More Metronome Musings

[Cressy Snr]

Thanks Chris I'll try out the port extension.

Steve, the DSP is just temporary, to give me an idea of where to go with the baffle step compensation circuit. The Metronome pages on the Frugal Horns site do give starting points for BSC so a few iterations, a few measurements and a lot of patience, should allow me to reproduce or better that pink noise plot without any DSP.

Well that's the theory anyway.

[Toppsy]

Steve,
I do wonder why you are playing around with filters and tweeks when we know the problem is the fostex driver and the only way you are gonna cure the frequency responce issue above 3K is to convert the speakers into a 2-way using a tweeter that will crossover around 2K to 2.5K prefereably the lower with something like a second order butterworth. That way you are only using the fostex in the region that you know you like and have no issues. A quality tweeter like some of the SEAS will give you all the quality of sound without the fostex nasty tweeks.

PM Scott as I know (having spoken to him earlier this week) he has a particular tweeter in mind that will suit your pocket and give you the sound you like. He also has a crossover in mind too.

Just my twp peneth as I have been where you are att now with the ribbon tweeters I was using on my Norge speakers. In the end a change to the Satori tweeter was a revelation that I should have done way before I did.

[Cressy Snr]

Aye Colin, I might just do that.

It's a bit of fun I'm having at the moment, trying to learn a bit about what does what and what affects what with speaker designs. I think I have done enough to make a reasonable silk purse out of the proverbial, but it's as you say, time to PM Scott and convert the damn thing to a two way.

OK I'll lose some sensitivity, but as I no longer have any time for flea-watt amps, the reduction should not be an issue.

I may be buggering about forever if I don't

[Cressy Snr]

Chris,
Have extended the ports to 2 inches long by using a bit of corrugated card, made into a tube and stuck around the inner diameter of the plastic tubes in the bottoms of the cabs.

It's surprising just how low these things will go in my room, without booming.

By gum, I'm enjoying myself today; more than I've done with this hobby, in long time.

[Cressy Snr]

OK:
The eq has been tweaked ever so slightly and I think I have hit the sweet spot.



I'll leave it at that I think, until the measuring mic I ordered, turns up.

[IslandPink]

Good work Steve. The only way to understand something about speakers is to get stuck in as you are doing. The proper kit will help a lot.
One thing though as I discussed before - the frequencies above 100Hz will vary greatly with regard to the measuring point ( or where you put your head ) and are not much affected by the port , it's design & what it's doing, as its contribution drops off quickly above 100Hz. When you get the microphone, do measurements a foot or two either side of the normal listening point and see what happens - you can easily see 5dB or even 10dB differences in some frequencies.

[Cressy Snr]

Well in the end I decided that after all the buggering about with, they have had to put up with, over the past couple of years, the Mets needed a complete strip, service and refurbishment:



So I took off all the 'andles, and the fings wot 'eld the candles and treated the cabs to several coats of beeswax until they regained the deep, rich lustre they had had when brand new.
A bi-wire terminal strip was fitted.



This has made crossover adjustments very quick and easy.

[Cressy Snr]

I restuffed the top third of each cabinet, reinstated the LPads, recabled the internals and coupled each Fostex driver to its holey brace via a thin bead of Blu-Tak.

I had decided to do something radical and return the Mets, as far as possible (given the absence of the FT17H tweeters) to the original spec.
The original crossover cap was specced at 4uF, but I used 3.3uF, giving a first order roll-off to the tweeter high pass, starting at 6KHz.

With this arrangement, and the Lpads at -1dB I have been able to switch out the eq. Nice, easy sound with plenty of top detail and a good overall balance.
In the longer term I think I will be able to live more happily with these. Full circle, but there you go. Sometimes you have to make a journey like this, to realise that what you had at the start was actually rather good.


I've been quite busy this weekend, just drawing, sketching and doing the maths, and in addition to fettling the existing Mets, I have come up with three fully worked out, brand new, two-way Met designs, based on SB acoustics ring radiator tweeters, coupled with some nice bass/mid units (not SB)
One uses an eight inch bass/mid another uses a 6 inch bass/mid and another a four inch full-ranger with no tweeter.

I've only decided on a name for the four incher; the "QT" (cutie...geddit)

I'll post a few pics of the finished articles over the next few months as each one gets completed.

[Cressy Snr]

A bit more on the sound of the refurbished and re-crossovered Mets:

It looks as if 3.3uF is the magic number for a Monacor RBT-95 ribbon tweeter, when partnered with a Fostex FF225WK.
It enables the tweeters to pull off a disappearing act and their resultant -6dB out of phase output at 3KHz, seems to flatten the peak from the Fostex driver, giving a nice even coverage through the crucial upper-mid/lower-treble region.

OK, so the Fs of the tweeter is 4.5KHz and generally it is not considered prudent to encourage output below Fs with any tweeter, let alone a ribbon, but it sounds absolutely fine to me, which is what counts in the end.

This, I feel, is the best the big Mets have sounded. They were almost as good with the 4uF crossover cap of the originals, as built by Colin in 2011. Evidence from these listening sessions appears to show that I went too far upwards with the crossover freq, causing both the Fostex horn tweeter and the later RBT-95 to become exposed and virtually impossible to integrate with the main driver unless DSP was used.

The Mets have been a PITA since I somehow persuaded myself that crossing ever higher in frequency was the way to integrate a tweeter with the FF225WK. The result of this misguided approach was that one minute the damn things had me waxing lyrical about how fab they sounded, the next they were utter shite; frustrating (just a lot)

As Bob Brines said, the FF225WK does need a tweeter with it, and not just to increase the bandwidth of the finished system. He never mentioned how he eventually got the FF225WK to integrate, when left to run wide open, but I think I have sort of stumbled upon how he got away with it.

I have played a wide range of stuff on them, ranging from hi-rez Linn downloads of jazz and classical, to Deep Purple/Led Zep/Who, to execrable recordings of Northern Soul classics and have so far, found a lot less to complain about. That's good.

[Cressy Snr]

Another effective little mod today:

A ring of blu-tak, placed right where the aluminium dustcap and the paper cone are jointed, mass-loading and damping the edge of the dustcap and the joint between it and the voice coil former.



This has put in a mechanical rolloff that complements, rather nicely I must say, the 6dB per octave, first-order crossover to the tweeter.
The combined mechanical/electrical system works well and provides a flat response across the midrange; at a stroke, removing the last vestiges of the peakiness.

Butt ugly visuals close-up, but perfectly acceptable at normal listening distances, and the resultant sound, complemented by the ribbon tweeter is well worth the deficit in the looks department. Despite the visuals, the blu-tak mod has turned the FF225WK from a reasonably good, but flawed "widebander" into quite an impressive bass/mid unit, able to work harmoniously with a high Fs ribbon tweeter.

Could be a long night ahead

[Cressy Snr]

Well it was 4am.

[IslandPink]

Good effort !
"Do it the way that sounds right" , is that a well-known quote ? ( yet ? )

[Cressy Snr]

Aye the effort has been worth it.
Below left is a third octave pink noise spectrograph, taken in-room at the listening position; no DSP. The one on the right is the response I was suffering from with the high crossover and no driver mods. No DSP on that either. The 2-8KHz region is pretty horrendous.



The flatness of the frequency response, on the left hand response set, between 2KHz and 8KHz, translates into a complete absence of nastiness on vocals, male or female.
The hump from 500 to 1KHz pushes voices forward of the plane of the speakers and gives a rock solid central image.
The treble extends nice and flat out to 16KHz before dropping off.

Taking on-axis plots 1m away gives a similar set of responses, just a touch of treble lift that brings up the HF a couple of dB, compared to the distant plot but the 2-8KHz region remains resolutely flat, from either position. Not bad for an amateur.

Those blu-tak rings are quite evidently doing their jobs.

The proof is, as always, in the listening, and the subjective experience reflects the measurement evidence, ie very good. Looks also like, although it is small, I do have a decent room, with a fairly even response over a wide listenng area.

[simon]

This is with that £30 mic Steve? Which app are you using?

[Cressy Snr]

This is with that £30 mic Steve? Which app are you using?

Yep, got the mic last Thursday and used it last Sunday, before I took the speakers to bits, to get the right hand side spectrogram. This meant I was able to compare the two plots as a "before/after" setup, once all my mods were completed.

The plots were obtained with two apps. The measuring app that produces the spectrograms is called "Audio Analyzer" and works on the iPad. The mic plugs into the iPad headphone socket, which is actually a four terminal input output socket.

The app that generates the signals lives on the Mac Mini and is called "Audio Test". It does lin and log frequency sweeps as sine square and sawtooth waveforms and generates various flavours of noise for analysis purposes.

It is a simple but effective system.