Problems with Scanspeak Illuminator 12MU/8731

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've recently completed a three way system using 12" woofer, 2 of the 12MU's per channel and dome tweeter in MMT configuration. All drivers properly counter sinked flush mounted.

My tweeters are the Scanspeak Illuminator D3004/6640

Being a novice I chose to use active crossovers (2 Pass Labs XVR-1) with multi amps to avoid problems designing passive crossover.

I used 2 midrange drivers per channel because I do like to play my music loud at times and my thinking was to bring the efficiency up a little bit closer to the Woofer and Tweeter which are both higher than single 12MU.

My baffle step on 16" width baffle ends a little over 300hz so I'm using my crossover point @ 360hz. I used the edge program for best location of drivers on the baffle.

On the top end I've experimented a lot between 2200, 2500 & 2900 using both LR 12 & 24 none of which makes any difference on the problem I'm having.

I used the proper value resistor/cap to flaten the voice coil inductance. I tried with or without but it made no difference either.

I also tried the midrange drivers in a much much smaller parts express curved box with only a 7" wide baffle just to experiment and it made no difference to the problem. I'm positive it's not a box or baffle problem. (at least on the outside)

Here is the issue:

On female vocals when certain sustained notes/pitch is hit it sounds like some harmonics or some type of ringing is being generated. This is occurring only in the midrange speakers as determined by listening only to the midrange drivers by themselves.

It's also very noticeable with certain instruments such as flute. With some kinds of music it's not noticeable and the drivers sound very good. You can't detect it with most male vocals. Even some female vocals sound good. It's only when they hit certain notes or frequency at little bit higher levels.

Prior to this I was using Focal midrange with the same electronics and I did not have this problem so I know it's not in my electronics anywhere. I did swap some amps and things to check as well but again no difference.

In my problem determination attempts I bought an OmniMic V2. With the mic and software I was able to make some better adjustments to levels and things but nothing bad is showing up that would indicate a problem. Frequency response looks fairly flat in the midrange. It has a slight rise starting approx. 750hz to about 1100hz but it's only about 3 db. I don't think that's my problem. The only other irregularity is a slight fairly sharp dip between 1700 to 1800hz. It's a very thin dip so I'm pretty sure that would not cause it either.

Over all response is fairly flat with most of the bumps/dips all in the bass due to room nodes and things which are expected. The tweeters start to roll off a little after 6k but that's another issue and they still sound good regardless.

So I'm about at the end of my rope.

I'm thinking there might be a design flaw in these midrange units. The flat ring that holds the spider looks like it's too close to the cone and to large around. I had to slope the inside of the holes and even with that it doesn't look like a lot of room for the back wave sound to flow around that ring on a 3/4 inch baffle. I think possibly I'm getting sound from that ring and spider bouncing back to the cone at certain frequencies but probably not enough which can be measured with my mic.

My next idea is to try to slope the inside of the holes some more but I'll be down to about 1/4 inch of wood left for the screws. I'm only speculating and that might not even be the problem.

Has anyone else successfully implemented these drivers? I've search and not really finding a whole lot.

Thanks!
 
Test for non-linear distortion with your OmniMic.

It's also useful to see the Impedance trace with a tool like DATS - and DATS has a "rub-buzz" test for problems with vc position and surround problems.

My guess is that it's either defective drivers or you are getting a baffle resonance due to a lack of proper chamfering. (..and more likely the later.)

Chamfering example around driver (..from Zaph's "Blog" section - note how even the inset on the test baffle uses a felt surround - something to think about in terms of the interior cabinet):
 

Attachments

  • Scan-Speak-10F-4424G00-rearchamfer.jpg
    Scan-Speak-10F-4424G00-rearchamfer.jpg
    142.1 KB · Views: 1,271
Last edited:
Test for non-linear distortion with your OmniMic.

It's also useful to see the Impedance trace with a tool like DATS - and DATS has a "rub-buzz" test for problems with vc position and surround problems.

My guess is that it's either defective drivers or you are getting a baffle resonance due to a lack of proper chamfering. (..and more likely the later.)

Chamfering example (from Zaph's "Blog" section):

Thanks,

Yeah I did the Chamfering already. My only thought was taking even more off but I won't have much wood left. I think that spider ring is really in the way of the back wave sound.

I did buy the DATS with my OmniMic so I could experiment with it. I haven't even opened up that box yet. I'll also have to learn more about testing with my OmniMic for the non-linear distortion and other tests it's capable of.

Okay with work it'll take me some time to get to this but I'll check into all three of these things.
 
I'm thinking there might be a design flaw in these midrange units.

My guess is that if there is anything wrong with it, it's the XO filter.
Ever tried building a 2 way with one 12mu and tweeter passively?

You ought to get a dsp based active filter product.

Once I had a chance to repair a passive 3 way diy speaker project
that had a sort of reverberating female voices at certain frequencies
and I've noticed it on Spanish Harlem by R.Pidgeon. It turned out
that midrange-tweeter filter was badly designed. The driver units
were Focal polykevlar.
 
For a TMM topology with both MM crossed at 2200Hz+ lobing and comb filtering effects of polar beaming and uneven power response can be audible. The sound stage "moves around". Some notes in a chords sound louder/softer as more power beams into a smaller room volume. To the listener, a slow sine wave sweep can change sound stage position and amplitude.

You might experiment with disconnecting the lower M and re-adjusting SPL levels for TMW.

You might decide on a 3.5 way crossover, where the lower M has a lower crossover frequency to supply the extra cone surface area needed at the lower midrange frequencies.
 
Thanks guys.

I just finished increasing the Chamfering behind the driver rear holes. I have it at least as good and maybe a little more than the pictures on Zaph's Blog that Scott showed.

I'd hoped that was it but sadly no change.

I think the next easiest thing to try is disconnecting the lower driver as LineSource suggested.

Gonna try that and I'll report back.
 
Update:

I ran only the single top midrange (MT instead of MMT) and problem is less pronounced but absolutely still present. I think it's so noticable to me because of using 4 drivers)

Hooked up the 12MU's full range fed right off my CD player to power amp. (eliminated the crossovers and all other electronics) Running the drivers full range with no crossovers the same problem is very much still present. This rules out any crossover problems which I was already pretty sure was not the case.

Played with the DATS'2. Doesn't really show anything and I didn't really expect it to. No voice coil or issues at all on any of the four driver with Buzz/test. The impedance sweep looks very flat from about 200 hz up. No irregularities to that.

Resonant frequency in the cabinets approx.90hz (doesn't really matter I measured it so thought I'd list)

I'll play with my OmniMic V2 more when I get some time. As I reported earlier though the frequency response sweep actually looks pretty damn good after playing around with various crossover and level settings.

I think the drivers is a bad design with that wide/flat ring completely blocking the back wave from the cones back. It's bouncing back to the cone at certain frequencies between that gap. I really think these are bad driver design and it's less pronounced when using only one so it could almost go unnoticed and chalked up to driver colorations.
 
Last edited:
I'd like to see the impedance sweep of both of the midranges
in the box without any filters or Zobels. From 10 Hz to 20kHz.

Got any stuffing to damp the standing waves in the box?

edit: the distance from the cone to the ring is less than 20 mm, right?
The standing wave would be like 8,6 kHz. You are crossing nowhere near.
 
Last edited:
I'd like to see the impedance sweep of both of the midranges
in the box without any filters or Zobels. From 10 Hz to 20kHz.

Got any stuffing to damp the standing waves in the box?

edit: the distance from the cone to the ring is less than 20 mm, right?
The standing wave would be like 8,6 kHz. You are crossing nowhere near.

I can do that impedance sweep for you later when I have a day off. I'll have to open the rear and remove the zobels to take it. Note that with the zobals it's pretty darn flat after 200hz.

Yes inside the box I have a combination of fiberglass lining the back and all sides with wool fiberfill in the center and behind the drivers.

I had also thought changing the inner stuffing as my next easiest thing to try. Perhaps lining some felt on all the walls prior to stuffing might make a difference.

Thanks everyone for the ideas. These drivers were expensive so I'd like to make them work. It's kind of hard for me to believe the design is bad but that damn spider ring sure is wide and right under the cone. I dunno......
 
Most spider designs are similar - it's almost certainly not the spider (at least when compared to almost any other driver).

What it might be however is a reflection off of the spider's support ring (driver frame) - which is a bit large. Consider applying a lossy ring of felt on it and see if that makes a difference.
 

Attachments

  • 12MU_4731T-rear.jpg
    12MU_4731T-rear.jpg
    144.7 KB · Views: 609
I would be very surprised if it was cabinet related or reflections from the driver's own basket. It's far more likely to bee the xover.

Designing active is no more trivial than designing passive. Both require good measurements for the end result to be excellent. And no textbook filters are going to give you good results.

A 3dB rise around the 1kHz mark can make things sound very forwards and fatiguing and is most certainly not something you should ignore. It would make my ears bleed.
 
5th element makes a good point on the crossover and can make or break a speaker no matter how good the drivers are.

One thing you haven't mentioned is your mid enclosure size and shape etc as this is one of the hardest thing to get right in 3-way speaker box design. Usually builders end up with 1-2 litres for a single driver which creates terrible backwave issues. I tend to go way oversize on the mid enclosure with a Qtc <0.5 if sealed. An angled back wall as well as non parallel sides and felt wall lining helps. Since the mid is supposed to be operating well above it's piston range, oversize enclosures are not detrimental to the driver's performance (unless it's operating too low if the crossover point has been chosen poorly).

I remember working with Andy G on a mid issue (P13 which is one of the smoothest around) on one of his speakers and it turned out to be a backwave issue and when the back of the mid enclosure was removed, the issue was rectified. The driver ended up being used in a TL type arrangement instead of sealed.

You could check by operating the mid driver in an OB situation in a baffle about 450 x 728 and have damping material over the rear of the driver to provide some acoustic loading as well as reduce the rear energy. Run this with the crossover so it's operating in it's expected range and any breakup or lower end loading is reduced. Running full range without the filter will give an incorrect impression.

For mid box designs, Troels Gravesen has some good ones on his site.

Some drivers can be helped with a coating on the cone but I would not suggest that on the expensive driver you are using. I use the 18W8531G00 as a mid in one of my designs and prefer it coated with 1 coat of Visaton LTS50.
 
Thanks again to everyone. I'll check into these additional suggestions when I get couple of days off work again. I got my work cut out for me but it sounds like I may eventually get to the bottom of this. Just take more tenacity on my part.

The midrange box is over sized at about 1.8 cubic feet. I noted earlier that I started out with the smallest Parts Express Cherry Curved box with only the 7 inch width front baffle. 7" by 11". I had this issue in that box which is why I went with the larger box initially thinking that was my problem. So basically I went from an undersized box to an over sized with no change in the issue I'm encountering. Regardless if it's the issue or not it does seem a good idea if I can make some modifications inside to eliminate parallel walls, line with felt etc. Currently it is rectangle shape so I need to work on that.

Again I appreciate all the feed back and will be looking into each of these suggestions as well as double checking some of my previous testing too.
 
Last edited:
This is probably a dumb question but something I've been thinking on. Would it make any difference at all if I tried wiring the drivers in Series (with required Zobal for series) for 16 ohm rather than in parallel at 4 ohm? I'm using a very expensive tube amp that does have 16 ohm taps. Is there anything detrimental to running two drivers in series vs parallel? If not it's another easy experiment I can try.

Another thing is looking at my boxes they are rectangle. I don't "think" it's a box issue but if I determine it is I'm not opposed to building new boxes again to get it right. I'm not a very handy wood worker which is why I kept it simple. I'm going to try adding some type of angled baffling in the midrange chamber to possible break up any nodes and also line the whole thing with felt too. I sort of don't feel this is the issue though because it's an identical anomaly I had in the very small Dayton Parts Express boxes 7.5" by 12". If it was box related I would think the anomaly would have changed in frequency somewhat with the vastly larger box size but it sounds exactly the same.

Thanks
 
If the mid enclosure is 50 litres, then that's not the problem. A bit of felt lining could help with some reflections.

It seems to be coming back to a crossover issue. I'd leave them in parallel if the amp can handle it

You could try testing with a SS amp to see if it's different.
 
Adding impedance curve (X-over frequencies: 290hz & 2500hz @ 24db)
Left Channel Sweep
Right Channel Sweep
Mono Sweep
Does anyone know why the Mono curve looks kind of mess up? (last one)
 

Attachments

  • Capture.PNG
    Capture.PNG
    154.7 KB · Views: 540
  • leftchannel.jpg
    leftchannel.jpg
    98.7 KB · Views: 518
  • rightchannel.jpg
    rightchannel.jpg
    98.7 KB · Views: 537
  • monosine.jpg
    monosine.jpg
    99 KB · Views: 445
Last edited:
I wonder if this could be a weird ringing in the active filters themselves. Try listening to the outputs of your crossovers with a pair of headphones or another speaker altogether to eliminate that as the problem.

I don't think so. All the electronics is exactly the same used in my prior tri-amp setup in which I had a Focal 5k4411 & an Eton dome tweeter. (same 12 inch woofers) No hint of any issues in that setup.

I'm pretty sure I got bad drivers. Something is resonating/ringing it isn't a response peak I'm hearing. It's less detectable with 2 midranges and I can see how it could almost go unnoticed or chalked up as slight driver coloration. The fact I'm using four magnifies the issue.

The above curve is bad at the midrange/tweeter crossover but it doesn't matter because it does it only driving the midrange with or without crossover full range. (I tried it again today)

Earlier today when I had the drivers out and taped the terminals with a nine volt battery (phase check) I notice they ring. They ring to nine volt battery. I'm like what the hell.

I've tried 6, 12, 18 & 24 db crossovers at low or med Q's between 1800, 2200, 2500 & 2900 with zero difference in the anomaly. No crossovers full range they still do it.\

Also yesterday I lined the midrange section of the cabinet thick with sheepskin wool and installed some sloped baffles to help break up and internal nodes. No change either.

I'm about to throw in the towel. I've already found about four other driver choices within only a few mm size wise which will fit in my current cutouts. I'm willing to put money on my problem being gone with new drivers.

I'll play with it a little longer but I'm about to order four new drivers from another manufacture. I'm looking at these choices:

Tang Band W4-1879
Morel EM 428
Fostex FE108EZ
SEAS Prestige MCA12RC
 
Last edited:
I thought that 1.5V max batteries were normally used for phase check (it was decades ago).

Drivers don't like DC voltages as witnessed when amp rails voltages come through from an amp failure.

Since you are running active, just try a DC blocking capacitor about 100uF in series to the +ve terminal of the mid driver just in case some DC is coming through from the electronics.

I can understand you wanting to throw in the towel as it's very frustrating. Try returning the drivers as they are too expensive.

You could have a look around for other designs using these mids and even talk to the designers for some opinions.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.