HalAir Aeralis - Fullrange Line Array (Vifa TC9-18-08)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Minidsp path is interesting indeed. The one thing that bothers me on the 2x4HD is the lack of any volume display indicator. I have checked and googled+++ but found no easy way to sort the volume-display issue ... have you found a way?

I am sure the minidsp will be way improvement :p
peq and digital xovers will probably do magic :D

Harald, if/when - welcome.

Ran the system for a few hours today, no failsafe triggered. BYRTT´s suggestion of DC offset may have been a good call :magnify:

Thx a lot :p
 
Last edited:
Pulled out REW to get a baseline before I start any serious EQ efforts.
The lines have not been EQ´d individually yet.
 

Attachments

  • L+R_EQ.jpg
    L+R_EQ.jpg
    134.4 KB · Views: 245
  • L+R_EQ_Dist.jpg
    L+R_EQ_Dist.jpg
    149 KB · Views: 253
  • L+R_EQ_Imp.jpg
    L+R_EQ_Imp.jpg
    103 KB · Views: 245
  • L+R_EQ_WF.jpg
    L+R_EQ_WF.jpg
    234.6 KB · Views: 239
  • l+R_EQ_spect.jpg
    l+R_EQ_spect.jpg
    148.1 KB · Views: 243
Did you check impedance plots with a fine tooth comb and look for (small) wiggles?
Each and every wiggle could lead to frequency anomalies. But I do not expect many below 100 Hz due to wavelengths at those frequencies. More absorption could be beneficial though. Stuff it to get the impedance peak as far towards the low frequencies and stop the stuffing if it moves back up. Leave some room for the driver to breath and use something between fiberglass (most effective over a wide band of frequencies) and the driver. Wool for instance, as it is excellent at absorbing higher frequencies.
 
No I havent that yet but will when I find the time :rolleyes: So far all focus have been to get them up and operational, now the work start to iron out as much as possible. As far at stuffing goes I have only 10g for each chamber = about 5g/l vs 12g/l for mk1.

Here is a non-EQ 10-500hz sweep, with 50 and 500ms IR gate, the dip at 90hz indicate problem with the room (depth at 3,8m):
 

Attachments

  • LS2019_500ms_10-200hz.jpg
    LS2019_500ms_10-200hz.jpg
    61.9 KB · Views: 97
  • LS2019_50ms.jpg
    LS2019_50ms.jpg
    69.6 KB · Views: 86
Last edited:
No I havent that yet but will when I find the time :rolleyes: So far all focus have been to get them up and operational, now the work start to iron out as much as possible. As far at stuffing goes I have only 10g for each chamber = about 5g/l vs 12g/l for mk1.

Here is a non-EQ 10-500hz sweep, with 50 and 500ms IR gate, the dip at 90hz indicate problem with the room (depth at 3,8m):

Is there a difference between left and right? If so, you might be able to use that.
 
Measurements may be void :trash:, I did an impendance measurement yesterday for Right (R) line, each individual speaker / series and whole group of 12 and found that one whole group of 12 was much higher in ohms than the others. Found a disconnected inter-driver wire.

It may have gotten disconnected during the individual driver measurement, because that is what I did first - but cant say for certain :(

@wesayso:
No, each chamber is stuffed equal between L & R so I cant use that as a comparison - good thinking but I didnt think of that when I closed #2 up
Impedance wiggles, REW specify that the resistance of the measuring circuit + resistor need to be a precise number. I dont have the ability to measure withing 0.1% but how detrimental to the information is a value with larger %-error?

Say if my test circuit have a 102ohm resistor and 0.4ohm circuit (102,4ohm total) and I enter 104ohm in REW - I assume wiggles etc will still appear, but the amplitude may be incorrect? Or may the "hill" be skewed?

@haraldo:
May have bad data - that would be nice in that regard :D
 
Must be bad data, I have no idea how you could possibly have any comb filtering at such a low frequency with your driver setup :D

Comb filtering at that freq. is physically not possible ;)
I tested again last nite briefly and with a shorter timewindow that dip is gone so its primarely the room itself.

If weather over the holidays allows it perhaps I'll be able to take them out in the garden and measure outdoors.

In terms of freq linearity my first job is to make sure all 24 drivers have a good and matching impedance response, this also means finalizing the internal damping of both lines :D
 
Comb filtering at that freq. is physically not possible ;)
I tested again last nite briefly and with a shorter timewindow that dip is gone so its primarely the room itself.

If weather over the holidays allows it perhaps I'll be able to take them out in the garden and measure outdoors.

In terms of freq linearity my first job is to make sure all 24 drivers have a good and matching impedance response, this also means finalizing the internal damping of both lines :D

Guess you may have to measure each individual driver then to make sure stuffing in cabinet is consistent across the line :D
 
Guess you may have to measure each individual driver then to make sure stuffing in cabinet is consistent across the line :D

Exactly, so it will take a little while :scared: :wiz:
  • Spend time with single driver, finalize the damping configuration
  • Implement same damping profile for the remaining drivers
  • Measure and tweak if wiggles are present + verify integrity of each volume (for instance leaks)
 
Last edited:
...In terms of freq linearity my first job is to make sure all 24 drivers have a good and matching impedance response, this also means finalizing the internal damping of both lines :D

:)... agree wesayso don't save on stuffing and as flat as possible resonance, know i told before and shared wesayso's zma file as a good target curve but pretty sure you will never regret in its directly related into improved analog like resolution how a keynote and its related harmonics line up in time domain relative to the original keynote of a recorded live sound, in lack of time to take a real study including documentation sorry i haven't any objective data proof on subject other than what ears told at wesayso's living room about real analog like resolution plus my own fast system and headphones run DSP correction for same time distortion problem. Know many is in love with keep it simple and what that sounds like but for this case we are not building a musical instrument where some varius resonance here and there can be a advantage to form a unique sound stage presentation, and really not shure arrays can be overstuffed but to be safe one can keep a eye on if frq response gets weird compared to datasheet and also give ears some little run time to accept or reconize the bit cleaner sound change it is when some distortion had been cured compared to what we used to, off course its problematic if some cured distortion here and there leads to a lifeless or cold clinic sound stage but instead of revert back suggest one hunt down what units or settings is responsible the unwanted soundstage.
 
Last edited:
Just out of curiosity, was making sections of say 6 drivers sharing the same box considered instead of a separate box for each driver?

No, not for this build. I wanted individual chambers due to chosen construction method. If I'd have gone with shared volume I'd give one volume per 3 drivers (one volume for each chain, series, of drivers)
See image for reference, this is the back side - since sealed off before completion.

:)... agree wesayso don't save on stuffing and as flat as possible resonance,......and really not shure arrays can be overstuffed but to be safe one can keep a eye on if frq response gets weird compared to datasheet and also give ears some little run time .

Yes, amount of damping is low right now, will compare to published data for verification - good tip :)
I assume when experimenting with single driver damping I need to measure nearfield (1cm away) for valid data?
 

Attachments

  • 74455028_539361996632115_1214425595199356928_n.jpg
    74455028_539361996632115_1214425595199356928_n.jpg
    142.4 KB · Views: 182
Last edited:
One of the things we used to say in R&D was that the worst thing that could happen is that when you first turned on a prototype, everything seemed to work. That made it easy to overlook some detail that wasn't quite right. But you aren't doing that!

I don't see a lot wrong with the initial measurements you posted, certainly not above 320 Hz. Below that it looks like room interaction, except I don't see compelling evidence for that in the spectrograms, but you don't have your spectrograms configured like Wesayso and I do to show the room better - swap the axes and zoom in on the first 100 ms or so.

THD down about 50db down above 400 Hz is really good. Interesting that it starts to rise so soon below that, given what we said earlier in W's thread about where to bring in the subs. But THD is down in the noise and that may be affecting the accuracy, making it look worse than it is. Is that 80 db calibrated?
 
I don't see a lot wrong with the initial measurements you posted, certainly not above 320 Hz. Below that it looks like room interaction, except I don't see compelling evidence for that in the spectrograms, but you don't have your spectrograms configured like Wesayso and I do to show the room better - swap the axes and zoom in on the first 100 ms or so.

THD down about 50db down above 400 Hz is really good. Interesting that it starts to rise so soon below that, given what we said earlier in W's thread about where to bring in the subs. But THD is down in the noise and that may be affecting the accuracy, making it look worse than it is. Is that 80 db calibrated?

Do you have a quick reference for a spectogram view sharing configuration you mention? I did a quick search but didnt locate one ;)

I have not calibrated the measuring output using a 2nd device, but could check with my iPhone/dB Checker app.

During the coming weekend I´ll try to do more sweeps, so lets see what I can get out of that.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.