calibrated yet its not completely accurate but it's in the ball park.

Attempting to adjust the Lpads with pink noise at 2 metres is not anywhere near as accurate as spot frequency net SPL measurement in the near field for adjusting the actual levels in small amounts. If you look at the scale of your RTA is was 6 db per division and the room is effecting the overall measurement.

We need 1.0 Db resolution when adjusting the Lpads and preferably 0.5 db resolution.

Can you imagine buying a phono preamp with -+ 1.00 db variation on the frequency response! That would be unacceptable.

So why why would not want to adjust your loudspeakers so at least both left and right are identical.

(The 2308 lense will require re adjustment of the horn).

Once you get the Dayton mic we can do some other measurements and you will be more informed about how the loudspeaker is performing.

[lordoflight]

thank you ian. I have solid bracing inside the box. I tried to replicate the 4346 from krs. double backbone and double struts between the baffle and the back above the small box.
my room is the worse kind of room, approx 60 sqm room with plasterboard walls which absorb the bass. Can i improve something here?

[lordoflight]

Ian,

i tried to do it as explained but my fear is that my 2123's are busted. at the same volume the same distance i have about 3 to 4 db difference between the right and the left speaker. Next if i'm trying to bring down the horn and the slot to the same spl as the one of the mid after the 3.5 db reduction i will have to go way down on the l-pads which translates into a very damped sound. So there is a clear imbalance between the mid speaker and the rest of the drivers.

I believe i should change the 2123 firat and repeat the test.

[Ian Mackenzie]

Okay,

That's a bitch.

Are you sure the amp balance for L+R is the same?

You could swap the input of the amp to check this.

Can you bypass the crossover and test each 2123 individually with the direct signal from the amp with the SPL meter.

This will prove if it's one of the drivers or a malfunction of one of the crossovers.

If the later we will need to measure the voltage drives of the mid filter and compare to each other.

We can investigate the horn filter following this step.

PS. If it's any comfort it's unlikely one of the drivers would have different sensitivity unless one was 8 ohms while the other was 16 ohms.

They are either working or blown.

[lordoflight]

left is 8 ohm right is 6.9 ohm measured at the posts.
difference in spl left is a full 5 db lower than right. The drivers were fed directly from my accuphase. no balance issue on the amp. I made the test with 800 hz sine wave.

I think i need to buy other speakers.

[lordoflight]

the membrane is ripped under the cap and was glued to the spider in the worse manner possible And with a ton of glue. 😡

No chance... need new speaker.

[Ian Mackenzie]

Okay,

"Also i realized that the left mid speaker was badly repaired and has a ton of glue at the neck of the membrane where it meets the spider on the outside. I am afraid to look below the cap. Thing is i will have to purchase another 2123 dont know if it worth reconing with aftermarket recone kit."

So it was the left reconed speaker that is faulty.

That is the voice coil resistance?

Have you compared how it sounds to the left speaker?

Suggest you do a full response measurement and compare both drivers to confirm the overall sensitivity is lower (not just one frequency).

[boputnam]

Suggest you do a full response measurement and compare both drivers to confirm the overall sensitivity is lower (not just one frequency).Uh, er, it might be a good idea to go back to Square One and determine the status of each element. Working hard on "perfecting" output is moot if the last elements in the signal path are questionable...:banghead:

[Ian Mackenzie]

Uh, er, it might be a good idea to go back to Square One and determine the status of each element. Working hard on "perfecting" output is moot if the last elements in the signal path are questionable...:banghead:


Yes agreed, l meant the driver only (sweep)

We are trying to determine if they are H or J ?

[Ian Mackenzie]

The weather is unsuitable for ground plane measurements outside so l have set up testing of the network (per hyprid cover) in my HT room.

I have obtained some parts and wound inductors with the LCR meter for the Giskard HF and UHF.

The plan is to see what the 4344 response should look like with the hyprid network and sort out the L pad levels.

I also want to validate the iPhone measurements against LMS.

If the phone is good we will try and come up with a best practise that can be adopted by other iPhone users.

[Ian Mackenzie]

Attached is the 4344mk111 (full crossover) will all L pads up full.

Green is near field, Pink is gated 2 metres.(LMS)

The lower end is the room effects.

Overall this is as expected for this type of system.

Once l have set up the Giskard filters I will measure and compare with the iPhone

[lordoflight]

Great work Ian,

Now we know how it should look like. looking forward to seeing the giskard graph.

how will you combine the two parts of the filter.

I am currently ordering the H version of 2123s from japan. one year guarantee.

I know that finally our struggle will pay off. Already the sound in my room has improved alot and I cannnot waot to see the final result of our efforts.

[Ian Mackenzie]

I will simply connect it like the other filter but disconnect the 4344mk11 filter in its place.

[lordoflight]

My Dayton mic has arrived. Now I can do some serious measurements. Awaiting instructions....

[lordoflight]

Any recommendation on the internal wiring of the speakers?

what should I use? solid core copper or stranded copper? what caliber?

[lordoflight]

latest measurements. pink noise full on lpads one meter on axis with the horn with the lenses on. calibrated dayton mic7200972010

[Ian Mackenzie]

That looks much better

[lordoflight]

I want to extend the feeling of gratitude to the members of this forum, and especially to Mr. Ian McKenzie who is a real professional and a great man. He instructed me remotely to succeed in getting an incredible sound out of my DYI 4346. Now looking back it was a long road to achieve the sound that I am enjoying right now.

There is still room for improvement but the speakers already sound divine.


I heard some so called hi end speakers at vendors and from where I stand I believe the vendors are ripping people off. My DYI speakers sound better than 90 % of what is in the vendor showroom where many of the speakers cost more than 10000 EUR.

Again Thank you guys and thank you Ian.

I will come back with a new and improved graph since although you cannot hear them at least you can see the sound :).

[Chas]

Good to hear things worked out with the 2123H. It's a great driver.

[lordoflight]

Is anyone aware of beryllium diaphragm for JBL 2425 1" I looked for it at Truextent and apparently the only make such diaphragms for bigger diameters speakers.

I understand that replacing the stock Aluminium diaphragm with a beryllium one will make a very significant difference.

Awaiting opinions.

[Chas]

Is anyone aware of beryllium diaphragm for JBL 2425 1" I looked for it at Truextent and apparently the only make such diaphragms for bigger diameters speakers.

I understand that replacing the stock Aluminium diaphragm with a beryllium one will make a very significant difference.

Awaiting opinions.

I changed mine out to TAD TD2002 Be years ago in my 4345's and never looked back. Not an inexpensive option, but well worth it.

[Ian Mackenzie]

http://www.proaudiodesign.com/search.php?mode=search&page=2


List price $1750 each

Looking at truexent diaphragm $650 each and second hand compression driver like 2450 $1000-1500 each and the risks (second hand drivers? redesign of the network, acquiring a 2311 horn) the Tad is a drop in replacement.

Radiian make a 1 inch driver Radian 475BePB that had use a Radian Be diaphragm $549 each . The bolt circle need to be checked and network may need re design.

[boputnam]

I changed mine out to TAD TD2002 Be years ago in my 4345's and never looked back. Dood. Oh, dood...:nutz::nutz:(two, for the stereo pair).

[Chas]

Dood. Oh, dood...:nutz::nutz:(two, for the stereo pair).

Yup, they continue to amaze me, Bo!

[focalguy]

I want to extend the feeling of gratitude to the members of this forum, and especially to Mr. Ian McKenzie who is a real professional and a great man. He instructed me remotely to succeed in getting an incredible sound out of my DYI 4346. Now looking back it was a long road to achieve the sound that I am enjoying right now.

There is still room for improvement but the speakers already sound divine.


I heard some so called hi end speakers at vendors and from where I stand I believe the vendors are ripping people off. My DYI speakers sound better than 90 % of what is in the vendor showroom where many of the speakers cost more than 10000 EUR.

Again Thank you guys and thank you Ian.

I will come back with a new and improved graph since although you cannot hear them at least you can see the sound :).


I agree.... I have gone with discrete amplification and am using the dbx venu 360 crossover and auto eq feature.... i will never go back to passive again. the dynamics, clarity and no more issues with crossover losses in power, phase, timing etc... the eq feature works awesome. check out my jbl 4345 comments in the diy section . I used the jbl 2123H driver also and it was so easy to integrate with the electronic crossover. Here are some videos ... https://www.youtube.com/watch?v=Djgjl8QTMyU


(https://www.youtube.com/watch?v=Djgjl8QTMyU)

[lordoflight]

Hello boys and girls,

This post is a heads-up and is written as a result of six months of extended testings and listening sessions, soldering and changing cables and modifying boxes and frustrations.
The story starts around March 2016 when I wanted to use a JBL 2123 in a 4343 configuration namely 2235 for bass, 2425J for HF and 2405 for UHF. Given this combination I was told to use a combination of filters. namely the 2123 section from 4344mkII and HF and UHF from Giskard schematics. I did just that. The result was a dull and veiled sound. Of course I refused to think that the crossovers are causing this and I started to investigate other potential causes. Cables, Lpads, Star grounding, boxes posts, type of wires, caps, resistors... and the list may go on and on. All this time my guide and guru was Mr. Ian McKenzie AKA the Doctor who tutored me and to whom I want to thank again for his patience and dedication resisted heroically to my avalanche of frustrations and tried to help little ol' me.

At some point he said he had enough of this and took the road of recreating my hybrid filters and compared them with a pair of pure and simple 4344MKII. The result was.... THERE IS NO COMPARISON. The sound of the hybrid is light years away from the sound of a HIFI system.

SO IF YOU THINK YOU WANT A HYBRID CROSSOVER, THINK AGAIN. OR OTHERWISE PUT DON'T DO IT. IT WILL ONLY FRUSTRATE YOU.

you may find here the result of his research http://www.audioheritage.org/vbulletin/showthread.php?35325-JBL-4344-Clones&p=399476&viewfull=1#post399476

Your brother in DiY,

Bogdan

[Ian Mackenzie]

Congratulations are in order.

Don't thank me.

Thank my wife for believing in what we were doing and the whole palaver as she quite rightly puts it.

[lordoflight]

Now I can say that I have managed to solve all my problems and listen to some music on my DYI 4346. All the thanks go again to Mr Ian MacKenzie who lend me his First Watt B$ until my handcrafted crossover is ready.
I just wanted to shar with you guys my experience and the results in a short post. My rig is as follows:
Source: Audirvana on MacMini
Digital USB to coax adapter Gustard
DAC: Theta DS Pro Basic 2 (a bit old but Excellent sound)
Preamp: Accuphase e360 preamp section (I use the Integrated Accuphase as separates)
Crossover: Currently Ian's First Watt B4
HF amp: Accuphase e360 Power section
LF Amp: QSC GX5 2x500W (Monster)
Crossovers : Handmade not charge coupled but boutique capacitors.
Speakers:
UHF:JBL 2405 original Dias (replaced by me)
HF JBL 2425J Original Ti dias (maintained by me)
MF: JBL 2123 Original No intervention
LF: JBL 2235 ( E140 chassis reconed with an aftermarket 2235 cone with added weight in order to reach proper response considering the stronger magnet)
Cables: Twisted coppers solid cores soldered on 8 poles Speakons.

Sound.... Incredible!!!

It is full and the bass is amazing. Thunderous I might say. Crisp imaging. And warmth.

[Earl K]

Nice!

:)