The importance of chamfering mid-range holes

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Ok thanks, so to simplify:
- near the driver, large section, more stuffing;
- far from driver, small section, less stuffing
Right?
Stuffing wants to be light around the cone. Near the cone you ideally want the sound travelling away to be guided by a constant cross sectional area (constant impedance) and with no stuffing (constant impedance). Alternatively it could be as open as possible with a hard reflecting surface to direct sound down the tube and again without stuffing. For cone drivers with big magnets this is not straightforward.

When the stuffing starts it wants to build up gradually. This is why anechoic chambers have wedges on the wall so that a sound wave first encounters the small tips of the wedges which gradually get bigger as it approaches the wall. If it was simply a solid block of foam then the the step change in impedance at the front face would reflect some sound. The gradual build up reflects less.

I suspect a tube would take some thinking about to do well. A large chamber with light stuffing near the cone and a bit denser further away is likely to work just as well at the price of taking up more space.

A possible solution for the small diameter would be to insert a cone modeled felt into the tube leaving free internal tube surface and occupying the center axis, with the tip of cone near the driver and the base far. Would it work?
Perhaps. I suspect success may depend on just what the short wavelength sound is doing in an around the driver.
 
I used a cylindric tube to insulate and load mid-range.
The tube's diameter is similar to membrane's diameter (about 7.5 cm) and its lenght is about 25-30 cm: what must I expect from this solution in terms of response..?

Damped TLs are my preferred mids cab loading and unless you plan to pack it with wool or other dense material to basically make it a leaky sealed back, then the tube needs to be tuned to the driver, so unless the driver's Fs is somewhere in the ~350-425 Hz range with a low Qts you may need to start over with a [much] larger, longer tube if heavy packing doesn't work well enough and of course need to pack it in a netting to keep it from getting into the driver.

GM
 
To my knowlegde, standing waves just has to fit - wavelength = free unupstructed dimension inside of cabinet.

They are 1/2 WL resonators, so wherever there's parallel surfaces and something to excite the air within its range there will be eigenmodes [standing waves].

Until cabs get quite large by this forum's standards, lining one wall, top, back with a modest amount of 1" acoustic fiberglass insulation is usually sufficient to quell them.

GM
 
Damped TLs are my preferred mids cab loading and unless you plan to pack it with wool or other dense material to basically make it a leaky sealed back, then the tube needs to be tuned to the driver, so unless the driver's Fs is somewhere in the ~350-425 Hz range with a low Qts you may need to start over with a [much] larger, longer tube if heavy packing doesn't work well enough and of course need to pack it in a netting to keep it from getting into the driver.

GM

Thank you. My mid-range is in reality the well-known full range VIFA TC9FD18, which has FS 125 Hz and Qts 0.89, but used as midrange in a three way system, filtered to cross at 300 and 3500 hz. My aim is to mount it out of box in a baffle-less configuration, and a TL is a really fascinating idea for me, but I haven't the necessary expertise to project and verify it, so it would be enough for me to use a closed box.
So I have just to avoid that problems / issues / waves-war that occurs in the inner part of the tube affect back driver's membrane and come out through it.
I'll try to slightly chamfer the near-driver hole to avoid forming a narrow duct with surfaces parallel to magnet's ones, and use a cylindrical felt to cover internal tube's surfaces and a truncated-cone felt placed in tube's axis forming a decreasing size duct going away from driver, this last even filled with poly.

TQWT_P830869_rev03_36L_38Hz#-Model.png
(in red "Geberit Silent Pro" thick plastic tube, in yellow felt and filling, in gray wood around hole)

Using lenght rather than diameter will I be able to fight internal reflections?
 
I'll try to slightly chamfer the near-driver hole to avoid forming a narrow duct with surfaces parallel to magnet's ones, and use a cylindrical felt to cover internal tube's surfaces and a truncated-cone felt placed in tube's axis forming a decreasing size duct going away from driver, this last even filled with poly.

View attachment 769253
(in red "Geberit Silent Pro" thick plastic tube, in yellow felt and filling, in gray wood around hole)

Using lenght rather than diameter will I be able to fight internal reflections?

You're welcome!

As a general rule, basic TLs need 4-10x Vas based on its Vas, Qts' specs* regardless of length.

Anyway according to HR, for your app: a tube i.d. area ~ = driver Sd sims fine for a ~300 Hz XO IF ~109 cm long. Try packing the area around the driver with open cell foam since felt may be too dense to deal with near reflections and pack the rest of it with your fave stuffing material to 'taste' since HR doesn't calculate total density for polyfil at high densities.

If you want to use felt on the driver, then put it on the inside of the frame legs; the rest of the driver's backside should be ~encased in the netted damping material.

* Qts' = Qts + any added series resistance [Rs]: HiFi Loudspeaker Design

GM
 

Attachments

  • TC9FDandreaemme.txt
    1 KB · Views: 29
You're welcome!

As a general rule, basic TLs need 4-10x Vas based on its Vas, Qts' specs* regardless of length.
...cut...

Thank you very much for your suggestion but 109 cm is really much for my project, probably (I think) due to the small diameter, so I'll go to a closed box. However, regardless of the loading typology I fear that the principle weakness of my idea is the ring portion of cabinet just around the magnet, forming a really strict duct just near the point of emission and possible source of harmful internal reflections as detected by Troel in his article (I can't chamfer this zone because there's not enough space): see red zone in this drawing:

TQWT_P830869_rev03_36L_38Hz#-Model.png

So there are two solutions:
- covering magnet's or hole surface involved in this ring with a thin layer of right damping, hoping this won't obstruct emission;
- or a more radical solution, inverting driver's mounting leaving magnet at sight! I'll be able to adjust phase mounting it further back and inverting poles; usign this second I'll be free to treat internal surfaces as I'll want. At this purpose, may there be any difference in emission quality between front and rear for a mid-range driver?

TQWT_P830869_rev03_36L_38Hz#-Model_.png

And:
Try packing the area around the driver with open cell foam since felt may be too dense to deal with near reflections
...cut...
please can you explain what do you mean for 'the area around the driver'? Maybe covering the first part of the internal surface of cylinder, let's say 10 cm?
 
Member
Joined 2008
Paid Member
You're welcome!

As a general rule, basic TLs need 4-10x Vas based on its Vas, Qts' specs* regardless of length.

Anyway according to HR, for your app: a tube i.d. area ~ = driver Sd sims fine for a ~300 Hz XO IF ~109 cm long. Try packing the area around the driver with open cell foam since felt may be too dense to deal with near reflections and pack the rest of it with your fave stuffing material to 'taste' since HR doesn't calculate total density for polyfil at high densities.

If you want to use felt on the driver, then put it on the inside of the frame legs; the rest of the driver's backside should be ~encased in the netted damping material.

* Qts' = Qts + any added series resistance [Rs]: HiFi Loudspeaker Design

GM

Hi Greg, 109 cm is actually a very long tube. If I understand this correctly, the required tube length would diminish as the tube diameter is increased (??)
 
Hello, finally I began to make some measurement, and I must say that neither frontward nor backward solution works well even with felt or other dampening, because in both I have a 1.0 khz dip which produces a sense of metallic taste in listening.
So here's what I'll be able to do without destroying everything:

TQWT_P830869_rev03_36L_38Hz#_temp-Model.png

I can build a bigger and deeper flange using multiple properly chamfered wood rings ending in an exponential polypropylene reflex tube, as visible in image, forming a path without parallel surfaces with a shape similar to what was previously suggested in this thread.
What do you think about?
Won't it be a problem the quite circular shape of the path section near the driver, in which however there's not any parallel surfaces?
Thank you very much.
 
Hi Greg, 109 cm is actually a very long tube. If I understand this correctly, the required tube length would diminish as the tube diameter is increased (??)

Greets!

Well, for a TL, its length will shorten due to a larger pipe end correction, which won't significantly reduce it, though why really large mouth horns will have an acoustical axial length [much] longer than it physically measures, so normally one uses a ~stopped pipe [high resistance vent], though still need a significant net Vb to offset high damping losses.

GM
 
Hello, finally I began to make some measurement, and I must say that neither frontward nor backward solution works well even with felt or other dampening, because in both I have a 1.0 khz dip which produces a sense of metallic taste in listening.

How long is the pipe, approximate net volume of the bulge [less driver]? I mean a dip at ~ 1 kHz is ~34400/2/1000 = ~17.2 cm/6.77".

GM
 
How long is the pipe, approximate net volume of the bulge [less driver]? I mean a dip at ~ 1 kHz is ~34400/2/1000 = ~17.2 cm/6.77".

GM

Hello, you're welcome! I didn't remember well, seeing saved graph in effect dip is at around 600 Hz and in effect your formula is perfect because 34400/2/600 = about 28 cm which is exactly my pipe lenght, as you can see here from the following gated measurements of midrange in the original project's straight pipe:
TQWT_P830869_rev03_36L_38Hz#-Model_.png

midrange_with felt.png
midrange reverse-mounted with felt all around internal surfaces of pipe;

midrange_without felt.png
midrange reverse-mounted without any damping;

tweeter.png
tweeter's response, just for a reference.

It's clear that mid's response isn't quite good and I can ensure that this is very very listenable..
Now I have to:
1) try a 'bulged/conical7 pipe as drawn in my previous post, or
2) try a normal closed rectangular box, or
3) try an open baffle solution.

What about the last OB idea, can you give me some dimensional suggestion please?
Thank you very much.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.