4355/3155 clones

[4313B]

Originally posted by boputnam
OK, a challenge for you, if you're in a goofy enough mood for it. You're the only one I know could get it done...

Can you show the two cases of, say, either the 4343 or 4345, in biamp mode with the LF and "others" both in phase and in opposite phase (polarity)?

I'm dang curious how much of the troublesome summing is obviated by having the LF and low-MF out of phase.

Ok, here is what I get for the 4343 / 4350 crossover card in the 5234A. The dip is with the 2235H and 2122H in phase and the hump is with the polarity on the 2122H reversed.

[4313B]

Originally posted by boputnam
OK, a challenge for you, if you're in a goofy enough mood for it. You're the only one I know could get it done...

Can you show the two cases of, say, either the 4343 or 4345, in biamp mode with the LF and "others" both in phase and in opposite phase (polarity)?

I'm dang curious how much of the troublesome summing is obviated by having the LF and low-MF out of phase.

Here is what I get for the 4344 / 4345 / 4355 crossover card in the 5234A. The hump is with the 2235H and 2122H in phase and the dip is with the polarity on the 2122H reversed.

[boputnam]

Originally posted by Giskard
Ok, here is what I get for the 4343 / 4350 crossover card in the 5234A. The dip is with the 2235H and 2122H in phase and the hump is with the polarity on the 2122H reversed.

You're not goofing on me, are you?

I would have expected the hump to be where the two were in-phase and additive... Oh NO!!!

Thanks, heaps, for the model. You do truly great work.

[4313B]

I'm too tired to check my work yet again tonight. Maybe I'll look at it again in the morning. It will take about 1/10th the time for you to measure the whole mess with an RTA as opposed to me trying to model an active filter in a software package that only supports passive filters. However, I was pretty tenacious and feel that I got damn close with the model. The graphs I posted earlier were very good at showing the phase of the two filters. Feel free to repost them if you want.

[boputnam]

Originally posted by Giskard
It will take about 1/10th the time for you to measure the whole mess with an RTA as opposed to me trying to model an active filter in a software package that only supports passive filters.

I knew - or had a feeling - it was that difficult. Thanks for the effort.

And, I'd gladly do the RTA thing, but Widget's gone "AWO4355" for a week-ish, so it will wait.

I think the models are excellent - I only question the text description you posted for the 4343 / 4350 - it is opposite that for the 4344 / 4345 / 4355 and of my intuition (which is often wrong, but never in doubt, as dear ol' mom used to remind... ).

However, thinking on the phasing plots you posted earlier today, it IS making sense - at least it hangs together. I take it back - your annotations are consistent with the phasing plots. But the reason for the difference between the 4343 / 4350 and 4344 / 4345 / 4355 boggles my feeble mind.

Separately, I agree an RTA test would be most revealing, when Widget beaches here, again. Will do.

"See ya" tomorrow, and thanks again.

[4313B]

I've posted the voltage drives of both filters elsewhere on the forum. The 4344 / 4345 / 4355 is an 18 dB / octave active filter with both knees set at ~ 290 Hz. The 4343 / 4350 is a 12 dB / octave filter with the low pass knee set at ~ 230 Hz and the high pass knee set at ~ 280 Hz with a mutual -4 dB downpoint at ~ 255 Hz. The phase plots I posted for you earlier should be helpful now that you see what else is going on.

*****

"But the reason for the difference between the 4343 / 4350 and 4344 / 4345 / 4355 boggles my feeble mind."

Don't let it, the 4344 / 4345 /4355 is a better solution in my opinion. I've posted this at least twice before. I would run it with a 4343 or 4350 as well. The 4343 / 4350 card is simply an older design.

[Mike C]

Hi Widget, A couple of year ago I owned A pair of 4350B's. I sold them along with a lot of other stuff hastely when I found out I had the big "C". Well I got over the cancer but not the sale of the 4350's, They too had a real muddy spot I'd call it around 250-300hz just above the crossover point with the urei network that came with the speakers. A friend of mine lent me a DBX active network to try and after some tinkering the muddyness disapeared. We ended up with the network set at 180 hz and the sound tightened up a whole bunch. I sold the whole system shortly after that so I didn't get to play with them too much. All in all thought I'm sure they didn't have anywhere near the mid bass of the stuff I'm playing with now. I think the benifit of active networks help tune these things and saying that my next purchase will be the DBX drive rack PA system to control the gear I have.

[Ian Mackenzie]

Giskard,

I appreciate your graphs and note with the 3145 network, these are 12/db slope graphs where in the active network as you state is 18 db slope, hence the assumed reversal of the dip in the graphs.

This is why I conclude the passive network 2122 conections are reversed and the other drivers are wired in phase with the 2122 as they are 18 db 3rd order end render the smoothest response in the passband and xover points.

Ian

[4313B]

Hello Ian,

As a follow up here is the passive 12 dB / octave 3144 /3145 filter. The dip is with the 2235H and 2122H in phase. The flat response is with the 2122H polarity reversed.

[Ian Mackenzie]

Wow,

I wonder whay GT would say to this...excellent stuff.

I have to say GT was certainly on the money with these designs, back then Xover design was so primative, he seems to have covered all the angles and those auto trannis chokes are way too cool.

Can't wait to try out that jasper Router and try some bi amping with the 700B!

Thanks again for these excellent graphs

Ian

[4313B]

Originally posted by Ian Mackenzie
I wonder what GT would say to this

As far as what? He thinks I'm goofy for messing around with all these old drivers.

Originally posted by Ian Mackenzie
I have to say GT was certainly on the money with these designs, back then Xover design was so primative, he seems to have covered all the angles and those auto trannis chokes are way too cool.

Yes, he really raised the bar and hasn't stopped since! I don't think he's too fond of tapped autotransformers though, at least that's what he's indicated to me on a couple of occasions

Originally posted by Ian Mackenzie
Can't wait to try out that jasper Router and try some bi amping with the 700B!

Very cool!

Originally posted by Ian Mackenzie
Thanks again for these excellent graphs

You're welcome I wasn't really in the mood at first but I got over it

[boputnam]

Originally posted by Giskard
Don't let it, the 4344 / 4345 /4355 is a better solution in my opinion. I've posted this at least twice before. I would run it with a 4343 or 4350 as well. The 4343 / 4350 card is simply an older design.

Agreed. The 4344 / 4345 /4355 solution seems an important evolution of JBL's four-way engineering.

And,

Originally posted by Giskard
I wasn't really in the mood at first but I got over it

Mighty glad you did! Glad I nudged you for it, to begin with! Proved a very interesting study.

[Ian Mackenzie]

Seems I have stired at the right time,

(just woke and have a nice pussy purring beside me muhhaha)

Testing my memory here, but I recall the 1977 JBL brochure said something about the mutual coupling of the 2121 & 2231.

By luck or good fortune the Nackamichi active xover we were using back then (1980) was in fact 6 db down at the crossover point so I neatly avoided this hump.

Ian

[4313B]

Originally posted by boputnam
Agreed. The 4344 / 4345 /4355 solution seems an important evolution of JBL's four-way engineering.

Well, if one absolutely has to have that exponential doohicky coupled to a plastic slant plate thingamajigger then I would think the 4344 / 4345 / 4355 would be the ticket What's quite interesting is the co-developement of the L250 Hello! Night and Day!

[boputnam]

Originally posted by Giskard
Well, if one absolutely has to have that exponential doohicky coupled to a plastic slant plate thingamajigger then I would think the 4344 / 4345 / 4355 would be the ticket What's quite interesting is the co-developement of the L250 Hello! Night and Day!

Yer RIGHT! I don't see how they get it done without the exponential doohicky coupled to a plastic slant plate thingamajigger

But, let's get back to who Ian's sleepin' with