250Ti - 4-way active experiment

[Dagfinn]

So, while my JBL 250 Ti passive crossovers are in for repair, I thought I might try an active setup - just as an experiment. I have a dbx 234Xs crossover and 4x200w amplification running fully balanced, with 2+2ch gain control, as a test setup. 31-band analog equalizer. Thinking I'll do one speaker first, and keep the second passive, as a model. Have a 12uF cap for the tweeter. I know DSP come to mind for many, but I want to keep it an the analog domain, hoping the dbx will do for the experiment. I have a UMIK-1 and REW, but sonic pleasure is the main goal . Anything else I should think about? Dagfinn

[Robh3606]

Hello

This should help you. The slopes are "wrong" but with the 31 band EQ you should be able to get something usable. Do you have an RTA to use to set the EQ?

This explains the Jubilee crossover but the native driver responses should be similar.

Rob

https://www.audioheritage.org/vbulle...work%20diagram

[Dagfinn]

Thanks, Rob, I'll see if I can decode something. That's quite different from the TI BQ network, btw. Did drivers stay the same for the Jubilee? To be honest, the theoretical side is weak here, so... I have the DAW software wich came with my Motu M4 - it has realtime FFT abilities which I hoped could work. Never used it with the Umik either. I'm guessing there may be better, dedicated softwares out there, is there something you'd recommend? Thanks again. Dagfinn BTW, look at what Greg Timbers has at home .

[Dagfinn]

I haven't used REW before, and I see it has what looks like a nice RTA and signal generator. Never mind the noise here (it's sitting on thwe workstation), but some single tones to each driver on my Genelecs, just to see it works. 400Hz is about on the c.o. border. To late anything more, but looks good .

[Robh3606]

If you have an RTA and a Pink Noise source you have a good start. The only difference in the drivers is the Jubilee uses 050 Gold vs 035Ti. The rest are the same.

Rob

[Dagfinn]

Great, I see the RTA will be very useful. I realize the dbx is limited, but the cost of next level is pretty steep. It seems building a custom active c.o. based on premade modules is possible, but above my skill level. I don't really know a readymade hifi product that fits the bill? Anyway, experimenting for now . Dagfinn.

[Odd]

Hei Dagfinn

If you want to learn more about REW, you will find a lot of useful information in Norwegian here
https://www.hifisentralen.no/forumet...5linger.79532/

and here
https://www.hifisentralen.no/forumet...-hjelp.101681/

[Robh3606]

Great, I see the RTA will be very useful. I realize the dbx is limited, but the cost of next level is pretty steep. It seems building a custom active c.o. based on premade modules is possible, but above my skill level. I don't really know a readymade hifi product that fits the bill? Anyway, experimenting for now . Dagfinn.

When you start experimenting remember JBL uses a flipped polarity on the woofers and if you look at the schematic you can see the polarity for proper hook up posted there for the standard network.

With a L/R 24db you should have them all the same polarity. You can check the individual drivers for polarity with a 1.5 volt battery less of course the 035Ti. And of course you can see a notch in the response at crossover using the RTA if the polarity is "wrong".

I use one of the DBX in my active 4 way and I like it! Not as flexible as a DSP but adequate to stay analog. I also use analog EQ Urei 539 1/3 octave to help "smooth" things out.

Using the figures in the referenced post looking at the Yellow curve on the tweeter you get an idea where you are going to need to use cut filters to flatten out the driver responses.

[Dagfinn]

Hi Odd, thanks, I will study them, for sure. Got my wires and connectors today, if they all fit I can make parts and perhaps connect up tomorrow .

[Dagfinn]

When you start experimenting remember JBL uses a flipped polarity on the woofers and if you look at the schematic you can see the polarity for proper hook up posted there for the standard network.

With a L/R 24db you should have them all the same polarity. You can check the individual drivers for polarity with a 1.5 volt battery less of course the 035Ti. And of course you can see a notch in the response at crossover using the RTA if the polarity is "wrong".

I use one of the DBX in my active 4 way and I like it! Not as flexible as a DSP but adequate to stay analog. I also use analog EQ Urei 539 1/3 octave to help "smooth" things out.

Using the figures in the referenced post looking at the Yellow curve on the tweeter you get an idea where you are going to need to use cut filters to flatten out the driver responses.

Thanks, I'll be careful. I think it's all well marked, they used Monstercable internally, clearly colorized and noted on schematic. Tidy. Looking at the raw driver data and what I've experienced with the 240Ti, I can adjust c.o. frequencies a bit without problems. Make them hit the notches on the equalizer, if needed . But REW looks great, no problems there. I have to ponder the power requirements, tweeter vs woofer asf. Class D vs class AB. I see it's popular to maximize power, but high mid and tweeter don't benefit much from 200w vs for instance 100w? Thinking about it, AW75 poweramp from Electrocompaniet should be able to drive the tweeters..!

[Dagfinn]

This is really useful. Very interesting to see how much the woofer is allowed to play, "star of the show" in the 240Ti.
Btw, here's the interview Erin's Audio corner did with Greg Timbers in 2022 https://youtu.be/afeTHLLp-EI?si=KUAV4dE7uNutrncw
Poor Erin, didn't know the L250... great interview nevertheless .

[Dagfinn]

So, things didn't quite go as planned, but no problem. The two mids are mounted in small boxes inside the baffle, and cables are glued. I'll cut the c.o. cables as it's the cleanest way.

I've looked at the c.o. from the Jubilee, and read a bit about the Linkwitz-Riley 24dB c.o. The dbx has Linkwitz-Riley24dB, which is ok for tweeter/hi-mid/mid, but not ideal for the woofer to low-mid. The Jubilee has a 6dB slope for the woofer, supporting the low.mid and certainly affecting sonic properties of the speaker. Trying to maintain this slope seems important. One solution is to keep the woofer-part of the passive c.o. in for now, and feed it a fullrange signal. This should work fine, it's already prepared for bi-amping. At least I hope so .

More challenging is the low/high mid c.o., but 24dB is what I have. Hard to find data on the 108H. I'll assume it is similar to the 2118H? Not much else to do. It's obvious that a more advanced c.o. would be needed to follow the slopes from the schematic. But if it sounds good, I'll have to ponder something for the woofer, perhaps just copy the passive c.o. and keep it in the circuit.

I found a JBL spec sheet for the 104H by G.T, perhaps webmaster could put it in the archives?

[grumpy]

or... https://groupdiy.com/threads/low-pas...e-level.34480/

[Dagfinn]

Thanks grumpy. Doing it line level would naturally be better and even looks within my capabilities to make... Didn't know it could be that simple.

[Dagfinn]

Got them up and running today, all drivers going via dbx crossover for now. I think this may work out really well, set up by ear is already very pleasing. The dynamic power is undeniably improved. I think 650w/channel will be sufficient. The current 200w for tweeters is a bit excessive, so I'll swap down to 125w for tweeters and 104H, while the 108H and woofer gets around 200w each. I changed my mind about the EC since it doesn't have a gain control, which I like to have.

Next I'll measure up the passive c.o., to have a proper record of it before full conversion. I may do the line-level lowpass too, I found I l aready have the needed parts for a 348Hz corner frequency. I forget, I'll need another dbx too... not too expensive, fortunately. Maybe a used shows up, no rush. So, now I look forward to fine tuning and listening to these amazing speakers. I already hear new stuff in well known music, I like that . It as if there is more space to let more instruments out....

(not sure why, emojis don't seem to work ???)