??
And you think that 200 "modern Beta-8's" wouldn't sell? (once word got out)
I rather think that anyone who wouldn't like to be trampled in the stampede had better stand back!
All the producer would need to do would be to have the specs and performance verified independently (ie that they match the originals), then step back and accept orders.
cheers
Doug
Ah, nice to see a thread about Coral's
I have a pair of Flat 8 II, a pair of Flat 8A, and have also worked with a bunch of the cheaper (OEM) 8A-100.
All 3 models look very similar but measure very differently above 2Khz. The best of the ones I have had are definitely the Flat 8 II.
What if anything stands out from other similar sized full range drivers ?
The Flat 8 II have a ferrite magnet and a fairly standard stamped steel frame (unlike the Beta series) so I don't think there is anything special there.
Surround is a reverse roll fabric surround - important for a full range design I think, but not unique, Fostex drivers have a very similar surround.
I think it almost entirely comes down to the cone design, especially the whizzer cone. Experimenting with various tweaks to the drivers to improve damping as well as comparing the differences in performance of otherwise similar looking drivers has given me some idea of what factors are at work.
The annular rings near the outside of the main cone seem to be very important in controlling upper midrange cone breakup, and make a considerable improvement compared to say a Fostex FE207E which doesn't have them.
On their own they're still not enough though, and I've tweaked mine with carefully placed self adhesive foam blocks on the rear of the cone in the area of the annular rings with a great deal of success. Correct placement of blocks (based on measurement) yields quite a dramatic improvement in frequency response flatness from 2-5Khz and also considerable improvement in CSD.
The next important factor is the whizzer cone design, which seems to be superior to most other whizzer cone full range drivers, certainly the ones I've seen.
The fact that they're curvilinear and extremely stiff and light seems to be key. If you bend the edge forward with your finger and release it snaps back into place with a loud "snap", that's how crisp they are. They seem to be made from a 3 layer laminate which is glued together with pre-tension to stay very rigid.
Most whizzer cones have a huge resonant peak somewhere between 2-3Khz (Fostex FE207E, I'm looking at you...) but the Corals do not, at least some of mine do not, while some which have cones in poorer condition do have a small resonance there, but nothing like the Fostex.
A curvilinear cone is naturally extremely resistant to bell modes, which is what the worst whizzer cone resonances are. Conical whizzer cones, even with the edge crimped are extremely susceptible to bell modes, and I believe this is why the whizzer cone resonance is so bad on the FE207E.
Curvilinear cones are naturally stiffer per weight in all axes than conical cones so cone breakup in general on a curvilinear whizzer cone that is sufficiently stiff and light is very minimal until well up into the treble.
For me a conical whizzer cone is a non-starter, if you are going to include a whizzer cone it must be curvilinear, so that eliminates many drivers from the running.
Finally there is the dust cap or phase plug. The Beta series have a phase plug, while the Flat series have an aluminium dust cap. I have never measured or heard a Beta driver so I'm only talking about the Flat series here.
I'm not sure how unique it is but the dust-cap on the Flat 8 is not glued to the whizzer cone, as it is on many whizzer cone full range drivers, (again, FE207E) but instead crimped directly onto the end of an extended voice coil former.
This ensures good transmission of high frequencies to the dust cap, as vibrations are not passing through a glue bond, but equally importantly provides some isolation between dust cap and whizzer cone, allowing the two to function independently at high frequencies.
The whizzer cone is actually glued to the voice coil former at the junction of the main cone, some 2-3 millimetres further down the former than where the dust cap is crimped, with the dust-cap not actually touching the whizzer.
The importance of this is two fold - the dust cap can operate by itself above about 10Khz, without being "weighed down" by the mass of the whizzer cone. This greatly increases the treble above 10Khz compared to having a direct glue bond between dust-cap and whizzer, and also improves the treble dispersion.
Secondly it avoids coupling the cavity resonance behind the dust-cap into the whizzer cone. One flaw the Flat 8 do have is they have no damping material between dust-cap and pole piece, so as in the case of many drivers there is a cavity resonance or two in there in the 1.5-2.5Khz region.
This is not normally a big problem though, because the cone area of the dust-cap is very small - you can measure this resonance when close mic'ing the dust cap, however at a normal listening distance it's drowned out by the output from all the rest of the cone, and is essentially inaudible.
However if the dust-cap and whizzer cone are glued directly together, the cavity resonance is coupled from the dust cap into the whizzer cone and radiated with many times the cone area of the dust cap alone. This puts a visible and audible peak in the response at about 1.5Khz, and a notch at about 2.5Khz.
Some Flat 8's are better or worse in this regard - one of my Flat 8 II has no trace of a notch at 2.5Khz at all, in this driver there is a clear gap between the edge of the dust cap and the whizzer cone.
The other driver does have a narrow notch of about 2dB at 2.5Khz - and in this driver there is a glue bond between dust-cap and whizzer cone.
The driver with the extra glue bond also has less treble in the 8-10Khz region. With this type of driver very small differences in assembly technique and glue application have rather large impacts on the final response, and as my pair are hand assembled units (from the new old stock cache found in the early 2000's) there is significant difference between them.
While speaking of tweaks, another simple tweak I've found for the Flat 8 is that putting a very small dent in the dust cap near the edge dramatically reduces the ringing of the dust-cap resonance at 10Khz (as seen on a CSD) with very little loss in over-all treble level. One of my two drivers already had a dent like this when I got it, which is how I discovered this, and then later duplicated the dent on the other one, measuring CSD before and after for confirmation. The dent is small enough that it can't be seen except during a close inspection.
As for modern drivers that compare to the Flat 8 ? I'm still looking for some...
Many years ago I bought some FE207E 's - big, big disappointment there, they don't even come close to the Flat 8.One driver that shows some promise but which I have not yet tried is the Tangband W8-1808. It doesn't have annular rings on the main cone, but it does appear to have a well designed curvilinear whizzer cone. It's a driver that I'd like to get and compare to the Flat 8 one day when I have some spare money...
Last edited:
Interesting. I haven't seen anyone publish measurements of the Beta 8 before.
I'm surprised at how peaky the response is, it doesn't seem to be as flat as the Flat 8. Below is a gated measurement of one of my Flat 8 II. (Only accurate down to about 500Hz due to window period)
To be fair, this is after my foam block damping tweaks, and some of my other Coral drivers measure considerably less flat.
Attachments
Will not agree with that, most of my studio monitor customers are still not 25 years old, the quest for quality never dissapeared, maybee took different routes from time to time.
But i would love to get my hands on a pair of Beta 8 or Beta 10, the first speakers John (3 years at that time) and I built together was a tiny set of Schmackshorns, just slightly above 500L, with Beta8´s, as we did this while Brittmarie, Johns mother My wife,
was away on a 3day conference we didn´t have the time for finish so when she arrived home sunday night these particleboardmonsters resided in the livingroom, she never uttered one word of protest as she loved the way the played music.
Not sure if you're referring to my Flat 8II measurement or the Beta 8 measurement, but my measurement is gated and does not include any room contribution.
Although its hard to be sure the CLIO measurement of the Beta 8 also looks like a gated measurement. Peaks and dips of that magnitude from 2Khz to 10Khz would not occur due to room effects even in an un-gated measurement if the measurement was done from about 1 metre. At those frequencies a Beta 8 will be very directional so there won't be much bounce from the floor/walls at the microphone.
I heard them a couple of time. very disappointing. the alpair 12p which Im also not really fond was much better
ive also tried to compare them with a low pass at 5khz to see if it was the highs that was problematic but no, the midrange just wasnt magic like I read. ive compared them to alpair 12p, jbl l8t and really dont see the hype.
ive also tried to compare them with a low pass at 5khz to see if it was the highs that was problematic but no, the midrange just wasnt magic like I read. ive compared them to alpair 12p, jbl l8t and really dont see the hype.
As soon as 1970, KEF lauched the "C-series" speakers consisting of Cadenza (1970-1976, B200 bass, BD139 passive, T27 tweeter), Chorale (1970-1979, B200 bass, T27 tweeter), Coda (1971-1977, B110 bass, T27 tweeter) and Cantor (1971-1978, B200 bass, T27 tweeter). The press reviewed them as very capable.
The word "coral" got associated to KEF in Europe.
The attraction for "real" Coral speakers vanished, especially their remarkable fullrange models, suddenly appearing outdated.
One must say that as soon as 1971, Coral underwent a revolution in marketing the CX-7, actually a 3-way closed box featuring a massive dome tweeter and a massive dome squawker, kind of Yamaha NS-655 NS-670 NS-690 or NS-1000.
Such radical move by Coral himself, indicated that multiway speakers were replacing fullrange speakers.
The Coral CX-7 got a first upgrade as CX-7/II, then a second upgrade as DX-3/II. A big brother came as DX-78, actually a 4-way closed box featuring a super dome tweeter on top of the dome tweeter and dome squawker, that got an upgrade as DX-11.
There were also the DX-5, CX-77 and DX-7, as Bass-Reflex variants.
Considering the magnets cost, membranes cost, assembly cost, dividing network (crossover) cost and shipping cost, those Coral multiway speakers were more costly than the Coral Beta-8 or Coral Beta-10 fullrange speakers.
Probably, those Coral multiway speakers never got equipped with a proper dividing network (crossover), for a cost reason.
Consider the cost of a proper 3-way dividing network, supposed to persuade all speaker drivers to deliver a 2nd order Linkwitz-Riley acoustic response, maintaining all speaker drivers in-phase whatever the frequency. You need four LC cells for the 3-way filtering, you need one RLC trap for damping the woofer unwanted high frequency resonance, and you need two or three RC traps for compensating the drivers impedance rise at high frequency. You may need to add more stuff like fixed resistive attenuators, or ou may prefer adding variable resistive attenuators of constant impedance (this is no cheap stuff).
Compared to fullrange speakers, multiway speakers need more science for working properly together, and they eat more resources (copper, magnets, baskets, glues).
People pretending that fullrange speakers were excellent because of being very costly to produce, are plain wrong.
Today, the advantage of fullrange speakers, is that even if you are a zero in electronics and audio science, not owning a proper audio lab (PC-based dual channel realtime FFT analyzer, decent USB sound card, decent mike preamp, decent mike), you still can feel some enjoyable "wow" effect after mounting a fullrange speaker in a cabinet of your choice, and listening to it. Clearly, less than a day of work, and less than $ 250 in spending, can bring you some self accomplishment satisfaction.
But, is the final result better than regular commercial multiway speakers ?
Not sure ...
The same is is not feasible in a multiway context. You'll waste one day in properly measuring the bare speaker drivers, you'll waste another day sketching a crossover, and you may waste a month in tweaks. Anyway, the bill will far exceed $ 250. After one month, your wife will ask you to wipe out the mess, stop the expense, save money during 6 months, and go with her buying a nice $ 1500 pair of speakers, better sounding than all your attempts, exhibiting some solid residual value, even after 5, 10 or 20 years. And for sure, such multiway speaker will sound better than any fullrange speaker.
Last edited:
magnet system incide a lot on final sound , field coils ,alnico, neo ,and ferrite all work different ,with different problem like the eddy current of ferrite ,the copper ring is used to tame this .all the best loudspeaker like altec are made in alnico,remember that al guitar pickup are made in alnico too, there is different grade (II,III,IV, V ,VI ,VIII) that sound different only few hight output for metal are in ferrite
Press - Coral Electronic see the pdf "History - Audio Car Stereo (IT)" can't read Italian but it looks to me this is where coral ended up
The equivalent of Coral Beta fullrange speakers, are those modern 3-inch speakers featuring an aluminium or beryllium alloy membrane, that's a direct extension of the coil former. This way all the highs get transferred to air. This way, the highs don't get absorbed by some glue. Of course, you need an extremely tight glue between the coil (source of movement) and the coil former.
I encourage you to experiment a 2-way MTM speaker made of :
- one Fountek FR-58 speaker covering 200 Hz to 20 kHz
- surrounded by two Rockwood DYA610-68 speakers covering 20 Hz to 200 Hz
- a 32-bit FIR filter crossover, linear phase, for ending up with acoustic 3rd-order slopes
- adequate volume controllers
- chipamp power amplifiers like LM3886 or TDA7293
This is inexpensive in case you already own (and can program) the digital crossover, own the volume controllers, and own the chipamps.
Programming the 32-bit FIR filter is not so trivial.
A 1000-tap FIR filter will feature a 44.1 Hz or 48 Hz frequency resolution, depending on the sampling frequency (44.1 kHz or 48 kHz) you are using.
The 1000 coefficients of the lowpass filter need to persuade the DYA610-68, to deliver a 3rd order lowpass acoustic response at 200 Hz. You thus need to adapt the 1000 coefficients, to the actual DYA610-68 acoustic response.
The 1000 coefficients of the highpass filter need to persuade the FR-58, to deliver a 3rd order highpass acoustic response. You thus need to adapt the 1000 coefficients, to the actual FR-58 acoustic response.
For determining the 1000 coefficients of the above FIR filters, you need to calculate (in gain and in phase) the required Bode diagrams, and apply an inverse DFT on them. The inverse DFTs deliver the impulse responses of the filters that you require, which also form the coefficients list of the FIR filters that you require.
Amazingly simple, isn't ?
Oh, by the way, you should add a 32-bit IIR-based equalizer, say two 32-bit IIR biquads in series, for straightening the bass response.
You can implement this using a Windows PC, Flowstone, and a 4-channel-out USB audio card operating under ASIO, possibly after installing ASIO4ALL after installing the required Windows driver, if no proper ASIO driver is supplied.
Flowstone : DSP Robotics Support • View topic - Speaker Lab
4-channel-out USB audio card : ESI - MAYA44 USB+
Consider this to be the modern and improved equivalent of a Coral Beta fullrange speaker, covering 40 Hz to 15 kHz in a 2 dB corridor (provided that you listen on-axis).
This is perhaps the reason, why nobody dares producing Coral Beta fullrange speakers anymore (at decent selling prices).
It takes an awful long time, a lot of specialised knowledge and huge investment to develop a good speaker at all and this cost is not easily recovered by short run or niche market numbers so the only people in position to develop these sort of drivers are established manufacturers who are quite reluctant to change the way they build speakers.
So, the only way new speakers come onto the market is from 'fresh start' manufacturers and there are quite a few of them but they too have to create their niche markets to survive and this is why there will possibly never be drivers like the Coral beta8 manufactured again - china manufacturers aren't especially interested in the small numbers, assuming that they could actually build drivers to these requirements.
With the accelerated development of the programs available to design and build speakers, it's only a matter of time until drivers that exceed the best features of the beta 8s become available - they were good and could easily be improved (reduced cone resonances, etc) but the market size is a huge barrier.
Another problem is the target market - much of the Asian market listen at far lower volumes than many of us and so high efficiency isn't a major factor for that market, as can be seem by the Mark Audio, El Ciao, etc designs
I haven't heard any of Katy Fertin's drivers (eMSpeakers) but they look quite promising on paper - no idea about prices
Some of the relatively easier drivers that respond to full or partial eNABL treatment are pretty impressive and improving all the time too - another approach
But, sadly, not a Coral beta8 replacement
So, the only way new speakers come onto the market is from 'fresh start' manufacturers and there are quite a few of them but they too have to create their niche markets to survive and this is why there will possibly never be drivers like the Coral beta8 manufactured again - china manufacturers aren't especially interested in the small numbers, assuming that they could actually build drivers to these requirements.
With the accelerated development of the programs available to design and build speakers, it's only a matter of time until drivers that exceed the best features of the beta 8s become available - they were good and could easily be improved (reduced cone resonances, etc) but the market size is a huge barrier.
Another problem is the target market - much of the Asian market listen at far lower volumes than many of us and so high efficiency isn't a major factor for that market, as can be seem by the Mark Audio, El Ciao, etc designs
I haven't heard any of Katy Fertin's drivers (eMSpeakers) but they look quite promising on paper - no idea about prices
Some of the relatively easier drivers that respond to full or partial eNABL treatment are pretty impressive and improving all the time too - another approach
But, sadly, not a Coral beta8 replacement
Today, nobody will appreciate vintage speaker like Coral, people buying it merely looking for re-sale value not for music.
I owned a pair of Coral Beta 8 for many years, thinking of letting it go but too bad no body appreciate it at all. Sad to say this.
Share the photo here.
I owned a pair of Coral Beta 8 for many years, thinking of letting it go but too bad no body appreciate it at all. Sad to say this.
Share the photo here.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.