Anyone made a waveguide for the TC9 (or 10F)?

[onni]

I have seen some people try the TC9 in horns (xrk in the tractrix synergy, weltersys on maltese horn), but how about in a waveguide?

Let's say I want to build a FAST setup with (preferably close to constant) directivity to below 1000 Hz, then a waveguide on a 3-4" fullrange driver should be a good start, no?

/Anton

 

[xrk971]

Please design one for us! 😀 The trick is to make the edge transition from driver bezel/surround to the WG so smooth that it doesn't cause the almost-certain-to-have dips at the driver-radius to WG wall half-wave frequency. The Bookshelf point source horn thread has some nice tricks that really helped the SB65WBAC-25 driver to put out a smooth output.

I would like to have a WG that is maybe 75mm (bezel to baffle) deep to allow me to get perfect time alignment when used with a woofer in a FAST with 1st order XO.

Note that TC9FD and 10F/84xx have the same bolt circle radius for mounting (shows they have same Danish DNA). The TC9 has a square bezel and 10F a round bezel. The WG itself would be identical, just the shape of the bezel frame is different. TC9 has bonus that it coems with a gasket foam on the front face already attached.

Now, I really need to fix the busted 3d printer... 🙂

 

[RajkoM]

Subscribe

 

[onni]

Since a 3.5" fullrange driver differs quite significantly from other things you usually put in a waveguide I'm guessing that there needs to be a lot of trial and error with prototypes to get rid of peaks and dips. Or, as my approach usually is, trying to simulate the response.

I started with measuring the TC9 with calipers and making a FEM model. First a sim of it being mounted (countersunk) on an infinite baffle:

Compared to data in datasheet:

An externally hosted image should be here but it was not working when we last tested it.

First cone resonance at 9 kHz I'm guessing when comparing to simulated response (pistonic).
Here is Weltersys polars for the TC9:

The directional behavior is overall correctly predicted by the simulation.

Here is SPL for the first 1 m at 1 kHz:

Monopole behavior.

At 4 kHz:

Starting to beam slightly.

At 8 kHz:

More beaming and slight lobing.

At 16 kHz:

Beaming and lobing. Here is a close-up:

You can se the shape of the cone.

I'm assuming pistonic motion (constant acceleration, linearly declining over surround).

I've started iterating a little to see what kind of waveguide could suit this driver. A question: is the cone of the 10F identical in shape? I'm guessing it is better suited for loading in a waveguide since the cone should be significantly stiffer with the added fiberglass.

/Anton

 

[xrk971]

Visually, the TC9 and 10F cone profile are very similar. Rounded dust cap in middle. Profile of cone is mostly conical, might be a bit curved. I can take some caliper measurements later when I get back to the lab. All you can do is assume pistonic for first cut. Make throat same dia as middle of surround roll and allow a smooth curved undercut below surround to bezel edge, Probably put dampening felt in that crevice. Here is the profile Bushemeister came up with after much trial-and-error:

http://www.diyaudio.com/forums/multi-way/285030-bookshelf-multi-source-horn-125.html#post4644522

 

[onni]

Please design one for us!

That's the plan! 😉

Visually, the TC9 and 10F cone profile are very similar. Rounded dust cap in middle. Profile of cone is mostly conical, might be a bit curved. I can take some caliper measurements later when I get back to the lab. All you can do is assume pistonic for first cut. Make throat same dia as middle of surround roll and allow a smooth curved undercut below surround to bezel edge, Probably put dampening felt in that crevice. Here is the profile Bushemeister came up with after much trial-and-error:

http://www.diyaudio.com/forums/multi-way/285030-bookshelf-multi-source-horn-125.html#post4644522

Thanks for the guidance, I've read that earlier as well.

Let's start with a few iterations (all 7.5 cm deep):

Iteration 1 - How about a simple conical that fits around the surround
Easiest to make. Here a 90 degree WG:

Response:

Quite decent 30-60 deg, but not so much 0-15 deg.

How about a 120 degree WG:

Response:

Nah, not better.

Iteration 2 - Conical above the surround
Alright, so let's follow xrks advice about starting the WG at top of surround (with a distance of 3.5 mm for excursion). The throat is 70 mm diameter.

Response:

Not really better than the 90 degree above.

Iteration 3 - Added initial roundover

Response:

That's a little better, decreased the on-axis dip 🙂 The radius is 2 cm.

/Anton

 

[xrk971]

Nice work Onni. I think a radius on the mouth edge will help as well. And instead of conical, make it expand a little faster on second half of expansion with a curved shape. Still for 75mm depth and similar angle of coverage. One way to approximate is apply a large radius round over tangent to angle of cone at 50% depth and tangent to baffle at exit. Use a single spline and tweak by hand. Then do swept revolution, like this:

 

[tmuikku]

Thanks for the effort, subscribed!🙂
ps. practical waveguide design tips/math by Rod Elliot Practical DIY Waveguides - Part 1

 

[onni]

Thanks for the effort, subscribed!🙂
ps. practical waveguide design tips/math by Rod Elliot Practical DIY Waveguides - Part 1

Interesting read! The method is however mainly applicable to dome tweeters, for fullrange drivers there needs to be a transition from cone to conical section. I'm guessing that it is difficult to calculate this analytically as the cone profile is quite complex, but iterating using software should be possible 🙂 I've got better results than shown already, I'll post them later.

/Anton

 

[tmuikku]

yeah Rod has some good articles that are somewhat scientific with literature references and all 🙂

Finnish pro audio company AuraAudio has speaker model F1 with interesting waveguide for inspiration
Aura Audio - Aura Audio F1 - F-Pro Esitystekniikka

Large picture http://www.auraaudio.fi/pics/F1.jpg

I'm guessing they have highly optimized waveguide for their application. It seems to have a bezel. And some weird wavy surface on the sides. Cool, must be computers crunching data to aid with design.

 

[wonderfulaudio]

Finnish pro audio company AuraAudio has speaker model F1 with interesting waveguide for inspiration

Array gives vertical control and wave-guide gives horizontal control.
Wouldn't it be easier to put a small line array of 3 or 4 drivers vertically and put straight walls on the sides ie a waveguide only horizontally?

 

[onni]

Array gives vertical control and wave-guide gives horizontal control.
Wouldn't it be easier to put a small line array of 3 or 4 drivers vertically and put straight walls on the sides ie a waveguide only horizontally?

Maybe the waveguide reduces lobing? I have seen that approach.

/Anton

 

[tmuikku]

sorry i didn't mean to point discussion to line array design. I just tried to give a hint for successfull waveguide design for FR driver since the Rod Elliot article I posted earlier was for dome tweeter which the tc9 is not.

 

[wonderfulaudio]

Just for reference, here is a TB 2" driver in a waveguide. Martinsson's Blog - 2" fullrange TB W2803SE on JBL 2386 upgrade.
This has back open, kinda dipole-ish, it may not work with a smaller wave-guide.

 

[xrk971]

If we are allowing big WG’s and 2.5in drivers here is SB65WB in LTH142 that I measured:

 

[onni]

If we are allowing big WG’s and 2.5in drivers here is SB65WB in LTH142 that I measured:

That is quite impressive and affordable! Is that without EQ? Might be suitable for Harsch XO that requires a large offset?

Here are some more results for the TC9:

Iteration 4
I tried going for the smoothest possible profile, so it starts around the surround and is a quadratic curve (2nd degree). There is a guide point 40 mm from the driver bezel and on the centerline.



Quite smooth! Not very constant directivity though. More like a tractrix maybe?

Iteration 5 - more conical
Let's try moving the guide point down slightly (10 mm) to get a more conical profile (and hopefully more constant directivity).



Now that's not so bad 🙂


Note: The downward tilt is exaggerated by assuming a pistonic motion with constant acceleration. As the waveguide causes the driver to couple better to the air it will cause the moving mass to increase and therefore the acceleration.

/Anton

 

[xrk971]

No, it is with PEQ applied. Although I recall Kees was able to get a super flat response on a tractrix and FR8S using a simple 10uF high pass cap. It was kind of unbelievably good and that is what I would aim for if I were to do it again and avoid all this DSP stuff. If you look at raw, it is a -6dB/oct fall off, so a high pass cap at +6dB/oct rise, if picked correctly, can null that.

Here is comparison of raw vs with PEQ:

Here is the PEQ applied:

Result with Phase:

More here:
A Bookshelf Multi-Way Point-Source Horn

 

[onni]

Ah, right. Would be interesting to see response with just a single capacitor (and maybe a parallel resistor to increase effectiveness of said capacitor at resonance).


Iteration 6
A slightly deeper WG (12 cm).




This is IMO decent. I might test this geometry... I'll try a few more iterations to see if I can get the on-axis response a little smoother (dip at 5.5 kHz), but off axis is already acceptable.

If I add a 50 mm fillet (roundover) at the mouth I get this:



Overall a little smoother, but no huge difference.

/Anton

 

[onni]

A few more figures for the last WG:

SPL at 1 kHz:

Slightly directional.

At 2 kHz:

Distincly directional.

At 4 kHz:


At 8 kHz:

Still no lobing.

At 16 kHz:

Lobing, but smoother than mounted on flat baffle (compare to post 4).

Animation of pressure distribution at 4 kHz:


And at 8 kHz:

Smooth.

At 16 kHz:

Lobing present (0 pressure along line).

EDIT: I can't get the animations to play :/

/Anton

 

[xrk971]

That’s looking really nice Onni.