Re-coating AUDAX dome tweeters w/measurements!!

[elnaldo]

There is a long time since I've noted a lower output in one of my tweeters (years perhaps), but never had time to take care of it. I'm talking about AUDAX HD100 textile dome tweeters in my DIY speakers.

They looked like this:

This are the measurements of both tweeters before the treatment: one of them was definitely bad.

Recently, I've found a local speaker factory that sells a "vinyl plasticizer" (google translation for "plastificante vinílico"), that they use to seal textile surrounds. They told me "it stays sticky", so I thought it would be a good material to treat my textile domes.

So, I opened them, cleaned the old material the best I could with isopropyl alcohol, and applied a thin coat of this "vinyl plasticizer". Started with the bad one, just in case something goes wrong, I had a good tweeter to compare.
It looks like this now:

AND, the most important thing, after doing both domes, the pair measures like this:
(I think the colors are swapped, the green line is the yellow line in the previous measurement)

Could it be better? I don't know. But the tweeters came back to life, every sound is clear and crisp again !!! I'm happy with the results, I think I'll try it again after a while, cleaning this coating and doing it again a bit better, perhaps using less material.

 

[Blue Shadow]

Good report. Interesting results. Possibly painting the bolts black would smooth out the response a bit more, but great so far.

 

[canuckaudiog]

You've piqued my curiousity. I wonder if the old Audax tweeters I have on my IMF Super Compacts suffer from the same problem - they sound rather dull at the moment.

 

[audiomagnate]

As I recall, they were made of butyl impregnated cambric, but you can't argue with your results. The butyl (synthetic rubber) was added to the cambric (the black material often found on the underside of upholstered chairs) to make the material air tight. It looks like your stuff does the same thing. It also looks like my dog was in the general vicinity of those tweets at one point in time. Good stuff!

 

[audiozaz]

Certainly not a subtle improvement, good idea and great job!

az

 

[elnaldo]

I bought those tweeters brand new in 1989 or 1990, and they followed me everywhere since there, I had a dog for some years before 1996... And my sister had a cat for some years, she had the speakers by that time.

So, if your tweeters look like this ones, and you think they are not performing like before, be sure they are not performing like before !!

I'll do some more tests, as I said, with less material, specially at the dome base.

It's a kind of sticky glue, water based, never dries completely.

 

[elnaldo]

Update !!

Managed to remove ALL the old material. It was easy with the new glue still sticky, and a piece of paper, sticking the paper to the sticky glue, and the help of hot water and some small pliers, removed the new glue and the old material came off, leaving the cloth almost clean.

Edited, added picture of the cleaned dome.

So I applied a very light coat of the "vinyl plasticizer"...

and the result is this: (measurement from the worst performing tweeter, the one that was really bad in the first post, and the green line in the 2nd measurement)

It seems this is the way to go...

I'm happy that I finally decided to fix this tweeters !!!

Doing the 2nd one tomorrow

 

[Stan man]

Hmm, I wonder if this would be the same fix for ADS drivers as well? Should be-right?

 

[elnaldo]

I think any textile dome tweeter can be "fixed" this way.

The trick is to find a suitable material for the re-coating. I've read about "shellac" in other threads, and PVA glue.

What I like of this material is that it stays sticky as the original.

 

[audiozaz]

Even better with the old stuff removed, brilliant!

az

 

[Blue Shadow]

It just gets better. Nice work. I know we can't get what you used here but if you have more information on the product you used, the group would love to see it and a picture.

 

[Nikko75]

What equipment and environment are you performing these measurements with? Can you drop the smoothing altogether? I would be particularly interested in seeing CSD, transient and impulse response are after this restoration.

 

[audiomagnate]

What's your measurement setup? I have 30 year old Morel MDT-28's that look fine, but they don't sound like they used to. It's probably just my ears that have deteriorated. Can you post some pictures of the actual product you are using? A search of "Vinyl Plasticizer" turns up pretty much nothing relevant.

 

[elnaldo]

I'm using ARTA software with my laptop computer, gated measurement to avoid reflections, cheap USB sound card, this DIY microphone
( http://gainphile.blogspot.com.ar/2008/11/cheap-and-accurate-speaker-measurement.html )

measuring small signal on a table, near the edge, covered with a thin wool carpet to minimize reflections too. I'll measure it with a simulated baffle, to see if it measures different.

The smoothing setting, 1/12 octave, I think is the minimum that ARTA allows.

But I don't know what CSD, transient and impulse response are. If you could point me to some tutorial, I could measure it to post here.

The tweeters sound very very good. I'm currently playing them for 12 hours to let the material dry and break-in. I'll post measurements tonight and a pict of the product.

 

[mikeybc]

Nice job, and thanks for the setup info.

 

[elnaldo]

Well, this is the best they can do: Measurement from 2 tweeters, small signal, and on bottom, measurement of both speakers from far field, more than 1 m.

Added this graphs, in case somebody find them useful. I can't understand that "cumulative spectral decay"...

And here, not too helpful, "The Thing". No name, no label, no brand, the seller said the word "polymer" in the talk... They sell it as "Vinyl Plasticizer", but a search for that (in spanish "plastificante vinílico"), only point me to the price list of that vendor !!! So perhaps he invented the name...

And finally, the speakers in question !!!! My corrugated paper speakers!! Peerless 831916 woofer + Audax HD100 tweeter You have to see the people's face when I play some loud music on them

 

[audiomagnate]

How does "WTF??!!" translate? I'm sure they sound great at low volumes, but cardboard? I KNOW wood, mdf, birch ply, concrete, ANYTHING almost would work better. Why cardboard? I have to ask, what do you use for internal damping/stuffing? Corrugated cardboard has a great stiffness to mass ratio, is that why you use it?

Anyway, how can I get my hands on some of your magic goo?

 

[elnaldo]

This was a prototype, but I don't have easy access to wood tools, and since everybody likes them, and I have some other main speakers, I leave them like this.

The box has some internal bracing, from cardboard triangles, a port at the back, and they sound awesome at near field. They literally disappear, the music comes from the air, from everywhere!

 

[Nikko75]

Now you know how to measure CSD (cumulative spectral decay) and impulse response. They are going to be very useful tools in speaker evaluation. CSD depicts a driver or speaker system's energy storage, ringing and native reluctance to the to settle back to the neutral position.Another good one is step response, which show the drivers ringing tendency in a higher resolution. You can also take polar power response curves on the horizontal or vertical axis by taking a measurement on axis, the moving the microphone 15 degrees in that direction to take another measurement, then 30 degrees for the next another measurement, and so on. When they are overlapped, you suddenly see how a driver or speaker fairs when the listener is off axis. If you live int he country, I would advise to experiment with taking measurements outside on a stand, and using several full bandwidth measurements you can get a summed average down to very low frequencies. This is the biggest problem for reviewers, because they tend to live in noisy environments, so most speaker measurements omit the low frequencies. If this is not possible, purchase some high NRC sound absorption panels to aid in your measuring. My advice is to standardize your measurement envelope in dB , such that all future measurements are taken at the same levels and over the same time windows. That will make comparisons very easy for all your drivers or speakers. I consider distortion and CSD down to 50dB below the fundamental to be of importance. I tend to do unweighted measurements that show all the peaks and little things that manufacturers and designers don't like to discuss. 1/24th and 1/48th measurements are also good when the surrounding environment tries to interact with the test results, and you may wish to experiment with that, too. I know a fair amount about precision speaker measurements, mostly using laser spectrometry, but I don't get "plus one" from around this place because I don't get involved in their little groupie cliques. You, on the other hand appear to have a good head on your shoulders.

 

[elnaldo]

What I'm understanding from that CSD graph, is that the fastest the cascade falls, the more accurate the speaker sounds... I understand those milliseconds means a sound still sounding in the driver after the signal finished... or something like that.

Anyway, without a deep knowledge, I like my Audax plot... it seems very smooth and fast compared with other graphs I've seen on the web, very irregular, and some from expensive tweeters that looks similar to mine. What do you see in this tweeter?

I'm sure I'll be able to standardize the graphs for easier comparison.

I've seen the "un-smoothed" response, and it was the same as the 1/24 plot, but that screen didn't allow to overlay the previous readings, so I'm using the 1/24 reading, accurate enough for an indoor reading and a cardboard speaker

The "step response" was something similar to the impulse response, but with a kind of very light "wave" at the end. I forgot about that since I didn't know what it was for...

To calibrate my microphone, I did the following: Set the amp output to 2.83 V with a 1000Hz test tone. Then, measured a known speaker, with a datasheet that states it's output level in dB, at 1 meter distance.

Then, started to tweak the microphone sensitivity setting in ARTA, until the measurement matches (approx.) the dB stated by the speaker manufacturer.

But, at the end, I do what the speaker manufactures do: I take the measurement that I like the best and hide the defects ...