[joshsehn]

Many thanks for the side view and the in-depth explanation of your delays. Both helped me very well to understand what you are doing, and I (believe that I) can follow you completely.

I was really barking up the wrong tree.

Rudolf
www.dipolplus.de/thema11.php

no worries...  .   

I wonder why the the music and design fellow doesn't consider the impulse delay of the 10" driver? Kind of defeats part of his otherwise good argument doesn't it?

BTW, I've read some of your work and quite enjoyed it.

cheers,
Josh

[matevana]

Niice. So are you planning any design changes or upgrades? What, if any, are you unhappy with?

[joshsehn]

Niice. So are you planning any design changes or upgrades? What, if any, are you unhappy with?

I'm not really planning any changes nor upgrades unless I can think of a good reason to.... I'm pretty happy with them.

But if I was to do it all over, I would:

1. Build a nicer baffle or go crazy and use no baffle. It started out as a little test of a 2 way with a relatively small baffle for 5 1/4" and 3/4" tweeter. The sound was so good I took the plunge to build full range so I didn't really put any work into making it look good.
- I would probably build smaller baffles or even no baffles
- find a better way to mount the midranges (i.e. decouple from baflle)
- I would also perhaps use opposing push-pull woofer configuration so as to minimize vibration OR separate the woofers from the main baffle to decouple them from vibration transmission.

However, the "no baffle" option would probably require a much greater investments in drivers and electronics.

2. a. better tweeters... the ones I'm using now are pretty cheap but they sound fine as far as I can tell as they are only playing the top 2 octaves so I'm not pushing them very hard.
    b. bigger subwoofers. The 10" sound great but 12" or 15" could be even better.

3. Better xovers. Although I have not found any that better meet my requirements of low budget and better performance. I have been thinking about going "all the way" with computer controlled setup using ProFire2626 as a soundcard to drive it all. (This would be similar to StigErik's crazy no baffle system)

I've found that the DCX boosts and cuts don't to do quite as much as they are supposed to do. I also believe that a more phase linear (i.e. perfect) crossover would be beneficial. Perhaps the only way this can be realized is with a CPU controlled setup or a much much more expensive active crossover.

4. Better amplifiers with XLR inputs.

But I'm not certain I could do much better within my limited budget constraints.

More than all of the above, my impression of the speakers thus far is that any great level of financial investment would likely be best spent on improving my room.

A friend stopped over today... he's not an audiophile or anything but the first words out of his mouth before he even sat down was, "It sounds incredible" and after 2 songs he asked if I would build him a pair so it looks like I'll have an opportunity to build it again anyways. I'll probably build 2 sets of baffles. 1 new set for him and 1 for myself. This may also present an opportunity for me to do the electronics differently.

[Rudolf]

I wonder why the the music and design fellow doesn't consider the impulse delay of the 10" driver? Kind of defeats part of his otherwise good argument doesn't it?

Well, Kreskovsky was only talking about correcting geometric delays, since that has other implications than pure electronic delays. Possibly my example can make that clear:



I look from above at two domes on the front and back of a 72 mm thick baffle. Now I delay each of them 36 mm. Looking from the front (left part of the diagram, direction of green arrow) they now should be time aligned in a mutual plane at the middle of the baffle. Sadly this works only for one position in space.

If I change my position from the dipole axis to the dipole plane (right part of the diagram) there are still those two delays. But now they move (literally speaking) the domes along the baffle surface in opposite directions. This will severly distort the dipole figure 8 pattern above 2 kHz. There is no longer a constant attenuation of SPL to the sides of the OB.

Such things have to be kept in mind when changing timings in a geometric context.

BTW: I very much like your notion that at a certain point looking at the loudspeaker and room as a system should have priority compared to fiddling with the loudspeaker alone.

Rudolf

[matevana]

Very cool.  Amplification plays a big role. I use a crown xls-1000 class D amp for the lows (built in 24db/oct x/o) and a QSC GX3 for the highs, class A/B which has more detail up top. Everything is balanced, Low-Z and is very quiet and hum free. I'm curious how you are providing power to all of those driver pairs?

I am now also sold on the sound attributes of an analog eq stage ahead of the amps. I use a Rane 31 band stereo eq and am happier with the impact over that of the Behringer DEQ. (Will probably put that up for sale now, despite it's other capable functions).

I do believe that more dipoles will start to incorporate coax or single point source drivers, as these options improve. The Silver-Iris 10" coax driver for example, is an impressive cast frame unit (made by Eminence) that can sound quite nice crossed and eq'd properly, and could easily handle 60Hz on up, with great dynamics and an authoritative midrange presence.  It is made especially for the dipole market.

I am currently experimenting with magnet mounting the drivers. The baffle could be nothing more than foamboard (light and anti resonant). I built a clamping device which stradles the magnet from top and bottom and isolates everthing else from vibration. This seems to have much promise.

[joshsehn]

Well, Kreskovsky was only talking about correcting geometric delays, since that has other implications than pure electronic delays. Possibly my example can make that clear:



I look from above at two domes on the front and back of a 72 mm thick baffle. Now I delay each of them 36 mm. Looking from the front (left part of the diagram, direction of green arrow) they now should be time aligned in a mutual plane at the middle of the baffle. Sadly this works only for one position in space.

If I change my position from the dipole axis to the dipole plane (right part of the diagram) there are still those two delays. But now they move (literally speaking) the domes along the baffle surface in opposite directions. This will severly distort the dipole figure 8 pattern above 2 kHz. There is no longer a constant attenuation of SPL to the sides of the OB.

The second part of your explanation only applies if one was to apply the different delay to front vs rear correct? If we apply same/constant delay to both the green arrows would both move to the left correct?

[joshsehn]

Very cool.  Amplification plays a big role. I use a crown xls-1000 class D amp for the lows (built in 24db/oct x/o) and a QSC GX3 for the highs, class A/B which has more detail up top. Everything is balanced, Low-Z and is very quiet and hum free. I'm curious how you are providing power to all of those driver pairs?

I am now also sold on the sound attributes of an analog eq stage ahead of the amps. I use a Rane 31 band stereo eq and am happier with the impact over that of the Behringer DEQ. (Will probably put that up for sale now, despite it's other capable functions).

I do believe that more dipoles will start to incorporate coax or single point source drivers, as these options improve. The Silver-Iris 10" coax driver for example, is an impressive cast frame unit (made by Eminence) that can sound quite nice crossed and eq'd properly, and could easily handle 60Hz on up, with great dynamics and an authoritative midrange presence.  It is made especially for the dipole market.

I am currently experimenting with magnet mounting the drivers. The baffle could be nothing more than foamboard (light and anti resonant). I built a clamping device which stradles the magnet from top and bottom and isolates everthing else from vibration. This seems to have much promise.

I am using 6 fairly cheap stereo amplifiers. Dayton 2 X 75Wrms.

I like the idea of the coax as well... I'll have a lot at those Eminence drivers you mentioned.

[Rudolf]

The second part of your explanation only applies if one was to apply the different delay to front vs rear correct? If we apply same/constant delay to both the green arrows would both move to the left correct?

The delay has a value and a direction.
The value is constant for all arrows: the travel time equivalent to 36 mm.
The direction is "+" (added time of travel) for the lower arrow and "-" (subtracted time of travel) for the upper arrow. This would lead (in the left picture) to a virtual "common origin of sound" at the meeting of the arrowheads - as seen by someone sitting in the direction of the large arrowhead.

If both arrows had the same direction, we would simply move both domes back or forth in unison. This would correspond to moving both domes left or right in the right picture:

[joshsehn]

The delay has a value and a direction.
The value is constant for all arrows: the travel time equivalent to 36 mm.
The direction is "+" (added time of travel) for the lower arrow and "-" (subtracted time of travel) for the upper arrow. This would lead (in the left picture) to a virtual "common origin of sound" at the meeting of the arrowheads - as seen by someone sitting in the direction of the large arrowhead.

If both arrows had the same direction, we would simply move both domes back or forth in unison. This would correspond to moving both domes left or right in the right picture:

I think we both understand.... finding the right words to describe our understanding is the most difficult part. Your English is very good especially if German is your native language!