Full Range TC9 Line Array CNC Cabinet

I found two tracks to be very useful in testing the low frequency capabilities.

One is "Hold Back the River" by James Bay.

https://www.youtube.com/watch?v=mqiH0ZSkM9I

I really like this song, the acoustic space is quite appealing and this sounds amazing with LX521. The double bass has some really difficult transients, this is the track I had the most trouble with clipping the DSP.

The other is "Free" by Rudimental feat Emeli Sande.

https://www.youtube.com/watch?v=KDPW_g2AhAU

This track has a significant amount of very low frequency energy in a synthetic base line. It really makes the dipole woofers move!

This is the spectrum from Audacity

View attachment 598756

Wesayso if you are reading this have you tried either of these two tracks on your arrays? If you have I would be interested to know how they handled them.

Found the time to listen. Love the James Bay track. Going to listen again. Very nice imaging and clear sound.

The other did less for me, though it is a workout for the bottom end it isn't reeling me in. A Perfect Circle with Lullaby is a bigger workout.
This one reminds me of the synthetic wobbling low lines in some Lorde tracks. I like those better.

I've send you a pair back, let me know what you think.

I use JRiver's analysing on all tracks, and use volume levelling to keep me from having low frequency accidents :). I try and balance to use as much room as I can without running into digital clipping. Keeping just a little headroom at all times. Volume control is capped to not be able to (over)boost volume with regards to the amount of EQ I use.
 
Last edited:
Found the time to listen. Love the James Bay track. Going to listen again. Very nice imaging and clear sound.

The other did less for me, though it is a workout for the bottom end it isn't reeling me in. A Perfect Circle with Lullaby is a bigger workout.
This one reminds me of the synthetic wobbling low lines in some Lorde tracks. I like those better.

I've send you a pair back, let me know what you think.

I use JRiver's analysing on all tracks, and use volume levelling to keep me from having low frequency accidents :). I try and balance to use as much room as I can without running into digital clipping. Keeping just a little headroom at all times. Volume control is capped to not be able to (over)boost volume with regards to the amount of EQ I use.

Glad you had time to have a listen, I actually like the Rudimental track but I primarily use it as a low end test. The dipoles probably suffer more as they have no cabinet rolloff and the boost goes all the way down.

I really like Get Lucky by Daft Punk especially Nile Rogers guitar parts. The remix and the album track have the same bass line which gets the woofers moving too.

I have listened to Lullaby on youtube but I don't hear a bass line at all so maybe that version is no good or it's so low my laptop speakers ignore it! Maybe you could point me in the right direction to a version that has it.

The James bay track and The Chain by Fleetwood Mac are two tracks that just sound magic on the LX521.

Getting Jriver setup on my media computer is on my list of things to do but it always seems to get bumped. The volume levelling is something that I think is very worthwhile. Once the arrays are closer to completion I will need to just get on with it.
 
Hi fluid that is very interesting hear a drummers views and some good points there regarding hi-hat, have a thought we never going to hear symbols be 100% realistic because of real speakers HF roll of makes squarer waves transients fade out above 4-6kHz area and diamond tweeters won't move much there in it will need far bigger steps to improve square waves up to say 20kHz area. My DSP run at 192kHz so can load a Rephase filter than turn phase up there and while i think it helps a bit isolated focus on symbols realism it ruins overall rhythm musicality for recording, compared to fully mic'ed live concert reproduce drums at home is good enough it gets harder to reproduce the non mic'ed drums. Heard new high resolution remastering "Led Zeppelin III" "Since I've been loving you" on weasayso's system, and think of that as being better one when focus on drum/symbols department.

Cymbals are another area where heavy processing tends to be used, most have the low end completely cut and then boosted around 10K. Real cymbals are a lot more raw and less shimmery. The volume of a band tends to drown it out though.

Getting a good phase response is definitely on the list of things to achieve with the processing of the arrays to see if I can appreciate the difference!
 
Almost no progress made here other than having to move all the parts from one room to another for construction to continue at some point.

Screws have arrived today, so I will be able to mount the drivers when I have the cabinets finished!
attachment.php
 

Attachments

  • IMG_0276.jpg
    IMG_0276.jpg
    414.7 KB · Views: 1,518
Due to a chronic lack of sleep and 40+ degree Celsius temperatures progress has been slower than normal but there has been some.

The rubber has a slight curl due to being rolled up for a number of years, unrolling and applying weight has made it better but there is still too much for it to just be glued down.

Staples seem to hold it quite nicely and should stop the curl from pulling the glue away over time.

IMG_2329.jpg

I made a few test runs to find the best way to mount the threaded nuts in the MDF as these will be holding the whole thing up so I don't want them stripping.

The two on the right I used an 8mm pilot hole and countersunk the edge as the wings on the nut were lifting the MDF at the surface.

The one on the left has an 8mm pilot hole with a 9mm one for the first few mm. This gives enough clearance for the body of the nut without lifting the top layers.

I have used superglue to harden the threads after the nut has cut them. This has worked better than I expected and the nuts go back in very tight and feel a lot more secure. You can see the one on the left is still wet and the two on the right have a white look to them where the thread has recut.

IMG_2330.jpg

I have got my Dowelmax out to cut dowel holes to assist with alignment and gluing. It is setup for 5/8" material so as long as I am careful about matching the faces together the slight difference doesn't matter.

IMG_2331.jpg

I dropped it and was worried that it wasn't perfectly true anymore. Glad to say that it is still accurate. In this picture the block was flipped end to end and there are two marks from a transfer punch in between the pencil marks, pretty hard to pick them apart.

IMG_2333.jpg

The CNC cut parts are very accurate and completely square which is important for the jig to work at it's best.

IMG_2334.jpg

My test pieces have gone together quite well but there is a slight gap which gets longer over the piece.

IMG_2335.jpg

IMG_2336.jpg

These test pieces are not accurately cut enough to work without this error creeping in.

Square at the corner but not all the way along.

IMG_2338.jpg

IMG_2337.jpg
 
I found an interesting article from Stereophile discussing measurements and methods to increase the reflection free period available making the measurement more room independent and to a lower frequency.

Time Dilation, Part 2 Page 3 | Stereophile.com

The technique filters the impulse before any gating which reduces it's length and allows a longer gate to be set, then an inverse filter after the gate is used to restore the frequency response.

The method used in the Dolby Labs paper seems to be something that could be achieved in REW with existing measurements. The KEF version would need to be in the measurement chain and seems less useful in practice.

The graphs seem to show that it can be quite effective. I think it is something that I will try when I get to measuring.

Maybe someone else has already tried it, or could try it on an existing measurement to see what happens?
 
The rubber has a slight curl due to being rolled up for a number of years, unrolling and applying weight has made it better but there is still too much for it to just be glued down.

Staples seem to hold it quite nicely and should stop the curl from pulling the glue away over time.

What kind of glue are you using? Will it really not hold if you allow the glue to cure? Must be a really mean curl! I can't remember... are you using rubber and another damping layer too? In considering my laminate of MVL and the foil lined dynamat style material, perhaps I'll try stapling the MVL to the dynamat and then using the pre-applied adhesive of the foil backed mat to stick to the baffle. Or just find a more aggressive glue!
 
What kind of glue are you using? Will it really not hold if you allow the glue to cure? Must be a really mean curl! I can't remember... are you using rubber and another damping layer too? In considering my laminate of MVL and the foil lined dynamat style material, perhaps I'll try stapling the MVL to the dynamat and then using the pre-applied adhesive of the foil backed mat to stick to the baffle. Or just find a more aggressive glue!

I've got Selley's 3 in 1 Silicone adhesive / sealer. It would likely hold it down but for how long is the question. I don't want to find them floating in the cabinet in a years time and have to take it all apart again.

As the rubber is curled the material wants to move that way. It will be easier to staple it down as that will hold it to the MDF while the glue dries and relieve some of the tension in the material. As you can see in the picture the staples really help to hold it nice and flat.

I am only using the one layer of rubber and it is more for mass and not really intended to be a constrained layer. With the silicone glue layer and limited direct connection from a few staples I hope it will still have a small lossy effect.

I would probably use a few staples to assist the MLV or Dynamat as it is quite heavy for it's size and vertically mounting it will put more strain on the pressure glue. Apart from the foil MLV and Dynamat are pretty much the same thing why a lamination of the two?
 
Combining MLV and Dynamat will do nothing but add weight. To create a CLD you need two stiff materials on the outside connected with a sheer (lossy) layer. Butyl or green glue can act as the sheer layer, but you'd need two stiff materials as a sandwich around it.
Dynamat works with it's foil layer for ~1 mm sheet metal. That's it's intended target to act as a CLD construction. It will add mass and lower resonances if added to a thicker material, but theoretically not be a real CLD anymore.

I have butyl rope between two layers of aluminium, that's a CLD construction. Green glue between a birch ply and MDF panel would act as a CLD too. (or between two pieces of Ply or MDF for that matter)

MLV is used to add weight, or it can be used as a sound blocker by decoupling it from the outer shell by putting a layer of neoprene between them as a decoupler. Adding the weight will still help reduce sound traveling trough it, but when you decouple the MLV from it's "wall" (with neoprene or sorbothane) the loss will potentially be greater. MLV cannot be seen as a sheer layer in a CLD construction i.m.h.o.

Mixing several materials will always do something, but it might not do what you want or expect. Read the quote again, I might have used different words, but the idea is still the same. In this sentence: "Neoprene (and maybe even better, sorbothane) is used as an isolator." I should have said decoupler instead of isolator to make myself more clear.

MLV is often used as a sound barrier. In cars it's often used with neoprene underneath to block (outside) sound.
Neoprene (and maybe even better, sorbothane) is used as an isolator.
Butyl damping mats with alu foil as a (CLD) damping material.

I used MLV on my baffle to weigh them down, make them heavier. It's not a flexible material and it doesn't work as a CLD layer between two stiff materials. Butyl would be way better at that. I've tried MLV and neoprene between both baffles I use. I stripped it all out again and replaced it with butyl rope to gain damping. A measurable improvement.

Look at MLV as being effective as a sound barrier. Or a means to add weight. If you compress it, it won't return to its original shape. It works even better if isolated (for instance with neoprene) from the "wall" it's used on.
 
Last edited:
I've glanced at it and think I understand what it is they do. I'd rather take multiple measurements at slightly different spots and averaging them to isolate the speaker from the room. And battle any reflection found as much as one can with old fashioned means.

I still prefer frequency dependent windowing as a means of gating. It's still gating, but a different gate at each frequency and gives a more clear view of the first wave front.

We want to correct the speaker and adjust to the room. Those are two different things. We cannot loose the room completely (at least not all of us) but can diminish it's influence.
 
Looking forward to the rest of your build and impressions Fluid! I found your thoughts on actual drum sets vs recordings very interesting. More and more I'm starting to believe that the ear/brain hearing mechanism is a waveform analyzer rather than a frequency analyzer. I think it's a possibility that the importance of a good impulse response (a very fast rise time is what I'm starting to believe is very important) is overlooked, and thankfully the TC9's look excellent for this. I came across a video today that the ear has a time resolution of 7 microseconds! I really hope the array gives you an impression of more realistic hi-hats :).
 
Why do you want the fiberglass on instead of paper?

The only upside is a 13Hz lower Fs.
Negatives:
Lower power handling
Lower Efficiency
Lower Xmax

TC9FD18-08 3-1/2" is still where it's at.

David

Anybody interested in a group buy of the ultimate Vifa/Peerless/Thympany driver TG9FD10-08 .

Price around $6 but there has to be many interested as minimum number of drivers for that price is >2000 pcs.
 
Why do you want the fiberglass on instead of paper?

The only upside is a 13Hz lower Fs.
Negatives:
Lower power handling
Lower Efficiency
Lower Xmax

TC9FD18-08 3-1/2" is still where it's at.

David

Those data seems to differ a bit between the various datasheet versions released over the years, have datasheets from 2009 where they branded VIFA and while power handling for both models was at 30W, TC9 had only 1,50mm Xmax where TG9 had 2,35mm.
Attached below is new version where they branded Peerless and Xmax is now 2,6mm for both models, efficiency is closer than ever, but TG9 have 5W lower power handling.
 

Attachments

  • TC.PNG
    TC.PNG
    275.3 KB · Views: 715
  • TG.PNG
    TG.PNG
    253.8 KB · Views: 202
I think I have at least 4 different versions of that datasheet of the TC9. :)
I chose paper, but I don't think the differences would be big after DSP FIR correction.

The power handling is a hard part to interpret. I've seen many numbers fly around. In these arrays they can handle a lot of power though! But x-max will be a limiting factor at the extreme low end. The different cones can make a difference in power needed. For instance an array of Scan Speak 10F's could produce the same curves as the TC9 array with only 70 watt, while the TC9 array would need 240 watt.
 
Last edited: