Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
24.184.225.216
I recently purchased a pair of Tympani IVa speakers, still in their factory sealed cartons after a complete rebuild at Magnepan last year. I've been having a problem getting centered images between the stereo pair, and after some diagnostics, discovered that one midrange driver is approximately 5dB lower in volume over its entire range than the other channel. I measured near-field to eliminate any room effects, and every other driver in the system (tweeter and both woofer panels) measures virtually identical in level left to right. My first thought was that one of the two runs of voice coil wire that make up the mid driver was not connected, causing the lower efficiency/volume. However, the left and right midranges both measure identical DC resistances (3.25 ohms including the 0.6mh coil), and if only one of the two runs of wire were connected, the DCR would be much higher (double) in the bad channel since they are connected in parallel.
I've checked the internal wiring and crossover (both external and external), and they all seem fine. I also jumpered the fuses, as I've had problems with intermittent fuseholders in the past, but that's not the case here.
Any suggestions for diagnosing/repairing the problem would be most welcome, short of shipping them back to Magnepan, which is my last resort.
Attached is a screenshot from OmniMic of the frequency response graphs showing the midrange channel to channel imbalance.
Mike
PS: I've communicated with Gary Edwards at Magnepan, who was helpful but admitted he doesn't have the technical expertise to help. He suggested I call Wendell, but I got no response there. Hoping someone here has some suggestions. Thanks.
PPS: This is actually my second pair of TIVa's - I foolishly sold my first pair many years ago. I posted the schematics of the TIV and TIVa crossovers in MUG Tweaks over 20 years ago!
Follow Ups:
Recap: One of my midrange drivers had lower output than the other channel, by 3 or 4 dB across its range. I determined the problem was an excessive bend in the perforated back panel of the driver (which hold the magnets), and the increased distance of the magnets from the wire (voice coil) on the mylar led to lower volume.
I've experimented with adding a variety of crossbars to reduce the bend in the midrange panel, and found that I only need 2 crossbars near the center (where the curve was greatest). I used 1/4" by 5/16 aluminum bars, secured with screws thru the MDF panel. I spent a couple hours building a jig to route notches in the MDF panel to mount the cross bars as well and new mounts (in place of the staples used by Magnepan), then routed 22 notches in about 5 minutes. After experimenting with the number of cross bars (I tried up to 4) and the pressure they exerted on the rear of the driver by spacing them away from the driver with flat washers, I found I only needed slight pressure on the center of the panel to get the level back up to match my other channel driver.
Attached is a photo of the bars in place showing my mounting technique.
Listening tests show I now have good left to right balance across the frequency spectrum, with solid center images for vocals (which I did NOT have before).
Thanks to everyone for their suggestions and help.
I'm now moving on to terminal/cap/fuse-holder upgrades.
Mike
.
Congrats !!!
Glad to hear you have a repair solution.
Progress Report: I've confirmed that reducing the bow of the midrange back panel clearly increases the output, so thats the problem that needs to be solved.
Methodology: I built a stand to hold the driver vertically, and measured the frequency response with the driver free-standing (and the back bowed out as I received it). I then clamped a couple of aluminum bars across the back, touching the middle of the back plate, but spaced out about 1/8" on each side, which reduced the bow a small amount, but did not bring it flat, so it still has some bow.
As you'll see in the attached FR graph, the output level increases significantly with just a small reduction in the bow (shown in blue).
These graphs were made with no baffle, and with a patch cord approximation of the original crossover, so don't draw any conclusions from the absolute frequency response, just the relative measurements. And those measurements confirm what I and several contributors here thought: the output level problem is a function of the bow in the backplate, which is significantly greater in this midrange than my other channel.
Next Steps: I've designed a mechanism to route the MDF panel to mount a number of crossbars as well as clamps to held the driver in place in the MDF baffle. Once that's done, I can adjust both the number of crossbars and the tension on the crossbars to match the output level of my other channel before buttoning it back up. That will take a week or so...
Mike
PS: I was checking the woofer panels for bowing, and noticed they in fact have metal cross bars, very similar to what I plan to use on the midrange, to keep those wide panels flat.
Edits: 04/11/21
.
Side clamp design proposal.
This so NO HOLES are drilled in the panel, which I've that others do. Panel should be floated in the frame and I'd use a special 3M Teflon Tape (not plumbers tape) to allow some easy movement.
2nd shot is a variation using a 'bowed' piece of wood as a spring. I've been expermenting with steam bending, so this is possible.
Too much is never enough
interesting thread I've been following; I can't contribute anything to your specific issue but I have experienced the 'unbalanced' Maggie before in the past. More specifically my MG1.6's back around the turn of the Century. Every blue Moon or so one panel would 'drop' about a decibel (I had no testing equipment) to the point where I had to adjust my balance to center the sound stage.I called my dealer when this initially occurred (bought new). He said he had no idea although he wasn't surprised since I wasn't the only one who had this issue. The only advice he could offer was to ship them back as they were still under warranty. I contemplated that for while and then without notice they popped back into normal operating mode- as if nothing had happened.
I never consider my upstream gear as being the culprits since this occurred over long periods of time (again, not often at all; every blue moon or so) with different gear.
Eventually I simply wrote it off as an anomaly attributed to either the XO or the Quasi ribbon in general (don't ask me how lol)..I haven't heard anyone with the 'i' series having this issue however.
Edits: 04/09/21 04/09/21
Temperature is a factor to keep in mind. If one of the speakers stands in a warmer spot, it will probably sound different to the one in a cold spot.
New data point: I removed the ribbon and the grille to inspect the mid panel. I found that the perforated metal backer panel of the midrange, to which the magnets are mounted, is bowed away from the mylar, over 1/4" in the center. That extra distance from the magnets to the mylar and voice coil is likely the cause of loss in efficiency.
I'll see if I can design some braces to hold that perforated panel closer to flat.
Photo attached.
Mike.
to allow for mylar excursion without the dreaded 'Maggie slap'?
But yours is certainly bowed waaay more than mine. :((
Andy
I'm not so sure the panel needs to be bowed (notwithstanding the fact that Magnepan says a small bow is normal). The woofer panels are much wider, and they have little or no bow, and surely the woofer diaphram moves farther than the mid.
Mike
Please, notice that if you try to straiten the perforated sheet metal, there is a risk the Mylar stretches or even break!
The T-IVa mids are the only Magnepan that I have with a curved back plate.
I think that is how they all are curved on the back from the factory. It is also the reason to why the Mylar leave the double-sided tape. Ask Magnepan if this is normal. I think mine, Josh's and Andy's have this curved back.
Roger-
Thanks, I've written to Gary at Magnepan. I'm thinking even if the mid panels normally have some warp to them, that this example may be a bit extreme.
My other channel (with the higher output) seems to be pretty flat.
Mike
Mine looks like this, about 1 mm warp/curve/bend.
Roger-
Thanks for the photo, very helpful. I just heard back from Gary at Magnepan, and he indicated a slight bow (about 1/8") was typical.
Mine is over twice that, so I plan to do some measurements this weekend comparing the panel as it is now and with some braces holding it closer to flat to confirm that raises the level.
Mike
My mids differ in output by about 3 dB. The distance Mylar to magnets is too large in one of them. I will rebuild mine, sand down the spacers so a flat surface will face the Mylar in order to make a secure contact with the double-sided foam tape.
Roger-
Since my mids were recently rebuilt at the factory, I'm planning on another approach (that doesn't require replacing or re-wiring the mylar) to move the magnet/perforated panel closer to the mylar (and closer to flat). This weekend I'll be testing to see if adding cross bars at 2 or 3 locations to the rear of the panel, pushing it closer to flat (but not quite there, per Magnepan). If that raises the output level, I'll route out some grooves in the MDF panel to mount the bars, which will also help hold the panel in place.
I'll post my progress.
Thanks,
Mike
I would've thought 4 positions would be necessary, for the stiffness & length of the mid panel?
And presumably these crossbars should not be made of steel?
Andy
andyr wrote:
"I would've thought 4 positions would be necessary, for the stiffness & length of the mid panel?
And presumably these crossbars should not be made of steel?"
Why not steel if stiffness is required? The slightly curved back of my T-IVa mids are stiffer than any other of my Magnepan drivers.
But now I think about it, Roger ... Magnepan used steel crossbars on my Maggie bass panels - so it can't be a problem!
Andy
Andy-
I'll experiment with the number of braces. Aluminum is usually my choice when doing metal work, as it's easier to machine. As a side benefit, it's not magnetic.
Mike
I'm afraid I will not be much help but looking at the schematic that you posted one score years ago, the mid-driver does not appear to be doubled. The mid impedance is marked at 5 ohms + the DCR of the coil. A better measurement is to remove the tweeter fuse so that the tweeter does not interfere. However, how this helps diagnose the problem is beyond me <(. I assume you already did the various cable and channel switches so you are convinced the problem is in the left channel speaker and not upstream. This is supported by the normal behavior of the tweeter and woofer.Addendum: Here's a thought. Pull the mid-fuse on the right panel and replace it with a 4 ohm resistor (the fuse is useless anyway). That will drop the output by ~ 5db : 20*log(5/9)=-5.1 db. Not a cure but a work-around. The remaining issue would be the balance of all three drivers.
I married the perfect woman - the downside is that everything that goes wrong is my fault.
Edits: 04/07/21 04/07/21 04/07/21
Thanks for your note. You are correct, that all upstream components have been ruled out. I checked my original TIVa schematic from years ago, and although the online copy is difficult to read, I believe the original shows the midrange resistance as 3 ohms (exactly what mine measures), not 5 ohms.
I like your work-around with the additional resistor to reduce the other channel, however the bad channel exhibits a suck-out in the midrange relative to the woofer and tweeter, so I'm afraid I'd then make both channels sound bad!
Best regards,
Mike
Correct! :-(( And the tweeter.
3 ohms is correct - that's what mine measure (well, actually 3.2 ohms).
If you convert your T-IVas to active, then you will be able to boost the faulty mid panel back to 'normal' volume.
Andy
The midrange driver without the crossover should measure about 3.6 Ohms. There have been a guy who let Magnepan re-wire the mids of the 3.6 with foil conductors. I did it myself, it was the smallest cross section foil, the one used in the mids of the 20-series. Will give the same impedance (resistance) but slightly lower output (as the foil is not sitting in strongest spot of the magnetic field). My 3.6 shows a difference between right and left speaker, tweeters are 2.5 dB hotter on one side. I am not impressed by the quality of work of Magnepan (I have four sets). One of my foiled T-IVa mids has also a lower output, the Mylar is further away from the magnets. It seems to be common that the Mylar leaves the double side tape, that will change the performance quite a bit.
I would not expect significant output from at 20 Hz and 10 kHz from the mid panel so I assume you are running all panels and the tweeter?
Since there is a crossover on each side it pretty unlikely the problem is a speaker.
On the plot you posted the right channel tracks the left channel so left and right speakers are tracking. Left channel is ~5 dB low across the band. That looks like the problem is the input to the left speaker, so amplifier, cable, interconnect, preamp, source.
Have you tried switching cables, left to right? Switch at one place at a time. I would start at the source and switch left for right. (Be sure to turn everything off when you do. LOL) Then the preamp to amp and then the speaker wires, those to the crossover box and from it to the speakers. If the low speaker follows the change you have narrowed down the problem. If not it must be the left speaker.
Steve-
Thanks for you suggestions, all those diagnostics have been done. Definitely not upstream, the FR measurements were in fact made from the same amplifier channel, same source, etc.
The problem is in the left mid/tweeter panel, no question.
Best regards,
Mike
That's a sad tale, Mike. :-(( Not only because it has left Magnepan in that state, after a re-wire - but also on account of the lack of response you have gotten from their people.
I had a similar problem of a mid driver on one side outputting less than the other one about 20 years ago, when I re-wired one of my mid panels (due to a slip with a screwdriver when I was screwing my IIIa panels into the new hardwood frames that I had a furniture maker construct). :-((
After glueing the new mid-range wire down, I found it had less output than the other one - which destroyed the central image.
I put this down to the fact that I had put too much glue on the mylar, when I was doing the re-wire and the extra mass of that mid reduced its output.
So I suggest you should have a close look at both mids and see whether the left one looks like it has more glue than the other.
If so ... re-doing the wiring is the only solution.
Good luck!
Andy
Andy-
Thanks for your note. Your suggestion is, unfortunately, exactly what I fear the problem may be - some error in re-furbishing the midrange at the factory. My next step is to dis-robe the good panel (another hundred or so staples!) and compare the two mids visually. Electrically they seem identical.
I've owned a number of Maggies over the years, and although I do a lot of DIY audio, I've managed to avoid having to re-wire a panel this far. Maybe my luck has run out!
Mike
It's not something I would ever attempt again, Mike! :-((
It's a shame that it was done by the factory .. but I suggest getting them to re-do it will be your best option.
Andy
PS: if you get this done ... maybe consider asking them to use QR foil rather than round-wire?
Andy-
I'll ask, but I've heard they'll only use the foil on speakers originally designed using it.
Thanks,
Mike
.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: