Problems with Scanspeak Illuminator 12MU/8731

[Lojzek]

Hi Mercury,

could you post a separate measurement of both midranges in parallel low
passed, then the tweeter high passed, and finally their sum without moving
the mic and leaving like 1 m away from speakers at the tweeter axis
(about the listening height)?

Measure only one channel. Try to present the mid LP and tw HP in the same
graph.

 

[Mercury3]

Hi Mercury,

could you post a separate measurement of both midranges in parallel low
passed, then the tweeter high passed, and finally their sum without moving
the mic and leaving like 1 m away from speakers at the tweeter axis
(about the listening height)?

Measure only one channel. Try to present the mid LP and tw HP in the same
graph.

Yep I can do that. I work 12 hour shift so might be a few days again when I have another day off.

 

[Mercury3]

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.

Okay I'll order a couple of 200uF caps. That'll give me low enough frequency I'll just leave them installed. I think can't hurt and the tweeters already have one.

 

[andy19191]

The driver is used in some good speakers and so the problems you describe are not going to be due to a design defect. You might have incorrectly manufactured drivers but it is unlikely. The most likely cause is an implementation issue.

"Very expensive tube amplifier" could be the source of a lot of problems. I would recommend replacing it with any reasonable class AB solid state amplifier while performing the diagnostic tests. If the impedance of your previous setup was different to the roughly 4 ohm load you now have then you cannot reasonably draw the conclusion that the valve amplifier is not the source of the problems.

You seem to be describing something like a resonance but are showing heavily smoothed plots which are not going to show resonances. I would suggest changing your measurement setup so that you can first see the problem and then try to work out what is causing it.

When you listen to the midrange alone do you include the low pass filter? The cone is fairly hard and so the breakup will need attention.

 

[Mario Pankov]

I have a few questions:

- have you got off axis measurements, up to +/- 75 degrees in 5 degree increments would be sufficient
- how close to walls or furniture are the loudspeakers in question
- do you have floor carpets or anything soft
- what crossover have you used for the tweeter

 

[eriksquires]

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.

This is the best experiment yet! I am not sure about the battery, but please leave them out of the cabinet and play full range music through them (at a reasonable volume) without the XO. If you still hear the issue then you have eliminated all of the cabinet, and the crossover. Also, just to be sure, you aren't using esoteric cables with high capacitance or inductance are you?

If you take the drivers out of the cabinets, and a battery as well as music still causes the issue then you pretty much know where the problem is.

If it wasn't for the 9V battery tap ringing in the same way I would suspect your amplifier as a possible culprit as well.

 

[Mercury3]

Okay thanks for all the additional suggestions. I want to be wrong about the drivers but I wish one of you could hear them right when in hand and just hitting the terminals with a battery. That ring sounds same frequency I'm hearing when female voice hits certain tone at high enough level. I've always done that batt. check just to make sure I'm phasing it right when I go to connect the wires and these are the first drivers I've ever heard do that.

However I am novice at speaker building. Many of you are experts so I'm hoping I'm wrong and think I could be wrong even though it looks like I'm not.

I need to play with and learn how to use the mic better which I will do. Then I'll try to get some additional graphs which have been suggested.

Incidentally when I very first encountered this I did think I had a tube ringing. Microphonics. I replaced all the tubes and bought some tube dampers. When that made no difference I did swap amps and checked with my transistor amp but it still made no difference. Not amp or tube related.

The crossovers are the Pass Labs XVR-1 and I can do a lot of adjustments with them. I thought if it might be ringing caused by the crossover I was hearing but low Q at 6db should almost eliminate the vast majority of crossover ringing. And the ringing is the same weather 6db or 24db slopes. It does sound like ringing/resonance rather than a peak.

I suppose I could be wrong but I sort of know what to listen for. I've been an audiophile for 35 years and used Altec and JBL horns/compression drivers back in the olden days so I've heard very bad peaks/irregular response curves and it doesn't sound like this. It's like a hiss/resonance being added to the sound but only when female vocals hits certain tone at high enough level. With the volume down it won't occur.

Tend to have to agree more people would be noticing something if it were the drivers though. I'll keep pursuing all of these suggestions and I appreciate everyone's efforts in attempting to help me with this.

By chance no one lives in the Denver area I suppose? It'd be nice if someone else could hear what my problem is.

 

[Lojzek]

Okay thanks for all the additional suggestions. I want to be wrong about the drivers but I wish one of you could hear them right when in hand and just hitting the terminals with a battery. That ring sounds same frequency I'm hearing when female voice hits certain tone at high enough level.

The 9V battery test is unnecessary high voltage just to test the phase.
That's 14 W of heat dissipation. 1,5V is enough.

You should hear only a short crackle sound the moment you touch the
terminals with battery and then the cone travels to some degree and
stops right there until you disconnect it.

Can you confirm that the voice coil is not rubbing while playing a very
low frequency sine wave signal (few Hz) of low voltage just to make the
cone move a couple of mm? The impedance graphs looks normal except
the high frequency part where the curve starts falling down (Zobel?).

edit: Try to simulate the TS parameters in your box at 14W of input
power and you should see how much of voice coil travel it would make
at certain frequencies. Could it be that you exceeded the mechanical Xmax
value?

 

[Mercury3]

The 9V battery test is unnecessary high voltage just to test the phase.
That's 14 W of heat dissipation. 1,5V is enough.

You should hear only a short crackle sound the moment you touch the
terminals with battery and then the cone travels to some degree and
stops right there until you disconnect it.

Can you confirm that the voice coil is not rubbing while playing a very
low frequency sine wave signal (few Hz) of low voltage just to make the
cone move a couple of mm? The impedance graphs looks normal except
the high frequency part where the curve starts falling down (Zobel?).

Thanks I'll add this to my list of things to check. One thing has been in the back of my mind. I ordered these four drivers from Madisound. One of the boxes was missing part of the box material and looked like it was opened. I realize madisound has to test for their LEAP and so thought maybe it was that but if so Madisound would have replaced all the cardboard/plastic in the box which was missing.

I need to test all four drivers and see if I possibly only have two bad drivers which may have been returned by someone prior to me purchasing them. However they did not show any signs of being mounted.

Question can the Zobel pull down the upper frequency response on the tweeter?

 

[Lojzek]

An RC filter in parallel with the tweeter with the right values
can indeed pull down the FR similar to being low passed.

 

[Mercury3]

An RC filter in parallel with the tweeter with the right values
can indeed pull down the FR similar to being low passed.

Thanks that's interesting. I'll have to check that response with zobal removed.

I'd like to note that these tweeters sound damn good. I ran them on female voice well into the midrange just to test them out. Actually I was checking to see if any of the anomaly came through to the tweeters but no it didn't. I'm very happy with the tweeters.

 

[Lojzek]

Thanks that's interesting. I'll have to check that response with zobal removed.

I have missed to add that in order for the RC network to have any
effect on FR, there has to be ohmic resistance before the network
in the sense of a high pass and an attenuation L pad or so. Sorry.

 

[infinia]

I have missed to add that in order for the RC network to have any
effect on FR, there has to be ohmic resistance before the network
in the sense of a high pass and an attenuation L pad or so. Sorry.

depends what you mean by "work"
properly designed they are nothing more than a conjugate match at the upper end of a drivers inductance Le or Le'.
I works but whether it helps sonics depends what precedes it, low DF / feedback amps sure.

 

[Mario Pankov]

I think the Denmark built drivers are all individually tested, I would expect this at this price level. Never heard anyone complain of defective Scanspeak drivers. But they may have had a rough life while they got to your address, hope this is not the case.

I asked you for offaxis response and placement because it might be your problem, if the drivers turn good. I noticed you`ve done only 0 degree measurements, while the response could be ruler flat, off-axis response may have a peak in the 4-7Khz region, which would usually cause some female voices to sound like you explain. The off-axis energy would bounce from walls and anything around the loudspeaker and depending on whether the drivers are physically time aligned, also from the floor, and the summed response at your listening position may have too much energy in the band you describe. Our brain is amplitude sensitive to reflections and at lower levels will ignore most of the bounced energy that does nto arrive in phase with the main signal but at high levels this changes.

I have no experience with this tweeter but notice it has a steep rise in the impedance at Fs probably due to the lack of ferrofluid in the gap. Similar Scanspeak and Vifa drivers tend to ring unless this is addressed and you are active so this has not been taken care of. Depending on how audible it is, a parallel resistor may do it, if not - an RLC network usually helps.

Do you get the same problem with tweeter disconnected?

 

[Mercury3]

I think the Denmark built drivers are all individually tested, I would expect this at this price level. Never heard anyone complain of defective Scanspeak drivers. But they may have had a rough life while they got to your address, hope this is not the case.

I asked you for offaxis response and placement because it might be your problem, if the drivers turn good. I noticed you`ve done only 0 degree measurements, while the response could be ruler flat, off-axis response may have a peak in the 4-7Khz region, which would usually cause some female voices to sound like you explain. The off-axis energy would bounce from walls and anything around the loudspeaker and depending on whether the drivers are physically time aligned, also from the floor, and the summed response at your listening position may have too much energy in the band you describe. Our brain is amplitude sensitive to reflections and at lower levels will ignore most of the bounced energy that does nto arrive in phase with the main signal but at high levels this changes.

I have no experience with this tweeter but notice it has a steep rise in the impedance at Fs probably due to the lack of ferrofluid in the gap. Similar Scanspeak and Vifa drivers tend to ring unless this is addressed and you are active so this has not been taken care of. Depending on how audible it is, a parallel resistor may do it, if not - an RLC network usually helps.

Do you get the same problem with tweeter disconnected?

Yeah the tweeter actually sounds great. I even ran it as low as 800 hz by itself low level just to test it. Nothing bad coming from the tweeter at all. In fact I'd go so far as to say these are the best tweeters I've ever owned.

Anyway just waiting for a few days off again and I'll get back to testing. As I said I need to play with the Mic more. It's not quite as intuitive as the directions say it is.

 

[Mercury3]

FYI to those helping me with this.

I am on hold for a little while because I sold two of my power amps the other day. Mean while I have two other nice vintage amps I'm going to tear down and replace the large power supply caps and if I find it's not to difficult any other caps I can get at as well. (The amps are 25 years old)

So I'm gonna be side tracked with this project and then get back to my speaker issues once that's completed.

Thanks!

 

[Face]

A distortion sweep could be helpful as well.

 

[Mercury3]

All right well probably some will be disappointed with me but nothing was working so I ordered four new drivers from Madisound to fit the the cutouts I already have. It's a radically different driver but I stuck them in and they are sounding fricking awesome and I haven't even done any testing yet. Just some by ear adjustments.

I used four of the Fostex FE108EZ. They sound incredible and I like them a lot. I can be happy with them. Only thing I worry is their low Xmax and power handling capability. Currently I have the crossover at 290hz 12db but I think I'll raise it up to 360hz which will give me a little safety factor when I crank up the volume. Gonna do that and get to testing/adjusting.

I don't know what the hell about the ScanSpeaks. I don't have countless hours to keep spending on playing with them and trying to figure it out.

The problem is 100 percent gone with the Fostex replacements. I didn't have any trouble with the Focal midrange drivers I had prior to the ScanSpeaks either. Honestly I don't know for sure but think something is wrong with them......

As time permits I'll put them back into my old boxes and see if I can get the proper measurements and still try to figure out.

Thanks to everyone.

 

[system7]

I think you may have driven the original drivers into a rocking mode by overdamping the voicecoils with direct drive (high damping factor) from the amplifier and active filters.

As I visualise it, the voicecoil doesn't have any freedom to move in and out, so cone energy escapes into a sideways mode. This can even produce voicecoil rubbing IMO. This happens less with traditional coils and resistive shunts. I could be wrong though.

 

[Face]

I'll gladly take these defective drivers off your hands if you're going to bin them.