Compression Driver Throat Screen

[Jamier]

I have been working with a couple of CDs that are 1 inch, 18 TPI screw mount,Ti diaphragm, 8 ohm
Drivers. Both have wire mesh throat screens. The gauge of the wire is probably close to 20. When I look at that screen I’ve got to believe that there is no way it is “acoustically invisible”. The upper end of both drivers looks pretty wobbly and I wonder if this might smooth out, if the screen was removed.

Jamie

 

[Paul Joppa]

There are a LOT of thoughts on this on the web! It's somewhat controversial, both whether it makes a difference and whether the risk of getting stuff in the voice coil gap is worth taking.

I have some thoughts but no direct experience or measurements, so I suggest you resign yourself to spending some time researching - places like Audio Asylum or DIYaudio for example.

 

[Jamier]

PJ, what are your thoughts ( other than the possible problems)? On one of these drivers I can easily remove the diaphragm and let any stufff fall through, lowering the chance of retaining debris. Another possible problem could arise from the gap that will remain between mounting seat on the horn and the flange of the CD when the screen is removed. I guess I could make an o- ring to fill that.

Jamie

 

[Paul Joppa]

My first thought was that if the screen has an effect, it would be an organ-pipe type of resonance, probably based on the diaphragm to screen distance - that would give a series of equally-spaced peaks and dips. So you could tell by looking at a response plot to see if the frequencies match the theory.

However, if there's a gap or other discontinuity left after ripping the screen out, you would not eliminate the problem - ANY kind of discontinuity will produce this problem.

The real problem with getting something inside occurs if the junk is magnetic, or has magnetic particles associated with it. You can't just shake that out, it's a very intense field in the gap.  The traditional trick is to wipe the gap with double-sided tape, risking replacing the particles with sticky tape residue.

It's the sort of thing that can go wrong in too many ways, and has only a theoretical argument that it would make an audible difference. That's why I suggest finding out what experiences others have had who have actually done it.  :^)

 

[Jamier]

PJ, after looking at the screen under magnification, I could see no evidence of a weld or solder holding it in, so I pulled on it with an instrument an it came out quite easily. It was glued in. I removed the excess glue ( with the diaphragm removed), blew it out with canned air an re- installed the diaphragm. Then I made an O- ring (silicone). I keep pre-made lengths of Silicone (injected into various diameters of Vinyl tubing) around. I re-installed the CD into one of the speaker horns. The results: Not Earth shattering but, an improvement. The top end is subtly smoother and more
complete, for lack of a better description. Vocal sibilance is no longer harsh and it just sounds more listenable. I listened for a while to discern these differences, so, as you can see this is not going to change anyone’s world. Bottom line: these screens do compromise SQ, but not hugely.

Edit: At least for this CD: Selenium 220 Ti

Jamie

 

[Paul Joppa]

Great! Thanks for posting your results. Your description of the effect matches pretty well with the pipe-resonance effect (a comb filter). Can you estimate the path length between diaphragm and where the mesh was?

 

[Deluk]

Yesterday I was reading this. https://sites.google.com/view/mpbs-ls35a-project
At the end there is comparison between the tweeter with and without the perforated dome. Having the dome looks to be beneficial. Yes, a different speaker and tweeter, but a similar situation.

Take care everyone. Obey the rules and stay safe.

 

[Jamier]

PJ, from the apex of the dome to the mesh, about 2.5 inches. Next time I have the baffles off I’ll remove the back covers and get a more exact figure. The mesh is not made from smooth round wire. It’s a flat wire, and has a pattern that makes it look like a miniature chain link fence. I’ll post some before/after pictures when I can. The total surface area of the mesh is fairly significant and, as such, I really expected a more audible effect from it’s removal. The other CDs I have are PSD-2013s. Those screens look to be more securely attached, so I don’t think that I will try to remove those.

@Deluk, I read that article, thanks for that. It’s kinda’ similar except that the LS’s have a conventional dome tweeter mounted on the baffle, so the dome is radiating some of it’s output across the baffle face. This CD dome is inverted and radiating it’s output into a horn. The dome is well behind the baffle and radiating almost none of it’s output across the battle face. In my mind this helps to explain why the grid of the LS’s could boost output in some frequencies.



Jamie

 

[Jamier]

PJ (and anyone else interested ), here are the with/without grid plus O-ring photos:
Edit: The O-ring doesn’t look that dirty in a live observation. I guess I ought to clean those things with some alcohol, but I can’t get any at the moment. The stores around here are all sold out. F#*%+’n hoarders!

Jamie

 

[Jamier]

PJ, how do you apply the diaphragm to screen measurement to determine the frequencies that might be filtered out by the screen?

Jamie

 

[Paul Joppa]

A discontinuity (such as the screen) will reflect sound back to the diaphragm. So if the screen is a half-wavelength from the diaphragm, the reflected wave will return to the diaphragm in phase, thus creating a peak on the sound level. This applies for all integer multiples of the half-wavelength frequency, resulting in a "comb filter" or "picket fence" spectrum.

So a 2.5" path length would have peaks around 5" wavelength, which is 2700 Hz, and multiples (5.4kHz, 8.1kHz, 10.8kHz, etc. Most horns have other resonances, too, so it's sometimes difficult to be sure what you are seeing - just a thing to be aware of.

This description ignores minor effects, such as reactive acoustic impedances including mechanical impedance of the diaphragm/amp, etc. but it should be close for a hard reflector like a metal screen. If the reflection is caused by a low acoustic impedance (relative to the plane wave impedance) it might reflect in reversed phase, giving resonances at 1/4 wavelength, 3/4, 5/4, etc. This is what gives the clarinet its characteristic sound - the reed end is high impedance, the open key is low impedance. But the spacing is the same.

 

[Jamier]

PJ,

  Here is a photo of the system freq. response with the screens in. I don't have the ability to measure (yet), so I can't include a curve of the system with the screens out/O-rings in, but eventually I will.Does this look close to what you predicted? Since the O-rings don't fill the space perfectly I would imagine that some of the wobble is still present.

Edit: PJ, this curve was generated by the designer, Michael Chua at Ampslab.I failed to mention that in the original post. There are irregularities very near the frequencies you predicted, it’s just that they look like notches and I’m trying to understand exactly what’s going on.

Jamie

 

[Jamier]

PJ, I reread the description in your last post and I still don’t get it. The notches still don’t align with your description, for me, but the measured frequency irregularities are pretty close to your prediction, especially since 2.5 inches was an eyeball measurement. Is my tweeter acting like a Clarinet?

Jamie

 

[Paul Joppa]

Sorry for the confusion - I got carried away. The whole bit about low-impedance reflections and clarinets was an aside, not relevant to what I think is happening in your situation.

There are some peaks around my predictions but there are too many other peaks and dips for me to be confident of the interpretation. It's plausible at best IMHO.

If you get measurements with the screen removed, to compare, then we can see whether these peaks change and the others do not.

 

[Jamier]

PJ (or PB),

    Will DATS do FR plots with a mic?

Jamie

 

[Paul Birkeland]

No.

 

[Jamier]

Since I don’t currently have the hardware/software to measure, I have offered one of my modded CDs to the designer to measure. I am hopeful that he will take me up on the offer.

Jamie

 

[Jamier]

PJ,
      I now have measuring gear and thought you might be interested in this. I made an FR response of the Osprey II, Throat Screen OUT (1/12 oct smoothing, which is what the designer used, I think ). To be fair, I need to get another Selenium Comp Driver to make a Screen IN comparison on my set up, to see if it looks similar to the designers data. It does look like the absence of the screen is measurable, even though it isn't dramatically audible, at least to me. When I get another D220 I will post the screen IN data that I get and see if this holds up.The Screen OUT curve is the CLIO, screen IN has the blue background.

Jamie