Oh you are totally right. My XO parts will come Friday, I have nothing to pad it with. Indeed, the low frequency elephant sized cat is gone with parts of the magic. But the mids are insanely good, I hear the studio backwall echo, some very previously unusual instrument separation, some instruments moving clearly back from the microphone. But too bright now, will be fatiguing soon. Also I think the woofer and twitter are overlapping too much, I hear some sort of weirdness when the same sound has two sources. I think the distortion in the mids is up.
It's not just padding, I will be building the XO from post #191
It's not just padding, I will be building the XO from post #191
Last edited:
I fixed the box leakes, if you are interested those are impedance files from 12 L sealed box:
Finished XO from post 191 except the tweeter is inverted. Here each driver separately and all together, measured at the actual listening position from a pic two pages ago.
The sound is better than anything I ever heard (however I never auditioned anything really high end, no idea how it sounds). Next level in details and separation. There is some very enjoyable smoothness in the sound, and yet everything is quite punchy - I don't understand how both can be at the same time. The amp is Aegir, they have synergy. I've made only one box for now, the second is still Dynaudio, somehow they play great together, much better than both identical Dyns. I'll get 200 hours of listening before deciding how I will do the 2nd box. They don't have to be identical, I'm past this stage lol.
Thanks everyone and especially @perceval for making it to happen!
The sound is better than anything I ever heard (however I never auditioned anything really high end, no idea how it sounds). Next level in details and separation. There is some very enjoyable smoothness in the sound, and yet everything is quite punchy - I don't understand how both can be at the same time. The amp is Aegir, they have synergy. I've made only one box for now, the second is still Dynaudio, somehow they play great together, much better than both identical Dyns. I'll get 200 hours of listening before deciding how I will do the 2nd box. They don't have to be identical, I'm past this stage lol.
Thanks everyone and especially @perceval for making it to happen!
I've used 18 gauge inductors in the woofer, and there are 2 of them serially. I think if one day I change them to 14 gauge. My current DCR is 0.5 Ohm + 0.22 Ohm. Theoretically I can move them to 0.25 + 0.11 Ohm. I am not sure what is the exact formula in Db gain is, if the impedance of the driver itself in the target area is 8 ohm, so the effect is about 5% reduction of the impedance. I am not sure adding resistance to the tweeter is sonically identical way. My Aegir amp kinda prefers sub-8 ohm load.
BTW I am using an autoformer based DIY volume control with separate knobs. I don't mind the drivers to have slightly different sensitivity. Right now my Dyn has the volume 1 click lower because apparently is has higher sensitivity than TB (I know they use very simple crossover). So I may slightly improve my TB sensitivity, just IMHO.
BTW I am using an autoformer based DIY volume control with separate knobs. I don't mind the drivers to have slightly different sensitivity. Right now my Dyn has the volume 1 click lower because apparently is has higher sensitivity than TB (I know they use very simple crossover). So I may slightly improve my TB sensitivity, just IMHO.
The woofer has definitely some distortion somewhere in the 1k - 3.5k area, my next quest is to determine exactly where. I know for sure because playing for few days with tweeter only, woofer off, I have zero ear fatigue. Playing only the woofer has exactly the same fatigue as with both. I will be gradually increasing the inductance until the fatigue is gone. Also will make it 2nd order (for now it's just post 191). You can guess each step will take a week.
My question is - is it possible to change the cone breakup (resonance) frequency by changing the box size?
This tweeter is designed to go as low as 1k, so the plan is to move the crossover lower, unless the breakup frequency is the box defect.
My question is - is it possible to change the cone breakup (resonance) frequency by changing the box size?
This tweeter is designed to go as low as 1k, so the plan is to move the crossover lower, unless the breakup frequency is the box defect.
Only a handful of tweeters can happily play down to 1k, and the TB coaxial 25mm is not one of them.
Here I measured only the woofer with low pass distortion using stepped sin 3 measurements per octave. I don't think it has any distortion on the sin wave. I am not sure if I can measure anything but sin with a microphone, perhaps an oscilloscope is needed for the square wave?
More comfortable explanation is that the fatigue is coming from the elevated 1k-8k level you can see in the post 443. I can try to keep the crossover frequncy the same but just push the curve 5 db down.
I'm happy you tried to measure if there are indeed distortion area, not just guessing.
My pair is distortion free through the range.
And yes, I already mentioned your curve is too hot.
Try to get the XO to follow my red line instead and you will have zero ear fatigue while still preserving details.
My pair is distortion free through the range.
And yes, I already mentioned your curve is too hot.
Try to get the XO to follow my red line instead and you will have zero ear fatigue while still preserving details.
but they are desktop speakers. Sitting quite close. Finding out the perfect angle might be a bit harder.
But easy, just extend your arm and pivot! 🙂
But easy, just extend your arm and pivot! 🙂
I was not totally happy with the crossover above and decided to finally try the clever @perceval serial, based on #430 but slightly tuned. This is how it models:
And this is how it really measures:
The quality is indeed better than the classic crossover, I've got very detailed, accurate sound, nice soundstage. I finally have better confidence that this could be an end of road driver.
However as you see the response curve is very incorrect and for now I'm lacking skills to fix it. I was swapping some spare parts and this is the best I've got so far, which is quite pathetic. I had a similar problem with previous XO, the woofer just can't get to the same loudness. Need some brain...
And this is how it really measures:
The quality is indeed better than the classic crossover, I've got very detailed, accurate sound, nice soundstage. I finally have better confidence that this could be an end of road driver.
However as you see the response curve is very incorrect and for now I'm lacking skills to fix it. I was swapping some spare parts and this is the best I've got so far, which is quite pathetic. I had a similar problem with previous XO, the woofer just can't get to the same loudness. Need some brain...
Attachments
@perceval your original circuit is dangerous because it makes the tweeter to play down to 1 khz. I've added a capacitor to bend it down below 3.5 khz. Also added an impedance reduction chain.
Now there is a distortion jump in the 6 khz area, I suspect this will be fatiguing, only just started to listen. The sound is pretty good but of course too bright, don't know how to fix it yet.
I just left home and can't measure or have access to my files for the next 2 weeks.
There is no distorsion nor fatigue when I listen to my setup. I have a Harmann drop of about 5 to 7dB. I will have to take a look at the tweeter XO again when I come back. 1kHz is not a problem being desktop speakers. You will not push the tweeter out of its comfort zone.
I pushed mine with some trance and deep house, for fun, and there was no sign of struggle. Even at some high levels.
The drop around 500Hz might be coming from your setup. My setup is a clear path in front of the speakers, and about 70cm off the wall behind. It looks like you may have a desk bounce between the speaker and your ears that might induce the 500Hz chasm you are experiencing.
Do you high pass your speaker at 110Hz? Because mine can reach 50Hz in my setup easily.
There is no distorsion nor fatigue when I listen to my setup. I have a Harmann drop of about 5 to 7dB. I will have to take a look at the tweeter XO again when I come back. 1kHz is not a problem being desktop speakers. You will not push the tweeter out of its comfort zone.
I pushed mine with some trance and deep house, for fun, and there was no sign of struggle. Even at some high levels.
The drop around 500Hz might be coming from your setup. My setup is a clear path in front of the speakers, and about 70cm off the wall behind. It looks like you may have a desk bounce between the speaker and your ears that might induce the 500Hz chasm you are experiencing.
Do you high pass your speaker at 110Hz? Because mine can reach 50Hz in my setup easily.
The area from 0 to 2 kHz is not of a concern at all because I plan to rebuild the box anyway after I fix the 2k+ area with xo. I tried to build a 12l box but I installed so many bracings and the speaker itself that it is 10l. I'll make it 17l, much larger odd shaped backwards tilted baffle.
The concern is that whatever I do the tweeter is much, much louder than the woofer, in any xo variant. And with 454 variant I've got a significant distortion at 6 kHz. 453 no distortion but the whole tweeter area is much louder.
The concern is that whatever I do the tweeter is much, much louder than the woofer, in any xo variant. And with 454 variant I've got a significant distortion at 6 kHz. 453 no distortion but the whole tweeter area is much louder.
Just had a look at post #430 and your XO.
You changed the values of many components, and your measurements differ from mine quite a bit. I'm not saying it's a bad thing, it's ok to experiment, but we have different output, that's for sure.
The tweeter should be at -15dBs around 1kHz, which should be fine for most applications. This is not a driver intended to play techno music at a dance hall in the wee hours of the morning! 🙂
You changed the values of many components, and your measurements differ from mine quite a bit. I'm not saying it's a bad thing, it's ok to experiment, but we have different output, that's for sure.
The tweeter should be at -15dBs around 1kHz, which should be fine for most applications. This is not a driver intended to play techno music at a dance hall in the wee hours of the morning! 🙂
Hi Perceval! Of course I've started with your XO almost exactly with the parts I had, I saved the measurements (attached) and moved interactively changing 1 component at a time. The box difference, placement difference, huge flat monitor in between, only 1.5 ft from the wall. I presume each driver is different a bit. This is why we DIY, we can build the speaker for the actual location and not in the middle of a huge room with dome ceilings. And I am still working on the left speaker to perfection, the right one is months away...
