hi those values [100dBSPL] will be @2-3cm from speker driver. at 1m from speaker it will be not higher than about 80db SPL
umik can withstand 133dbSPL
so not a problem
you can check it with REW - try to measure at 1m and then move towards speaker. spl will rise the closer you get.
umik can withstand 133dbSPL
so not a problemyou can check it with REW - try to measure at 1m and then move towards speaker. spl will rise the closer you get.
So, I figured I could use cheap multichannel amplification with a 4 channels TPA3255 amp: expected configuration was 2 hypex nc400 for bass, and this 4xTPA355 for mid and high.
Well, the hiss coming from the horn with the TPA3255 is extremely loud, and completely disappear when I use a nc400 instead. Guess there is some truth about some amplifiers being better than others.
Yet, going to 6 hypex nc400 seem overkill and sort of a waste of energy. Are the smaller ucd hypex modules as quiet as the nc400?
If I had known I was going to build three way active speakers a few years ago, I would have bought 2x FA503!
Well, the hiss coming from the horn with the TPA3255 is extremely loud, and completely disappear when I use a nc400 instead. Guess there is some truth about some amplifiers being better than others.
Yet, going to 6 hypex nc400 seem overkill and sort of a waste of energy. Are the smaller ucd hypex modules as quiet as the nc400?
If I had known I was going to build three way active speakers a few years ago, I would have bought 2x FA503!
Yesterday, I went with the described "pro" routine :
I EQ'ed each driver separately (from near field measurement, not sure if this is the correct way of doing things, but since these speakers will move to another room, I am not sure I should EQ them from 1 meter ?)
Then I set some gains to level the three drivers.
And I programmed three filters in the DSP
I think I am missing the time delay correction, but my impulse tab in REW is unreadable with one large block of graph which I cannot seem to be able to filter to a meaningful impulse graph, I must have messed with the options somewhere...
I then measure from one meter, at listening ear height, and proceeded to measure my 8031C at the same position. My crappy speaker is orange and the state of the art Genelec is green.
As you can see, there is still some work to do! Can anyone point me to some documentation about setting up the digital filters? I just went with LR24 everywhere, and sometime overlapping (for instance, I think I put the bass filter cutoff is at 400Hz, while the mid begins at 300) trying to get something flat with vituixcad, but some hints or systematic approach would help!
I EQ'ed each driver separately (from near field measurement, not sure if this is the correct way of doing things, but since these speakers will move to another room, I am not sure I should EQ them from 1 meter ?)
Then I set some gains to level the three drivers.
And I programmed three filters in the DSP
I think I am missing the time delay correction, but my impulse tab in REW is unreadable with one large block of graph which I cannot seem to be able to filter to a meaningful impulse graph, I must have messed with the options somewhere...
I then measure from one meter, at listening ear height, and proceeded to measure my 8031C at the same position. My crappy speaker is orange and the state of the art Genelec is green.
As you can see, there is still some work to do! Can anyone point me to some documentation about setting up the digital filters? I just went with LR24 everywhere, and sometime overlapping (for instance, I think I put the bass filter cutoff is at 400Hz, while the mid begins at 300) trying to get something flat with vituixcad, but some hints or systematic approach would help!
Hi, read vituixcad measurement manual (for REW). While you read it try to understand why the guide is what it is
Since the speaker is big you should ideally measure even farther than 1m, which meas windowing gets shorter at home in small room, so accuracy gets worse. In general there is some error with home measurements. Making EQ and crossover with errorneus data will yield some error in sound as well. If you know what the error is, you can somewhat compensate / expect, and try use some other method verifying.
Since the speaker is big you should ideally measure even farther than 1m, which meas windowing gets shorter at home in small room, so accuracy gets worse. In general there is some error with home measurements. Making EQ and crossover with errorneus data will yield some error in sound as well. If you know what the error is, you can somewhat compensate / expect, and try use some other method verifying.
Last edited:
Your path to greatest success will be combining good measurements with simulation. I recommend VituixCad for this, but if you are not going to bother with optimizing the polar responses / directivity index, then there may be other simulation programs which are easier to master.
It will be much easier for you to dial in the lower crossover if you have a good quasi-anechoic response down into the bass region. You do this by making a good near field measurement of a driver and merging it with the gated far field measurement. This process removes the effects of room modes.
You are making great progress so far.
It will be much easier for you to dial in the lower crossover if you have a good quasi-anechoic response down into the bass region. You do this by making a good near field measurement of a driver and merging it with the gated far field measurement. This process removes the effects of room modes.
You are making great progress so far.
It should be possible to use ground plane measurements instead of near field. VituixCad has an option in the Merger tool to use several types of baffle step simulations, and one of the options is "none"... which is what we would use when doing ground plane.
The nice thing about ground plane is that there is no upper frequency limit in theory. I have successfully measured woofers up to 2k. Above that I was seeing small differences between the gated response and the ground plane response. But others on this site have been able to measure all the way to 20 kHz with ground plane... It simply requires extremely good measurement setup and technique.
https://www.diyaudio.com/community/...-close-enough-to-anechoic.353347/post-6175663
j
The nice thing about ground plane is that there is no upper frequency limit in theory. I have successfully measured woofers up to 2k. Above that I was seeing small differences between the gated response and the ground plane response. But others on this site have been able to measure all the way to 20 kHz with ground plane... It simply requires extremely good measurement setup and technique.
https://www.diyaudio.com/community/...-close-enough-to-anechoic.353347/post-6175663
j
what would be the ideal size for the mid woofer (faital 10pr320) sealed chamber, it was about 25 liters in the prototype but it was more or less sized with whatever scrap ply I had lying around. Operating range is 300-1500 more or less (1200 to 1800?)
it has a Vas of 45.9l, should I aim this? Any rules?
it has a Vas of 45.9l, should I aim this? Any rules?
Depends on your definition of 'ideal'..........
a = (Qb/Qts')^2 - 1
Vb = Vas/a
Qb = 0.707/max flat, 0.5/transient perfect or choose something in between
(Qts'): (Qts) + any added series resistance (Rs)
