12" OS zvukovody (Stránka 5 z 10)

I'm all ears, Jack Any competent competition? How much is the final crossover different from the first one?

Well I'm actually quite happy with DE250s so far. I'm sure the foam plugs also do some good. For bass I have prepared four 15" for open baffles from the last project - it should suffice for a while, it will be DSPed by another digital crossover, here on a picture together with miniDSP modules:

Need drawing some power supplies now. I still play from PC with filter emulator on a cheap receiver... and like it very much.

I don't think it's a step in the right direction if you plan subwoofer(s) (which is a good idea anyway). With subs you don't need extension and can save much heat from being generated without any benefit - more current, more heat, more distortion and less dynamics. I haven't tried it, but can't imagine using any low sensitivity 12" woofer for this kind of loudspeaker with otherwise such unlimited capabilities.

Did you try to decrease the level above 8-10kHz even more? To match a regular dome tweeter's power response (i.e. what we are used to) it should IMHO copy it's aprox. 30 deg response, maybe even more, i.e pretty fast falling. Also, throat in our WG is not perfect, Jack - I've smoothed it a bit by precise template created with laser printer, putty and a sandpaper. I don't know if it improved anything but I don't hear any harshness. I would try foam plugs - just to see if it makes a difference, but well it can be anything, including the driver.

In my case it sounds definitely much "sharper" without the plugs, but I didn't make any comparsion with/without with the correct level matching, so can't really say anything regarding plugs **. I just don't have the time to test everything. But I will do it one day, 'cos I'm simply curious...

** Well, I wrote earlier that "I'm sure the foam plugs also do some good." I should have written "From what I've heard so far, I would say it does some good".

You problably won't recognise these things on a frequency response data. I tried to lookup differences with and without plugs directly in impulse response, but didn't find anything suspicious. Maybe it was just the level - it would mean that the crossover is really 90%, maybe more.

- To the throat precision - at one place it just seemed to be a little sharper then it should, nothing that dramatic. I then tried to fit it better to 1" without much conical part and smoothly.

Is there any feedback from others who bought our WG? I remember some went to Scandinavia at the same time my were made.. Jeez, it was at Christmas last year and I'm still not finished

J. Gerhard hanging around would be a good promotion Geddes won't find a time? Then I would set out...

- Honestly I admit I've never heard of Ines Adler before.

"Initially just as a hobby, Inčs Adler began experimenting with widebanders in the 1980s. After she finished her engineering studies with a master's degree, she went to work for Daimler's Mercedes Benz diesel section where she developed engine electronics, ECUs and software. She holds 14 patents on diesel combustion, injection and exhaust treatments."

Wow!

Check your mailbox.

Is this the final measurement? How was that measured (and before)? This way it would sound very, very thin, bright and fatiguing, lacking "body". It would almost look like you measured and then simulated with mismatched transducers levels. Or? Can't believe you would seriously listen to this..

Otherwise the directivity looks very well behaved. But at 90° it should be about 20dB below the axial response so I suppose it's really less then 90 (at 45° it should be about 6dB down) (?).

- Once you have correct data, including impedances, the simulation should be pretty close to reality Jack, Xover is good as anything else...

Now I'm really looking forward to hear from you when this is corrected

As I said, if you had all the input data right, all you would see in Xover would be pretty excatly all you would get, i.e. what you would see if you measured it as a whole (except crossover parts tolerances etc., but this is usually quite small effect). There must be something wrong and I would bet on input data if the result is so much different. If you have to switch polarity to match the simulation, then the input data must be completely wrong somewhere (level, phase, impedance scale, a switched polarity, at least -  I don't know).

As the first step, I'd try to eliminate all the reflections from the data.

- You can easily check real network voltage gains with Arta or whatever. It should closely follow the "Net Gain" plot in Xover. If not, then most probably you have the impedance(s) wrong or some parts are different values. You can check real part values in Arta also.