[WerTicus]

interesting... well i think ill give this analogue 24db oct crossover a go

ill use only the highest quality components... the real question then becomes which op amps to use!?  since the rest is easy

they are obviously going to have to be pin for pin compatible with the ones the schematic is currently designed for.... dvv would your recommended AD826AN suit this application?  and where would i buy this (i have a wholesale RS account btw)

[sfdoddsy]

What drivers are you using?

Steve

[DVV]

interesting... well i think ill give this analogue 24db oct crossover a go

ill use only the highest quality components... the real question then becomes which op amps to use!?  since the rest is easy

they are obviously going to have to be pin for pin compatible with the ones the schematic is currently designed for.... dvv would your recommended AD826AN suit this application?  and where would i buy this (i have a wholesale RS account btw)

I don't know about RS, I buy mine direct from AD via Internet, using my MasterCard for payment. Last time, they cost me like $3.68 per piece, and I bought 40, since my current stocks were sort of depleted.

Yes, they are pin-to-pin compatible with TL072/82 series, but make sure you order with the "AN" suffix - this denoted 8 pin DIL package, which is what you want, just "AD826" might get you the surface mount version. So, make sure you use the full name - AD826AN.

Its advantages are that its entire output stage runs in pure class A (most others use their output stages in class AB), that it has a more than sufficent slew rate at 350 V/uS (you actually need less than 20 V/uS, but it's nice to have a LAAAAARGE overhead margin), and, most important, its SETTLING time is below 45 nS (which gives you the so much desired ambience capability).

If you ever use it in its linear mode (which is what I do most of the time), always bypass the feedback resistor with a silver mica low value capacitor (e.g. for 4K7 resistor, use 3 pF, for 10K resistor use 5 pF, etc). This rids you of some glitches which can even be seen on a 'scope and anyway, what what you want to be doing above 1 MHz at full power output? By then, your 20Hz-20kHz phase shift is below 0.1 degree, in practical terms totally meaningless.

A suggestion, if I may - before proceeding any further, try to search out a book, published in 1982 (and probably with later editions) by National Semiconductor, called the Audio Handbook. It's one of the best of its kind I have ever seen, with detailed explanations of various circuits, a section on electronic crossovers, with all types explained, all maths provided for, actual examples draw out and ready-made tables of components for various crossover points already supplied.

If you should fail, I think I have it scanned somewhere, so I could send it over, but those will be kinda large files. Anyway, if required, I will do it.

Cheers,
DVV

[DVV]

There's always conflicting information in audio. That's what makes it interesting.

 

I understood you were being recommended 12dB slopes, hence the question mark with the drivers. 18dB would cause feewer problems, and 24dB even less.

Not really, my feeling is that 18 dB/oct provides the best set of compromises between cut rate and phase linearity.

Now that you mentioned it, there is one case where I would give 12 dB/oct my firmest possible recommendation. That is when you use an op amp as a line buffer, and the input differential pairs of the power amps as both that AND filters for the drivers. What you lose out on cut rate, you regain and then some on simplicity and ellegance of the design.

However, this is not something for a novice and/or the feint of heart, because you have to design, test and tweak discrete components, always much harder to do than using tried and tested op amps. I've done it only once and was pleased with the initial results, but unfortunately, I never had the opportunity to really develop it. That was in 1986, my son was born, so instead of fiddling with trannies, I had my hands full of diapers, milk bottles and baby powder. And there was this planar speaker of unbelievable efficiency (heck, it ran without any power at all!) in my hands, made piezo electrics look distinctly crippled.

And after tweaking him for 17 years and 11 months, all I got out of it was that he nicked my AR speakers last month.

Cheers,
DVV

[WerTicus]

sfdoddsy:
I have yet to decide on the drivers but I will probably be using the ring radiator scan speak 70000 tweeter.

the 13M/8636 kevlar mid range 4.5" or the sliced paper 12m4631g00 2.5"
for mids   ( i keep changing my mind as to which one)

and then then if i choose the sliced paper mid range ill go with the 6.5" sliced paper woofer the 18w8531 to match or if i go with the kevlar mid ill go with the kevlar woofer to match ... probably the 8" 21w8554g00 which needs to be crossed below 1khz... but that wouldnt be a problem as the 4.5" mid will go down to 500hz or so no problem

that will be the three way system - with crossovers at 4khz/ 500/ 100 to a subwoofer... my adire shiva.  Ill be using a similar design to this http://home.hetnet.nl/~geenius/Diary.html

....

dvv: thanks ill look into that op amp when i go to build this thing!  hehe i just fixed up my dad's old ar's for him.... they are not quiet good enough for my tastes!  wood is too thin for them to ever been good really.

in any case with the design, these drivers, some quality amps, some quality construction and im gonig to be certain to blow myself away

[DVV]

...
dvv: thanks ill look into that op amp when i go to build this thing! hehe i just fixed up my dad's old ar's for him.... they are not quiet good enough for my tastes! wood is too thin for them to ever been good really.

Don't bet on that. I agree, their wooden sides are VERY typical of industrial production runs, in my case at 19 mm thickness, and surely enough, I would have preferred to have seen 22, or even 25 mm walls. However, those are air suspension or compression speakers, and they work a little differently from bass reflex.

Also, there are techniques which can be used to minimize resonance, such as for example bracing, which can be used to stiffen speakers.

As for taste, well, that's not to be discussed, to each his own. All I can say is that AR, like everybody else, had their outstandingly good and bad products. I do like their general sound, but in all honesty, there were models which I wouldn't be caught dead with.

in any case with the design, these drivers, some quality amps, some quality construction and im gonig to be certain to blow myself away ...

True. Going active improves literally everyhting, the only question being to what extent, i.e. was it worth the time and trouble. It usually is, if you can keep the electronics prices down to a corresponding level to the quality of the speakers. But there is no fixed rule.

Cheers,
DVV

[WerTicus]

well thanks for your advice dvv ill let ya know how it goes... though im still in the design phase for the speakers atm so it might be a while - heck when i start the construction ill start a thread about it with step by step pics for the fun of it