AR3a tweeter dissected...

Copa1934

My ears are bleeding
Subscriber
This was NOT my goal. I purchased a "non working" driver, cheap (far as I'm concerned) as another sample for my tweeter rebuild. Well this got deeper than the other experiments. I performed all the usual checks, tests, etc. and concluded I wasn't going to get it working. Used a blade to separate the "surround" from the plate and then carefully removed the driver assembly. Since the leads are "in" the adhesive I can't verify if the break is in the adhesive, but from the looks of things there are no breaks "outside" the adhesive. So moving forward I wanted to remove the black plastic housing. These are held in place with, likely, two part epoxy. Looks a lot like varnish after it's aged many years. Since varnish can be removed with impact, due to it's brittle nature, I began tapping around the enclosure in hopes of breaking it free. Oh, part of the reason was the flange was badly distorted and I was hoping to come up with some replacement "solution", later down the road of course. Well my tapping didn't go over well. It chipped away a lot of the epoxy, but in places it cracked the housing, so I simply busted the housing off. I know, some of you are shaking your heads, others have their hands in their heads and hopefully others are LAO. I didn't have a functional tweeter to start with, so not really a big loss. But, I did learn some interesting things about the construction of these buggers. I'm attaching images of the tweeter from the original listing, don't worry not a link to eBay. Also will be a number of images of the different phases of the deconstruction. This will be my third tweeter taken apart, but NOT to this degree. The others I just removed the dome/VC assembly.

First two images are reversed, chronologically speaking. The first image is what happens when tapping breaks the face plate from the magnet assembly. Note the third image, a blob of epoxy. There were NO imperfections in the magnet or the face plate. Both were very flat and smooth. Can't understand why that blob, and a couple not visible, would be there. Also of note is the corrosion (oxidation?) on the pole piece. If this is not uncommon (near impossible to determine without taking apart), it may explain some failures or why some tweeters do not perform as well as others. I test fit the VC assembly (will post image later) and it was snug, not free play to speak of.

Anyone who has questions, please feel free to ask and I'll answer as best I can. I can't get too technical, I just don't know enough to formulate opinions on much, but will try where I can and feel comfortable doing so.

I do plan to re-assemble, and additional images will show my efforts thus far. I did all this work just this afternoon.
 

Attachments

  • $_57.jpg
    $_57.jpg
    60.7 KB · Views: 312
  • IMG_1179.jpg
    IMG_1179.jpg
    72.3 KB · Views: 301
  • IMG_1181.jpg
    IMG_1181.jpg
    54.5 KB · Views: 287
  • IMG_1186.jpg
    IMG_1186.jpg
    88.8 KB · Views: 299
  • IMG_1189.jpg
    IMG_1189.jpg
    64.6 KB · Views: 287
Note: I was wrong. The attachment tool had the images out of order, but they are show chronologically correct. If you examine the third image and look closely at the pole piece you can see epoxy. Besides corrosion, there seemed to be "excess" epoxy. I found the only way to remove it was to scrape it away. When I test fit the plate it would NOT line up properly. It seems reasonable to assume that the three notches are to aid alignment when assembling. If I'm wrong, I'm sure I'll be corrected. The last image shows how the pole piece is attached.
 
More images

First image shows the cleaned up pole piece. Next image a different angle. Looking closely you can see the epoxy at the base. It also appears the pole piece is flared at the base, presumably to better secure to the bottom plate and to keep it flush and the pole perfectly perpendicular. Using a blade I was able to easily remove the "film" of epoxy. Give the strength of the magnet it must squeeze all excess epoxy out, and I discovered the hard way just how strong. Nearly pinched a couple fingers when I lost grip of the plate. Next image shows the nice cleaned up surfaces. The last is the dome/VC rested on the pole. Not I repaired the dome. Pushed the dents out from the underside, then applied a light coating of superglue to strengthen the dome. Doing it from the underside also concealed the repair so the dome looks more natural.
 

Attachments

  • IMG_1197.jpg
    IMG_1197.jpg
    45 KB · Views: 212
  • IMG_1195.jpg
    IMG_1195.jpg
    46.8 KB · Views: 148
  • IMG_1193.jpg
    IMG_1193.jpg
    72.1 KB · Views: 149
  • IMG_1191.jpg
    IMG_1191.jpg
    46.8 KB · Views: 126
  • IMG_1190.jpg
    IMG_1190.jpg
    63.7 KB · Views: 137
Note: I was wrong. The attachment tool had the images out of order, but they are show chronologically correct. If you examine the third image and look closely at the pole piece you can see epoxy. Besides corrosion, there seemed to be "excess" epoxy. I found the only way to remove it was to scrape it away. When I test fit the plate it would NOT line up properly. It seems reasonable to assume that the three notches are to aid alignment when assembling. If I'm wrong, I'm sure I'll be corrected. The last image shows how the pole piece is attached.

The magnet structures are originally assemble BEFORE the magnets are magnetized. To get that top plate centered when you are ready to re-glue it to the magnet, you will need to shim the pole piece with the right thickness of shim material to center the plate around the pole and still allow you to remove the shim after the new epoxy has cured.
 
Nice work! Can you measure the the center pole diameter with decent precision? I've been thinking about rebuilding this tweeter and if the original voice coil is working I'd probably use a very thin sheet of flat foam as a suspension to reattach it back to the structure. The original VC looks to be wound on a paper former and will not have good thermal capacity.

It would be good to find a modern replacement 4 ohm dome that fits but I don't know of one.
 
This was NOT my goal. I purchased a "non working" driver, cheap (far as I'm concerned) as another sample for my tweeter rebuild. Well this got deeper than the other experiments. I performed all the usual checks, tests, etc. and concluded I wasn't going to get it working. Used a blade to separate the "surround" from the plate and then carefully removed the driver assembly. Since the leads are "in" the adhesive I can't verify if the break is in the adhesive, but from the looks of things there are no breaks "outside" the adhesive. So moving forward I wanted to remove the black plastic housing. These are held in place with, likely, two part epoxy. Looks a lot like varnish after it's aged many years. Since varnish can be removed with impact, due to it's brittle nature, I began tapping around the enclosure in hopes of breaking it free. Oh, part of the reason was the flange was badly distorted and I was hoping to come up with some replacement "solution", later down the road of course. Well my tapping didn't go over well. It chipped away a lot of the epoxy, but in places it cracked the housing, so I simply busted the housing off. I know, some of you are shaking your heads, others have their hands in their heads and hopefully others are LAO. I didn't have a functional tweeter to start with, so not really a big loss. But, I did learn some interesting things about the construction of these buggers. I'm attaching images of the tweeter from the original listing, don't worry not a link to eBay. Also will be a number of images of the different phases of the deconstruction. This will be my third tweeter taken apart, but NOT to this degree. The others I just removed the dome/VC assembly.

First two images are reversed, chronologically speaking. The first image is what happens when tapping breaks the face plate from the magnet assembly. Note the third image, a blob of epoxy. There were NO imperfections in the magnet or the face plate. Both were very flat and smooth. Can't understand why that blob, and a couple not visible, would be there. Also of note is the corrosion (oxidation?) on the pole piece. If this is not uncommon (near impossible to determine without taking apart), it may explain some failures or why some tweeters do not perform as well as others. I test fit the VC assembly (will post image later) and it was snug, not free play to speak of.

Anyone who has questions, please feel free to ask and I'll answer as best I can. I can't get too technical, I just don't know enough to formulate opinions on much, but will try where I can and feel comfortable doing so.

I do plan to re-assemble, and additional images will show my efforts thus far. I did all this work just this afternoon.

Pretty interesting. I've been curious what the guts of these look like, since I have 6 dead ones from my LST project. I don't know what to do with them. Trash them, sell them if there's even any value for someone to restore, or just stick them on a shelf to collect dust. Are these in fact repairable at all?

Anyway, great pics Copa!
 
Magnetized after? I didn't know ceramic magnets were made that way. Interesting.

Pole dimensions? According to my dial calipers 95/128", so shy of 3/4" Last time I measured a VC it was 3/4", but that is very challenging due to flex. So 3/4" likely is correct.

I have sheets of Kapton that I believe will work for new VC. I also have magnet wire that appears to be the correct gauge.

I would be willing to pay shipping for dead tweeters, the more samples the better. This is a very time consuming process, but I'm hoping the time and effort payoff.

GD70, sometimes they are, but most have been repairing the lead *from* the surround. Any attempt to remove the motor usually results in minor damage, and then restoring it a nightmare as there is NO simple or documented method to centering the motor and then applying an appropriate "surround" material. The latter part is one thing I'm working on and may have a solution for the surround. But building a new VC, well that's where I've been slowed down some.
 
You might be better off trying to find an existing voice coil and trimming it to fit the dome at the right depth.

Your Kapton may or may not be thin enough to accommodate the coil windings and still fit in the gap. Those inside and outside tolerances are pretty tight. Magnet wire is moving in the right direction, but making it stick to the former and stay intact on the former would be a vey tricky proposition. Purpose-made voice coil wire is a special animal. It is of course coated with enamel, but on top of the enamel is a thin coat of acetone activate adhesive. The wire is sprayed with acetone while it is being wound, activating the adhesive, and it basically glues itself together while being wound.
 
I have access to some VC through MWA, but they're all 8-ohm. They've expressed an interest in making 4-ohm, but tooling isn't cheap, so figuring out how many could be sold is a terrible challenge. I can't imagine they'd have to do more than figure the number of windings.

As for the Kapton, already measured thickness and it's actually a tad thinner than the paper (I'm assuming it is) used with the original tweeter. I may test the "spray" idea with the motor winding wire anyway, just to see. But I do agree, securing the wire has been one of my concerns. It's all fine and dandy doing the wrapping, but securing it so it doesn't come apart AND doesn't add too much to the final thickness. This is not fun, but I'm hoping I can get it done. If successful I plan on doing some midrange drivers as well. I have one or two of those as well, that are already apart.
 
More images...

I don't know how visible it is, but I've noticed a common issue with these, and likely just age OR abuse? Note on the wire (part of the winding) that is over the lip of the former? Not sure how the images will be sequenced.

One image was taken on an HPM tweeter, this shows the inside where I have superglued the underside to strengthen the dome. Worked pretty good.

The distant image is the one that the wire should be visible.

The close up shows not only the wire but the indentation in the former from the wire. Looking inside there is a slight tear, this was caused by me removing the assembly. I have shaky hands, but intermittent. Adds to the challenge.
 

Attachments

  • IMG_1198.jpg
    IMG_1198.jpg
    110 KB · Views: 143
  • IMG_1203 small.jpg
    IMG_1203 small.jpg
    66.9 KB · Views: 141
  • IMG_1201.jpg
    IMG_1201.jpg
    35.2 KB · Views: 126
I'd try to wind on an aluminum former if there is enough clearance, nearly all, quality, modern tweeters that I've looked at use aluminum. Will probably increase the mass but I suppose we could calculate by how much to see if it is significant. Most hobby shops have thin sheet, not sure if it is thin enough.
 
That complicates things for me, being "out in the sticks". I live in farm country. Nearest "city" with a population greater than 50k is over 50 miles away. They have more places to shop, but not sure for what I need, possibly.

I've read where Kapton is used quite a lot, but mostly it seems with smaller drivers where the larger ones seem to use Aluminum, Copper, etc. I know the 5b and Mach One woofers have copper, though with Mach One I think had Aluminum in early models.

Of course, the whole point of this thread is getting input, not just conveying my "discoveries" so to speak. So input is appreciated.
 
The way the windings split indicate just enough heat hit the voice coil to cause failure of the adhesive holding the windings together. That's not really an age issue, but in a roundabout way, it is. Newer voice coil wire has a higher temp enamel on it. A better insulation to keep heat from causing adhesive failure.

If you can obtain an 8 Ohm coil that is the right diameter with a winding length that is about twice as long as your 4 Ohm coil, just remove the top half of the windings.

Aluminum formers were a popular replacement for paper formers for a few decades. They would absorb some heat without burning. In some applications, especially in high power drivers, that heat would transfer up the former to the cone/spider joint, and while rather rare, the former got hot enough to start the cone and/or spider on fire before the voice coil failed to the point of becoming an open circuit. Aluminum formers also introduce eddy currents into the picture.

Kapton or Nomex are the two materials used today to make the majority of voice coil formers. Nomex is good for low power applications like mids and tweeters, lousy for woofers because it will bubble long before the coil burns. Kapton is good for both low and high power apps. It too will bubble, but only under much higher temperatures than Nomex.
 
Kool. I'll stick with the Kapton for the moment. Should there be a gap in the former? I've pondered this as well. As components heat up gaps contract, however, in this case if the objects expands with heat, the gap may not shrink? Advise here would be appreciated. Oh, other than the many youtube vids I've examined is there any practical technical info on "making" VC's? Online preferred.
 
Should there be a gap in the former? I've pondered this as well. As components heat up gaps contract, however, in this case if the objects expands with heat, the gap may not shrink?
Yes, but it only needs to be wide enough to keep the edges from overlapping. When you consider the inside tolerance of the voice coil to the pole piece is likely to be less than .008", any expansion or contraction would cause the former to hug the pole piece before it would ever cause the edges to hit each other or overlap.

Oh, other than the many youtube vids I've examined is there any practical technical info on "making" VC's? Online preferred.
I don't know. I bought a coil winding machine and 52 different diameter expandable mandrels to start winding my own voice coils maybe 18 years ago. I still needed to go see someone who knew what they were doing with the same or similar equipment actually winding coils to fully understand the entire process. It is a specialized field meant for mass production. The tolerances and specialized components involved do not lend themselves to DIY one-off production....or reproduction in your case.
 
I have access to some VC through MWA, but they're all 8-ohm. They've expressed an interest in making 4-ohm, but tooling isn't cheap, so figuring out how many could be sold is a terrible challenge.
Would a .761" diameter, .500" former, .100" winding width, 4 Ohm coil work?
 
I wound 6 voice coils like these years ago for 4 woofers. The first 2 were a bit too large
on the OD and I made a few adjustments: It was a 4 layer coil and I made it dual VC
just in case I wanted to try some experiments.

The woofers were 8" Philips/Norelco AD 8060/W8 with paper formers. I used them in pairs
in huge home made TLs, they worked well and we drove them with about 250W/ch for
several years. It smelled like toast cooking when we really cranked the system and
indeed the paper formers were burnt when they finally failed. Amazingly, the insulation
slowly degraded until the VCs were a dead short - never saw speakers fail like this
before.

attachment.php
 
Last edited:
Would a .761" diameter, .500" former, .100" winding width, 4 Ohm coil work?
95/128 works out to 0.7421875. I think that is close enough, since it's larger and not "tighter". Oh, with a DMM they need to test around 3.5, but I'll have to look it up. The "nominal" rating is 4-ohm. Once installed the rating will change, I've noticed this with a couple rebuilds. I've also noticed a change in T/S specs when something as simple as the surround separates from the cone or frame.

Are you saying you have some?
 
Back
Top Bottom