Problem is that I have a single EQ for each band, that does speaker EQ, XO, and room correction all together, but I will post them with comments.
XO points are 140, 500 and 1722 Hz.
Tweeter:
Some diffraction at 10kHz, and a resonance at 1278 Hz are the main issues here. Apart from that its very well behaved.
Midrange:
Dipole peak EQ is centered around 1520 Hz. The peak is a bit higher in frequency actually, around 1700 Hz. Also some EQ at 2396 Hz to flatten the overall response because of the dipole peak. The EQ at 700 Hz is room related (floor reflection)
Woofer:
Now it gets worse... a lot of room correction EQ here! Some are room modes, and some are to compensate for floor reflections. The large dip at 1 kHz is to compensate for the dipole peak and is also a part of the dipole EQ itself.
Subwoofer:
Finally we see what dipole compensation can look like! There are various things going on there making it necessary to compensate approx 12 dB/oct from 20 Hz and up. I get some additional roll-off below 40-50 Hz because the driver have Fs at 33 Hz and rather low Q. And then we have a large dipole peak around 150 Hz. Some room modes around 50 and 80 Hz are EQ'ed as well. A lot of EQ around 20 Hz, but the sensitivity at 20 Hz is above 90 dB, so I dont worry.....
Now try to do this EQ with passive XO....
Tweeter:
Some diffraction at 10kHz, and a resonance at 1278 Hz are the main issues here. Apart from that its very well behaved.
An externally hosted image should be here but it was not working when we last tested it.
Midrange:
Dipole peak EQ is centered around 1520 Hz. The peak is a bit higher in frequency actually, around 1700 Hz. Also some EQ at 2396 Hz to flatten the overall response because of the dipole peak. The EQ at 700 Hz is room related (floor reflection)
An externally hosted image should be here but it was not working when we last tested it.
Woofer:
Now it gets worse... a lot of room correction EQ here! Some are room modes, and some are to compensate for floor reflections. The large dip at 1 kHz is to compensate for the dipole peak and is also a part of the dipole EQ itself.
An externally hosted image should be here but it was not working when we last tested it.
Subwoofer:
Finally we see what dipole compensation can look like! There are various things going on there making it necessary to compensate approx 12 dB/oct from 20 Hz and up. I get some additional roll-off below 40-50 Hz because the driver have Fs at 33 Hz and rather low Q. And then we have a large dipole peak around 150 Hz. Some room modes around 50 and 80 Hz are EQ'ed as well. A lot of EQ around 20 Hz, but the sensitivity at 20 Hz is above 90 dB, so I dont worry.....
An externally hosted image should be here but it was not working when we last tested it.
Now try to do this EQ with passive XO....
Ah well, I like to listen to music, not look at drivers....
but I see your point! I did have an idea to make a glass frame and suspend the drivers in nylon fishing line.... that would be rather cool.Ah well, I like to listen to music, not look at drivers....
That wouldn't be cool,...
... it would be cool times 10!!
Mundorf AMT update
I see Mundorf is updating the AMT line up with stuffed back chambers. No more dipole?
http://www.hs-devices.com/english/amt.htm
I see Mundorf is updating the AMT line up with stuffed back chambers. No more dipole?
http://www.hs-devices.com/english/amt.htm
LOL - passive XO guys get warned !
XO points are 140, 500 and 1722 Hz.
Tweeter:
Some diffraction at 10kHz, and a resonance at 1278 Hz are the main issues here. Apart from that its very well behaved.
An externally hosted image should be here but it was not working when we last tested it.
Midrange:
Dipole peak EQ is centered around 1520 Hz. The peak is a bit higher in frequency actually, around 1700 Hz. Also some EQ at 2396 Hz to flatten the overall response because of the dipole peak. The EQ at 700 Hz is room related (floor reflection)
An externally hosted image should be here but it was not working when we last tested it.
Woofer:
Now it gets worse... a lot of room correction EQ here! Some are room modes, and some are to compensate for floor reflections. The large dip at 1 kHz is to compensate for the dipole peak and is also a part of the dipole EQ itself.
An externally hosted image should be here but it was not working when we last tested it.
Subwoofer:
Finally we see what dipole compensation can look like! There are various things going on there making it necessary to compensate approx 12 dB/oct from 20 Hz and up. I get some additional roll-off below 40-50 Hz because the driver have Fs at 33 Hz and rather low Q. And then we have a large dipole peak around 150 Hz. Some room modes around 50 and 80 Hz are EQ'ed as well. A lot of EQ around 20 Hz, but the sensitivity at 20 Hz is above 90 dB, so I dont worry.....
An externally hosted image should be here but it was not working when we last tested it.
Now try to do this EQ with passive XO....
Beautiful distortion plots for the new AMT 2310c and AMT 2510c!
From reading in a German DIY magazin, the front plates are no longer removable either – though we still see screws at the AMT 2310c in this pix...
Might be they make a nude dipole version available, as was with the older line - haven't checked out
Michael
Last edited:
Stig Erik,
The crossover between bass and mids is rather high (500 hz). The distance between the highest mid and lowest woofer is more than 1.5 wavelength. Does this not degrade integration? Mind you that at the crossover you have 5 units generating 500 hz tones together.
In my own Unbaffled Dipole I try to cross as low as possible, without overloading the midwoofer electrically. To be able to cross even lower I am contemplating using a more sensitive driver such as the B&C 6NDL38. It has sufficient excursion capabilities and is able to withstand a lot more power.
Considering the fact that your crossover is done on a computer and you are much more flexible in your settings than I am, you could also use one of the W22's up to 500 hz and let the other two roll off earlier.
The crossover between bass and mids is rather high (500 hz). The distance between the highest mid and lowest woofer is more than 1.5 wavelength. Does this not degrade integration? Mind you that at the crossover you have 5 units generating 500 hz tones together.
In my own Unbaffled Dipole I try to cross as low as possible, without overloading the midwoofer electrically. To be able to cross even lower I am contemplating using a more sensitive driver such as the B&C 6NDL38. It has sufficient excursion capabilities and is able to withstand a lot more power.
Considering the fact that your crossover is done on a computer and you are much more flexible in your settings than I am, you could also use one of the W22's up to 500 hz and let the other two roll off earlier.
That would make it a 5-way... but why not?
What integration problems are you thinking of? Frequency response (at the listening positition) is not a problem, I have full control over that.
The integration in the vertical plane probably could be better. Don't you get different results when you move the mic a bit up and down? The change I proposed will probably decrease this effect. I'm not sure if it will lead to better sound though.
Last edited:
I've been think a bit more about this... I will build a prototype frame to see if moving the posts out of the dipole null causes any problems. There is significant tweeter output at the location of the new frame posts, so it could be a problem. Lets see.
This is interesting:
While I'm waiting for my finished H-baffles to arrive, I just had to test a few things on the flimsy prototypes. They are build from 19 mm MDF without any bracing, so they are not very stiff. They vibrate a lot! The finished baffles will be much stiffer. I dont quite hear any resonances or colouring of the sound, but it cant be good, so I made a brace from lumber between the long side walls. What a difference! It sounded almost like I gained 2 dB sensitivity, had to lower the subwoofers 2 dB for them to integrate with the rest. There's also much more transient attack and "punch". Interesting experience. At the same time I changed the wiring from 2.5 sq mm to 6 sq mm, which probably contributed a little as well (maybe).
As you probably understand, I'm very exited to hear the finished H-baffle subwoofer towers with very stiff baffles and four 21" woofer per side!
While I'm waiting for my finished H-baffles to arrive, I just had to test a few things on the flimsy prototypes. They are build from 19 mm MDF without any bracing, so they are not very stiff. They vibrate a lot! The finished baffles will be much stiffer. I dont quite hear any resonances or colouring of the sound, but it cant be good, so I made a brace from lumber between the long side walls. What a difference! It sounded almost like I gained 2 dB sensitivity, had to lower the subwoofers 2 dB for them to integrate with the rest. There's also much more transient attack and "punch". Interesting experience. At the same time I changed the wiring from 2.5 sq mm to 6 sq mm, which probably contributed a little as well (maybe).
As you probably understand, I'm very exited to hear the finished H-baffle subwoofer towers with very stiff baffles and four 21" woofer per side!
Last edited:
Be interesting to see if anything shows on the measurements.
Makes sense; reducing box flexion would decrease energy absorption and produce a sharper attack. Do time or frequency domain measurements show any difference?
How long are the cables? I had a similar experience of a 1dB adjustment when changing internal speaker wiring runs of about one meter from 1.3 to 16 sq mm. That was by ear some years ago and I didn't have measurement gear at that time.
Haven't done the mock-up frame posts yet. What you say about box flexing and cable makes sense. The woofer cables are around 1 meter long, the amps are placed just behind the baffles.
What I have done this evening is to suspend the tweeter and midranges separately. Seems like a good idea, listening tests will tell! Also shifted from MTM to TMM layout.
What I have done this evening is to suspend the tweeter and midranges separately. Seems like a good idea, listening tests will tell! Also shifted from MTM to TMM layout.
An externally hosted image should be here but it was not working when we last tested it.
Well... lets consider that the subs themselves are also suspended, the entire baffle "floats" on a rubber tire tube, so our friend Newton is here too!
The vibrating baffle is a different problem than the drivers moving back and forth. The vibrating baffle is distorting the signal since it also happens in the time domain (resonances), the moving drivers is (in theory) not, only affecting the output level, although very little. There are some good explanations on this from Michael (mige0) earlier in this thread.
The vibrating baffle is a different problem than the drivers moving back and forth. The vibrating baffle is distorting the signal since it also happens in the time domain (resonances), the moving drivers is (in theory) not, only affecting the output level, although very little. There are some good explanations on this from Michael (mige0) earlier in this thread.
Last edited:
