JBL 4344 Clones

[christo]

Here some picture of the JBL 4344s I recently finished.

I’d like to start by thanking all the DIY’s who post in this forum as without your contributions my project would not have been possible!

Components 2235, 2123, 2426H, 2405, DIY charged couple cross-overs.

I used the plans from Ayaboh – fabulous saved many hours/days/weeks of planning. I did not really deviate from them.

Wood used 1” MDF for all panels 3/4 in Baltic birch for the baffle, 5/8 Baltic birch for the 2123 dog box and panel bracing, walnut veneer and trim. Stainless Steel for the LPAD focial, 2405 mounting plate and 2308 lens. White birch veneer stained black for the baffle.

[christo]

Everyone like tools, here’s the machine I used for the project members from down under may recognize it. I bought it back in the late 80’s for my last speaker project. It’s as accurate as any real saw/router table and folds down and can be put away for those of us that can’t dedicate space to a permanent workshop. Drawback is that no third party jigs will fit; you have to make your own. The extension table is wonderful as ripping a 4x8 is easy. Well it was when I was younger for this project I went to my local RONA and had all the pieces cut close to the size and then made all the final cuts at home – my version of quality control. Beside do you know how much a 1” 4x8 of MDF weights?

[christo]

The dog box ready to be assembled.

[christo]

On the back panel I inserted dowels where the screws for the Speakon connector and the battery box will be located.

[christo]

I clamped the back access panel frame pieces in to place to get the correct fit.

[christo]

I disassembled the mock up without disturbing the frame pieces and then drilled the dowel holes through the frame and in to the panels. Somewhat unorthodox but it was late in the fall and I was in a hurry.

[martin2395]

Lovely!
It's only a pity that people use that ugly unit layout ;)
I went with the 4343 just because of that, the other ones were giving me a slight OCD attack. ;)

[christo]

Not sure why I was in a hurry as I stalled over the winter doing the cross overs. I have no experience with building crossovers and muddled my way through it. So I started with mock ups (left) and progressed to final version (right).

Much thanks to Chas who provided much guidance with the cross overs!

[christo]

This is an easy way to get the lacquer off the coils. Let the wire sit in the furniture stripper (water based) 15 to 25 minutes depending upon how thick the lacquer is and then brush it off with a brass bristle brush. Worked great!

[christo]

I stacked the 2123 xover on top of the 2405 and 2426 circuits.

[christo]

First pass at testing the circuits

[christo]

Second pass with all the wiring complete and ready to be installed, Note the terminal block installed on top of the dog box. I also color coded all the wiring with shrink wrap to ease the correct wiring so even if you lose the wiring diagram if you follow the color coding things will still work.

[christo]

Battery Box

I used a 9 volt guitar battery box mounted on the back panel to make things easy.
http://www.guitarpartsresource.com/electrical_other.htm

The box itself is deeper than one inch so I had to manufacture an internal cover with connections to inside the cabinet.

[christo]

Here I’m gluing the battery box cover on the back cabinet panel.

[christo]

To get the wires in to 2123 I used speaker binding posts

[christo]

Here’s a fun picture, I used this set up get the correct length for the internal wiring harness

[christo]

This is a cut-away side view of the cabinet; I included it to show how much room (NOT) there is inside the box. Once the ports are installed you’ll never be able to reach to the top of the cabinet without the top back access panel.

[christo]

As I manufactured my own 2308 lens I also had a template made and pre-drilled the screw holes.

[christo]

I attached the fiberglass prior to the final cabinet assembly. Taped off all the areas where the glue is not to go. I used 3M Super 77 glue.

[christo]

Finally assembling the cabinet.

[christo]

Installing the xover prior to installing the front baffle it’s a lot easier working though this gap than a woofer hole

[christo]

Wiring up the LPADs to the terminal block prior to dropping the baffle in to the cabinet

[christo]

Inserting the baffle in to the cabinet, I taped off the edges with masking tape to keep the glue off the finished baffle.

[christo]

Once the dowels holding the baffle in have been trimmed off, it’s time to assemble the walnut trim. If you look at the left front piece of trim you can see I cut a recess out of the bottom of the trim and left a 1/8 inch overhang. This allowed me to square up the trim (overhang) against the side of the cabinet panels. When the trim is cut trimmed flush to the sides the overhang is removed.

[christo]

This is the saw I used to cut the 45 degree angles for the trim - NOBEX excellent tool from Sweden.
http://www.nobex.se/GB/engelska.html

It is also big enough to cut the 4 inch ABS pipe (sewage pipe) for the port tubes and actually has a blade specially made for ABS. I had to be careful cutting the trim I had no spares for any screw ups.

[christo]

Cabinets with the excess walnut trim removed and ready for the veneer. I placed a piece of plywood on top of the trim to allow for a stable platform for the router when cutting the walnut trim flush to the sides of the cabinet. Its impossible to hold the router square on just the trim alone.

[christo]

Top left – Tools for the job note no veneer roller required, just start the iron in the middle and work to the outside.
Top right – Using a sponge roller to apply the glue, ensure a uniform thin coat is applied. I applied three coats per surface.
Bottom left – As the glue dies it changes color the glue on the cabinet in the background is almost dry.
Bottom right – Applying glue to the veneer, I taped the veneer to wax paper you don’t have to worry about the outside edge as they will be cut off.

[christo]

The cabinets are ready for finishing; I used Watco Danish Oil - “Natural”. I applied four coats of the Danish oil, first coat wet sanded the first coat and then three coats of clear varathane and a final coat of satin varathane.

[christo]

The only three screws used in the entire cabinet basically to hold the cross brace while the glue dried.

[christo]

I used Silicone to glue the ports (ABS or sewage pipe) in to the cabinet

[christo]

To make the grills I used a pine frame with birch corner braces. The grills worked out better than I thought they are light as a feather and ridged.

[christo]

The top frame of the grill.

[christo]

Finished, don’t for get to make finger grips to ease the removal of the grill. I use the large ball and socket type pins they really lock together and you really only need the corners the middle pins can be skipped.

[christo]

Once they were finished I listen to them for a few months and was not very happy with the 2426H driver with a Ti diaphragm. Chas suggested I upgrade to TAD-2002 – WOW best move ever. This upgrade made the whole project worthwhile. The speakers actually do image (to a degree) and when sitting in the sweet spot the sound stage has some depth. I highly recommend this upgrade for JBL four way monitors, you will not regret it.

[christo]

The TAD-2002 driver is tall, so tall I have to cut a 6x6 ½ inch deep recess in the back panel to get them to fit.

[christo]

I’m tri-amping the system with KRELL EVO 600’s for the 2235s and a KRELL EVO 302 handling the Hi-Pass duties with a Marchand XM9 active cross over. The Monster power filter is only used for the Marchand crossover and the Dayton Subwoofer amplifier.

[Odd]

Fantastic job you have done here. And nice pictures.

Thank you for sharing.

[JBLP]

RESPECT!!!! :applaud:
What a nice job, nice tools, nice finishings, nice......, nice......., nice......
I'm jealous.

Can you show how you trimmed the walnut trim in angle on the front?,
You just started making the X-overs....??, how did you know what to buy (quality)?,

Really nice!

[wpod]

But what I really want to see is some pictures of how you assembled that beautiful stainless steel lense. Did you weld the pieces together, tap and thread them, or use regular old nuts and bolts? You might want to consider making a few pair for some of us 4343/4344 owners , I know that I'd certainly be interested. Great work! regards, Paul from Providence

[cvengr]

Outstanding work. Thanks for sharing. Beautiful veneer work.

[ivica]

Hi christo,

Nice WORK.
Many thanks for the info You have shown us.

Regards
Ivica

[Mr. Widget]

Congratulations on a fine project.

Beautiful work and excellent documentation!

I know what you mean about inserting those TD-2002s... Bo and I heard the same thing when we popped a pair into his 4345s. It was a surprising difference!


Widget

[just4kinks]

Nice work!

I would also like to see more details about your stainless lenses. I daydreamed about it, and my father-in-law even offered to help with the metalwork, but in the end I went with the Kenrick lenses.

[christo]

I got my inspiration from this thread and I had a friend who works in the SS fabrication industry. I described what I wanted done (individual lens plates that attached to the baffle with screws) and gave him an old L91 lens to bring to work his CAD engineers did the rest.

http:// (http://www.audioheritage.org/vbulletin/showthread.php?8635-2308-Lens/page2)www.audioheritage.org/vbulletin/showthread.php?8635-2308-Lens/page2 (http://www.audioheritage.org/vbulletin/showthread.php?8635-2308-Lens/page2)
Post 22

61329

The lens is made up of two pieces a top plate (right) and all the other plates (left) which have a notch cut out to allow the sound through. I kept the design simple.

6133061331

[christo]

The pieces were laser cut and then bent in to shape it is 14 gauge SS and once bent will not lose its shape. I believe the downward slate is 32 degrees. As the gauge of SS used is quite heavy each lens weighs 200 grams 11 lens used per horn equals 4.8 pounds!

[christo]

Each piece is individually attached to the baffle with screws. To ensure the lens line up uniformly I had a template made to allow for the predrilling of the plate screw holes. The template is centered on the baffle by using the existing bolt holes for the horn. That was my original plan but it didn’t end up that way (wrong size holes for centering), for me it did not matter as I had not drilled the bolt holes in the baffle yet so I was able to just use the pilot holes.

61333

Here is the layout of the template

61334

Once the lens mounting screw holes have been predrilled it is just a matter of attaching the plates one at a time from the bottom up.

61336

[christo]

For the SS LPAD foilcal I used a vinyl sticker, I had sections of it made clear so the SS is still visible beneath it.

[christo]

But what I really want to see is some pictures of how you assembled that beautiful stainless steel lense. Did you weld the pieces together, tap and thread them, or use regular old nuts and bolts? You might want to consider making a few pair for some of us 4343/4344 owners , I know that I'd certainly be interested. Great work! regards, Paul from Providence

Thanks Paul

If any members are interested in the Stainless Steel 2308 lens please PM me, if there is sufficient interest I'll see what can be done – not promising anything at this point.

Chris

[christo]

Can you show how you trimmed the walnut trim in angle on the front?,


The trim is chamfered at 30 degrees. Scrap wood is clamped at each end to allow the router bit guide wheel to roll straight off the edge of the box. Once each side has been chamfered you can’t do this for the top or bottom you have to use a straight edged to guide the router. Cause if you don’t the router guide wheel would follow the chamfered edge and a grown man will cry. The following pictures are from my B380 project but the technique is the same.

6133761338

[christo]

You just started making the X-overs....??, how did you know what to buy (quality)?,


Well let’s just say “I just started assembling X-overs” I’ll leave the making to the experts.

I did my research; Google Advanced Search is your friend when looking for information in this web site. Search using “Giskard's equivalent network”.

There is tons of information on X-overs within this web with the associated part list.

I used the 2405H and 2425J circuits from “Giskard's equivalent network” and the JBL 4344 MKII for the 2123. I have redrawn the schematic from the scanned imaged in the web site. To any experts out there if I have miss-drawn the four 36.0 µF Caps that represent the two 72 µF Caps on the original JBL schematic let me know and I’ll fix. I’m no expert in this area.
61339

This is the parts list I used and were sourced from www.solen.ca (http://www.solen.ca).

61340

[Amnes]

Inspiring thread - thank you for posting.

[Chas]

Congratulations on a fine project.

Beautiful work and excellent documentation!

I know what you mean about inserting those TD-2002s... Bo and I heard the same thing when we popped a pair into his 4345s. It was a surprising difference!


Widget
;) Even better with cc or upgraded caps, too.

[christo]

Well let’s just say “I just started assembling X-overs” I’ll leave the making to the experts.

I did my research; Google Advanced Search is your friend when looking for information in this web site. Search using “Giskard's equivalent network”.

There is tons of information on X-overs within this web with the associated part list.

I used the 2405H and 2425J circuits from “Giskard's equivalent network” and the JBL 4344 MKII for the 2123. I have redrawn the schematic from the scanned imaged in the web site. To any experts out there if I have miss-drawn the four 36.0 µF Caps that represent the two 72 µF Caps on the original JBL schematic let me know and I’ll fix. I’m no expert in this area.
61339

This is the parts list I used and were sourced from www.solen.ca (http://www.solen.ca).

61340

The network diagram has a typo with the capacitors in the 2405 circuit. The original schematic had a pair of 3.3 uF this is incorrect. The pairing should be 3.3 uF and 3.9 uF. The parts list in the original post is correct.

Below is the corrected schematic.
73823

[Dr.db]

I used Silicone to glue the ports (ABS or sewage pipe) in to the cabinet

Once I was told to never use silicone in a speaker enclosure due to it´s agressive degassing.
It could dissolve the glue of driver-cones and suspensions!

I have never tried, so I cannot confirm this quote but I just wanted to let you know...

[Woody Banks]

Beautiful work and documentation. Thanks for sharing.:thmbsup:

[fencki]

hi.
can i use the 2420 instead of the 2426 (TAD) for the HF?
or do i have to adapt the R´s, C´s and L´s?

thx

[christo]

hi.
can i use the 2420 instead of the 2426 (TAD) for the HF?
or do i have to adapt the R´s, C´s and L´s?

thx

The 2420 = LE85 which will work fine with the crossover, I have tried this.

[christo]

Beautiful work and documentation. Thanks for sharing.:thmbsup:

Thanks Woody I have always admired your work !

[christo]

Once I was told to never use silicone in a speaker enclosure due to it´s agressive degassing.
It could dissolve the glue of driver-cones and suspensions!

I have never tried, so I cannot confirm this quote but I just wanted to let you know...

I have never heard that, the boxes sat around for a week before I put the 2235s back in so hopefully the initial degassing would have occurred. It been two year's now with no issues.

PS Like your work on your tribute project!

[fencki]

The 2420 = LE85 which will work fine with the crossover, I have tried this.

thx for your answer.
you helped a lot!

i have one more question for my JBL 4345 clones.
beacause i have no 2122h i got one 2123h.
can i use this crossover from 4344 MK2 with chassis 2123h and build up the 3145 crossover?:

http://www.audioheritage.org/vbulletin/showthread.php?35325-JBL-4344-Clones&p=397270&viewfull=1#post397270
(#53)


please help with that too...
so i can start to order the parts for the crossover.

big thx!!!



btw you did a REALLY GREAT job with ur JBL´s. just phantastic!

[christo]

thx for your answer.
you helped a lot!

i have one more question for my JBL 4345 clones.
beacause i have no 2122h i got one 2123h.
can i use this crossover from 4344 MK2 with chassis 2123h and build up the 3145 crossover?:

http://www.audioheritage.org/vbulletin/showthread.php?35325-JBL-4344-Clones&p=397270&viewfull=1#post397270
(#53)


please help with that too...
so i can start to order the parts for the crossover.

big thx!!!



btw you did a REALLY GREAT job with ur JBL´s. just phantastic!


My understanding is that the 4344 MKII crossover will get you the smoothest sound out of the 2123 for this type of application. The issue would be the crossover point of the 2245 would have to be raised to 340Hz from 290Hz as in the stock 4345 with a 2122.

Most people in the forum don’t like having the 2245 play this high.

There is a lot of information on the differences between the JBL vintage 10” speakers here,

http://www.audioheritage.org/vbulletin/showthread.php?16962-Field-Guide-to-the-JBL-10-inch-Midbass-driver-2121-2122-2123

as they are not considered to be that interchangeable.

There much discussion on the 4344 MKII crossover you’ll have to search the forum on your own as it is spread across multiple threads.

Chris

[ivica]

thx for your answer.
you helped a lot!

i have one more question for my JBL 4345 clones.
beacause i have no 2122h i got one 2123h.
can i use this crossover from 4344 MK2 with chassis 2123h and build up the 3145 crossover?:

http://www.audioheritage.org/vbulletin/showthread.php?35325-JBL-4344-Clones&p=397270&viewfull=1#post397270
(#53)


please help with that too...
so i can start to order the parts for the crossover.

big thx!!!



btw you did a REALLY GREAT job with ur JBL´s. just phantastic!

Hi fencki,
As I can understand there is an error in the drawing presented on the #53
C1, C3, C5 and C7 form equivalent C=36/2=18uF, but from Giskard L=1.8mH, and C=14uF......
as hi-pass network -6dB relative to the LF part is round 1.3kHz

regards
ivica

[christo]

Hi fencki,
As I can understand there is an error in the drawing presented on the #53
C1, C3, C5 and C7 form equivalent C=36/2=18uF, but from Giskard L=1.8mH, and C=14uF......
as hi-pass network -6dB relative to the LF part is round 1.3kHz

regards
ivica

Ivica

I curious as to why you consider the schematic wrong as the section you are referring to comes from the JBL stock 4344 MkII schematic which has the C5, C7 as 72uF which in a CC network would be 2 x 36uF. Or am I missing something?

I was not aware that Giskard had done a schematic for a 2245/2123 crossover if he did could you point me to it I would be very interested in looking at it.

Chris

[Ian Mackenzie]

http://www.audioheritage.org/vbulletin/showthread.php?24283-4344-Mk-II-network


The capacitor values in the mid filter are for charge coupling

The values are correct when simulating the voltage drive in LEAP.

The measured acoustic response in LMS is correct for the 4344mk11 schematic

The use of the 2245 requires an active 3rd order filter

Some users have reported using an active 4th order filter

[fencki]

http://www.audioheritage.org/vbulletin/showthread.php?24283-4344-Mk-II-network


The capacitor values in the mid filter are for charge coupling

The values are correct when simulating the voltage drive in LEAP.

The measured acoustic response in LMS is correct for the 4344mk11 schematic

The use of the 2245 requires an active 3rd order filter

Some users have reported using an active 4th order filter


hi.
can you please explain that, for somebody not that good in crossovers?
are the corrected values right, what ivica suggested? or the original ones former posted?
how should the crossovers for the 2245h and 2123h look like if using the 2123h and the crossover for 4344 mk2?
i mean the crossover frequencies are different (290HZ in 4345 and 340HZ in 4344)

using a 3rd or 4th order filter?
can u show a schematic for the usage of 2245h and 2123h?

i don´t have such simulating programs so i have to ask the pros here...
:)

thanks a lot and greetings from vienna
slavko

[Ian Mackenzie]

http://www.audioheritage.org/vbulletin/showthread.php?37868-JBL-4346-DYI-with-2123H&p=390896&viewfull=1#post390896

This post read please.f
Ft
Original values are correct

TWhen you go off the reservation and mix different schematic for mid and HF we cannot guarantee the outcome but the filter values for the 2123 are correct in the 4344mk11 schematic.

The 2122 and 2123 are slightly different drivers.

As l recall the 2123 is slightly f sensitive and tthe low end response is different.

The 2122 was specifically designed for the 4345 monitor.

Unfortunately l am not in a position to provide details of the 2245 filter right now.

NFilters follow a mathematical transferfunction denoted by filter poles usually 1-4 poles in analogue filters.

1 pole is a 6 db slope...usually a very gentle slope but no good for the 2245.

3 pole is a 18 db filter and is better for the 2245.

4 pole is 24 db and is steeper slope and can be used for the 2245.
Do a search of the 4345 monitor.

The concept of bi amping is simple enough.

DO NOT use a passive crossover for the 2245.

I suggest acquiring a adjustable active filter like the Ashley and adjust the level until it seems subjectivity correct.

Anywhere from 250-340 Hz is going to work.

[fencki]

http://www.audioheritage.org/vbulletin/showthread.php?37868-JBL-4346-DYI-with-2123H&p=390896&viewfull=1#post390896

This post read please.f
Ft
Original values are correct

TWhen you go off the reservation and mix different schematic for mid and HF we cannot guarantee the outcome but the filter values for the 2123 are correct in the 4344mk11 schematic.

The 2122 and 2123 are slightly different drivers.

As l recall the 2123 is slightly f sensitive and tthe low end response is different.

The 2122 was specifically designed for the 4345 monitor.

Unfortunately l am not in a position to provide details of the 2245 filter right now.

NFilters follow a mathematical transferfunction denoted by filter poles usually 1-4 poles in analogue filters.

1 pole is a 6 db slope...usually a very gentle slope but no good for the 2245.

3 pole is a 18 db filter and is better for the 2245.

4 pole is 24 db and is steeper slope and can be used for the 2245.
Do a search of the 4345 monitor.

The concept of bi amping is simple enough.

DO NOT use a passive crossover for the 2245.

I suggest acquiring a adjustable active filter like the Ashley and adjust the level until it seems subjectivity correct.

Anywhere from 250-340 Hz is going to work.


ok. thanks for your reply.
i will do more research and think about all the stuff. my head is full of infos. mixing all up...
thx

actually i am thinking about an mini-dsp 4x10 HD... or 2x4 HD...
maybe fully active. but 8 amlifiers will be a real overkill!
:)

br
slavko

[Ian Mackenzie]

That's it

Do some reading to get a general idea

The lovely people over in the USA are a little pre occupied right now so apologies if they are not around

[fencki]

That's it

Do some reading to get a general idea

The lovely people over in the USA are a little pre occupied right now so apologies if they are not around

i really can imagine! i was really (let´s say) supriesed about the election and the final result...

[fencki]

did you ever try out the mini-dsp?

[Ian Mackenzie]

No only analog at this point

You would need to emulate the voltage drives in the mini dsp which are optimised for the specific drivers and baffle placement in the Xy axis domain.

The key advantage of dsp is phase alignment of the mid and horn crossover point but this is only for one point in space.

I there recommend you bi amp the woofer and mid start with and use the dap to EQ the woofer response below 150 hertz where room placement impacts on the subjective outcome.

[Chas]

Based upon my experience using the 2123 with the 4344 MKII crossover it does, in fact, work quite well when used within a hybrid 4344/4345 set up. I've done this myself and I have heard Christo's set up. No, it might not meet the ultimate technical analysis parameters, but to my ears, based upon my own 4345 experience and having A-B'd my own 2123 franken version....it simply works quite well.


No, the 2123 doesn't sound like a 2122, but both are really great drivers in this context and I am hard pressed to pick a favourite.

[fencki]

hi
thanks for your input here...

br
slavko

[Ian Mackenzie]

For those interested l have recently implemented the 4344mk11 full passive crossover using the 2245.

I used
2245H
2123H
2420 with aquaplas dusted diaphragm
2405

There has been much discussion on the best approach to using the 2123H with the equivalent Giskard filters.

I propose to measure the hyprid MK11 (2123H filter) with the Giskard HF and UHF filter and compare to the complete 4344mk11 filters.

I have not compared the filters subjectively.

But l must say the complete 4344mk11 crossover using the dusted diaphragm is spectacular.

Gone is the live razers edge voicing of the 4343.

The transition of the 2123H to the horn is seamless and natural while the 2405 compliments the top end with crystal like clarity.

I plan to post further analysis of the above soon including biamping

[Ian Mackenzie]

The pic below is the amplitude response comparing the Giskard hyprid (mk11 mid filter for the 2123H) and the Giskard horn and slot filters

The violet curve is the Giskard hyprid filter, green is the MK11 filter .

The Lpads were at maximum position and the test mic was located in the near field on axis with the bottom of the horn lens.

Given this measurement was indoors and ungated the response within 5 db from 40 Hz to 15khz is remarkable and a testament to the subjective performance .

Despite attempts to adjust the horn and slot Lpad l was unable to achieve an acceptable result compared to the MK11 filter per attachment.

If anyone is considering the 4344 as a diy project, given the 2122H mid is NLA the 2123H is a stellar midrange driver with the MK11 network. I have owned the 2212H using the Giskard equivalent network and would describe the MK11 based 2123H as a comparable system.

The MK11 network was constructed with Hovland capacitors only because they are readily available.

The charge-coupled Solen fast cap is a well known option. I am completely agnostic about the subjective attributes of capacitors and let the quality of implementation do the talking.

Subjectively l would describe the hyprid as hideously coarse sounding with a veiled presentation.

The response is an indicator of overall smoothness but manner in which carefully chosen slopes compliment the individual characteristics of the drivers is outstanding.

The 2123H in the MK11 crossover with dusted diaphragms and 2405 is the smoothest and best voiced 4 way JBL l have heard to date and is up there with the 4338 and other JBL Blue baffle monitors.

Cutting and pasting schematics is a risky as the above test indicates.

(Do not confuse with the other Giskard filter that works for the 2123/2202 drivers )

You may be wondering if the MK11 works for the 2235 and the 2245 drivers.

I have not tried the 2235H yet but the 2245 seems to work very well in full passive mode.

I will post details of the 2245 passive and active networks soon.

For those who are interested l can provide fabrication of the MK11 network, testing with LMS/LEAP 5.

[fencki]

hi.
thanks a lot for those measurements and test.
very interesting for me, because i am still not sure about my decision which crossover to use.

because my englisch is not the best i have one or two questions.

does that mean, that you tried the 4344 mk2 schematic with the mentioned drivers (with no change of any values or routing)? (= green line)
does that mean, that you tried the giskard schematic (with taking the 2123h section from MK2 for MF) and all other values from giskard? (violett line)

if i understood that right, then the original 4344mk2 schematic resullt (=green) looks better for me (maybe not the "hole" at 10kHz.) :eek:


thx a lot for your expertiese...

best regards from vienna
slavko

[fencki]

and YES, i am very interested in reading your future post regarding to this topic.

[Ian Mackenzie]

They are independent measurement of actual component values where people have used the mid filter of the MK11 crossover and then used the Giskard filter for the horn and slot. I refer to this as hyprid filter (violet)

The MK11 filter uses if all values in the MK11 schematic for the mid, horn and slot (green)

The reason for people using a hyprid filter alludes me.

But someone essentially complained about the outcome after being given advice to use the hyprid filter so l investigated the reason.

Of course it is entirely up to the individual what they decide do with their own project.

I have posted this information as a heads up.

If you want or need individual "advice" pm me.

I'm am usually available with donations of fine red wine for my amazing wife!

[fencki]

i like your explainations.
they help me a lot and i am thinking more about the aspects of all sides of the topic.
please keep up with that. it is very interesting for me.
i am a newbie for crossovers and their effects and thankful for everything i can learn.

and this graph was very helpful for my project.

thx

[fencki]

They are independent measurement of actual component values where people have used the mid filter of the MK11 crossover and then used the Giskard filter for the horn and slot. I refer to this as hyprid filter (violet)

The MK11 filter uses if all values in the MK11 schematic for the mid, horn and slot (green)

The reason for people using a hyprid filter alludes me.

But someone essentially complained about the outcome after being given advice to use the hyprid filter so l investigated the reason.

Of course it is entirely up to the individual what they decide do with their own project.

I have posted this information as a heads up.

If you want or need individual "advice" pm me.

I'm am usually available with donations of fine red wine for my amazing wife!


in austria we do have VERY good wines!
;-))

[Ian Mackenzie]

I will do a near field and ground plane measurement of the 2123H impedance and acoustic amplitude response on the baffle and of the 2245H with LMS for analysis in LEAP5.

Other than the original JBL technical data sheets l don't think this had been published recently so it will be quite interesting.

[Ian Mackenzie]

Attached is a voltage drive charting the Giskard and MK11 filters

Refer to the legend for the horn and slot voltage drives.

I have also attached the individual response curves of the native 2122H and 2123H drivers for comparison

Note this NOT the acoustic response of the individual drivers with the filters.

As can be seen the 2123H is somewhat more extended in the midrange and more sensitive than the 2122H.

The MK11 mid filter is more complex as it must attenuate the additional sensitivity while equalising the native driver to form a flat response and provide the crossover slope.

In this respect the 2123H and the horn have a somewhat more gentle overlap before attenuation slope.

My subjective view of this is closer voicing or blending of both the 2123H and the horn which was always obvious in the prior 4343-4344.

Likewise the horn is allowed to extend with a gentle roll off slope in a systematic and controlled manner into the upper octaves. The acquaplas dusted diaphragm provides more depth meaning blacker backgrounds improved resolution.

The 2405 lights up in the upper octaves and provides remarkable speed, air and spatial clarity.

[fencki]

very very good job! very interesting post!

[Ian Mackenzie]

I have done some initial work.

At this point l would say the voltage drives for biamping like the earlier 4343 are not simply 24 db LR 340hz or 12 or 18 db slopes.

The woofer is not straight forward either.

I need to do more measurements before proceeding .

[fencki]

awaiting your research. very interested in such things...

[Ian Mackenzie]

Okay

If the weather is okay today l will do some external measurement using the 2245H.

The question or what might be term the issue is how will the 2235H and 2245H work with the MK11 passive filter and what is the active filter requirement?

The reason for external measurement if you are not familiar is that a measurement in a room and particularly a smaller rooms are problematic to be of any use at 500 Hz and below. Reflecting surfaces and room mode come into play.

In the case of the 2245H as mentioned earlier it is best used with an active 18 db 3rd order or a 24 db LR filter.

A 12 db filter would not be suitable in this application.

The 2235H can be used with a 12 db filter slope but l would tend to standardise on a defacto active filter for both drivers in this application.

Arriving at what could be described as good best practice for using either the 2245H or the 2235H in passive mode is a challenge.

The convenience of less equipment a consideration.

Will either of these drivers work in the stock MK11 filter requires some further investigation.

We firstly need to understand the function of the MK11 passive filter.

Looking at the MK11 passive filter schematic in detail the slope of the 2123H with L3 and R8 with the switch in bi amp position is approximately 6 db per octave.

The inclusion of the capacitors C1, C3 R1 and R5 is approximately a 12 db per octave acoustic high pass acoustic slope where these components for the additional pole.

So in theory at least to arrive at an 18 db per octave 3rd order high pass filter for the 2123H the active filter would be 12 db 2nd order 2 Pole filter.

However the asymmetric nature of such an active filter 18/12 db per octave may not be convenient in practice.

JBL may have considered a custom card for their DX1 active crossover when they were available. This unit is NLA.

The other option is a 24 db LR active crossover with a degree of adjustment like the Ashly.

This crossover has an an adjustment at the crossover point to arrive at a flat summed response. The actual crossover point is a matter of adjustment but in the region of 300 hertz.

http://ashly.com/wp-content/uploads/2015/08/xr-1001-2001-4001-crossovers-r01.pdf

You can score this crossover on EBay at quite a reasonable price.

The question of whether the Ashly is deemed a hifi qualify active crossover is questionable but it's a good starting point before you decide to spend US$1250 on a First Watt B4 crossover which l use. The B4 model has a range of custom filter adjustment that make it entirely suitable.

The importance of utter transparency in the active crossover cannot be overstated if your quest is to maximise the performance of the MK11 system.

I have used a LR 24 db filter with the 4345 in the past.

The traditional well executed LR24 filter is industry standard but may require careful adjustment as the driver overlap is narrow.

Note too that an asymmetric filter function will apply (24/36 db per octave) with the MK11 if a non adjustable 24db LR filter is used. The impact of this is unknown at this stage.

I will investigate measurement of the passive and active mode as time permits.

The attached curve below is near field for illustrative purposes of the discussion using the 2345H.

The smooth black curve is the high pass 12 db 2nd order slope "objective".

The 2123H tracks the objective reasonably well.

Because the near field measurement was on the 2123H axis the result of the 2245H is not accurate.

[bldozier]

😶

[BMWCCA]

The question of whether the Ashly is deemed a hifi qualify active crossover is questionable but it's a good starting point before you decide to spend US$1250 on a First Watt B4 crossover which l use. The B4 model has a range of custom filter adjustment that make it entirely suitable.Not that I would know anything about a $1250 crossover, but I did consider the Marchand prior to purchasing my first (used) Ashly XR1001. It's worked great for me for over eight years. I had a bit of a problem with dropouts from one section at some point which I traced to a pot on the Ashly that had a dead-spot. I contacted Ashly and they quoted me a very reasonable fee to repair it but also offered to send me a new potentiometer free-of-charge, which I took them up on. I'm still using it without repair just bumping the setting slightly and I've since purchased another excellent condition used Ashly to put in the system while I effect the repair on the original. Then I'll either have a replacement, should I need it, or use the spare to try bi-amped L7s!

[Ian Mackenzie]

Thanks for chiming in.

Like any piece of audio equipment the final choice depends on your own situation and the considerations associated with that such as your technical understanding of how to set these crossover up. Some manufacturers offer assistance.

But otherwise the user adopts a Monet approach to passive/active crossover execution with insights obtained from various forum posts. The key is in collecting quality data.

As l said the Ashly works, it does get you going and they are accessible .

I started with the JBL 5232 off eBay.

http://www.jblproservice.com/pdf/Vintage%20JBL-UREI%20Electronics/JBL-5231_5232.pdf

Other forum members have moved on and purchased the Bryson active crossover, DEQX , .Passlabs and reported audible improvements.

There is nothing exclusive about diy or vintage audio.

Do what you.

[fencki]

i have a tannoy xover and will get an JBL 5235 with teh right cards for 4345...
will have to do a lot of test.
:D

but first the JBL has to be finished.

[Ian Mackenzie]

When you are ready let me know

[Chas]

Veiled? I didn't hear that at all and in fact, quite the opposite. Ian, the "hybrid" came from a mod to a change of the Giskard version and was based upon the lack of 2122 availability. No mystery there....

The combination, properly implemented, works really well up here in this hemisphere! But then, our drainage water flow runs clockwise!:D

Now, that certainly doesn't mean other ideas and methodologies don't do well and possibly exceed the intention on paper, at least.:)

Furthermore, given that there are LF room interactions influencing at the bi amp crossover frequencies, the bi-amp crossover frequency, give or take 50 Hz and accounting for slope, in the real world, are really moot, IMHO. Fourth order LZ at~290 Hz works fine here!

[Ian Mackenzie]

It comes down to specifics and saying something in a post is a loose conversation without controlled conditions where you are in fact comparing one filter set up to another and measurement to look at it objectively.

What l am saying is you may not have done exactly what we are on about here

It would be useful if you posted your actual schematic and some valid measurements

If you want to be a fly in the ointment no problem.

It's interesting that designs despite there apparent simplicity of using two drivers as for example the M2 have taken on a level of precision not seen before as a diy thread

Why?

Because they want it to be authentic and the want get it right.

"Furthermore, given that there are LF room interactions influencing at the bi amp crossover frequencies, the bi-amp crossover frequency, give or take 50 Hz and accounting for slope, in the real world, are really moot, IMHO. Fourth order LZ at~290 Hz works fine here!"

Well l am not sure jbl would agree with you on that statement

Jbl realised that a custom bi filter was required for the 4343 and then 18 Db for the 4344//4345.

Please refer to my comments concerning the 4344mk11 filters.

[Chas]

Ian, I certainly appreciate the work you are doing and always have done. But there is always a slant to DIY that should not be discounted. GT himself commented about the individual subjective appraisal of speakers.

There are no perfect solutions..:)

[Ian Mackenzie]

Based upon my experience using the 2123 with the 4344 MKII crossover it does, in fact, work quite well when used within a hybrid 4344/4345 set up. I've done this myself and I have heard Christo's set up. No, it might not meet the ultimate technical analysis parameters, but to my ears, based upon my own 4345 experience and having A-B'd my own 2123 franken version....it simply worksheets quite well.


No, the 2123 doesn't sound like a 2122, but both are really great drivers in this context and I am hard pressed to pick a favourite.

Chas

If l read your post correctly are you saying you used the 434k11 schematic in its entirety for biamp use with the 2245?

If this is the case that is a different scenario that in post 60 and 62 whe there seems to be the belief that you can splice the filter of an earlier model in the 43xx series with another.

What Giskard did was create a different crossover that could be used with the 2123 or the 2202 so far as the 4344/4345 is concerned.

However, others have just spliced the mid filter of the 434k11 filter into the equivalent 3145 crossover

The acoustic response with the filters of the mid drivers and HF are different in both the 4344mk11 and the equivalent 4344 are not the same by any means.

The slopes and the levels are different and that why l have raised this point and posted graphs.


Then there is the transition of the woofer and 2123.

The 4344mk11 used an ME150H which was more sensitive than the 2235H but similar sensitivity to the 2345H

The 4344mk11 schematic is arranged do allow either bi wiring and bi amping.

In biamp mode the low end of the 2123H is limited by the action of the band Pass filter form by the remaining elements and the driver raw response.

As l mentioned earlier the 4344mk11 is more advanced and incorporated a permanent trap LR across the driver.

When the biamp switch is position to bi amp the effect is the driver low end is rolling off at 6db per octave in the Pass Band.

In passive crossover position the 36uf and series 1.2R resister forms an additional 6db of slope creating the 2nd order filter slope.

Now, to ensure we don't screw up the wonderfull job of the crossover designer it would be useful to evaluate what bi amp filter will work best for the pure 4344mk11 in the case of the 2235H and the 2245H

If a conventional symmetrical active filter is applied the actual acoustic response of the 2123H may not sum flat and may have phase lobes in the vertical plane that impact on the voicing of the system.

Okay you might say what does it matter what l do?

We have no control of that but it's useful to publish valid information about what should work as opposed to what l think l did!

Multi way systems like this are complex and sometimes you only have to change or have one thing out of whack and the whole shabang turns to an annoyance.

This is not the first time bi amp filters have been discussed

Previously when some decided to investigate the Bryson active crossover the question of filter crossover points and slope came up.

[Ian Mackenzie]

Curves
1.Measurement of Passive system crossover curves (near field) (top)

2. 2245 5 Db/ division 1 octave smoothing
Raw driver
With passive low passfilter (3145)
The peak in the bass can be ignored as is an environmental issue.

3. System response 10 Db/division 1 octave smoothing
The gentle rise in the lower end is a function of the ground plane measurement technique.

4. 2123 5 Db/division 1 octave smoothing
Raw driver
With mid filter bi amp mode

5. Simulation of the passive 2245/2123 crossover using the 3145 low pass filter values and the 4344MK11 mid filter .As can be seen this filter sums flat and has the correct phase relationship at the crossover point. The simulation is derived from real data using measurement of the drivers. The data is conditioned and then applied to different scenarios.

This is more efficient than building trial networks.

Ground plane measurement- enclosure inverted - mic distance 1 metre

LMS - precision calibrated mic


I have done some preliminary bi amp simulations

The 3rd order 290 Hz active crossover does not sum flat with an +3 Db hump unlike the passive network. No surprises there.

It is important to understand that these humps are phase errors and cannot be reduced or removed with equalisation without causing more problems than it solves.

The better way is to design the crossover point to sum flat at a reference point on the baffle(at the bottom of the lens) from a given distance to the test mic.

TBC

[Ian Mackenzie]

I have simulated both the 3rd order 290 Hz and the 290 Hz LR bi

In practice the 3rd order filter may prove the best overall option

Why?

It provides sufficient attenuation and may provide better phase coherence in the crossover region.

This may defy the text book attributes of the LR filter but because of the offset of woofer and mid driver voice coil the typical forward Lobe points to the floor and not the listening space.

What this means is the power response in the vertical domain an octave either side of the crossover point is not ideal.

In this respect the LR filter is imperfect and there is no clear in phase/ out of phase relationship when the driver polarity of the woofer is reversed.

The 3rd order filter exhibits a more clearly define in phase/ out of phase relationship when the woofer polarity is reversed.

I plan to do a ground plane measurement later in the week to confirm this.

The diagram below illustrates the ideal LR polar response in the crossover region.(right)

Imagine the lobe skewed downward and the direct sound is the null.

Of course much of the sound is indirect but we ideally don't want the main lobe pointing away from the listening position
On the left