New Tweeters for my Mets

Dedicated to those large boxes at one end of the room
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#31

Post by Cressy Snr »

I hear you Ed.
There is a big problem, trying to get a speaker sounding right simply by ear alone.
As Lynn was saying earlier on in Mark's thread, you can eq the hell out of a Lowther EX4 to get it sounding even across the frequency range, but then you don't have a Lowther anymore, so why did you spend all that dosh when you could have saved yourself a whole load of cash and bought a bass unit, a mid and a treble, used a three way crossover and still had change?

I'm going on female speech and singing voice, to get the sound right, but again, I don't know whether the voice on the recording hasn't been eq'd itself, so I am trying to get the vocal right across a range of female voices.
But then I get all the voices right and the sound bores me to tears on everything else.
There is a setting somewhere that keeps things reasonable without sucking all the life and dynamics out of the music. I have not found it yet and I suspect I never will. I don't have the experience/skill yet, to be able to balance life and dynamics with neutrality.

I think maybe I am wanting the impossible and need to pick a sound I am comfortable with, on most of my music, stick with it and stop expecting everything I listen to, no matter how badly recorded, to sound acceptable.

I know you tried to tell me a few years ago about the futility of this kind of stubborn approach and I am beginning, not before time, to realise that doing things that way can only drag everything else down to the level of the lowest common denominator.

I have taken the eq out and my Mets sound like My Mets again except with a gorgeous treble. Maybe we'll leave it at that. Dusty and Petula will just have to shriek on a couple of their singles. I listen far too loud half the time.

Food for thought. I think.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#32

Post by Cressy Snr »

OK now a very interesting development.
I began to think back to my first experiments with my original concept quadratic tapered, quarter wave resonators, eg my original Fostex FE108ES driven Metronome cabs from way back in 2006.

I began to think about when it was that the original design actually fell into place, ie when the open end, with Scott's help, was fitted with a port to mass load the terminus, turning the cabinet into a hybrid of a 1930s, quarter-wave Voigt horn and a mass-loaded transmission line, due to it being open at the wide end.

At that time, the open ended cabinet had been stuffed and lined to damp the ripple. Scott had suggested mass loading the open end and i had carried out the mods, only to find tat the speaker was quite frankly bloody awful; shouty, screechy and lifeless. It was only when I removed most of the damping from the cabinet, leaving only the back panel lined with carpet underlay that the speaker began to sing. The hybrid, horn/ transmission line nature of the cabinet makes stuffing the cabinet a counterproductive exercise, at the most only the top third should be damped with acoustic wadding, with the back lined.

To cut a long story short, I have taken out the all of the stuffing I specced for Colin's commission build from the top half of the cabinet, leaving only the back lining. The sound is now in a completely different league. I am speechless to be honest. Bass detail and tone is fantastic, and ironically is more even-handed from the top to the bottom of its range than it has ever been.
No need at all for any digital hocus-pocus. A purely analogue, acoustic solution.

Thanks Ed for prompting me to think a bit more deeply about the Metronome design as a whole, rather than a collection of parts. I've lived with my design concept so long, I just took it for granted that everything was right with it except the crossover; not so.

Yes the Fostex FT17H still got on my nerves and the Monacor ribbon is a big improvement, but allowing the cabinet's horn characteristic to have a free reign has done wonders for the sound of the whole thing.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#33

Post by IslandPink »

I'm elsewhere at the moment ... quick reply - but I must say that getting female vocals right is a nightmare - not just with speakers but with the whole audio chain !
One thing I did notice about the (passive) notch filter on Spice, is that the phase shift around the notch can be quite low ( <15 deg ) if the notch is shallow ( eg. 3-4 db ) and adjusted for the inductance of the driver.
"Once you find out ... the Circumstances ; then you can go out"
chris661
Shed dweller
Posts: 2556
Joined: Mon Feb 22, 2010 7:29 am
Location: Sheffield

#34

Post by chris661 »

If its still a bother, use a notch filter at the speaker end, but use a variable resistor (half an L-pad will do) in place of the fixed-value resistor.

Dial it in for female voices, see how it goes, lower the resistance, see how you like it, lower a bit further....

You get the idea. There'll be a balance that'll do everything more-or-less right. Hopefully.

Chris
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#35

Post by Cressy Snr »

I think I have found the sweet spot with the FF225WK when mounted in the Metronome cabinet.

Applying a parametric eq filter of -2.5dB at 3600Hz, with a Q of 1, via JMRC's DSP tools, gives just the right amount of cut to tame the midrange peak, without sucking all the life out of the sound.
I went from -5dB/3600Hz with Q=0.6 to the present setting, over the course of this week. Amazingly you can hear the effect of as little as 0.5dB alteration in this critical region.
-2dB and there is a slight hard edge on one or two female vocals, -3dB and the edge goes, but so does the life, -2.5dB and you hit the sweet spot nicely.

All that needs doing now is to make the equivalent passive LCR, parallel dip filter. That is going to need a bit of trial and error around the theoretical component values to get the right voicing.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
chris661
Shed dweller
Posts: 2556
Joined: Mon Feb 22, 2010 7:29 am
Location: Sheffield

#36

Post by chris661 »

Might be worth using a Zobel network to get the notch filter to play nice with the driver's inductance - sound mean the theoretical values are pretty much bob on.
FWIW, measurements I've seen of the Fostex unit show a ~6dB peak at 3kHz. Not sure how it looks off-axis, though, but ~3dB sounds somewhere near.

If I make it to Owston, I'll bring my speaker measuring stuff - its quite handy to see what's going on, if the notch filter's still a work in progress at that point.

Chris
User avatar
Scottmoose
Needs to get out more
Posts: 1802
Joined: Sat Sep 08, 2007 11:03 am
Contact:

#37

Post by Scottmoose »

Out of curiosity, what measurement gear are you running at the moment Chris?
chris661
Shed dweller
Posts: 2556
Joined: Mon Feb 22, 2010 7:29 am
Location: Sheffield

#38

Post by chris661 »

It's the cheapie Dayton phone/tablet measurement mic. Come with calibration file, so no complaints from me.

Some reading here: http://www.diyaudio.com/forums/multi-wa ... nture.html

Chris
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#39

Post by Cressy Snr »

OK a bit more progress on the Mets.
The crossover cap value has been increased to 1.68uF and the parametric eq turned off.
As Romy has hinted at, getting a first order crossover, right is a difficult job, but not impossible; you have to juggle the shallow slope and the relative driver efficiencies until the two drivers blend together.
1.68uF gives a 12KHz corner frequency with this pair of drivers, which means it is 6dB down at 6KHz. The widebander is 94dB efficient and the tweeter is 98dB which means that the difference at the crossover frequency is +1dB, but the drivers are out of phase, which smooths oout the crossover region nicely.

We are 12dB down at 3KHz and out of phase which offers a degree of cancellation of the peak in the widebander's 3-4KHz region.

To my ears, the FR seems flat, through the crossover region and HF sparkle is nice and bright. For example, harpsichord continuo in Vivaldi, Johann Friedrich Fasch etc chamber music is much more clearly defined and we are offered a nicely spacious insight into the actual room the music was recorded in.

I have decided to use exclusively classical music to set up the crossovers in my speaker projects, as getting classical right seems to have improved the presentation the speaker gives of other kinds of music too.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
User avatar
Scottmoose
Needs to get out more
Posts: 1802
Joined: Sat Sep 08, 2007 11:03 am
Contact:

#40

Post by Scottmoose »

I'd be a trifle careful on assuming the manufacturer sensitivity ratings are accurate Steve. The RBT-95SQ was measured by K+T at about 95dB 1m/w. Also, it's naturally rolling off at ~3rd order < 4KHz, so at 3KHz, you'll be down about 18dB from the baseline ~95dB sensitivity through it's own acoustic rolloff combined with the 1st order electrical high pass.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#41

Post by Cressy Snr »

Cheers Scott,
that's actually further down than I expected, but that's even better, if you see what I mean.
However in theory the peak from the widebander is not pushed down as much.

Yet it does sound like it is down, unless the reason is that everything above it has gone up, so we now have a shelf lift above 4KHz instead of a load of peaks; which might make more sense.

I like playing with speakers :)
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Cressy Snr
Amstrad Tower of Power
Posts: 10552
Joined: Wed May 30, 2007 12:25 am
Location: South Yorks.

#42

Post by Cressy Snr »

Post Owston, I've been tweaking up the Mets and meeting with great success.
I've slightly altered the bass/room alignment by trimming a half-inch off the port lengths.
The mass loading ports have been reduced from 1.5"x 3"to 1"x 3"and the results have been encouraging. The improvement in the quality and tone of the bass has been very useful and has had a knock-on effect on the midband resolution, which has been kicked up a serious notch.
Voices are more focussed, reverb tails go on for longer and the central image had firmed up nicely.

That's the nice thing about Metronomes, they are eminently tweakable. Small changes at the down-firing port end have a far from subtle effect on the overall presentation.

I was prompted to sort the speakers out, following the thrashing they got from DTB, Mark and Thomas. Despite the music they played not being exactly my cup of tea, it was useful as it pointed up a slight cloudiness to the low bass, that I had been aware of, but had done nothing about as TBH I have found it too easy to just play music and enjoy them since they were built three years ago, especially since the Monacor planar magnetic HF drivers were fitted. That and I'm a bit of an idle b'stard, when it comes to fine adjustments where speakers are concerned.

As is often the case with these sorts of tweaks, sorting out bass issues seems to improve the rest of the spectrum, especially up top, where there had been a peak that could be set off by certain notes played on chime bars.

Chris (Stratmangler) had noticed this at Owsters during my dem and after he had pointed it out, I too noticed the problem on, of all things, King Crimson played by Mark, rather loudly, through DTBs amp; or maybe it was Thomas's. There were one or two guitar notes from Robert Fripp, that had me wincing.
Anyway, this peak seems to have gone.

Speakers are funny things and often, changes made in one area have unexpected results in another. Quite how tweaking the bass alignment, could tame a treble peak, I'm not sure, but these things happen all the time with speakers, which as a breed are much more difficult to get right than is often appreciated.

The sum total of these speaker mods and the balanced PP pentode amp developments is that my system continues to move further towards my ideal of neutrality, even-handedness and absence of character. IOW letting the music through unmolested.
Sgt. Baker started talkin’ with a Bullhorn in his hand.
Post Reply