Crossower suggestion

[Zeljko]

I have Peerless midbass HDS 164 id:850439, ScanSpeak tweeter model D2010/851300 and finally passive radiator SEAS SP17R. Can you help me with making crossover? I have this one (my friend was maid it with simulation software), but I wish that someone check it.
Any suggestion?
Regards!

Zeljko

[Hugo Rodriguez]

Crossover design is the most difficult part in the system design.

Time ago I worked them by trial and error, that means moving tha value in one component while remaining unchanged the others, and ploting the frequency response, with FFT technics to save money and time.

All this job requires a lot of coils and capacitor values, and also very important: to measure in an anechoic field.

We realized the tests using the final assembly and having both units mounted in the box but one disconnected, to analize the other.

What I think is that you can evaluate the performance using just your ear if lacked instrumentation. I do not know if networks designed with simulators can give good results.

Try reproducing non electronic music as violins, flutes, guitars, etc. The best systems reproduce them musically and also give good resolution when playing several instruments at a time.

Human voice is also a good test signal, then opera areas sung by divas will tell you how much your crossover works (or fails).

Regards

[Sjef]

Originally posted by Hugo Rodriguez
I do not know if networks designed with simulators can give good results.

Yes, simulation can give very good results. Specially when you are dealing with a complex crossover circuit every component has got it's influence on the other components. Changing one component often means that you have to re-design the whole filter, otherwise you really don't know what you get anymore. (strange phase errors, large dips or peaks in response etc.)

I am using CALSOD for years now and it's amazingly accurate. When you measure the speaker with a Calsod optimized filter it looks almost exactly the same as the simulation. The real problem is that you really have to know what you are doing. Making good measurements, making the right choices for you filter design, knowing how to read the measured curves, specify your own target curve etc, there is no software that can do that for you. The end result totally depends on your own skills, the software is just a very accurate calculator, it can be very educational though.


As for checking the supposed filter design. I, and I think, nobody else can give any good advise on this, just because we don't know how the frequency and impedance response of the speakers in your enclosure.

The only thing I can tell you is that it doesn't look strange. The design looks oke, the component value are not strange. It looks like a 18db/oct crossover with a crossover frequency of somewhere in between 2,5 and 3 kHz, nothing wrong with that, but again, that's the only thing I can tell.

[Lionel]

I'm not a specialist but we will see what will tell us crossover guru...

I will said that it is :

1 - a second order Linkwitz-Riley for the bass section
2 - a third order Butterworth for the tweeter section
3 - There's a Lpad attenuator (2 unknown resistances) on the tweeter section

I'm sure it doesn't help...

[Zeljko]

First, thank you all for reply.
OK, I think that I can make crossover without any midbass or tweeter measurements but I see it is wrong way.
Once more thing, can I make .frd (for frequency response) and .zda (for speaker impedance) files without measurement, only if I have frequency response and impedance diagram (I mean read diagram and put value in file).
Regrads!
Zeljko