January 29, 2012

Rythmik dual opposed tapped horn sub - sims

Black: simulated
Magenta: measured sealed sub (right corner)
Orange: measured sealed sub (rear corner)
Grey: measured sealed sub (left corner)

January 26, 2012

80L CSS SDX15 sub - part 4

Edges trimmed:

Edges were trimmed with the hand plane. You might notice the gouges where I've used the plane. Unfortunately I was in a bit of a hurry at the time and didn't check the blade. On closer inspection later I found that it had been damaged and needed grinding back. The gouges aren't a serious concern because they will soon be covered with bog anyway. 

Filling holes:

Bog filler applied over butt joins, then sanded:

Next up, I applied a roundover to the edges, 3mm on the bottom edge, 6mm to all others. 

I also cut a groove at the join. The idea here is to conceal the join so that if any cracks appear over time, they will be hidden in the groove.

Clamping the feet. Fixed with polyurethane glue, chosen for its very high strength. It is four times as strong as ordinary PVA and expands when curing to ensure a good bond. The oozing glue was wiped over the edges to seal them up.

Preparing for painting over MDF butt joins

Have you ever seen an MDF box where paint has been applied over butt joins? In virtually every case where great care isn't taken, a hairline crack appears. It is caused by the ends of the MDF soaking up paint and then expanding more than the rest of the material. How am I handling this?

Firstly, I've cut a groove at the join so that any cracks will be concealed if they appear. Secondly I have buttered up all the edges with two part bog. This will tend to avoid the edges soaking up paint. I've also applied polyurethane glue over the edges as a second layer. The plan is to spray the box with automotive paint with a satin black finish. 

January 21, 2012

Rythmik dual opposed tapped horn sub

I have always been happy with the sound of my Rythmik subs. I have a pair of sealed 12" servo kits that have been in 70L sealed boxes. They are equal to the best I have heard. I achieve 15 Hz extension in room without any more EQ than is included in the servo system. I have found that I can find their limits thermally with music and have managed to drive them to thermal shutdown twice (since 2006). This is a limitation inherent in nearly all subs. Low efficiency combined with amps intended for home theatre with a lighter duty cycle results in subs that can overheat with music turned up.

I'm very reluctant to let these go so my solution is to start with active woofers running down to 40 Hz. I have 18" pro woofers driven by a pro amp - no chance of running out of steam. The Rythmik subs will then simply add extension and carry a lighter thermal load. Tapped horns will make the box carry more of the load and increase the sensitivity so that instead of around 88 dB from a pair I get around 94 dB.

Here is the concept:

These are around 400L and yes that is big! Boxes will sit under the floor and be attached to it, hence the need for dual opposed mounting. The box is around 400 x 600 x 2250mm. 

1w1m simulated:

94 dB with some sacrifice of anechoic extension to achieve higher efficiency. I need only to get down to 25 Hz and room gain gets me down to 15 Hz with a sealed box. I'm aiming for sensitivity and headroom over extreme extension. 

Max SPL as limited by the amp:


The excursion peaks at 31 Hz (16.7mm) - xmax is 18mm. 

How to avoid over-excursion

The driver will bottom as with a vented enclosure if not carefully protected. The trick is to add a high pass. The amp includes a first order HP which won't be enough when pushed. I will add a 2nd order HP, probably around 20 Hz. I can then test with tones, using a 30 Hz sine wave and noting excursion. Then applying a 10 Hz sine wave with a check on excursion. The excursion below the cut-off should never be able to exceed that above. When that is achieved, the output will be limited only by amplifier clipping and I will be sure that xmax will not be exceeded. This means when unexpected heavy bass hits in a movie, I won't have to worry about damage to subs. I will have set up the system so that the drivers are protected from all conditions. 

Achieving headroom

I don't want to create the impression that Rythmik subs are lacking output. However, asking 12" subs to match the headroom of a high sensitivity horn system isn't realistic. To get down to 20 Hz it is necessary to move a lot of air. This could mean one giant super hero woofer with a very powerful amp, but I prefer a number of more modest subs. So for those who want SQ loud, a number of Rythmik subs is a good way to go. I may in fact build two of these, one in each corner.

80L CSS SDX15 sub - continued

Cut sheets:

Circle jig for the router:

A piece of plywood forms the jig. Since this isn't a plunge router, it's useful to have a flexible material.

Cut outs for bracing:

Double layer thick baffle:


Next >

January 20, 2012

80L CSS SDX15 sub

Driver: Creative Sound Solutions SDX15 15" sub driver
Box: Sealed 80L

This is a basic matrix-braced sealed box with a down-firing orientation.

It's good to have a cut plan so that you can cut efficiently. I like to set the table saw up to cut to a width, then do all the cuts of a given width. Otherwise you end up setting it to 450mm, then 414, then 450 then 564 then 414 and so on. It takes time to set it up accurately and you will never get it quite the same twice. 

Pieces cut:

January 16, 2012

S2 point source horn vertical directivity measurements

Previously I measured horizontal directivity but ran out of time to measure vertically.

Horizontal measurements >

Vertical measurements are much more challenging in this case. On my measurement rig, it means rotating the horn 90 degrees so that it is now tall instead of wide. This allows me to use the angle markings, but I found that without assistance the horn was unstable and my quickly rigged up struts were not up to the task. I would need a person to hold it in place carefully, or spend a lot more time making sure it stays up there. You will get some idea what I'm talking about if you see the setup for horizontal.

In the end I settled on a different setup where the horn was placed on the ground with thick foam on the first reflection point. So the horn was placed on the ground firing horizontally, and then the horn itself was rotated to the angles required, with the mic in place. I printed out a guide for the angles. I also placed foam around reflection points. This is certainly not ideal and the results show obvious problems, but still give some idea of what is going on.

Measurement steps are 10 degrees. So the first line (black) is on axis, the next is 10 degrees, then 20, all the way up to 90 degrees. 45 degrees is shown in black also. So the first 3 lines are within the defined 40 degree coverage angle. All other angles are what will be seen by the floor and ceiling. 

The peaks and dips are related to reflections - the horn itself is much smoother than shown. Yet you can see observe a trend in dispersion. Notice how the lines are bunched together at the bottom? This means it is an omnidirectional source down there. However, above about 260 Hz the dispersion does start to narrow slightly, shown by the lines becoming further apart. Overlooking the peaks and dips, the overall trend is a gradual broadening of dispersion up to about 1k. 

What does this mean?

Shown here is the vertical dispersion of this 90 x 40 degree horn. What this shows is that you want to arrange the horn it so that you are sitting with your ears within that 40 degree vertical window where the response is consistent. The floor and ceiling see the other angles. 

Horizontally, S2 controls directivity down low, but vertically the smaller mouth dimension means that this can't be maintained. So the ceiling will see more lower midrange than the side walls. I'd be inclined to leave side wall reflection points along (no treatment) but consider ceiling treatment that might absorb more lower midrange, like say an RGP BAD panel.

Fury's JBL waveguide speakers

Last Saturday we had a measuring session for fury's (Danny) new DIY active speakers, shown above. 

Rack includes a Tascam CD player, Behringer Ultramatch and two Behringer DCX crossover units, as well as an Altronics multi channel volume control below, a power conditioner/sequencer and three beefy Yamaha pro amps. 

We measured the speakers outdoors on my rotating measurement rig.

You can see that it looks very much like a Geddes speaker. It features a BMS 4550 compression driver in an Autosound Oblate Spheroidal waveguide and reticulated foam that was cut by me. I have posted about this previously (a search will show previous posts). The midwoofer is a 12" JBL 2206H.

BMS 4550 compression driver + Oblate Spheroidal waveguide

The waveguide provides good control of dispersion even down as low as 600 Hz, below which point pattern control is quickly lost. There is some freedom here to cross low if desired. Above 15k the pattern narrows slightly.

JBL 2206H

This measurement was perhaps a little surprising in that the woofer measurements tended to indicate there was some flexibility with crossover point as well. 

Measurement laptop + DCX + power amp + audio interface.

Crossover measurements:

Black - response 22 degrees off axis
Red - axial response (note the diffraction dip around 5k)
Blue - the result of letting auto align mess things up

After some experimenting and measuring, a crossover at 900 Hz was chosen. We took some extra measurements to confirm that the filters were set correctly and ran auto align only to find that in this case it might be called "auto mis-align!" We used the EQ prediction capability of REW to work out the settings to EQ the top end, resulting in the flat response shown above. The axial response must not be used because there is a diffraction dip.

We ran out of time and didn't get a chance to listen with the new crossover settings, so we heard them without CD EQ applied. The midrange was clean and controlled and I could not help but feel that it was partly due to the 50mm thick cabinets as well as very good drivers.

January 12, 2012

S2 update

More filling and sanding with bog. The goal is to get a smooth transition, very much like a JBL progressive transition waveguide or an oblate spheroidal waveguide applied to a different cross section. 

Baffles now glued on, more filling and sanding ...

S2 in a corner on top of a 60L sealed bass box for my 18" driver.