Diagnosing & Fixing Shifted Polk MW Drivers

So, your speaker fell over, or you had it shipped, or you just turned on your system one day, and it starts sounding awfully quiet. You press gently on the cones of your MW drivers and find one no longer moves in & out freely. It’s stuck.

First, stop playing music! Don’t just crank the volume and hope it forces the driver to move! Cranking the tunes with a frozen but repairable driver can lead to a melted voice coil, which kills the driver permanently.

(I’ve seen various drivers come to me over the years with melted voice coils. I wish I had taken pictures. Voice coil windings are insulated so that they cannot conduct electricity other than along the wire as it coils into a cylindrical shape. Thus, creating a speaker, see...)

https://www.youtube.com/watch?v=048tBZMt3eY

When the driver seizes and can’t move, the electrical current through the voice coil creates heat. Heat melts the voice coil winding’s insulation which allows electricity to arc (or short) from one coil to another. (In extreme cases, the whole coil unwinds from the voice coil former like a Slinky!)

Before getting into repairing it, let’s talk about what happened. Speaker 101 is a pre-requisite to understanding this; thankfully, basic understanding of transducer construction is pretty simple … see https://blog.landr.com/how-do-speakers-work/

jzpma8suh3k8.png

The phenomenon we experience with our Polk MW drivers is commonly called “magnet shift” … which is literally true but not the whole truth. The magnet shifting only causes a seized driver because the pole piece shifts with it locking the voice coil in place:

r6ja2r64f6kg.png

I have never seen a seized driver caused by the magnet shifting relative to the pole piece.
In fact, I don’t think I’ve ever seen a magnet / pole piece assembly that came perfectly centered relative to each other from the factory.

lkf471g5m4ma.png

In fact, the magnet / pole piece gap has PLENTLY of room for the voice coil even if it looks a little shifted.

While these dimensions may very from model to model, this MW6500 magnet I.D. is around 32.4mm. The O.D. of the voice coil is around 26.4mm. This means that there’s 3mm of room (32.4 – 26.4 = 6mm diameter / 2 = 3mm radially) for the voice coil inside of the magnet.

(Of course, ideally the magnetic field is perfectly even around all 360 degrees, but as long as the cone can move without interference in & out, it’s my belief that the rubber cone surround and spider do enough to center the cone’s movement that the effects of slight unevenness in the magnetic field from a slightly off-center magnet are completely negligible.)

wtdjn465asve.png

qod601lla29b.png

The CRITIAL dimension is the voice coil I.D. to pole piece O.D. … the pole piece needs to be perfectly within the voice coil so that the voice coil can move freely up and down.

This one is tricky. The voice coil I.D. is around 25.75mm …

io1pk5xjeoqo.png

… and the pole piece O.D. is around 25.25mm.

l7e8ooylxm92.png

This means we only have around 0.25mm of freedom radially for the pole piece inside the voice coil (25.75 – 25.25 = 0.5mm diameter / 2 = 0.25mm radially). Not much room!

So, this is what happens: the glue holding the magnet to the fails, and thanks to the magnetic field, the pole piece will be attracted to the inner wall of the basket. It only has to move 0.25mm and you’ve got a seized driver!

Here’s what the pole piece to voice coil looks like in a seized driver …

rawjkr8127xs.png

You now need to see if you can repair it. To be continued ...
<3 my 3.1TLs

I will fix your shifted magnets for free. :)
«134

Comments

  • lawdogg
    lawdogg Posts: 455
    edited November 2022
    Step 1: Remove the driver from the cabinet, and disconnect the speaker wire leads. Mark what lead goes to what terminal. Some (mostly later) Polks have spade connectors, some (mostly early) Polks have wire soldered to the driver’s terminals.

    Step 2: Check for voice coil continuity and correct resistance by measuring with a multimeter. It doesn’t matter which leads goes to which terminal. You can look up the spec resistance (DC resistance or “DCR”) of each driver somewhere in the forum (see, e.g. https://forum.polkaudio.com/discussion/118471/current-production-drivers-vintage-support/p1).

    Driver DC Resistance (ohms)
    MW6500 6.8
    MW6501 7.75
    MW6502 3.502
    MW6503 6.54
    MW6509 8.97
    MW6510 6.57
    MW6511 3.13
    MW6512 3.47

    In this case, I am working on an MW6500 which measures 6.8ohms. This is fine. So far so good.

    0b3qqv09hub3.png

    Step 3: Disassemble the driver. Because the glue holding the magnet to the basket has already broken, it will only be held together by magnetic force, which isn’t that strong. I simply grab the magnet with my hand, the basket with my other hand, and pull them apart. The trick here is not to damage the voice coil! So the pole piece must come straight out, more or less. If it comes out crooked you could hurt the voice coil former, or worse.

    If you don’t feel comfortable using brute strength to take it apart, use flat head screwdrivers or razor blades to separate the magnet-to-basket surface. Get it started on one side, then separate it on the other side. At some point you will need to use your hands to grab and pull, but do it gently.

    Now you’ve got two pieces: the basket w/ the cone, and the pole piece/magnet assembly.

    sxxink4p4abt.png

    Many times, the magnet and pole piece are still centered relative to one another. If this is the case, you can choose to leave it together. Or, you can take it apart as well. (If you do, mark the orientation of the magnet, as you’ll want to re-assemble without flipping the magnet upside-down.) You’ll have three pieces on your workbench like so.

    2smwi5qe9l3w.png

    Step 4: Inspect the spider to make sure it’s nice and flat. A deformed spider causes the voice coil to move in & out crooked, which can cause voice coil rub. (This step comes after disassembly because sometimes the driver seizes while the cone is in or out, which makes the spider look like it’s sagging or flexing.)

    Here’s a bad example: this one doesn’t look too bad, but it’s bad enough that it couldn’t re-center correctly.

    eqo4r3vx3bce.png

    (You can fix slightly crooked / sagging spiders using a spray bottle of water, two chopsticks, a rubber band, and a heat gun, but I won’t get into that here.)

    Here’s a perfect, flat spider.

    rp5xnkjgyvcf.png

    Step 5: Get all that old glue off! I use a sharp razor blade to scrape the old stuff off. Be careful, go slow. It’ll all flake off.

    eenjtco0cq5w.png

    mkglwzm36sas.png

    I finish by wiping with a soft cloth & isopropyl alcohol to remove any lingering oils or contaminants.

    It should feel nice and smooth. The glue can be tricky and hide from you, if you notice any bumps, it’s probably old glue that escaped your scraping the first time. Scrape it off until it’s perfectly flat, feels smooth, and it shiny in the light.

    10y8pi2mowi6.png

    Do this with the pole piece mating surface and basket mating surface too. Although I find most of the glue stays on the magnet, I do find patches of glue still on the other surfaces too.

    When you’re scraping the basket surface, you’ll be awfully close to the voice coil. Here’s a little trick, hold the basket in one hand, and use two fingers to flex the cone up, which removes the voice coil from harm’s way as you scrape with the other hand.

    uxx3lqb4ni11.png

    To be continued ...
    Post edited by SeleniumFalcon on
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • lawdogg
    lawdogg Posts: 455
    Step 6: If you removed the magnet from the pole piece, put them back together, being careful to re-assemble with the magnet facing the same way it was originally. You can do this by hand, carefully. Once together, slide the pole piece around so it’s perfectly centered inside the magnet. Eyeballing this is OK, because of the notes I made above re: magnet-to-voice coil clearance.

    It’ll look like this:

    s1ur964neodu.png

    rm2lz5qydg6n.png

    Step 7: You’re ready for the first round of adhesive. I use 2-part epoxy, although there’s much discussion about other adhesives that could work. Pick your poison.

    h877wo862pua.png

    You don’t need much. This was enough for me to do two magnet/pole piece assemblies.

    z7jlvc4jdcmg.png

    Apply the adhesive (in my case, mixed 2-part epoxy) to the shoulder between the magnet and pole piece. I ladle it on and spread it evenly with the pictured wooden stirrer. You should have a nice even bead around the magnet. (I like having a whole bead all the way around, thinking that the epoxy holds to itself, but you probably don’t need that much. You only need enough to prevent further shifting. Up to you.)

    27net1joc187.png

    This epoxy needs to be CURED before you move to the re-centering step!

    Step 8: While the epoxy is curing, turn to the dust cap.

    If it’s an MW6500 or another early driver, it’ll have a felt dust cap that needs to be removed so you can see the pole piece when re-centering. If it’s a later driver with a mesh dust cap, I’ve found success using a small flashlight to see the pole piece through the mesh, and don’t have to remove it.

    If you’ve got a felt dust cap, use a sharp / new razor blade to remove it. If you’re careful you’ll be able to align it perfectly when gluing it back on and it’ll be very hard to notice that it was removed.

    7tzm7vqua8ej.png

    Step 9: Inspect your voice coil. It should be perfectly cylindrical, the edge should be nice and straight. It should be free from debris inside (where it might rub on the pole piece) or outside (you can see when it’s extended, this’ll rub on the basket).

    2juk2954k46l.png

    Sometimes I see smashed bottom edges (from overdriving the driver), sometimes I see out-of-round voice coil formers (when the pole piece shifts too violently and smashes into it). You can massage these back to round gently by hand, or using a round paintbrush handle or something similar.

    Should be nice and smooth and clean and circular …

    30ul2budsoy6.png

    nzyf6cgg7rsz.png

    To be continued ...
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • lawdogg
    lawdogg Posts: 455
    Step 10: Game time – now it’s time to re-assemble the magnet/pole piece assembly and the basket. Hold the basket and use two fingers to retract the voice coil, like you did when cleaning it.

    Gently put one edge of the magnet up to the basket. It will want to snap together. Do what you can to slowly lay the magnet flat on the basket, allowing the pole piece to slide into the basket opening. The pole piece will be shifted to one side, and the voice coil will be sitting on top of it. This is normal.

    Let go of the cone, and your driver should look like this.

    kmyxh5pe889n.png

    Step 11: Mix up some more epoxy. Spread a bead on the magnet/basket shoulder just like you did on the magnet/pole piece shoulder.

    (This step depends on your re-centering jig. Mine, as I’ll explain below, doesn’t let me access the magnet after it goes into the jig, so I need to apply the adhesive before it goes in. This forces me to race against the clock to get it centered. I’ve spoken with some folks who have made re-centering jigs with access holes so they can apply adhesive once it’s already centered. 100% up to you how you want to do it.)

    1hjalo4l0i0q.png

    Step 12: Into the re-centering jig! Here’s mine. It’s very simple.

    1s0viei7aqj2.png

    I used a PVC pipe (I can’t remember the exact size, I got it from Home Depot … it may have been 3.5” or 4”?). Thick sidewalls are good to have, so the screws can have enough thread to grip while re-centering. Cut a slice down the length with a cutoff wheel so it can expand to accept the O.D. of a magnet, and small enough so it can be tightened down snugly around the magnet. Then you just need screws and a hose clamp.

    As you have probably guessed (correctly), the hose clamp goes around the magnet, which fixes the pole piece in place by extension since the magnet is already adhered to the pole piece. The screws are positioned so they push on the basket and by extension the voice coil … to re-center the driver, you adjust the screws which locates the voice coil relative to the pole piece.

    Slide the driver all the way in, so the basket is sitting on the pipe, the hose clamp is aligned with the magnet, and the screws are in position to push on the basket without hitting the magnet.

    k8rbulmczejw.png

    Tighten the hose clamp snugly around the magnet.

    hujx01kp4wc9.png

    To be continued ...
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • lawdogg
    lawdogg Posts: 455
    Step 13: Time to re-center! Look down on the pole piece from above, and remember that you’re moving the VOICE COIL relative to the POLE PIECE! So, start by tightening the screw where you see the biggest gap between the voice coil to pole piece. You will find that the pole piece likes to run away from you as you close the gaps, so normally I’ll tighten the screws in a circle with the last screw locking it into place.

    g2u9x660h1yo.png

    The whole process normally takes me under a minute. It’s very quick. Some tricks …

    -Be gentle.
    -Don’t overtighten – remember you need to leave around 0.25mm of a gap at each screw.
    -Let the voice coil free up on its own. You will hear it slide down as soon as it’s centered well enough to move. Don’t force it.
    -Gently move the cone in and out to test for rubbing. Apply even pressure. If you apply uneven pressure the cone will move unevenly, and these clearances are way too tight for any misalignment. If you hear rubbing, listen for where it’s coming from. Wherever you hear rubbing, the gap is too tight, so you probably need to tighten the opposite screw.

    4pvo9ny88jwv.png

    The first time, it will be intimidating. Some drivers still give me a heck of a time. Some don’t … the ones I’ve pictured were in great shape and only took one lap around the screws to center up. Less than 30 seconds each.

    Ta-da! Here’s a successfully centered driver.

    al0p1npq1s55.png

    Step 14: LET THE ADHESIVE CURE! If you remove it from the jig before it’s locked in place, it’ll probably shift again, and then you’ve gotta start over. (Learned this the hard way!)

    Step 15: Once it’s cured, remove it from the jig. I usually loosen the hose clamp first, and then I can spin the screws out by hand. My stomach still jumps when I hear it loosen up, but as long as the adhesive is cured, it won’t shift again.

    Step 16: Inspect your work, trim any excess adhesive. Your magnet should have two beads of adhesive.

    c8ne3v8u81ji.png

    (Note: I use beads of adhesive around the outside, and not any adhesive between the faces themselves, because I want everything to be perfectly parallel. If you use adhesive between the faces, especially thick glue of some kind, there’s a chance it’ll be uneven which sets up the pole piece crooked to the voice coil. As covered earlier, there’s very little room in there and any misalignment will make it that much more difficult / impossible to re-center the driver without interference. Don’t ask me how I found that out!)

    To be continued ...
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • lawdogg
    lawdogg Posts: 455
    Step 17: If you have a felt dust cap that you cut out, look at your cut marks and align it. Use a very small amount of PVA glue (I use Aileen’s) and drop the dust cap back into place. It’s taken me a lot of practice but these days I can get away with barely a sign that the dust cap was ever cut out.

    djvp2fbb43n6.png

    I find it helpful to apply a small amount of pressure while the glue cures. I find a small trinket like a socket and put it on the dust cap until the glue is dry and turns invisible.

    3347ct4r7bnv.png

    Step 18: Measure your voice coil to make sure it’s still good, hook up a test tone to make sure it sounds fine, marvel at your handiness, brag to your friends, ask your wife / kids / dog / neighbor / fellow Club Polk members to listen as you push the cone in and hear no scraping at all as it moves freely, re-install the driver, pour a beverage or you know, whatever, and enjoy!

    On a personal note, here are the first Polks I ever had, a set of Monitor 7 Series 2 that got the whole treatment. They are part of the inspiration for offering the free re-centering service. This picture was taken as my freshly-assembled Polks were stomping all over my previous king of the music room, JBL L26, and giving me that epiphany about Polk you’ve either already had, or about to have with your operational driver!

    zhz7je5grqcm.png

    Any questions don’t hesitate to ask. I've also saved this as a PDF, if there's any kind of database I can upload it to, or if you'd like me to email it to you, please let me know.

    Any improvements to my process and/or equipment are welcomes and encouraged, I only ask that you share with the group to advance our collective knowledge!

    -Ryan

    (All done)
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • lawdogg
    lawdogg Posts: 455
    P.S. Woops, I forgot to post the pictures of the PVC jig measurements. I can't edit my post so here they are ...

    apc8168i2jip.png

    qlyrbdrjhy4k.png
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • pitdogg2
    pitdogg2 Posts: 25,576
    Nice write up Ryan!
  • lawdogg wrote: »

    Step 11: Mix up some more epoxy. Spread a bead on the magnet/basket shoulder just like you did on the magnet/pole piece shoulder.

    (This step depends on your re-centering jig. Mine, as I’ll explain below, doesn’t let me access the magnet after it goes into the jig, so I need to apply the adhesive before it goes in. This forces me to race against the clock to get it centered. I’ve spoken with some folks who have made re-centering jigs with access holes so they can apply adhesive once it’s already centered. 100% up to you how you want to do it.)



    Very detailed write-up Ryan! Thanks for walking me through the process when I did mine. Having never done it before I was worried about completing the process before the epoxy setup, so I made my jig with 4 windows to allow the epoxy to be applied after the magnet was centered. Thanks again!

    j5harnj0jmb7.jpeg

    d6kha0g40g8k.jpeg



  • Astrl55
    Astrl55 Posts: 64
    edited September 2022
    Beautiful job of explaining and illustrating this problem -- Many thanks!
  • xschop
    xschop Posts: 5,000
    Great writeup. What do you do with the burnt voice coil drivers?
    Don't take experimental gene therapies from known eugenicists.
  • pitdogg2
    pitdogg2 Posts: 25,576
    xschop wrote: »
    Great writeup. What do you do with the burnt voice coil drivers?

    PULL!!
    :D:D:D
  • lawdogg
    lawdogg Posts: 455
    Thanks all - it was more than time to write this out. I'll be psyched if it helps.

    @joebass3 thanks for the pics! Yeah it was your jig I had in mind when I wrote it out!

    @xschop burnt coil drivers become parts sources. As it turns out, the magnets are pretty fragile, and I've had a couple drivers show up on the doorstep with fragmented magnets. Spare parts to the rescue. If I had a source for replacement voice coils, I could probably save the burnt drivers ...
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • xschop
    xschop Posts: 5,000
    pitdogg2 wrote: »
    xschop wrote: »
    Great writeup. What do you do with the burnt voice coil drivers?

    PULL!!
    :D:D:D

    Ok, Larry Silverstein.
    Don't take experimental gene therapies from known eugenicists.
  • Thanks for taking the time and effort to put that together. So much to learn.
  • Thanks for taking the time and effort to put that together. So much information.
  • lawdogg
    lawdogg Posts: 455
    For reference, here's a picture of a more extremely deformed spider. The funny thing is, I did get this one back straight enough to re-center.

    5h2kovw067lv.jpg
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • Sitka
    Sitka Posts: 74
    Thanks again - me thinks this particular repair is going to be needed more and more as these drivers age. Those MW 6500's featured are 42 years old! Agree with F1 - should be a sticky.
  • TennMan
    TennMan Posts: 1,266
    Thank You!! Great repair tutorial that is badly needed. Agree with this needing to be a sticky.
    • SDA 2BTL · Sonicaps · Mills resistors · RDO-198s · New gaskets · H-nuts · Erse inductors · BH5 · Dynamat
    • Crossover upgrades by westmassguy
    • Marantz 1504 AVR (front speaker pre-outs to Adcom 555)
    • Adcom GFA-555 amp · Upgrades & speaker protection added by OldmanSRS
    • Pioneer DV-610AV DVD/CD player
    • SDA CRS+ · Hidden away in the closet
  • lawdogg wrote: »
    For reference, here's a picture of a more extremely deformed spider. The funny thing is, I did get this one back straight enough to re-center.

    5h2kovw067lv.jpg

    Put me down as interested in a tutorial on fixing distorted spiders. Yes, nice work!
    George / NJ

    Polk 7B main speakers, std. mods+ (1979, orig owner)
    Martin Logan Dynamo sub w/6ft 14awg Power Cord
    Onkyo A-8017 integrated
    Logitech Squeezebox Touch Streamer w/EDO applet
    iFi nano iDSD DAC
    iPurifier3
    iDefender w/ iPower PS
    Custom Steve Wilson 1m UPOCC Interconnect
    iFi Mercury 0.5m OFHC continuous cast copper USB cable
    Custom Ribbon Speaker Cables, 5ft long, 4N Copper, 14awg, ultra low inductance
    Custom Vibration Isolation Speaker Stands and Sub Platform
  • lawdogg
    lawdogg Posts: 455
    Sounds good - I'll take some pics next time I get one with a wacky spider.

    In the meantime, here's a good case in point re: how off-center the magnets can be from the factory. This is an UNSHIFTED magnet / pole piece assembly, just how it left the factory.

    kkrlfp1t1dk5.jpg
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • xschop
    xschop Posts: 5,000
    Polktastic!
    Don't take experimental gene therapies from known eugenicists.
  • odcics2
    odcics2 Posts: 339
    lawdogg wrote: »
    Sounds good - I'll take some pics next time I get one with a wacky spider.

    In the meantime, here's a good case in point re: how off-center the magnets can be from the factory. This is an UNSHIFTED magnet / pole piece assembly, just how it left the factory.

    kkrlfp1t1dk5.jpg

    Yikes!

  • xschop
    xschop Posts: 5,000
    lawdogg wrote: »
    Thanks all - it was more than time to write this out. I'll be psyched if it helps.

    @joebass3 thanks for the pics! Yeah it was your jig I had in mind when I wrote it out!

    @xschop burnt coil drivers become parts sources. As it turns out, the magnets are pretty fragile, and I've had a couple drivers show up on the doorstep with fragmented magnets. Spare parts to the rescue. If I had a source for replacement voice coils, I could probably save the burnt drivers ...

    Let me know if you have an MW basket with a good base-plate. Doesn't even need a cone.
    Don't take experimental gene therapies from known eugenicists.
  • lawdogg
    lawdogg Posts: 455
    I actually think I have one. You want it / want any measurements? LMK what you need.
    <3 my 3.1TLs

    I will fix your shifted magnets for free. :)
  • xschop
    xschop Posts: 5,000
    I need one to make another jig from. I'll PM thanks.
    Don't take experimental gene therapies from known eugenicists.
  • I dunno...

    Can these be massaged?

    The dust caps on my midwoofers pulled off easily.

    Using my 5b cab to smack out the pole piece was easy peasy.

    t7huxyuo95aq.jpeg
    etxdivashuti.jpeg
    e155ztkta1wq.jpeg
    0g9d2ip4r2eb.jpeg
    b4uxnivaye0i.jpeg



    The Thrifty Setups in Mah House Big thrifty stereo in the basement w/ my custom SDA-1C (built with help from kind forum members) * Beautiful 1966 MCM GE console upgraded w/ Bluetooth, Dual turntable, and Paradigm speakers in family room * Swanky 1980 Realistic system and great TEAC eq with dancing colored lights in the living room * custom 5B on a system for my Dad * Ye Olde college stereo in the garage
  • pitdogg2
    pitdogg2 Posts: 25,576
    Doubt it. The VC is enamel coated to prevent it from shorting on the wire next to it.
  • How does it ohm out? Hard to tell in pictures but is the voice coil cocked in addition to rubbing on one side?

    Damage to the former and coil windings is obvious but might not be fatal.

    I'd check carefully to see if the spider is cracked.
    George / NJ

    Polk 7B main speakers, std. mods+ (1979, orig owner)
    Martin Logan Dynamo sub w/6ft 14awg Power Cord
    Onkyo A-8017 integrated
    Logitech Squeezebox Touch Streamer w/EDO applet
    iFi nano iDSD DAC
    iPurifier3
    iDefender w/ iPower PS
    Custom Steve Wilson 1m UPOCC Interconnect
    iFi Mercury 0.5m OFHC continuous cast copper USB cable
    Custom Ribbon Speaker Cables, 5ft long, 4N Copper, 14awg, ultra low inductance
    Custom Vibration Isolation Speaker Stands and Sub Platform
  • Spider looks fine, coil isn't cocked.

    Dunno how to ohm it.

    I hate that it's lookin' fubar. But willing to tweak under guidance.

    Shame, the sister speaker is doing awesome =(
    The Thrifty Setups in Mah House Big thrifty stereo in the basement w/ my custom SDA-1C (built with help from kind forum members) * Beautiful 1966 MCM GE console upgraded w/ Bluetooth, Dual turntable, and Paradigm speakers in family room * Swanky 1980 Realistic system and great TEAC eq with dancing colored lights in the living room * custom 5B on a system for my Dad * Ye Olde college stereo in the garage