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.
Todays efforts, baking time!

1660g EG, 125degF for several hours :)
 

Attachments

  • 68241201_2167392336716610_2173702022455361536_n.jpg
    68241201_2167392336716610_2173702022455361536_n.jpg
    540.2 KB · Views: 516
  • 68283080_457991565046631_7091038074242596864_n.jpg
    68283080_457991565046631_7091038074242596864_n.jpg
    520.9 KB · Views: 505
  • 67953922_662111217632780_7738389331940737024_n.jpg
    67953922_662111217632780_7738389331940737024_n.jpg
    414.4 KB · Views: 509
No pictures yet but the test cabinet i complete and playing :)

To-do list:
  • Make the diffuser
  • Create plugs/circuit for impedance testing, using REW
  • Get creative with the stuffing... :rolleyes:

I took the test cabinet over to my old LA´s and performed an informal knuckle test to them, there is no comparison :D
 
Okay, test rig completed and seem to be operational - this is no quality test performed just a quick scan to get baselines and data to learn from.
The cabinet is not fully sealed, neither at the driver side or rear wall.

Time to read up on how to "read" impedance plots, I mean they read resistance but how it all connects :confused::D
 

Attachments

  • Impedance.jpg
    Impedance.jpg
    54 KB · Views: 472
The Magenta trace, open back, is showing extra peaks at harmonics of the fundamental resonance.

That is how I interpret the plot. The two sealed plots dont show this cavity resonance. Stuffing the open back version to varying degrees may help damp this resonance out.

That would be a good test...

How much damping to eliminate those peaks?

Seeing as there are two magenta traces, its confusing to look at.

The Fb has dropped from Fs due to air mass loading the rear of the cone, but the column of air is resonant. Damping rear should help damp that out

Epic build by the looks of it - I've considered sand resin before and never had the nerve to try it!
 
Last edited:
Thanks for the comments Mondo :) I have never done this kind of testing and appreciate all input!

Okay, the driver have been sealed properly and same is the rear (removable) back wall.
I briefly tested with diffuser with/without other stuffing (interior lined with felt on all)

This should be considered the first "valid" set of sample data.

attachment.php
 

Attachments

  • Impedance_diff2.jpg
    Impedance_diff2.jpg
    65 KB · Views: 628
Last edited:
To really know/learn what's going on, don't be afraid to zoom in on the plots.
You're looking to get the least amount of wiggles from the peak value and above.

Open air looks nice and clean, now zoom in on the 150-2 Khz to see from 6 to ~16 Ohm.
The better damping will show the least amount of wiggles. That's what we're after.Every wiggle could possibly lead to a FR anomalie (dip/peak). Use the Vifa open air as the reference. The Max stuffing seems to look good, what stuffing? I'd expect a max stuffing to shift the peak to the left (lower in HZ) a bit. The itchy fiberglass does that quite well. While a layer of felt generally lowers the wiggles by quite a bit. That combination worked best for me. While on this scale everything looks fine, don't be afraid to zoom in a little closer. Like I said, felt on walls and itchy fiberglass (loose but kind of filled) in the open middle works wonders. Keep the back side of the driver free. I have a layer of felt there between back of driver and the fiberglass fill. About ~ 1.5 to 2 cm free air between back of driver and that felt layer.
 
Thank you Ronald, I will visit this aspect of the build several times before they are completed :D
Tested a little back and forth today and the diffusor dont seem to smooth out for me, it causes the peaks to be a little "loopsided". In terms of peak symmetry and lack of wiggles it seems to be a KISS=better scenario.
Interiour fully covered with felt and various degree of "itchy stuff" placed inside seemed to be the cleanest. As each driver will be a closed chamber I can later test each driver individually, so if I decide to go all-in I might opt to tune each one individually so they all are as close as possible. Its gonne be a FUN fall/winter :hphones::eek:

A couple of pictures from today test run forming the casting tubes:

The last one is especially enticing, resembelling a "top fuelled" compressor intake :D Can´t wait to stack 24 pieces of these up... :p
 

Attachments

  • 68756663_384099998930551_4963574720508198912_n.jpg
    68756663_384099998930551_4963574720508198912_n.jpg
    380.6 KB · Views: 220
  • 69118941_915071405508350_4175304812734185472_n.jpg
    69118941_915071405508350_4175304812734185472_n.jpg
    431.2 KB · Views: 188
  • 68949181_893352134357817_8122695944660582400_n.jpg
    68949181_893352134357817_8122695944660582400_n.jpg
    388.9 KB · Views: 166
Last edited:
Minor update, image wise not much progress to show for but:

  1. Currently tweaking the pipe shaping to get a uniform thickness of each section, I have 15 done, 9 to go.
  2. Bought hardwood flooring, which will be stacked horisontally and act as cabinet sides. 15mm thick vs MDF 19mm = close to a 10m slimmer cabinet.
  3. MDF sheet purchased so I can CNC the front baffles during september.

Test piece of the stacked floor tiles look good, so I'll do the remaining 3 in the next few days. Gross dimension is 340x2340mm which will allow final cut-down later.
End-result will hopefully be something similar to Sonus Faber Venere :worship:
 

Attachments

  • LA2019-FB.jpg
    LA2019-FB.jpg
    41.3 KB · Views: 225
  • 0_134110.jpg
    0_134110.jpg
    91.7 KB · Views: 136
  • images.jpg
    images.jpg
    3 KB · Views: 167
After being away most of the past couple of weeks I am back on it :)

Decided to redesign the baffle with shallower (larger radii waveguides) pockets for the drivers and add some angles to the outside edge vs the continous radius I had previously.

Yesterday I also finalized the three CNC-programs needed for the long, tiled carve. Top/Mid(x4)/Bottom section.
A jig and alignemnt pins will make sure the intersections are carved at correct length offsets and kept straight.

If I am lucky I may get the first done this weekend :)

Video of the CNC-simulations at high speed, real carve time is approx 30min per section (6 needed for each baffle):
YouTube
(First two motions are pecking operation for the alignment pins)


Image show front baffle rendering, I am leaning towards satin black or satin dark grey:
attachment.php
 

Attachments

  • LA2019-v4.jpg
    LA2019-v4.jpg
    53.4 KB · Views: 80
  • LA2019-v4-black.png
    LA2019-v4-black.png
    157.2 KB · Views: 580
BTW - while I have been away/preoccupied the slow boat have done a couple of deliveries :)

  • 30m red and 30m black 18AWG Silver coated wire
  • 200 gold plated speaker connectors, 2,8mm and 4,8mm wide
  • Cable stripper / crimping tool

Small items that amount to a lot of work when the wiring needs to be done :shutup:
 
Thanks Ronald :)

The router died due to a machine crash (bad code or glitch) going full depth in MDF (38mm) and smoke escaped the RPM governor inside the palm router. Its been returned under warranty and hopefully back in operation soon.

If not I will get a new one identical one as the mounting brackets are specific to the router (Makita RT0700CJ)

After a machine crash like this the mechanics of the CNC have been checked throroughly, those checked out okay so I am now just waiting the the router itself... :wrench:
 
Operational again :cheerful:

Yesterday I started a new baffle, hopefully it will come out good but I am anxious due to the mishaps which very quickly can ruin hours of work :Ohno:

Only had time to do the top 4 driver positions (1 carve with 3 tool operations)
On the machine bed I have attached a gude plate (second surface with rails) that ensure the MDF-stock stay parallell to the Y-axis of the machine.

Pilot and guide holes ensure I can do partial segments with equal spacing, simply lifting the stock-MDF up, slide to the right and align it with the guide hole. Clamp down and machine next segment. Like mentioned before one baffle consist of 6 segments, one top, 4 repeats of middle and one bottom.

1 down, 5 to go...!
(and repeat for 2nd baffle...if all goes well)

  1. First operation, doing the pilot and guide hole with 6mm flat tool. CC spacing is 4x84mm (20s)
  2. Second operation, machining the driver openings and mounting flange with 6mm flat tool (10min)
  3. Third operation, contouring the waveguide surface with 6mm ball tool (20min)
I could do finer contouring but that = more machine time. Earlier tests show that the visible contouring lines quickly dissapear with light sanding in preparation of finishing. I will seal the MDF with epoxy resin which will serve as the primer/base coat.
 

Attachments

  • LA-slider.jpg
    LA-slider.jpg
    33.4 KB · Views: 384
  • LA-top.jpg
    LA-top.jpg
    243.4 KB · Views: 396
  • LA-closeup.jpg
    LA-closeup.jpg
    142.8 KB · Views: 390
Last edited:
Partial progress, another issue to deal with :eek:

I ran a full baffle yesterday, with no surface contouring for the lower 5*4 sections to nail down my workflow.

My machine code however have a bug I need to deal with - most likely a motion command that is misinterpreted by the controller. See picture.
It happens for the 2nd driver on the right side on each section of 4. Since the error is relative small, and hidden by massive amounts of MDF-chips I didnt notice untill I test-fit the Vifa's.

Good thing I didnt do the surface contouring :D
 

Attachments

  • LA-feil.jpg
    LA-feil.jpg
    230.7 KB · Views: 383
Another picture highlighting the flaw, it occured for each of the 6 sections that I carved.

I also have a 2nd bug that happened between the two lowest driver openings.

Already preparing a new set of machining files that will be individually tested on some scrap MDF :D
 

Attachments

  • LA-feil2.jpg
    LA-feil2.jpg
    239.9 KB · Views: 382
  • LA-feil3.jpg
    LA-feil3.jpg
    433.6 KB · Views: 152
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.