How to improve the crossover parts in the 1400 Array?

[Ian Mackenzie]

"If not, what we do, how should we do since there's not much information about changing xo points while keeping the great sound from the superior models' xos."


I see two skills for this:

Design
Proficient in crossover theory and Leap simulation

Execution
Design and construction of active and or passive networks.

Experience
Even if you have the above you need Experience so you don't make costly mistakes.

Equipment and Tools

Even if the above was story booked in a how to guide unless you had the above 4 areas down pat it's not an option.

The notion of up picking someone's brain in a forum thread is an absurdity not to mention the time required to respond to these questions.

I mean it looked like you are reading into snippets of information that it was simply A+B=C

Unfortunately passion often becomes an obsession and then crashes with reality.

We have previously seen people come along blinded by obsession only to see them feed up two or three years later because they could not make it work.

[4313B]

+1

Thank you Ian.

[Ian Mackenzie]

Well it had to be said

One day when l retire l might load up what l know and have done in a series of iBooks that someone can buy online and doodle their lives away

[Mr. Widget]

+1

Thank you Ian.

Yep...

Greg Timbers makes it look easy. It isn't.


Widget

[JuniorJBL]

Yep...

Greg Timbers makes it look easy. It isn't.


Widget

Spot on Widget, B and Ian!


That was easy to say! And just about where it stops!


Thanks for all you do Greg!

[martin_wu99]

Where to buy 1400 Array that U.S made?

[johnhere]

Ian, your language indeed has style, not like anything I've read before, a little similar to Vollmann, but more coherent and more distant at the same time. Who are your favorite writers, may I ask?


I know Leap, good stuff, few can afford it. Actually I find out this is the most-leap-users speaker formus on the planet, this makes me head-ache for almost one hundred seconds.

[Ian Mackenzie]

The reference to Vollmann is amusing! I share nothing in common with this writer.

As far as this non fictional vacation is concerned l like Doug Self's books.

He'an Englishman, quite concise, well researched, presents well and frames the facts in a logical manner.

I don't go out of my way to post "directions" or spot light in every thread.

But sometimes it's more efficient to call it out sooner than later with a moment of "truth" or absolute clarity when the dialogue starts to turn into a "can't be type of situation".

We are all unique with our thought processes.

I recall in my first conversation with someone here many years ago when we were trying to make the crossover network for 2344 bi radial horn work and this NY'er lost his shit when l explained a solution. (surprise surprise)

Well about a week later the penny dropped and we have grown old friends since.

[Robh3606]

Ian is offering you very good advice. It's not easy and you can easily make a mistake while wiring the whole mess together. As an example this is what a CC network for the arrays looks like in schematic form. This is incomplete and does not include the 9 volt battery and wiring to the capacitor pairs to apply the DC voltage.

I built mine with Beryllium drivers so I had to make a couple of changes to the stock networks.

I have both LEAP and Clio to measure and run simulations so I could measure the voltage drives as 4313B said and run acoustic measurements in Clio.

This isn't something where you can just read a few quick posts on the internet and do. It has a bit of a learning curve and the right tools to get it right.

And yes Greg sure does make it look easy!

Rob

[honkytonkwillie]

...when I build or modify networks I run voltage drives on the whole mess to ensure that nothing is shorted or open and that the voltage drives look the way they are supposed to before I start hooking them up to expensive transducers and amplifiers.

I have both LEAP and Clio to measure and run simulations so I could measure the voltage drives...

Can someone describe this practice of measuring voltage drives a little bit?

Is it "just" a process where a known signal is measured/scoped across each component to verify proper attenuation?

[4313B]

You take a network and terminate it with the intended transducers or with resistors of the appropriate values. 6 ohms for something like a 2216Nd, 12 ohms for something like a 1501AL-2, 16 ohms for something like a 476Be, etc.

CLIO then makes it easy.
You hook up the CLIO box like in the diagram below but instead of a microphone on the input to the CLIO box you use test leads. Those test leads are connected to the output of each filter of the network and the resulting voltage is measured as CLIO makes a sweep.
For a two way network you would run two drives, one on the high pass and one on the low pass. CLIO can store up to ten sweeps.

Back in the old days we had to do it all by hand writing down the values at various frequencies; There was an HP plotter available back then that would do it with a plug in module but I personally didn't have one, I had to use various test equipment in the physics lab at Denver University. They thought it was funny how I repurposed their gear for testing loudspeakers. They ended up gifting me a very nice (but huge) HP VT audio oscillator and VTVM. Awesome stuff back then for a teenager!

Anyway... once you have those drives you can do neato things with them like import them into LEAP and use them as targets to make new networks, for example, you could measure the voltage drives of the active M2 loudspeaker, import those voltage drives into LEAP along with the impedance curves and frequency response curves of the 2216Nd and D2430K (also measured with CLIO) and then have LEAP come up with a passive filter equivalent.

[srm51555]

So your saying looking at the parts express catalog's crossover chart to find the values for the high and low pass on my M2's isn't going to cut it.

Really though, thanks for posting this. I really had no clue on how the design process worked.

For the record the M2's will stay active.

[Robh3606]

Here is a voltage drive for a network and the raw measurement for the compression driver in Blue and Grey on the top of the graph. The black line is the target curve and the paired lines are the simulated response in LEAP for both drivers in the set. If you look closely you will see that the voltage drive from the network is the inverse of the raw compression driver measurement. This gives you the simulated flat response for the network for a simple point and shoot on axis measurement. Basically you would design the network to give you a voltage drive that gives you your desired response.

Rob

[honkytonkwillie]

How cool! It almost looks like cheating though. I'll be doing it the old fashioned way for the time being.

[Ian Mackenzie]

It's not nearly as simple as it looks in practise.

Drivers are real world devices and they all have their quirks.

Sometimes the passive crossovers can interact with the driver passive EQ so you need lots of experience.

It takes skill and experience and a lot of time to get it right.

Then you listen and make adjustments from there.

People like GT do this stuff for a reason.